JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 103, NO. B6, PAGES 12,413-12,433, JUNE 10, 1998
The tectonic history of the Tasman Sea: A puzzle with 13 pieces CarmenGaina, • Dietmar R. Mtiller,•Jean-Yves Royer,: Joann Stock, 3 JeanneHardebeck, 3 and Phil Symonds4
Abstract. We presenta new model for the tectonicevolution of the TasmanSea basedon dense satellitealtimetry data and a new shipboarddata set. We utilized a combinedset of revised magneticanomaly and fracture zone interpretations to calculaterelative motions and their uncertaintiesbetween the Australianand the Lord Howe Rise platesfrom 73.6 Ma to 52 Ma when spreadingceased. From chron31 (67.7 Ma) to chron29 (64.0 Ma) the modelimplies transpressionbetween the Chesterfieldand the Marion plateaus, followed by strike-slipmotion. This transpressionmay havebeen responsible for the formationof the CapricornBasin south of the Marion Plateau. Anothermajor tectonic event took place at chron27 (61.2 Ma), whena counterclockwisechange in spreadingdirection occurred, contemporaneous with a similarevent in the southwestPacific Ocean. The earlyopening of theTasman Sea cannot be modeledby a simple two-platesystem because (1) riftingin thisbasin propagated from southto northin severalstages and(2) severalrifts failed. We identified13 continentalblocks which acted as microplates between 90 Ma and64 Ma. Our modelis constrainedby tectoniclineaments visible in the gravityanomaly grid and interpretedas strike-slipfaults, by magneticanomaly, bathymetry and seismicdata, and in caseof the SouthTasman Rise, by the ageand affinity of dredgedrocks. By combiningall this informationwe derivedfinite rotations that describe the dispersal of thesetectonic elements during the early openingof the TasmanSea.
1. Introduction anomaly24) for the TasmanSea. They describedthe opening of the basin in terms of a two-plate spreading system but About two decadesago the opening of the TasmanSea was consideredthe possibility of a subductioninterval to account modeled by several authors as a simple two-plate spreading for the truncated pattern of magnetic anomalies near SE system,active betweenchron 33old (80 Ma) and post-chron24 Australia. After calculating finite rotations based on (52 Ma). New geophysicaldata, however,offer an opportunity reidentified magnetic anomalies, Weissel and Hayes [1977] for reconstructingthe tectonic history of the Tasman Sea in concludedthat the assumptionof an episodeof subductionof muchmore detail than previously possible. We present a new Tasman basin crust at the east Australian margin is not tectonic model based on gravity anomaly data derived from necessary, because an extension of the two plate model satellitealtimetry and on both existingand new magneticdata, describedthe evolutionof mostof the basinquite well. as well as on seismicand geologicaldata. Shaw [1979] was the first author to interpret magnetic The Tasman Sea is an ocean basin boundedto the west by anomalies north of 30øS in the Tasman Sea based on new continental Australia, to the east by the Lord Howe magnetic data. Following Ringis' [1972] idea that initial Rise/ChallengerPlateau and New Zealand,and to the southby a rifting had takenplace in the Lord Howe and Middleton basins major discordancethat separatesit from younger oceanic crust (Figure 1), Shaw [1979] attempted to identify magnetic generated at the Southeast Indian Ridge and the extinct anomaliesin thesebasins. He arguedthat the northernTasman MacquarieRidge spreadingcenter (Figure 1). Its northernpart Sea morphologycould be explainedonly by strikeslip motion contains an elongated segment of continental crust, the and not a pull-apartmechanism as suggestedby Weissel and Dampier Ridge, separatedfrom the Lord Howe Rise by two Hayes[1977]. Stockand Molnar [1982] interpretedmagnetic small basins: the Lord Howe and the Middleton basins. anomalyand fracturezone locations in the SW Pacific region, Hayes and Ringis [1973] identified for the first time a and computednew rotations for anomalies 28 and 32 in the complete magnetic anomaly set (from anomaly 330 to TasmanSea, including partial uncertaintyrotations. Another tectonic model for the central and southern Tasman Sea, based • Departmentof Geologyand Geophysiscs,University of Sydney, New South Wales, Australia. on a revisedmagnetic data set and the gravity anomaly grid 2 Unit6 Mixte de Recherche6526, G6osciencesAzur, Villefranche publishedby Sandwelland Smith[1992] hasbeen proposed by Sur Mer, France. Rollet [1994]. Rollet [1994] used the inversion method 3 SeismologicalLaboratory, California Institute of Technology, developedby Royer and Chang [1991] for computingfinite Pasadena. 4 AustralianGeological Survey Organisation, Canberra, Australia. rotations,and the resultsare very similar to Shaw's [1979].
Copyright1998 by the AmericanGeophysical Union 2. Magnetic Data
Papernumber 98JB00386. In this studywe compiledmagnetic anomaly crossings from 0148-0227/98 JB-00386509.00 new ship data (HMAS Cook (1985-1989), R/V Melville
12,413 12,414 GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA
l TASMAN SEA magneticanomaly • 24o, 28y, 33y, 25 øS • 25y c• 260, 31y, 34y o 27o, 32y • 29y v 30y
30'S
35'S
40'S
45'S
• . 50øSl"_ I I I I I I I I I I I ll 150øE 155•E 160øE 165•E 170øE
Figure 1. Tasman Sea, magnetic anomalies plotted along selectedship tracks. Shadedareas indicate continental crust;AUS, Australia;TAS, Tasmania;LHR, Lord Howe Rise; DR, Dampier Ridge; STR, South TasmanRise; GSM, Gilbert SeamountComplex; NZ, New Zealand;Chest. P1, ChesterfieldPlateau; Chall. P1, ChallengerPlateau; Bellona Tr, Bellona Trough;Lord Howe, Lord Howe basin; Middleton, Middleton basin; ETR, East Tasman Plateau; Marion P1, Marion Plateau; dark grey line represents extinct ridge axis. Bold profilesare presentedin Figures3 and 10. GAINAET AL.:TECTONIC HISTORY OF THE TASMAN SEA 12,415
sw ridgeextinctaxis bmr14 /•_^ )•,,•/• I I .I I I I I , I I • , , ' ' , ; ' ; •; • , mv9512 ,',' I • • : ' ; ' • '" II ' ' ' I I
I I ' at941 r . ß ' • ' , . I I !
! , ,, - ' , , I , ,' ,, synthetic
• ,t ,,' ,I ,I .,, :{l I a, . , profile
• '.' , , ' •. • , , , ! ' I ' { , I ' I I I I • I , , , , , , , , , , , , , • • , • , , , , • ',
34y 33o 33 32 31 30 29 28 27 26 25 • • 25 26 27 •29 30 31 32 33 330 •y 4m•yr q, . 22mm/yr • 16•yr • 20•yr _,.4mm/yr h•f spread•g ute
83 79,1 73.6 71.0 67.7 64.0 61.2 55.9 53.3 I I I I I I I I I Ma Figure 2. Selectedmagnetic anomaly profiles of theTasman Sea compared with synthetic magnetic anomalies(see locations in Figure 1). Thesynthetic profiles are based on the geomagnetic timescale of Cande andKent [1995], a depthto the upper surface of themagnetized layer of 3 km,a thicknessof the magnetized layerof 1.5km and an azimuth of 150ø. Thepresent inclination and declination ofthe magnetic field were calculatedfrom the IGRF90 reference field and the paleoinclination and paleodeclination were calculated based onthe position of thepaleomagnetic polefor the Australian plate at 52 Ma[from Irving and Irving, 1982]. Australia'slatitudinal position did not change substantially while the Tasman Sea was opening; hence we foundthat it makeslittle difference for the phase shift of syntheticmagnetic anomalies whether we use a late Cretaceousor an earlyTertiary paleomagnetic pole.
(1995), N/O L'Atalante(1994), and R/V MauriceEwing wherewe were able to identify a considerablenumber of (1996)) andreinterpreted previously existing data (Figure 1). magneticanomalies. However,a detaileddiscussion of Using the Candeand Kent [1995] reversaltimescale, we spreadingasymmetry is beyondthe scopeof this paper. identifiedeither young or oldends of magneticchrons (Figure Specialattention has been paid to the northernpart of the 2), unlikeprevious authors who chose the centerof magnetic TasmanSea where volcanic activity has partly obliterated anomalies[Weissel and Hayes,1977; Shaw, 1979] or only the magneticlineations on the seafloor;we succeededin olderends [Rollet, 1994]. Anomalieson the westernflank lie identifyingmagnetic anomalies in the northernmostTasman fartherfrom the ridgethan correspondinganomalies on the Sea but not in the Lord Howe or Middleton basins, which are easternflank. Asymmetryin crustal accretionoccurred likely to be underlainby extendedcontinental crust. We especiallybetween chron 270 andchron 33y. This observation identified anomalies330 and 34y (young) in the central is mostlybased on magnetic data from the central Tasman Sea, Tasman Sea. However, the scarcity of data hinder the 12,416 GAINA ET AL.: TECTONIC HISTORY OFTHE TASMAN SEA
interpretation of Late Cretaceousanomalies in the southern The selectedfracture zone limb displaysthe samepattern as Tasman Sea. the other fracturezone in the TasmanSea (i.e., a changein spreadingdirection shortly before the ridge becameextinct; 3. Fracture Zone Data Figure3). As the morphology of this fracturezone resembles the Atlantic type (a central valley boundedby a high wall on Images of the marine gravity field [Sandwelland Smith, the older side), we usedthe troughsin the gravity anomaly as 1997] from satellite altimetry allow improved interpretations the actual location of the fracture zone. To better constrain of fracture zones and other tectonic features (an extinct reconstructionsfor chrons25y and 240, which imply a change spreadingridge, continent ocean boundary, V-shaped features). in the direction of spreading,we includedadditional points The generalNNE trendof the fracturezones interpretedfrom from other relevantfracture zone segments. densesatellite gravity datadiffers from previouslyinterpreted fracturezones by about 10ø (NNE versusNE). Many fracture 4. Method zones show a distinct counterclockwisechange in spreading direction which occurredshortly before spreading ceased, The new magnetic anomaly picks and fracture zone traces possiblyat chron25. The newgravity anomalydata allowed digitized from satellite datahave been combinedto calculate us alsoto bettermap the fracturezones in the northernTasman new finite rotations and their 95% uncertainty confidence Sea (fracturezone valleys are characterizedby negativegravity regions for the opening of the TasmanSea. They have been anomalies, see Figure 3), previously mappedonly from computedusing an inversion method based on Hellinger's magneticanomaly and bathymetricdata. The fracturezones [1981] criterionof fit, developedby Royer and Chang [ 1991]. undulatewhich reflectsseveral changes in spreadingdirection. In this method both magnetic and fracture zone data are Becauseof insufficient magnetic anomaly identifications in regardedas points on two conjugateisochrons consisting of the northern Tasman Sea and because of large offsets of great circle segments. The best fit reconstructionis derived northern Tasman Sea fracture zones, we did not use these from minimizing the sum of the misfits of conjugate sets of fracture zones to constrain our reconstructions. Only one magneticanomaly and fracturezone data pointswith respectto fracture zone (which is located at about 40øS) in the entire individual great circle segments. TasmanSea fulfilled our requirementsto be includedin the Since the quality of the determinationof the best fitting inversion:namely, it showsa mediumoffset (about 120 km), rotationsand their uncertaintiesdepends on the errors assigned appearsto follow a flow line, andits locationis constrainedby to the inverted data, we carried out an analysis of data magneticanomaly data on both sidesas well as by satellite dispersion. For assigningnavigation uncertainties to the ship gravity data. data, we faceda choice of either estimating navigation error The location of fracturezones depends on the type of the through time since the 1960s, basedon changing navigation ocean floor spreadingregime. Large offset Pacific fracture methodsin the last 30 years (Rollet's [1994] approach) or zones(i.e., a fast spreadingregime) exhibit a depth/agestep rather assumingthat the navigationerror is roughly similar for (the old side is lower due to the depth-agerelationship of all data we have used. The latter approach was adoptedin a oceaniclithosphere) [Sandwell and Shubert,1982], whereas detailed analysis by Royer et al. [1997] in the Indian Ocean. small to medium offset Atlantic and western Indian Ocean (slow To investigatewhether navigationerrors do vary substantially spreadingregime) fracturezones are characterizedby a central resulting from different navigation methods, Royer et al. valley [Van Andel, 1971; Fox and Gallo, 1986]. Mediumand [1997] divided magneticdata into four categoriesbased on the large offset fracturezones observed in the North Atlantic are quality of the navigation: the A category included data since asymmetricin crosssection, with a high wall on the old sideof 1985 (basedon Global PositioningSystem (GPS) navigation), the fracture zone [Collette and Roest, 1992]. the B categoryincluded data from 1980 to 1985 (basedon GPS To determine the correct location of the eastern limb of the and Transit satellitesnavigation), the C categoryincluded data fracture zone that we chose for our inversions, we combined from the 1970s which relied mainly on Transit satellite observations of magnetic anomalies and gravity anomalies navigation and the D category included pre-1970 cruises from shipand satelliteand seismicdata. The gravity anomaly without satellite navigation. Their expectation was that the grid from satellitealtimetry displaystwo possiblelocations navigationalerrors shouldincrease for older data. Becausethe for the easternlimb of the selectedfracture zone (Figure 4a). A numberof data points in individual categorieswas not large seismic profile (Eltanin 34), which crossesthe two possible enoughto be inverted separately,they combinedtwo or three locations indicates an obvious change in topography (about groups of data for their dispersion analysis. Their results 1000 m) at 40øS, 160øE (intersection with the northern fracture surprisingly show that navigational uncertainties for data zone) and a high wall on the old side of the northern fracture includedin categoriesA, B, and C differ by a small amount zone as described for Northern Atlantic fracture zones by only, ranging from 3.0 km for the new data to 3.9 km for the Collette and Roest [1992] (Figure 4b). The gravity anomaly olderdata. The D datawere assigned an uncertaintyof 5.2 km. from the Eltanin 34 profile showsa dramaticchange in values It is important to point out, however, that uncertainties (--20 mGal)coincident with the changein topographyvisible resultingfrom this analysisof dispersiondo not strictly reflect on the seismic profile. The two possible fracture zone navigational errors alone but rather reflect a sum of locationsare intersectedby three magnetic profiles: two Verna uncertaintiesin navigation and in locating the magnetometer profiles (v3215, v1812), and one Cook profile (cook60_0) with respect to the ship and in identifying the magnetic (Figure 4a). The shapesof the easternmostanomalies observed anomalies. If the sum of these errors were dominatedby on profile v1812 more closely resemble anomaly 33 (as navigationerrors, then magneticanomaly identifications from interpretedon profile v3215) than anomalies 32 and 33 (as GPS navigatedcruises would be expectedto show errorsin the identified on profile Melville-mv9512, north of v3215), rangefrom 0.2 to 1 km, ratherthan 3.0 km, as foundby Royer indicating that the correctinterpretation of this fracture zone is et al. [1997]. The discrepancylikely reflectsuncertainties in the northern alternative. locating the magnetometerwith respect to the GPS antenna, GAINAET AL.:TECTONIC HISTORY OF THE TASMAN SEA 12,417 38'S
90
80
39'S 70
60
50
40
30
40'S 20
10
0
-10
-20
41'S -30
-40
-50
42'S
156'E 157'E 158'E 159'E 160*E 16I*E 162'E 163'E
shiptrack gravity 15-
10-
5- 0-
-5 -10 - ' \ I• /- x. ../ /\ i •1 ,,,I \'..-'x I \ ^ / \ -15 - / \ / \ i..../ .x./ \ / -20 - • i \ I -25 - \ I
7600 7700 7800 7900 distance[km] Fz Fz
6-
7-
A 21:00I" 0:00 04:00 07:00I 11:•0 !
FZ FZ
Figure 4a. Greyshaded contours of gravityanomalies from satellite altimetry (Mercator projection); superimposedare magneticanomaly wiggles from ship dataalong severalship trackswhich cross this area. Thin solid lines are two possible fracturezone locations (see text for discussion),thick solid line is the extinctridge axis. Magneticanomaly identifications are representedby soliddots accompanied by their identificationnumbers. The profileAB is from Eltanin34 anddisplays time annotations. GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA 12,419
shiptrack gravity
10 5 0 -5 /'"- • • ,.. /\ -10-15 //•/ ',,,/\¾.• \ / // " \\\ /'-/ ^•'/ / \\ / -20 / • -25 •/ x/ \\ / / \ /
7600 7700 •7800 7900 I ' I distance[kin]' •z I '\•'{ I '
8
21 .'00 0:00 04:00 I 07:00i 1170 I B
FZ FZ
Figure 4b. Gravity and seismic profiles along segment AB from Eltanin 34 ship track show step in basementdepth boundingthe fracturezone valley. Shipboardgravity gravity anomaly of the same profile extractedfrom satellitealtimetry grid is plotted(dashed line). Arrowsindicate the two possible fracturezone locations. which is usuallytowed 300-400 m behindthe vessel,as well as Theparameter •' indicateswhether the assigneduncertainties uncertaintiesin locatingmagnetic reversals along track, which are relatively correct (t• = 1), underestimated(•'<<1) or are never substantiallybelow 2-3 km. The implication is that overestimated(t•>>l). We calculatedthe averagevalue of t• recentlyacquired magnetic data cannot be locatedas accurately basedon each individual magneticanomaly data set with the as expectedfrom assessingthe quality of the GPS navigation following formula: systemalone, and the navigationquality of ship datafrom the 1960s may be better in some areas than expected from technical arguments. i._L__ i=1•, tCi._•__•} (2) Becauseof fewer data in the Tasman Sea comparedto the K'avg tt centralIndian Ocean, we could not computerotations from such subsetsof data. Instead,we calculatedthe best fitting rotations wheret• i is the estimatedquality factor calculated for eachset using all data (magnetic data and fracturezones) and then of magneticcrossings and n = 9 (numberof data sets). appliedthese rotations to the magneticcrossings only in order Weobtained •avg of 2.01 with a 95%confidence interval of to assesstheir dispersion. We assignedan initial uncertainty 1.03 to 3.28 (Figure 5). Thus the true 1-sigmastandard error of of 7 km to all magneticanomaly crossings. This was done for themagnetic dataiso' = •' / • =4.94 km. This value has eachof the nine setsof magneticanomaly data corresponding to chrons24-32, and a qualityfactor t• wasdetermined for each been roundedup to 5.0 km. These errors includenavigation chron.The quality factor •' relatesthe uncertainties assigned errors and uncertaintiesin locating the magnetometerand in interpreting the magnetic anomalies. For each data set, Table tothe data (6') to their true estimates (0'),t•=(•'/o') 2. 1 showsthe valuesof t•, the trueuncertainties (O'), the total Althought• is unknown,it canbe estimatedfrom the misfit, misfit (r), the numberof degreesof freedom(dF) and the number the geometry of the plate boundaryand the numberof data of datapoints (N). Thevalues of •' aregenerally larger than 1 [Royer and Chang, 1991] as follows: (exceptfor anomaly32, where• = 1, andfor anomaly33 where t• I I I Most reconstructionsare characterizedby an approximately linear distribution in qq plots. For the older reconstructions, the curves formed by the histogram peaks differ from a Gaussiandistribution, mainly because of few datapoints or, as in the case of anomaly 33, slight underestimationof uncertainties in the data. We show the finite rotation pole positions and their uncertainties(Table 2a) in comparisonwith poles of rotation obtainedby other authors(Figure 8). We presentcovariance matrices which describe uncertainties in rotation locations which are computedin a referenceframe fixed to the Australian ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: plate(Table 2b). They werecomputed as following: i i o a24y a25y a26o a27o a28y a29y a3Oy a31y a32y (1!/•')X d x10 -6 (3). Figure 5. Estimationof standarderror assignedto magnetic anomalycrossings. Confidence intervals for •' have been computedfor each anomaly data set (excluding fracturezone Finite rotationsfor chrons34y (83.0 Ma)and 33o (79.0 Ma) crossings). The resulting has been averaged, in order to werecalculated assuming that the pole of the oldest derived obtainan averagedtCavg value which is includedin each stagerotation (betweenchrons 33y and 32y) remainedstable confidenceinterval (stippled area showsthe uncertaintyfor the since chron 34y. The numberof datapoints for these two averaged•'). Thisestimated •' hasthen been used to inferthe reconstruction times is insufficient to derive total uncertainties in the magnetic data locations (see text for reconstructionsindependently. The path defined by our discussion). rotation poles showsthe sameSE-NW trend as the poles previously calculated, but it is situated more westward and 5. Results and Discussion includestwo cuspsdefined by polesof anomalies34y, 32y, and 29y and poles of anomalies270, 260, and 25y. The We presentlocations of anomalypicks usedin this study differencesbetween the locationsof the new poles and other (open symbols) along with their restoredpositions (rotated rotationpoles are the resultof usingan improvedmagnetic usingthe new set of finite rotations, solid symbols), for both dataset (a largernumber of magneticanomaly identifications, the easternand westernflanks of the Tasman Sea (Figure 6). which includesmagnetic data in the northern TasmanSea), a Generally, the fitted data match fairly well, except for the different interpretation of fracture zones orientation, and a reconstructionat chron 33y which is representedby few data differentcomputation method. The pole determinedfor chron points and is less constrained. 240 hasthe largestconfidence ellipse dueto its very small We analyzedthe variancebased on the assumptionthat the rotation angle. data are selectedfrom approximatelynormal distributions. The Using the new set of finite rotations,we derivedstage normality of the distributioncan be verified using a normal rotations and several synthetic flow lines in the Tasman Sea. quantileplot (qqplot). This representationarranges the data The stage poles clusterin one group located in northern residualsin increasingorder and thenplots each datavalue at a Australia,except for stagepoles between chrons 33-32, 29-28, position in a normal distribution. If the data residuals are and 27-26 which have more remotelocations (Figure 9). normally distributed,all points in the plot shouldlie on a Taking into account the sparse data used in older straight line. Figure7 showsqq plots and histogramsof the reconstructions,the stagepoles determined for the more recent weightedresiduals. The histogramsof dataresiduals show the intervals are more reliable than those determined for older distributionof normalizedresiduals between best fitting great anomalies. The synthetic flow lines and the trends of the circle segmentsand magneticanomaly and fracturezone data. fracturezones observed from gravity anomaliesmatch well Table 1. StatisticalParameters for FiniteRotations Derived From Inversion of Magnetic Anomaly Data Only Chron ?mag O'mag Fmag dFma• Nmag 24o 3.88 3.55 0.51 29 54 25y 2.77 4.20 0.60 29 48 26o 2.33 4.58 0.65 29 52 27o 1.34 6.05 0.86 32 57 28y 1.62 5.49 0.78 35 64 29y 3.82 3.58 0.51 30 55 30y 4.39 3.34 0.48 17 40 31y 1.57 5.59 0.80 22 41 32y 1.07 6.76 0.97 15 32 33y 0.68 8.50 1.21 13 32 The•' representstheestimated quality factor, O' is the true unknown uncertainty (inkm) estimated fromt• obtained byinverting themagnetic crossings onlywith a •' =7.0 km, cr = •' / •/-•avg,r isthe total misfit,dF is thedegree of freedomand N is numberof datapoints in eachdata set. 25 øS 30øS 35øS 40øS 45 øS 50øS 1500E 1550E 1600E 1650E I I I I I I I I I I I I I -70 -60 -50 -40 -.'-0 -20 -10 0 10 20 30 40 50 Figure 6. Gravityanomaly from satellitealtimetry. Open symbols are magneticanomaly picks (see legend from Figure1) usedin determinationof the finite rotations(Table 2a); solidsymbols are magnetic anomaly crossings after rotation (from eastto westand from westto east);white lines are synthetic flow linesderived from the finite rotations. The seedpoints (represented by largeopen circles along the extinctspreading ridge) mark time of seafloorextinction in the TasmanSea (52 Ma). Becausethe fracturezone used in our reconstructionhas a mediumoffset, we constructedtwo flow linesfrom seedpoints located at the ridge-transformintersections. The fracture zonelocation lies inside the envelope defined by thetwo flow lines.The extinctridge axis is drawnin darkgrey, COB's are thin solidlines. The fracturezone used for derivingthe finite rotationsis shownas a dashedsolid line. 12,422 GAINA ET AL,: TECTONIC HISTORY OF THE TASMAN SEA 24o 25y 26o 27o 28y 0 i ....I ....I .... • .... • .... • .... •- 0 •.... • .... • .... • .... • .... i ....! 0 •.... • ...... •.... 0 'l.... I ....• .... • .... • i"'• ....i ....• .... i ....• .... -3-2-1 0 1 2 3 -3-2-1 0 1 2 3 -3-2-1 0 1 2 3 -3-2-1 0 1 2 3 -3-2-1 0 1 2 3 weightedresiduals 2 2 2 2 . 2 ... o ..... o ..... o .....ß o •.... o •.' -2 -2 ' -2 ' -2 ' ,2 ß . ' I ' ' ' I ' ' ' I ' I ''-' I •' ' I ' ' I ' ' ' I ' ' ' I ' I '' '• '' ' I ' • ...... -2 0 2 -2.. •. 2 -2. ....-' 0 2 -2. 0 2 -2.... 0 2 29y 30y 31y 32y 33y ß i 0 I....i illlI ....illill i .... iiiI[ill .... illl.i .... iiI Iii1•1 ....I 0 I.,i,,'"i I,,,,...... I,,,, I,,,,I,i,,,I,I.... ,,,I 0 i....lli I.Ii I....IIIIIi .... Illli ....Illl i'"*iIIII I I....Ill I.I 0 Illll I.... IIIll,l,l•I .... I ....II"11i .... III ....IIl,I, I •"' III I 0 II ....IIIIII,111l .... I .... ItIIIllIll .... II.•lI .... IIfilIll .... If -3-2-1 0 1 2 3 -3-2-! 0 1 2 3 -3-2-1 0 ! 2 3 -3-2-1 0 1 2 3 -3-2-1 0 1 2 3 weightedresiduals Figure 7. The qq-plotsand histogramsof the weightedresiduals for all reconstructions.The histogramsof data residualsshow the distributionof the normalizedresiduals between bestfitting great circle segmentsand magneticanomaly and fracture zone data. (Figure6), especiallyin the northern TasmanSea where the complicated compressional pattern along some transform undulatingshapes of fracture zones correspondwell to the faults. synthetic flow lines (Figure 3). Counterclockwisechanges in Changesin spreadingrates and spreadingdirections are spreadingdirection and right lateral ridge offsets produceda calculatedfor a point situatedat the intersection of the extinct Table 2ao Finite Rotations Between the Australian and the Lord Howe Rise Plates With Australia-Fixed Chron Age, Latitude Longitude Angle, r t• dF N s Ma øS øE deg 24o 53.3 14.19 130.41 -0.723 16.13 2.29 37 66 13 25y 55.8 15.93 133.47 -2.112 21.38 2.11 45 68 I() 260 57.9 16.93 136.23 -3.792 21.60 1.53 33 58 I I 270 61.2 4.65 131.51 -4.432 27.75 1.26 35 62 12 28y 62.5 4.71 132.68 -5.168 26.39 1.55 41 72 14 29y 64.0 0.19 130.37 -5.461 10.28 3.31 34 61 12 30y 65.6 3.99 131.80 -6.735 5.65 3.72 21 46 l l 31y 67.7 9.04 134.46 -8.83 29.36 0.99 29 50 9 32y 71.1 14.72 139.04 -13.078 26.11 0.80 21 40 8 33y 73.6 9.53 137.20 -12.937 32.89 0.61 20 41 9 Parametersarer istotal misfit, •' is estimatedquality factor, dF isnumber of degreesof freedom,N is number of datapoints, s is numberof greatcircle segments. We assignedCY =5 km for the magneticand fracture zone crossings. GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA 12,423 26 10'S 20'S © this study ¸ Rollet, 1994 © Stock andMolnar, 1982 •, Shaw, 1979 ß W½tss½l& Hayes, 1977 120'E 130'E 140'E 150'i:t. Figure 8. Locationsof the finite reconstructionpoles (Table 2a) and their 95% confidenceregions for the relative motion of Australia/LordHowe Rise from chron 34y (83 Ma) to chron 240 (53.3 Ma). Previously determinedpole pathsare shownfor comparison. ridge with the fracturezone that we constrains our plate model slow half spreadingrate of 3.1 mm/yr. At 79 Ma (chron 33o), (Figure 10). Errors in the spreadingdirections and rates for the half spreadingrate increasedto about 20 mm/yr and has chrons240 to 33y are derived from the covariancematrices of been approximately constantuntil chron 25y (53.3 Ma), when the stage rotations. Seafloor spreadingin the Tasman Sea it decreasedto 16 mm/yr. The errorsin the half spreadingrates beganbefore chron 34y (at ~ 84 Ma) in a NE-SW direction at a are within +2 mm/yr; whereasfor stages 31y-30y, 30y-29y, Table 2b. Australia-Lord Howe Rise Rotation Uncertainties Chron a b c d e f _. 240 5.098 -2.583 4.238 1.593 -2.344 3.775 25y 8.027 -3.805 6.735 2.422 -3.710 6.265 260 1.120 -5.325 1.027 3.133 -5.384 1.003 270 1.119 -5.135 9.585 2.897 -4.798 8.719 28y 1.128 -1.965 7.455 1.605 -2.308 5.889 29y 9.889 - 1.695 6.104 1.664 -2.127 4.769 30y 3.076 - 1.839 3.426 1.748 -2.682 4.47i 3 ly 2.100 -1.014 1.981 5.707 -1.028 1.954 32y 3.149 - 1.661 3.058 9.136 - 1.634 3.006 33y 2.838 -1.489 2.804 8.307 -1.490 2.800 The uncertaintiesin the finite rotations are describedby covariancematrices, here computedin a referenceframe fixed to the Australianplate. 12,424 GAINA ET AL.:TECTONIC HISTORY OF THE TASMANSEA al., 1989]. Therefore, initiation of rifting in the basin between about 68 and 64 Ma resulting from left-lateral transpression,followed by strike-slip, is supportedby the sedimentarysection recovered at Aquarius-1. Seafloor spreadingceased shortly after chron 240 (53.3 Ma). The tectonic events which started at chron 270 and culminatedin the cessationof spreadingreflect a major change in plate drivingforces of the Australianplate duringthat time, that predatesthe beginning of the India-Eurasiacollision at chron 24 [Rowley, 19961. 6. Tectonic Elements The early openingof the Tasman Sea cannot be modeledby a simple two-plate system because rifting in this basin propagatedfrom south to north in several stages and several rifts. The boundaries between continental and oceanic crust -31 (COB) used in this study have been defined by combining satellite gravity anomalies,bathymetric contours(ETOPO-5), magneticdata and, where possible,seismic data. We identified 13 continental blocks which actedas microplatesbetween 90 Ma and 64 Ma, namely, the Lord Howe Rise, split into three blocks; the ChallengerPlateau; four fragmentsof the Dampier Figure 9. Stagepoles and their 95% uncertaintyellipses for Australia-Lord Howe Rise relative motion (Australia is fixed). Ridge; the ChesterfieldPlateau; the east and west SouthTasman The dashedline showsthe stagepoles path. Rise; the East Tasman Plateau; and the Gilbert Seamount Complex (Figure 1). Our model is constrainedby tectonic lineamentsvisible in the gravity grid interpretedas strike-slip and 28y-27o, the errors are larger. Uncertaintyin the faults, by magnetic anomaly, bathymetry, and seismic data, directionsof spreadingis large,especially for stagepoles 33y- 32y, 31y-30y, 30y-29y and28y-27o (Figure10). However, the changesin spreadingdirection exhibit three successive time (Ma) trends. From chron 33y to chron 29y, spreadingdirection 55 60 65 70 75 80 85 changedcounterclockwise, from chron 29y to 270 this trend was reversed, and then relative motion changed counterclockwiseagain. Becauseof the relativelysparse data :...,.-,-....,-.:...... : ?.::.::.i•..:::.: ...... weused in computingthefinite rotations, individual changes inspreading directionare not significant withinthe 95% confidenceinterval. However, changes in spreadingdirection at chron29y and 270 are supported by other observations, e.g.,changes infracture zone directions visible on satellite altimetrydata in theTasman Sea. The chron 270 tectonic eventhas also been recorded in fracture zones of thesouthwest Pacific Ocean[Cande et al., 1995]. From chron 31 (67.7 Ma) to chron 29 (64.0 Ma), our model suggeststranspression between the ChesterfieldPlateau and ¾y ¾y ,4y the Marion Plateau(Figure 1), followedby strike slip. This partof ourmodel is well constrainedby magneticanomaly and responsiblefracturezonefor identifications. the formation Theof thetranspression Capricorn Basinmay havesouth been of theMarion Plateau. We propose an"edge-driven" mechanism [Schoutenand Klitgord, 1994] where the counterclockwise rotationof thesoutheastern part of theMarion Plateau is the resultofforces applied toits margins bythe NE movement of theChesterfield Plateau. Thus the southeastern partof the Marion Plateau has been temporarily transformedinto a microplatewhose rotation determinedextension and opening of the CapricornBasin. The axis of the basin is roughly perpendicularto the southernstrike-slip margin of the Marion Plateau. The oldest dated marine sediments in the Capricorn Figure 10. Spreadingrates and directions of spreading Basin are of Late Oligocene age [Davies et al., 1989]. computedalong a flow line correspondingto the major fracture However, they are underlainby over 1000 m of claystone, zone at 40S. Grey areasindicate the 95% uncertaintiesin the likely Early Tertiaryin age,which are underlainby sandstones spreadingrate and direction values. Spreadingrates are based drilledat Aquarius1 in the northernpart of the basin[Davies et on the reversal time scale of Cande and Kent [ 1995]. GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA 12,425 and, as in case of the South Tasman Rise, by the age and (Figure 11) showsnormal faults which bound the main block of affinity of dredgedrocks. the GilbertSeamount. This block is also characterizedby low Lord Howe Rise, the major continentalfragment [Willcox et magneticanomaly values. The unpublishedinterpretation of al., 1980], whose separationfrom Australialed to the opening new multichannelseismic data (collectedby Rig Seismicin of the Tasman Sea, is interpreted as a conglomerateof four late 1996) shows a more detailed structure of the Gilbert microplates(including the Challenger Plateau). The tectonic SeamountComplex and presentsadditional evidence for its lineamentsvisible on the gravity map, interpretedas strike- continental origin. slip faults, suggestthat the northern and middlefragments of Lord Howe Rise acted as separate tectonic blocks. Three 7. Tectonic History seismic profiles display changesin basementcharacter that correspondto the observedgravity anomalies. Seismic data A major tectonic event at 95 Ma that changed the indicate that the boundary between the northern and middle configuration of plates around Australia also affected its Lord Howe Rise is relatively wide, but for the purpose of eastern part [Mailer et al., 1997]. Therefore rifting started modeling the tectonic history between these blocks we between Australia and the Lord Howe Rise after a slow consider it as a continuous lineament. The southern extension spreading system was established in the south between of the Lord Howe Rise, the Challenger Plateau, is separated Australia and East Antarctica. from the middle Lord Howe Rise by the Bellona Trough and a To evaluate our tectonic model for the opening of the small fragment of continentalcrust whoseboundaries represent TasmanSea in a larger framework,we testedtwo positions of strike-slip faults interpretedfrom the gravity map. The two East Antarctica relative to Australia given by finite rotation faults (visible on seismiclines as shown by Bentz [1974]) are poles calculatedby Royer and Rollet [1997] and Tikku and likely to representthe changein strike of the westernmargin Cande [1997] (Figure 12). Royer and Rollet's [1997] tectonic of the ChallengerPlateau from north to northwest. model of relative motion between Australia and Antarctica in The origin of the Dampier Ridge hasbeen much debated,but Cretaceoustime is based on magnetic anomalies identified dredgedrock samples have revealed its continental nature betweenthe two plates,continent ocean boundaries interpreted [McDougallet al., 1994]. Some of the N-NE fracturezones from the satellitealtimetry data in the Great AustralianBight, pass acrossthe Dampier Ridge, dividing it into four main west of Tasmania and east of East Antarctica. The elements. Another continental plateau, the Chesterfield reconstructionfor 95 Ma modeledby Tikku and Cande [1997] Plateau,is locatednorth of the Dampier Ridge and was first is constrainedby two NW-SE fracturezone crossings (Perth and consideredby Shaw [1979] as an active tectonic block in the Vincennes)and the Cretaceousquiet zone magnetic anomalies opening of the Tasman Sea. The boundary between the identifiedin the Australian-Antarcticbasin. The pole for this Chesterfield Plateau and the northern Lord Howe Rise is not reconstructionhas been obtainedby fitting a stagepole to a clearly defined, and we have constrainedit only by gravity fracturezone and closing it until the Cretaceousquiet zone derived from satellite altimetry data. magneticanomalies in the Antarctic-Australianbasin matched. The South Tasman Rise is a continental fragment which For the time of rifting and opening in the south Tasman Sea belonged to Eastern Gondwanaland [Kennet et al., 1975]. (95-83 Ma), Royer and Rollet's [1997] modelpredicts NW-SE Recently, Rollet et al. [1996] suggestedthat the South Tasman motionbetween Australia and East Antarcticawhich changesto Rise is composedof two major continental terraenesseparated NNW-SSE after chron 34. Tikku and Cande's [1997] model by an N-S transformfault. The western fragment was attached implies NNW-SSE motion from 95 Ma until 61.2 Ma (chron to Antarcticaduring the early opening of the SoutheastIndian 27). We presenttwo reconstructions(Figures 12a and 12b) for Ocean, while the easternfragment belonged to the Australian theTasman Sea area for prebreakuptime (90 Ma) includingall plate. The East Tasman Plateau, a circular plateau separated the tectonic blocks recognizedas parts of Gondwanalandthat from Tasmaniaby a deepsaddle floored by oceaniccrust [Shaw, were dispersedduring opening of the TasmanSea. Subsidence 1979], is underlainby continentalbasement rocks [Exon et al, analysis of the SE margin of Australia demonstratesthat the 1996]. This supportsthe Soela Seamountguyot formedas the Bass, Otway and Gippslandbasins began to form during resultof Paleogenehotspot volcanism [Duncan and McDougal, Mesozoicrifting and the mostrapid andwidespread subsidence 1989]. The SouthTasman Rise is separatedfrom Tasmaniaby endedduring Early Cretaceoustime [Hegarty et al., 1988]. a deep basin which seems to be underlain by extended Consequently,we assumethat extension in these basins was continentalcrust. There is evidenceof oceanfloor formedby almost completed at our reconstructiontime (90 Ma) with spreadingbetween the South Tasman Rise and East Tasman Tasmaniain its present-daylocation relative to Australia. We Plateau: for example, magnetic anomalies and seismic data interpretedthe boundarybetween continental and oceaniccrust collected on L'Atalante [Rollet et al., 1996] indicate the (COB) of East Antarcticafrom the gravity grid derivedfrom presenceof a triple junction between East Tasman Plateau, satellite altimetry north of 72øS [Sandwell and Smith, 1997] SouthTasman Rise, and Lord Howe Rise. Royer aa•dRollet and from the gravitygrid derivedfrom ERS-1 satellitealtimetry [1997] propose a tentative interpretationfor four magnetic southof 72øS[McAdoo and Laxon, 1997]. The bathymetric profiles crossing the basin between the East Tasman Plateau contoursagree well with the distinctpositive gravity anomaly andSouth Tasman Rise and dated the end of this triple junction that correspondsto the COB, as observedfrom severalseismic activity at 65.6 Ma (chron30y). lines which crossit [Cande et al., 1998]. Another tectonic Seismicevidence as well as magneticanomalies suggest a featurewe includedin our reconstructionsis the Iselin Bank, a continentalnature of the Gilbert SeamountComplex [Ringis, continental block that is part of West Antarctica. Several 1972]; it separatedfrom the Australian continent by early analyses[Stock and Molnar, 1982, 1987; Candeet al., 1995] rifting. Recent geophysicaldata (R/V MauriceEwing, 1996), of the tectonic framework of Pacific, Antarctic and Australian including single-channel seismic data and magnetic data, plates requirean additional plate boundarybetween East and confirmedRingis's theory. Interpretationof seismicprofiles West Antarcticawhich may have been active since chron 27 12,426 GAINA ET AL.: TECTONIC HISTORY OFTHE TASMAN SEA Gravity anomaly Magnetic anomaly Distance (kin) Figure 11. Gilbert SeamountComplex: gravity, magnetic,and seismicprofiles from Ewing 9513 cruise (profile outlinedin bold on Figure 1). The low valuesof magneticanomalies along with some normal faults observedon the seismic profile strongly suggest a continental nature of the main block for the Gilbert SeamountComplex. (61.2 Ma). Cande et al. [1996] suggestedthat at the time someoverlaps between the tectonicblocks that accountfor the seafloor spreading staffed in the Tasman Sea, the Iselin Bank continental stretching during this time. The second may have beenpart of East Antarctica. Rocksdredged from two reconstruction (Figure 12b), which uses the rotations for sites on Iselin Bank (crest and east flank) show similarities Australiaand Antarcticaderived by Tikku and Cande [1997], with dredgedmaterial from northern Victoria Land. Since the locatesAntarctica in a more eastwardposition, reducingthe dredgedsamples consist of moderatelyrounded grains and there space left for the two South Tasman Rise blocks. Partial are no known processeswhich could have transportedthe closingof embaymentsand reductionof plateausproportional dredgedmaterial eastwardof the Victoria Land Basin, these to their probableextension would lead to a better reassembly. similarities likely reflect correlation in lithologies [Wong et It has been suggested that Wilkes Basin, an Antarctic al., 1987]. In addition, basement rocks underlying Iselin depressionsituated near the coast, is the result of Mesozoic Bank have a two-layer velocity structuresimilar to that of the crustal extension [Steed, 1983] and microfloras are similar to Victoria Land Basin [Cooper et al., 1987]. Multichannel these in the Otway Basin of Australia [Quilty, 1986]. seismic reflection data across the Iselin Bank revealed an Accordingly,Bradshaw [1991] proposeda reassemblyof the asymmetry between the western flank (borderedby numerous southwest Pacific-Antarctic region before breakup which normalfaults) and the easternflank (borderedby a steepnormal reduced the area of the Wilkes Basin and Ross Sea. We fault) [Davey and Cooper, 1987]. mentionedbefore that the Otway Basin likely formedbefore In the reconstructionbased on Royer and Rollet's [1997] breakupin the TasmanSea [Hegartyet al., 1988]; this implies model (Fig. 12a) the West SouthTasman Rise lies west of East no significant extension in this basin and its Antarctic South Tasman Rise and Tasmania, overlapping the extended counterpartsubsequent to 90 Ma. Comparedwith the Royer crust of East Australia and Australia. The reconstructed and Rollct [1997] basedreconstruction (Figure 12a), the Tikku positionof the Iselin Bank fills most of the gap betweenthe and Cande[1997] basedreconstruction (Figure 12b) requires SW Challenger Plateauand the easternmostpart of Antarctic that the easternSouth Tasman Rise was in a more northerly continentaloceanic boundary. This reconstructiondisplays a position, resulting in substantialoverlap among this block gap among the eastern South Tasman Rise, East Tasman and the neighbouringmicroplates, in particularamong the two Plateau,Gilbert SeamountComplex, and Lord Howe Rise and blocks of the South Tasman Rise and the extended crust of GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA 12,427 cob 90 Ma [Royerand Rollet, 1997] cob 90 Ma [Tikku andCande. 1997] Australia Australia 30øS 30øS COB Antarctica Antarctica b 150 ø E 150 øE Figure 12. Reconstructionsfor prebreakuptime (90 Ma) usingtwo differentrotation poles for Australia- Antarctica:(a) rotationpoles derived by Royerand Rollet [1997] and(b) rotationpoles computed by Tikkuand Cande [1997]. Shadedareas indicate overlaps. See text for discussion. The 13 tectonic blocks active during the openingof the TasmanSea are CsP,Chesterfield Plateau; Drl-4, fragmentsof the DampierRidge; LHR_N and LHR_m, northernand middle of the Lord Howe Rise; ChP, the ChallengerPlateau; W, small wedge between the middle Lord Howe Rise and the ChallengerPlateau; Gsm, Gilbert SeamountComplex; WSTR and ESTR, western and eastern blocks of the South Tasman Rise; ETR, East Tasman Plateau; IB, Iselin Bank; COB, the boundarybetween the continentaland oceaniccrust; Tas, Tasmania. Australia and Tasmania. In addition, this model exhibits a (attached to the Australian plate at the time) and the South large overlap (~ 130 km) between Australian and Antarctic TasmanRise and to E-W seafloor spreadingbetween the South continental crust west of Tasmania. Tasman Rise and East Tasman Plateau. These events were Startingwith the fit reconstructiona (rotationsfor Australia contemporaneouswith the onset of seafloor spreading in the and Antarctica derived from model (1)) and incorporating Tasman Sea. We cannotestimate the exact timing for seafloor geologicaland geophysical observations discussed earlier in spreading between the eastern South Tasman Sea and East this paper for times between90 Ma and 83 Ma and finite TasmanPlateau because magnetic data are sparse, commonly rotationsfor youngerchrons (33y-24o), we derivedadditional obscuredby seamounts, and inaccurately navigated. The reconstructionsusing the PLATES software(Table 3) for key eastern South Tasman Rise became attached to Australia periodsin the openingof the TasmanSea (Figure 13). shortly before chron 31y (~ 70 Ma); the west South Tasman The magnetic fabric of the Tasman Sea shows that the Rise became fixed to the east South Tasman Rise at 40 Ma, present continental margin does not represent a single after about 70 km of left- lateral strike-slip between the two prebreakupisochron. Becauseknowledge of the preanomaly tectonic blocks. 330 tectonichistory is sparse,the prebreakupconfiguration is Seafloor spreading in the Tasman Sea started first in the basedon constraintssuch as strike-slipfaults (as in the case of southas a resultof rifting between the Challenger Plateau (the the Lord Howe Rise and the Dampier Ridge) and failed rifts southernpart of the Lord Howe Rise) andthe middleLord Howe (present between Tasmania and the South Tasman Rise, Rise. Short-lived left-lateral transtension between the Tasmania and the East Tasman Plateau, and Gilbert Seamount Challenger Plateau and the middle Lord Howe Rise led to the Complex and the Challenger Plateau) (Figure 1). Prior to formation of the Bellona Trough before the onset of seafloor opening of the Tasman Sea, rifting and a short period of spreadingin the TasmanSea (Figure 13b). The Bellona Trough seafloor spreading occurredbetween Australia and the East appears to have 0.5-1 s two way time (TWT)of Cretaceous Tasman Plateau (Figure 13a). At the same time, the South sediments[Wood, 1993], and its axis correspondsto the NNE Tasman Rise (divided into two blocks by a transform fault) alignment visible on the gravity map (Figure 6) and detected startedmoving slowly southwardrelative to Australia. The on seismic profiles to the north of the Challenger Plateau. southward movement led to extension between Tasmania This evidenceled us to presumethat the openingof the Bellona 12,428 GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA Table 3. Finite Rotations for Different Tectonic Elements of the Tasman Sea Age, Latitude Longitude Angle, Ma +øN +øE deg North Lord Howe Rise Relative to Australia (90-84 Ma) 84.0 3.51 - 43.6 12.27 North Lord Howe Rise Relative to Middle Lord Howe Rise(84-64 Ma) 84.0 8.10 -35.62 -2.72 Middle Lord Howe Rise Relative to Australia* 79.0 0.37 133.82 -13.00 83.0 1.57 133.42 -13.04 84.0 4.50 - 42.26 14.96 86.0 4.06 -42.35 15.51 90.0 3.27 -42.59 16.61 ChallengerPlateau Relative to Middle Lord Howe Rise(90-77 Ma) 86.0 60.83 63.50 -1.00 90.0 72.27 63.13 -1.39 North Dampier RidgeRelative to Australia (90-71.4 Ma) 71.4 11.84 - 44.20 11.20 Middle 1 Dampier RidgeRelative to Australia(90-72.4 Ma) 72.4 13.05 -41.52 12.49 Middle 2 Dampier Ridge Relative to Australia (90-74.2 Ma) 74.2 7.25 - 44.15 11.46 SouthDampier Ridge Relative to Australia( 90-79 Ma) 79.0 5.28 -43.88 11.87 Gilbert SeamotmtComplex Relative to East SouthTasman Rise (90-79 Ma) 79.0 0.52 131.47 -13.10 90.0 2.30 135.09 -13.10 ChesterfieldPlateau Relative to Australia (90-71.6 Ma) 71.6 7.57 112.18 - 5.03 ChesterfieldPlateau Relative to North Lord Howe Rise (71.6-64 Ma) 71.6 22.48 -24.74 -8.83 East TasmanRise Relative to Australia (90-84 Ma) 90.0 60.20 -38.13 -2.70 East South Tasman Rise Relative to Antarctica (90-70 Ma) 70.0 11.50 33.02 -26.35 West South Tasman Rise Relative to Antarctica(90-40.IMa) 40.1 26.19 22.65 -30.63 SmallPlate Between Middle Lord Howe Rise and ChallengerPlateau Relative to ChallengerPlateau (90-86 Ma) 90.0 56.51 64.76 0.59 SmallPlate Between Middle Lord Howe Rise and ChallengerPlateau Relative to middleLord Howe Rise (86-77 Ma) 86.0 65.08 108.98 -0.83 Antarctica Relative to Australia 40.1 14.32 31.75 23.77 53.3 12.65 32.76 25.24 65.6 11.79 32.96 26.05 73.6 11.27 33.08 26.59 83.0 10.67 33.22 27.22 95.0 8.14 33.34 27.83 GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA 12,429 Table 3. (continued) Age, Latitude Longitude Angle, Ma +øN +øE deg Antarctica Relative to Australia 61.2 7.86 35.81 25.54 95.0 5.50 36.58 28.14 * See Table 2 for youngerrotations from chron 240 - 333,. p Rotationscalculated by Royer and Rollet [1997]. -I:Rotations calculated by Tikku and Cande[1997]. Troughpreceded opening of theTasman Sea basin. Following gravity grid are consistent with similar featuresobserved on this event, the ChallengerPlateau, attachedto the middleLord the juxtaposed Chesterfield Plateau gravity grid in its Howe Rise, and Australia separated. We identified anomaly reconstructedposition. Gravity highs present on the 33y very closeto the Gilbert SeamountComplex (Figure 1) and Challenger Plateau appear to follow the direction of other estimated that 250 km of oceanic crust between the Gilbert gravity highs from the middle of the Lord Howe Rise. Since SeamountComplex and the ChallengerPlateau is the resultof the middle Lord Howe Rise and the Challenger Plateau oceanspreading between 83 and 77 Ma. Thereforewe suggest experiencedtranstension during the opening of the Tasman that early rifting commencedat about84 Ma, after short-lived Sea,this reconstructionreveals the prebreakupposition of the strike-slipmotion between the Gilbert SeamountComplex and elevated tectonic features observed on the Lord Howe Rise. the SouthTasman Rise, and stoppedprior to anomaly33y time when the Gilbert SeamountComplex was transferredto the 8. Conclusions ChallengerPlateau (Figure 13b). At 84 Ma, extensionbegan between the subsequentDampier We presenta new, comprehensivemodel for the opening of Ridge and the northernpart of the Lord Howe Rise creating the the Tasman Sea, based on a new gravity anomaly grid, an Lord Howe andMiddleton basins. At that time, the fragments improved magnetic data set and other geophysical data. We which would later constitutethe Dampier Ridge were part of the have combineda qualitativemodel for the early, prechron 3 3 y Australian continent. Before chron 33y (79 Ma) the southern openingof the TasmanSea basedon all available geophysical Dampier Ridge becameattached to the northernLord Howe Rise and geologicaldata with a quantitativemodel for the postchron following a ridge propagationfrom the Lord Howe basin to the 33y opening of the Tasman Sea from inverting magnetic Tasman Sea (we interpreted anomaly 33y very close to the anomaly and fracture zone data to obtain finite rotations and southernmostpart of this tectonic block). As a consequence, their uncertainties. transtension began between the two fragments south of the The early openingof the TasmanSea cannot by modeledb y DampierRidge. This motionceased when the secondblock of a simple two-plate system, because rifting propagated the Dampier Ridge joined the northern Lord Howe Rise and northwardin several stages, creating several failed rifts and strike-slipmotion occurredbetween the two middle elementsof strike-slip faults. We identified 13 tectonic units, based on the Dampier Ridge. At 72 Ma, after short-lived transtension tectonic lineamentsvisible in the gravity grid and interpreted betweenthe northernDampier Ridge and the adjacentfragment as strike-slip faults, by magnetic anomaly, bathymetry, and to the south, the entire Dampier Ridge becameattached to the seismicdata, and, as in caseof the SouthTasman Rise, by the Lord Howe Rise and rifting in the Lord Howe and Middleton age and affinity of dredgedrocks. These 13 tectonicblocks and basins ceased (Figure 13c). The northern Dampier Ridge the Australian continent gradually separated due to either moved along the Australianmargin shortly before the onset of extensionalor strike-slip movements, generating the Tasman seafloorspreading in the northernTasman Sea. The last stage Sea, the Lord Howe and the Middleton basins, and several failed of rift propagation at chron 28 separatedthe Chesterfield rifts. We suggestthat strike-slip motion occurredbetween the Plateauand the Australian plate (Figure 13d) and establisheda northern and the middle Lord Howe Rise along a fault visible continuousspreading center in the TasmanSea. At the same on the gravity grid and seismic reflection data. We also time, the northern Lord Howe Rise became attached to the consideredthe Challenger Plateau as a separate microplate, middle Lord Howe Rise after 260 km of left-lateral strike-slip whose motion relative to the middle Lord Howe Rise constrains motion. the plate circuit amongAntarctica, Australia, and New Zealand. To better evaluate the fit of all tectonic blocks in this area The openingof the northernTasman Sea is well explainedby and correlations between tectonic lineaments on the gravity the gradualstepwise separation of four Dampier Ridge tectonic grid of the 13 tectonicelements that were part of the Australian fragments plus the Chesterfield Plateau and Australia in the continent prior to the opening of the Tasman Sea, we frameworkof a northwardpropagating rift. Transtensionor constructeda gravity grid mosaic for prebreakuptime (Figure strike-slip motion characterizedthe motion of each tectonic 14). The gravity grid polygons for each tectonic block were block of the Dampier Ridge relative to adjacentblocks. converted to ASCII format and rotated about their fit rotation Our postchron 33y model describesa number of tectonic poles relative to Australia, averagedby a block mean filter and eventsin the spreadinghistory of the Tasmansea basin. The regriddedusing a spline interpolationalgorithm with a tension general trend of the youngest fracturezones observedin the value of 0.25 [Wessel and Smith, 1991]. Clear SW-NE oriented satellite gravity grid differs from previously interpreted lineations on the southern Queensland Australian margin fracture zones by about 10ø (NNE versus NE). Hence the 12,430 GAINA ET AL.: TECTONICHISTORY OF THE TASMAN SEA be -• +•'•- t•'•--cOs Chron33(73.6Ma) 30 øS 30 øS 150 øE 150 øE Ce de 30 øS 30 øS 150 øE 150 øE f, 30os 30 *S 150 *E 150 *E Figure 13. Set of reconstructionsof the TasmanSea opening: (a) chron 34y (83 Ma), (b) chron 33y (73.6 Ma), (c) 31y (67.7 Ma), (d) chron28y (62.5 Ma), (e) chron260 (57.9 Ma), (f) chron 24 (52 Ma). Small dots are magneticanomaly and fracturezone identifications. For tectonic blocks identification, see Figure 12 caption. GAINA ET AL.:TECTONIC HISTORY OFTHE TASMAN SEA 12,43! -25 ø -30' -35 ø -45' _50 ø 145' 150' 155 ø 160 ø I I I I I I I I I I I I I I I I I I -890-790-690-590-490-390-290-190-90 10 110 210 310 410 510 610 710 810 Figure 14. Gravitymosaic for prebreakuptime (90 Ma). Black lines representboundaries between continentaloceanic crust. Overlapsamong different tectonic blocks account for synriftcrustal extension. All overlapsbetween tectonic elements and Australia show Australian plate gravity values. Other areas of gapor overlapsbetween different tectonic blocks show averaged or interpolatedgravity values, respectively, and can be ignored(the Australian continental gravity grid used in this reconstructionwas produced by Australian GeologicalSurvey Organization). See Figure 1 for continentalblock identifications. directionof spreadingis inferredto be moreSSW-NNE than Marion Plateau,followed by strike-slip, from chron 31 (67.7 SW-NE, as previouslysuggested, and spreadingdirection Ma) to chron 29 (64.0 Ma). This previously unrecognized changedseveral times most notably counterclockwisejust transpressionis well constrainedby magneticanomalies and beforespreading ceased, possible at chron25. Our model fracture zone crossings and may be responsible for the predictstranspression between the Chesterfield Plateau and the formationof the CapricornBasin southof the Marion Plateau. 12,432 GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA Another major tectonic event took place at chron 27o (61.2 Davies,P.J., P.A. Symonds,D.A. Feary,and C.J. Pigram,The evolution Ma), when spreading direction changed counterclockwise, of the carbonateplatforms of northeastAustralia, in Controlson correlating well with a synchronousevent in the southwest CarbonatePla(form and Basin Development,edited by P.D. Crevello et al., Spec.Publ., Soc.Econ. Paleontol. Mineral., 44, 233- Pacific Ocean[Cande et al., 1995]. This event, a subsequent 258, 1989. counterclockwisechange in spreadingdirection at about chron Duncan,R., and I. McDougall,Volcanic time-space relationships, in 25y (55.9 Ma) accompaniedby a further decreasein spreading Intraplate Volcanismin Eastern Australia and New Zealand, edited rate, lead to the cessationof spreadingsoon after chron 24o by R. Johnson,pp. 43-54, CambridgeUniv. Press,New York, 1989. (52 Ma) and reflect a major changein plate driving forces of Exon, N.F., J.P. Kennett, J. Mascle, P.J. Hill, J.-Y. Royer, G.C.H. Chaproniere,and S. Shafik, The "SouthernGateway" between the Australian plate, which predatesthe beginning of India- Australia and Antarctica:A proposalfor ODP paleoclimatic, Eurasiacollision at chron 24 [Rowley, 1996]. (Magnetic and paleoceanographicand transformmargin drilling, A GSO Rec. 34, fracture zone crossings(in digital format) usedin this studyare Aust. Geol. Surv. Org., Canberra,1996. available on the Worldwide Web site: www.es.su.oz.au.) Fox, P.J.and D.G. Gallo, The geologyof the North Atlantic transform plate boundariesand their aseismicextensions, in The Geologyof the North America, vol M, The WesternNorth Atlantic Region, editedby R. Vogt and B.E. Tucholke,pp. 157-172,Geol. Soc. of Acknowledglnents. The resultspresented here are part of a Ph.D. Am., Boulder, Colo., 1986. project sponsoredby the AustralianGeological Survey Organisation. Hayes, D.E., and J. Ringis,Seafloor spreading in the TasmanSea, The first authoracknowledges support from the Departmentof Geology Nature, 243, 454-458, 1973. and Geophysics,University of Sydney. We thank the Royal Australian Hegarty,K.A., J.K. Weissel,and J.C. Mutter, Subsidencehistory of Navy for makingdigital magneticanomaly data collectedby the HMAS Australia's southernmargin: Constraintson basin models,AAPG Cookavailable to us. We benefitedfrom someinsightful discussions with Bull., 72, 615-633, 1988. and unbiasedviews (at the 95% confidencelevel) from Ted Changand Hellinger,S.J., The uncertaintiesof finiterotations in platetectonics, J. Richard Gordon while they were visiting Sydney University. For Geophys.Res., 86, 9312-9318, 1981. digitizingmagnetic anomaly identifications we useda programdeveloped Irving, E., and G.A. Irving, Apparent polar wander paths by AnahitaTikku fromthe ScrippsInstitution of Oceanography.Most of Carboniferousthrough Cenozoic and the assemblyof Gondwana, the figures were drafted usingthe GMT software [Wessel and Smith, Geophys.Surv., 5, 141-88, 1982. 1991]. This work was partially supportedby National Science Kennett,J.P., et al., InitialReports of theDeep Sea Drilling Project, vol. Foundationgrant OCE-9416779 to J. Stock. JYR acknowledgessupport 29, 1197pp., U.S. Govt.Print. Off., Washington,D.C., 1975. from the Centre National de la RechercheScientifique and from the McAdoo, D.C., and S.W. Laxon, Antarctic tectonics:Constraints from FrenchMinistry of Educationduring his stayin Sydney. Marine gravity a new satellitemaxine gravity field, Science,276, 556-561, 1997. data southof 72øS suppliedby SeymourLaxon, UniversityCollege McDougall,M.A., M.A.H. Maboko,P.A. Symonds,M.T. McCulloch, London, and Dave McAdoo, NOAA. G6osciences Azur contribution I.S. Williams,and H.R, Kudrass,Dampier Ridge, Tasman Sea, as a 174. strandedcontinental fragment, Aust. J. Earth Sci., 41, 395-406, 1994. Mtiller, R.D., D.T. Sandwell, B.E. Tucholke, J.G. Sclater, and P.R. Shaw, Depth to basementand geoid expressionof the Kane References fracture zone: A comparison,Mar. Geophys.Res., 13, 105-129, 1991. Bentz,F.P., Marine geologyof the southernLord Howe Rise,southwest Mtiller, R.D., A. Tikku, D. Mihut, C. Gaina,J. Stock,and S. Cande, The Pacific,in The Geologyof ContinentalMargins, edited by C.A. Burk 95 Ma and 61 Ma (Chron 27) Tectonic events around Australia: and C.L. Drake, pp. 537-547, Springer-Verlag,New York, 1974. The Relationshipbetween absoluteand relative plate motion, Bradshaw,J.D., Cretaceousdispersion of Gondwana:Continental and abstractpresented at the ChapmanConference on the Historyand oceanic spreadingin the south-westPacific-Antarctic sector, in Dynamicsof Global Plate Motions,AGU, Marshall,Calif., 1997. Geological Evolution of Antarctica; Proceedings of the Fifth Quilty,P.G., Australia'sAntarctic basins: Identification and evolution, International Symposiumon Antarctic Earth Sciences, Int.Symp. Pet. Explor. Soc.Aust. J., 8(1), 93-101, 1986. Antarct. Earth Sci., vol. 5, edited by M.R.A. Thomson,J.A Crame, Ringis,J., The structureand history of the Tasman Sea and the and J.W. Thompson,pp. 581-585, CambridgeUniv. Press,New southeastAustralian margin, Ph.D. thesis, Univ. of New South York, 1991. Wales, Sydney, 1972. Cande, S.C., and D.V. Kent, Revisedcalibration of the geomagnetic Rollet,N., Ouverturede l'oc/•an australdans la region du South polarity timescale for the Late Cretaceous and Cenozoic, J. TasmanRise (SE Australie):apport d'un lev• de bathymetrie Geophys.Res., 100, 6093-6095, 1995. multifaisceaux( campagneTasmante 1994)et des donn•es de Cande,S.C., C.A. Raymond,J. Stock,and W.F. Haxby, Geophysicsof gravin•triesatellitaire (GEOSAT GM), Raportdu stagede the Pitmanfracture zone and Pacific Antarcticplate motionduring Magist6re,Univ. Paxis,1994. the Cenozoic, Science, 270, 947-953, 1995. Rollet, N., J.-Y. Royer, N.F. Exon, and P.J. Hill, Le Plateau Cande, S., J. Stock, R.D. Miiller, C. Raymond, and A. Tikku, Early Sud_Tasman(Australie): collage de deuxfragments du Gondwana Tertiary motion between East and West Antarctica, Eos Trans. Oriental?, C. R. Acad.Sci. Oceanogr.Geophys. Mar., Ser.lla, 323, AGU, 77(46), Fall Meet. Suppl.,F647, 1996. 865-872, 1996. Cande, S., J. Stock.,C. Raymond,and R.D. MiJller, New constraintson Rowley, D.B., Age of initiationof collisionbetween India and Asia: A an old plate tectonic puzzle of the southwestPacific, Eos Trans. review of stratigraphicdata, Earth Planet. Sci. Lett., 145, 1-113, AGU, 79, 1998 1996. Collette, B.J., and W.R. Roest, Further investigationsof the Noah Royer,J.-Y., T. and Chang,Evidence for relative motionsbetween the Atlanticbetween 10 ø and 40øN and an analysisof spreadingfrom Indianand Australianplates during the last 20 m.y. from plate 118 Ma ago to Present,Proc. K. Ned. Akad. Wet.Set. B, 95(2), 159- tectonicreconstructions: Implications for the deformationof the 206, 1992. Indo-Australianplate, J. Geophys.Res., 96, 11,779-11,802,1991. Cooper, A.K., F.J. Davey, and G.R. Cochrane, Structure of Royer,J-Y., and N. Rollet, Plate tectonicsetting of the Tasmanian extensionallyrifted crust beneaththe western RossSea and Iselin region,Aust. J. Earth Sci., 44, 543-560, 1997. Bank, Antarctica, from suonobuoyseismic data, in The Antarctic Royer,J.-Y., R.G. Gordon,C. DeMets,and P.R.Vogt, New limitson ContinentalMargin, Geology and Geophysicsof the WesternRoss the motionbetween India andAustralia since chron 5 (11 Ma) and Sea, Earth Sci. Set., vol. 5B, pp. 93-119, Circum-Pac. Counc. for implicationsfor lithosphericdeformation in the equatorialIndian Energy and Miner. Resour.,Houston, Tex., 1987. Ocean,J. Geophys.Int., 129, 41-74, 1997. Davey, F.J., and A.K. Cooper, Gravity studiesof the Victoria Land Sandwell,D.T., andG. Shubert,Lithospheric flexure at fracturezones, Basin and Iselin Bank, in The Antarctic Continental Margin, J.Geophys. Res., 87, 4657-4667, 1982. Geology and Geophysicsof the WesternRoss Sea, Earth Sci. Ser, Sandwell,D.T., and W.H.F. Smith.,Global maxine gravity anomaly vol. 5B, pp. 119-139, Circum-Pac.Counc. for Energy and Miner. ERS-I, Geosatand Seasatreveals new tectonicfabric, Eos Trans. Resour., Houston, Tex., 1987. AGU. 73(43)Fall Meet. Suppl.,133, 1992. GAINA ET AL.: TECTONIC HISTORY OF THE TASMAN SEA 12,433 Sandwell,D.T., and W.H.F. Smith,Marine gravity anomalyfrom ERS- Tasman Sea-preliminary geophysical results and petroleum 1, Geosat and satellite altimetry, J. Geophys.Res., 102, 10039- prospects,BMR J. Aust.Geol. Geophys.,5, 225-236, 1980. 10045, 1997. Wong, F.L., P.J. Barett, J. Gamble, and D.G. Howell, Petrographyof Schouten,H., and K.D. Klitgord,Mechanistic solutions to the opening rock samples from Iselin Bank, Ross Sea, Antarctica, in The of the Gulf of Mexico, Geology,22, 507-510, 1994. Antarctic ContinentalMargin, Geology and Geophysicsof the Shaw, R.D., On the evolution of the Tasman Sea and adjacent WesternRoss Sea, Earth Sci. Ser., vol. 5B, pp. 231-253, Circu•n- continentalmargins, Ph.D. thesis,Univ. of Sydney,Sydney, N.S.W., Pac. Counc.for Energy and Miner. Res., Houston,Tex., 1987. Australia, 1979. Wood, R.A., The ChallengerPlateau, South Pacific SedimentaryBasins Steed,R.H.N., Structuralinterpretation of Wilkes land, Antarctica,in of the World, vol. 2, pp. 351-364, F. Ballance,Elsevier, New York, AntarcticEarth Science,edited by R.L. Oliver, P.R. James,and J.B. 1993. Jago,pp. 567-572,Aust. Acad. of Sci.,Canberra, 1983. Stock,J., and P. Molnar, Uncertaintiesin the relative positionsof the C. Gaina and R.D. Mtiller, Departmentof Geologyand Geophysics, Australia,Antarctica, Lord Howe, and Pacific plates since the late EdgeworthDavid Building,F05, The Universityof Sydney,New South Cretaceous,J. Geophys.Res., 87, 4697-4714, 1982. Wales 2006, Australia. (e-mail: [email protected]; Stock,J., and P. Molnar, Revisedhistory of early Tertiary plate motion dietmar @es .s u. o z.au) in the south-westPacific, Nature, 325, 495-499, 1987. J.-Y. Royer, Unit6 Mixte de Recherche 6526, G6osciences Azur, Tikku, A. and S. Cande, Reconstruction of Australia and East Quai de la Darse-BP 48, Villefranche Sur Mer, France. ( e- Antarcticafrom total fit to the Early Tertiary, abstractpresented at mail:[email protected]) the ChapmanConference on the History and Dynamicsof Global J. Stock and J. Hardebeck, SeismologicalLaboratory, MS 252-21, Plate Motions, AGU, Marshall, Calif., 1997. California Institute of Technology, Pasadena, CA 91125. (e-mail: Van Andel, T.H., Fracture zones, Comments on earth sciences, [email protected], jhardebeck @gps.caltech.edu) Geophysics,1, 159-166, 1971. P. Symonds, Australian Geological Survey Organisation, Cnr Wessel, P., and W.H.F. Smith,Free software helps map and display ConstitutionAve. & Anzac Pde, Canberra A.C.T., Australia. ( e- data, Eos Trans. AGU, 72, 441,445-446, 1991. mail:psymonds @ agso.go v.au.) Weissel, J.K., and D.E. Hayes, Evolution of the Tasman Sea reappraised,Earth Planet.Sci. Lett.,36, 77-84, 1977. (ReceivedAugust 20, 1997; revisedDecember 23, 1997; Willcox, J.B., P.A. Symonds,K. Hinz, andD. Bennett,Lord Howe Rise, acceptedJanuary 27, 1998)