Hawkins, J., Parson, L., Allan, J., et al., 1994Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 135

47. PALEOMAGNETIC CONSTRAINTS ON TONGA ARC TECTONIC ROTATIONFROM SEDIMENTS DRILLED AT SITES 840 AND 841l

William W. Sager,2 Christopher J. MacLeod,3 and Niels Abrahamsen4

ABSTRACT

The geometry of the Tonga Arc implies that it has rotated approximately 17° clockwise away from the Lau Ridge as the LauBasin formed in between. Questions have arisen about the timing of the opening, whether the arc behaved rigidly, and whetherthe opening occurred instead from motion of the Lau Ridge, the remanent arc. We undertook to address these questions by takingpaleomagnetic samples from sediment cores drilled on the Tonga Arc at Sites 840 and 841, orienting the samples in azimuth, andcomparing the paleodeclinations to expected directions. Advanced hydraulic piston corer (APC) cores from Holes 840C and 841Awere oriented during drilling with a tool based on a magnetic compass and attached to the core barrel. Samples from Hole 84IBwere drilled with a rotary core barrel (RCB) and therefore are azimuthally unoriented. They were oriented by identifying faultsand dipping beds in the core and aligning them with the same features in the Formation MicroScanner (FMS) wireline logs, whichwere themselves oriented with a three-axis magnetometer in the FMS tool. The best results came from the APC cores, whichyielded a mean pole at -69.0°S, 112.2°E for an age of 4 Ma. This pole implies a declination anomaly of 20.8° ± 12.6° (95%confidence limit), which appears to have occurred by tectonic rotation of the Tonga Arc. This value is almost exactly that expectedfrom the geometry of the arc and implies that it did indeed rotate clockwise as a rigid body. The large uncertainty in azimuthresults from core orientation errors, which have an average standard deviation of 18.6°. The youngest cores used to calculate theAPC pole contain sediments deposited during Subchron 2A (2.48-3.40 Ma), and their declinations are indistinguishable from theothers. This observation suggests that most of the rotation occurred after their deposition; this conclusion must be treated withcaution, however, because of the large azimuthal orientation errors. Poles from late and early Miocene sediments of Hole 84IBare more difficult to interpret. Samples from this hole are mostly normal in polarity, fail a reversal test, and yield poles that suggestthat the normal-polarity directions may be a recent overprint. Late Miocene reversed-polarity samples may be unaffected by thisoverprint; if so, they imply a declination anomaly of51.1°±11.5°. This observation may indicate that, for older sediments, Tongaforearc rotations are larger than expected.

INTRODUCTION

The Lau Basin is an active backarc basin located in the southwest-ern Pacific between the Tonga Island Arc and the Lau Ridge (Fig. 1).Geophysical data (shallow bathymetry, thin sediment cover, high acous-tic backscatter, high heat flow, seismicity, and young magnetic anom-aly lineations) and geologic samples (fresh MORB-like basalts) attestto the recent occurrence of a process akin to seafloor spreading in thebasin (Karig, 1970; Sclater et al., 1972; Hawkins, 1974; Hamburgerand Isaacs, 1988; Parson et al., 1990). Apparently, during Miocene timethe Tonga and Lau ridges were combined as a single island arc, butsubsequently the Lau Basin opened between the two.

Although the Lau Basin and environs have been extensivelystudied, a number of questions about its tectonic evolution remainunanswered. When did the opening begin? Did the Tonga Arc rotateclockwise away from a relatively fixed Lau Ridge, as suggested bygeometry? Can the opening be described by a rigid rotation?

The timing of the opening is uncertain. Magnetic lineations havebeen identified in the basin, but because they are unclear, these anoma-lies have been interpreted variously. Estimates for the time of openingrange from about 2.5 to 5.9 Ma (Weissel, 1977; Malahoff, Feden, andFleming, 1982; Kroenke, 1984; Malahoff et al., in press). Moreover,initial drilling results from Leg 135 indicate that the opening of thebasin may have occurred in two stages: a period of slow extension,from late Miocene to about 2.5 Ma, followed by seafloor spreadingand rapid opening (Parson, Hawkins, Allan, et al., 1992).

Hawkins, J., Parson, L., Allan, J., et al., 1994. Proc. ODP, Sci. Results, 135: CollegeStation, TX (Ocean Drilling Program).

2 Departments of Oceanography, Geophysics, and Geodynamics Research Institute,Texas A&M University, College Station, TX 77843, U.S.A.

3 Institute of Oceanographic Sciences, Deacon Laboratory, Wormley, Surrey, GU85UB, United Kingdom.

4 Geophysical Laboratory, Aarhus University, Finlandsgade 8, DK-8200, Aarhus N,Denmark.

The Lau and Tonga ridges form a tilted "V", pointed southward,with the Lau Basin in the middle (Fig. 1). The western side of theTonga Arc is relatively linear and has an azimuth of N21°E; the eastside of the Lau Ridge is less so, but the southern part has an azimuthof about N4°E. This configuration implies that the basin has notopened uniformly. On the one hand, the Tonga and Lau ridges mayhave rotated relative to one another by about 17° around an Euler polelocated near the j unction of the two trends (Fig. l)at28.5°S, 178.7°W.Most authors have favored a rigid rotation for the opening of the basin(Weissel, 1977; Malahoff, Feden, and Fleming, 1982; Auzende et al.,1988). An alternative explanation is that seafloor spreading in the basincould have started in the north and progressed southward, "unzipping"the two ridges. This would imply that the Tonga Arc and Lau Ridgedid not separate by simple rigid-body rotation.

To investigate the tectonic rotation of the Tonga Arc, we studied thePaleomagnetism of sediments drilled during Ocean Drilling Program(ODP) Leg 135 on the Tonga Arc at Sites 840 and 841. The idea wasto calculate paleomagnetic poles for the Tonga Arc and to comparethem with poles from the Indo-Australian Plate, to which the Lau Ridgehas been attached since Oligocene time when spreading ceased in theSouth Fiji Basin (Watts et al., 1977). Because the pole of rotation forthe opening of the Lau Basin is nearby, no significant differences inpaleomagnetic inclinations are expected. Instead, the rotation shouldappear as a difference in paleomagnetic declinations. Paleodeclinationdata are difficult to extract from ODP cores because most have anarbitrary azimuthal orientation. Attempts to orient samples from ODPcores fall into three categories: (1) APC cores oriented with a compasstool, (2) samples reoriented by assuming that a normal-polarity over-print on the cores was acquired during the Brunhes Chron and pointsto north, and (3) orientation of samples relative to azimuthally orientedFMS wireline logs. On Leg 135, oriented APC cores were drilled fromthe Tonga Arc at Sites 840 and 841. Furthermore, FMS logs wereobtained at both sites. Consequently, we used Methods 1 and 3 todetermine Tonga Ridge paleomagnetic poles.

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

15°S

17C

19C

21'

23C

25C

^Samoa?

179°W 177° 175° 173°

Figure 1. Map showing location of Leg 135 sites and generalized bathymetryof Tonga Arc, Lau Basin, and Lau Ridge. Heavy dashed lines show trends ofwestern Tonga Ridge and eastern Lau Ridge; the difference in azimuths is 17°.Islands include Tongatapu (T), 'Eua (E), Vava'u (V), 'Ata (A), and Niuafo'ou(NF). CSLC = Central Lau Spreading Center, ELSC = Eastern Lau SpreadingCenter, VF = Valu Fa Ridge, and MTJ = Mangatolu Triple Junction. Contourinterval in kilometers. (Modified from Parson, Hawkins, Allan, et al., 1992).

Previous Paleomagnetic Studies

Few paleomagnetic studies have been done in the islands of thesouthwestern Pacific and most of these have focused on the islandarcs of the New Hebrides (Falvey, 1978) and Solomons (Musgrave,1986; 1990). The Fijian island of Viti Levu has also been studied(Tarling, 1967; James and Falvey, 1978; Malahoff, Hammond, et al.,1982), and a counterclockwise rotation of 21 ° since 4 Ma and as muchas 70° for late Miocene time has been documented. This rotation isthought to be localized in the region of Fiji and to be related tospreading in the North Fiji Basin (Malahoff, Feden, and Fleming,1982; Auzende et al., 1988). It is not known how far to the south,down the Lau Ridge, that the Fiji islands rotation extends, but thechange in the trend of the ridge at 21°S may be related.

Leg 135 Tonga Arc Sites

Drilling was undertaken at two sites (840 and 841) on the TongaRidge during Leg 135. Site 840 is located at a water depth of 754 mon the summit of the shallow carbonate platform of the Tonga Ridge,45 km east-northeast of Ata Island (Fig. 1). After an abortive attemptat piston coring, Hole 840B was rotary cored to a depth of 597.3 mbsf,recovering sediments as old as late Miocene. Although the boreholewas imaged with the FMS log, few features could be reliably oriented.

Faults were not common and the dips of sedimentary strata weretypically only 2°-3° (Parson, Hawkins, Allan, et al., 1992), whereasdips in excess of 5° are necessary for reliable orientation (MacLeodet al., 1992; MacLeod, this volume).

Hole 840C was drilled with the APC in depth intervals where softsediments were encountered in drilling Hole 840B. In all, 13 APC coreswere drilled, 11 of them oriented with the Eastman-Whipstock multi-shot tool. Recovered sediments typically consisted of ashy nannofossilooze layers within massive beds of unconsolidated vitric ash (Parson,Hawkins, Allan, et al., 1992). Of 74.8 m of core recovered from Hole840C, only 10.7 m consisted of nannofossil ooze. Moreover, onlyCores 135-840C-3H through -7H contained more than 0.5 m of nanno-fossil ooze. Dating of these sediments was mainly by microfossils;however, shipboard paleomagnetic measurements showed that mostHole 840C samples are normally polarized and this was interpreted asan indication that they were deposited during the Gauss polarity epoch(Subchron 2A) during the late Pliocene (Parson, Hawkins, Allan, et al.,1992). Biostratigraphic markers indicate a sedimentation rate of ap-proximately 39 m/m.y. (Parson, Hawkins, Allan, et al., 1992).

Site 841 is located at a water depth of 4810 m on the Tonga forearc,140 km south of Site 840 (Fig. 1). Hole 841A used a combination ofthe APC and extended core barrel (XCB) to drill to a depth of 186.6mbsf, bottoming in late Miocene vitric siltstone and sandstone. Theupper part of the hole contained barren clays with thin ash interbeds.From this section, four oriented APC cores were obtained (Parson,Hawkins, Allan, et al., 1992). Dating of these cores was primarily bymagnetic stratigraphy because no diagnostic microfossils were found.Although the magnetic stratigraphy was hampered by a pervasivedrill-string overprint that was impossible to remove completely dur-ing shipboard analyses, paleomagnetic results suggest the sedimentsin the oriented cores were deposited during early and late Pliocene atrates of about 8 m/m.y. (Parson, Hawkins, Allan, et al., 1992).

Hole 84IB penetrated to a depth of 834.2 mbsf using the RCB.The sediments recovered from this hole consisted mainly of vitricsand- stone and siltstone, but basalt, andesite, rhyolite, and carbonateswere also encountered. At the bottom of the hole, Eocene carbonateswere cored overlying rhyolites of unknown age (Parson, Hawkins,Allan, et al., 1992). Tectonism was clearly a major component of thegeologic history of sediments from Hole 84IB. The hole passedthrough two significant normal fault zones and several minor faults(Parson, Hawkins, Allan, et al., 1992). One of the large faults, ata depthof 449 mbsf, separates late and early Miocene sediments. The other, ata depth of 606 mbsf, separates the Eocene calcareous volcanic sand-stone from the underlying rhyolite. Microfaults are common in Hole84IB sediments and tilting of the strata increases downhole to anglesin excess of 45° (MacLeod, this volume). Hydrothermal alteration wasnoted in many cores and appears to have been most pervasive belowthe major fault at 449 mbsf (Parson, Hawkins, Allan, et al., 1992).Dating for the sediments of Hole 84IB was principally accomplishedby biostratigraphy aided with magnetic stratigraphy. Sedimentationrates of 35-160 m/m.y. were found in the sections sampled for paleo-magnetic study (Parson, Hawkins, Allan, et al., 1992).

METHODS

Samples were taken from the split faces of cores from Holes 840C,841 A, and 84IB. Soft sediments in APC cores from Holes 840C and841A were cut with a spatula and placed into 7-cm3 plastic boxes.More indurated sediments in the RCB cores from Hole 84IB weresampled either by cutting 10-cm3 cubes with a two-bladed rock sawor drilling 2.5-cm diameter cylindrical minicores of about the samevolume. All samples were partially demagnetized to isolate the char-acteristic remanent magnetization (ChRM) by removing spuriousoverprint magnetizations. Isothermal remanent magnetization (IRM)acquisition studies indicated that the magnetic grains within thesesediments have low coercivities (Parson, Hawkins, Allan, et al., 1992),so alternating-field (AF) demagnetization was used. Approximately

764

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

50% of the samples were treated to detailed AF demagnetizationroutines, with 10-12 steps up to 100-mT field strength. The remainderof the samples were treated to an abbreviated AF demagnetizationseries, typically with 7 steps up to 60 mT.

The ChRM directions were found from an orthogonal vector plotof the AF demagnetization data for each sample. Straight-line sectionstoward the origin indicate univectorial decay of the magnetizationvector and were assumed to represent the ChRM. Three or more(typically 4-5) directions from the straight line section were averagedusing Fisher statistics (Fisher, 1953) to obtain the mean ChRMdirection and the circular standard deviation (CSD) of the directions,the latter to be used as a measure of the repeatability and reliabilityof the ChRM direction.

Despite efforts to identify and discard unreliable sample data, inmost analyses a few percent of the samples gave virtual geomagneticpoles (VGP) that did not fall near or within the main cluster of VGPs.Excising such data is a common problem for paleomagnetic studies,but there is no universally accepted analytical technique for doing so.We used Θ 9 5 , the radius of the circle around the Fisher mean, withinwhich 95% of the sample population should lie (McFadden, 1980) asan outlier test. Samples outside this circle were considered outliers.According to McFadden (1980), the use of θ 9 5 as an outlier criterionis not strictly correct. Nevertheless, it provides an objective measurefor rejecting poor data and typically those data discarded are thosefew that would have been chosen "by eye".

The analysis of APC and RCB core samples naturally fell into twodifferent divisions. This was mainly a result of the different methodsused in orienting the two types of samples, but also because the APCcores contain younger sediment than the RCB cores, so there was alsoan obvious division by age.

APC Core Orientation

The APC cores were oriented with the Eastman-Whipstock multi-shot orientation tool. This device is a miniature camera that photo-graphs a compass and pendulum-bob inclinometer. Through a slot-keyarrangement on its pressure case, the multishot tool is attached andaligned with the sinker bar-core barrel assembly that contains the coreliner. The core liner is positioned so that the double line on the workinghalf is aligned with the fiducial mark on the compass and a nonmag-netic drill collar is placed in the drill string bottom assembly at the levelof the tool so that the magnetic field of the drill string does notsignificantly affect the compass reading when the core barrel is inplace. Data from the multishot tool give the azimuth of the double linerelative to magnetic north as well as the off-vertical tilt of the core andthe tilt direction. Sample ChRM declinations were first corrected forthe azimuth angle and the present magnetic declination at the site. TheChRM vectors were subsequently rotated about a horizontal axis per-pendicular to the core tilt direction to correct for the off-vertical angle.The core tilts were small, ranging from 1° to 2° (Leg 135 APC coreorientation data are given in Parson, Hawkins, Allan, et al., 1992).

Nonhorizontal strata were cored in Hole 841 A, so bedding correc-tions were made for samples from this hole. Two different types ofbedding correction were tried. Both used the dips and dip directionsof strata in the APC cores reoriented using the multishot orientationdata (Appendix A). One method reoriented samples to the mean dipand dip direction calculated from all bed forms in that core. The othermade the assumption that samples from a given sediment layer werelaid down horizontally upon the stratal boundary at its base, so sam-ples were corrected for the dip and dip direction of the nearest bedform below.

APC Core Paleomagnetic Pole Calculations

Because only one set of orientation data was taken per core, meanpaleomagnetic directions were calculated using a two-tier approach.First, a mean direction was determined for each core using Fisher

statistics (Fisher, 1953). At this first step, we used the 0 9 5 value toreject outlier data. Typically this method only rejected a few percentof the samples. For the second tier of calculations, that of determininga mean pole, we grouped all of the oriented APC cores from Holes840C and 841A because these cores are few in number and of aboutthe same age.

Fisher statistics were not used to determine the mean pole locationand its error limits because the underlying assumption of this methodis that errors in direction are isotropic. Instead, the scatter in absolutedeclination values was larger than that of inclination because the errorfrom core orientation is mainly in azimuth. Consequently, we used aleast-squares method (Gordon and Cox, 1980) that can calculate amean pole using declination and inclination values independently.This method finds the pole location that minimizes the distancebetween the average pole location and that predicted by each datum.It also can determine an elliptical 95% confidence region around thatpole by linear propagation of errors from the estimated precision ofinclination and declination (Gordon and Cox, 1980). Another favor-able feature of this method is that the weight given to each datum isinversely proportional to its error; so that the more precisely deter-mined data place more constraint on the pole location. Finally, fromthe misfits of the data to the model pole, the method can calculate aX2 statistic that can be compared to 95% confidence limits in standardstatistical tables. If the statistic falls outside the confidence limits, thenthe estimated data errors are inconsistent with the distribution of data(Gordon and Cox, 1980).

In calculating the pole location and its confidence ellipse, wewished to account for the systematic error inherent in the APC orien-tation process, but there exist no reliable estimates of this error. Weassumed that orientation error only affected declination values. This isnot strictly true, but should be a good approximation because theoff-vertical tilt of these cores is small. To calculate the mean variancecaused by orientation error, So

2, we assumed that the observed Univari-ate variance in mean-core declinations, Srd

2, has two parts:

2 _d

where Sb

2 is the between-core declination variance in the absence oforientation errors. We estimated Sb

2 by the variance of core-mean incli-nations (assuming within core errors to be isotropic), calculated Std

2

from the observed variance of core-mean declinations, and solved forS2. We wanted the estimated errors of each core mean direction to re-flect the precision of the core mean, so we used α6 3, the two-dimensionalangular standard deviation from Fisher (1953) statistics, to calculateUnivariate standard deviations in paleomagnetic inclination (5, ) anddeclination (Sd) for each core (Table 1). To arrive at Univariate equiva-lents, we used the relations derived by Demarest (1983):

and

Sd=f sin 1[sin(α6 3 ) / s i n(δ)],

where d is the paleomagnetic colatitude and / is a factor between0.70 and 0.80 that depends on the number of independent samples(Demarest, 1983). Finally, the standard deviation of a core-meandeclination was calculated by adding the within-core errors and theaverage orientation error:

ç _ /-ç 2 , ç 2Λ1/2òdc - Kòd + òo ) •

RCB Core Sample Orientation

Pieces from RCB cores are typically unoriented azimuthally be-cause the core barrel tends to spin as it advances. In some holes,however, dipping sedimentary bed forms can be imaged in the bore-

765

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

Table 1. APC core-mean paleopole data.

Corenumber

135-840C-2H3H4H5H-15H-26H7H10H

135-841A-4H5H6H8H

Paleopole

Latitude(°N)

-39.1-57.9-63.7-10.7-83.9-61.6-53.6-85.3

-81.2-50.7-62.6-73.2

Longitude(°E)

106.0128.8102.7121.0*168.4254.4*119.0294.0

95.6103.8129.8118.2

135-841A- (with dip corrections)4H5H6HXH

-79.6-60.9-80.0-70.1

149.298.5

168.3111.2

N

31111638

123

191479

191479

20.410.18.24.X

15.66.58.3

13.4

6.15.3

28.324.1

6.94.7

26.122.0

k

37.521.531.7

199.863.573.228.283.7

30.958.3

5.55.5

24.372.2

6.36.4

Dec

55.433.228.689.7

1.9329.2

39.6355.3

9.643.028.017.7

7.131.6

2.421.6

21.920.419.218.921.119.119.620.0

18.918.929.724.8

19.318.826.923.4

Inc

-39.8-56.2-A1.0-47.5^16.9-W.I-51.7-36.8

-40.7-41.1-56.0-48.0

-51.0-39.1-52.4-46.2

s i

8.94.63.72.26.73.03.85.9

2.72.4

12.710.8

3.22.1

11.810.0

Bedcorr.

NNNNNNNN

NNNN

AAII

Notes: Paleopole = mean paleomagnetic pole calculated with Fisher statistics; N = number of samples used for calculation;(x95 = radius of 95% confidence circle around mean pole; k = precision parameter; "Dec" and "Inc" = mean declinationand inclination (calculated from pole); Sdc, Sf = Univariate standard deviation of declination and inclination (see text);Bed corr. (bedding correction) abbreviations as follows: N = none, A = average bedding dip and dip direction in core,and I = individual layer dip and dip direction. Values labeled with an asterisk (*) indicate pole not used to calculatemean direction.

hole wall by the FMS wireline log, and these bed forms can berecognized in the recovered core. Because the FMS contains a three-axis magnetometer, features in its images can be oriented relative tothe magnetic field within a borehole walled with weakly magneticrock (e.g., most sediments). If the same features can be recognizedwithin the core, they can be reoriented into geographic coordinates.

The methodology of reorienting samples with reference to FMSlogs is discussed elsewhere (MacLeod et al., 1992; MacLeod, thisvolume); here we only briefly outline the steps involved. Featuresamenable to orientation should be tilted and planar, so that a well-defined dip angle and dip direction can be determined from the FMSlog and core. For Leg 135, the most suitable features were tilted bedforms and faults, found principally in Hole 84IB.

We examined the Hole 84IB FMS logs for faults and bed forms.Tilted planar features appear as sinusoids on FMS borehole images,so we used a computer workstation with proprietary Schlumbergersoftware that determines the best-fitting sinusoid to an FMS logfeature and then calculates the implied dip and dip direction.

Dips of tilted planar features within the cores were measured ontwo perpendicular axes, usually along the split face of the workinghalf and on a cut made perpendicular to the split face. These directionsare along the y and x axes, respectively, of ODP-standard corecoordinates (with the z-axis positive downward). With the two meas-ured angles, α and ß, we used the cross product of the two perpen-dicular dip vectors to calculate the dip and dip direction of the planein core coordinates:

dip = cos"1[cos(α) cos(ß)],

dir = tan-![tan(ß) / tan(α)].

Here, α is the dip angle along the x axis (toward the double lineon the working half), positive downcore, and ß is the dip angle alongthe positive y axis, positive downcore and to the right side of the core(see MacLeod, this volume).

The challenge was to then match features seen in the cores on a1:1 basis with the same features in the FMS logs (or vice versa). Thereis uncertainty in the relation of FMS log depths and core depths owingto the arbitrary placement of pieces within a core when recovery is

incomplete and because of inaccuracies in determining the physicallocation of the drilled interval relative to the wireline logging tool. Asa consequence, given a feature in the core we had to search for a cor-responding feature in the FMS log within a "window" of probablelocation. In many locations it was possible to make a convincingmatch. Of 120 samples we oriented, 93 (78%) were matched withFMS log features; 31 of these were faults, 51 were bed forms. Inaddition, 4 samples were oriented by inference from adjacent orientedsamples (e.g., a sample between two bed forms matched to the FMSlog with similar dip directions is given the same dip direction).

All of the remaining 33 oriented core features were bed forms fromsections of the borehole in which the bedding dip and dip directionwere relatively constant (Fig. 2). For these samples, we used anaverage of the dip directions derived from values determined with theSchlumberger workstation. The mean dip directions were variable indirection and consistency, with 95% confidence cones of the meandirections in Hole 841B ranging from 3.5° to 35.6° (Appendix A).

RCB Core Sample Statistics

Unlike the samples from the APC cores, each of those orientedusing the FMS log is independently oriented, so we did not calculatecore-mean directions for the RCB samples. Instead, all of the incli-nations and declinations for samples of a given group were used todetermine a mean paleomagnetic pole using the Gordon and Cox(1980) least-squares method. The same standard error values for dec-lination and inclination were used for each sample, and these weredetermined by calculating the Univariate standard deviations of theseparameters for the group after discarding outlier directions. Outlierswere identified by the θ 9 5 test, as for the APC core samples. Onceagain, this method rejected only a small number of directions, typically<5%. The x2 statistic values for all poles fell within 95% confidencelimits (Table 2), implying that the estimated errors are reasonable.

We tried analyzing the directions of the RCB core samples inseveral different groups. There was a natural division by age into lateMiocene (Cores 135-841B-2R to -3IR) and early Miocene (Cores135-841B-32R to -4IR) with a few older samples (from Cores 135-841B-42R to -44 R). We also compared the directions of normal andreversed samples for a "reversal test". Furthermore, to make sure that

766

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

Dip0

500

510

(mbs

f)D

epth

540

5 5 0 -

0

. . . .

V

i

A

\

i

3C

"JO

Vs

Q

S

i

o-

...

60°

Λ,

s

... ...• •

90° δz

• •

p

36R

37R

38R

39R

40R

41R

Figure 2. Example of dip angle and dip-direction values from FMS dipmeterlog, Hole 841B. Graph plots dip magnitude on horizontal axis vs. depth onvertical axis. "Tails" on data points show dip direction: north (0°) is up andsouth (180°) is down. Solid symbols show more reliable dip values than opensymbols. This section of Hole 84IB has relatively constant bed dips and dipdirections. Data are from the automatic-dipmeter log, which is derived fromFMS microresistivity images and averages over about 0.5 m in depth.

no bias was being introduced by the method of orientation and bed-dipcorrections, we compared sample virtual geomagnetic pole (VGP)clusters with the samples reoriented in four different ways. First, thesamples were rotated in azimuth only, with no correction for bed dip.The second correction method was to correct the sample azimuths andthen make individual corrections for bed dip using the dip of the stratawithin the sample. The third method was to use the average stratal dipand dip direction, determined from the FMS log, for the reorientationand bed-dip correction (Appendix A). For the final correction method,the samples were reoriented to the average FMS bed-dip direction,but the individual sample stratal dips were used for making the bed-dip correction.

RESULTS

APC Core Magnetic Properties

A total of 127 samples from oriented APC cores were measured;111 samples from 11 cores gave good results: 58 samples from Hole840C and 53 samples from Hole 841A (Appendix B). The number of

samples per core ranged from 3 to 19, the lower numbers a result oflow recovery. All of the APC core samples were of nannofossil oozewith varying amounts of volcanic ash (Parson, Hawkins, Allan, et al.,1992). Because of the volcaniclastic component, sample natural rema-nent magnetizations (NRMs) were high, ranging from 15.45 to 894.71mAm~1 (Appendix B). These samples typically displayed pervasiveupward-directed overprints thought to have resulted from exposureto magnetic fields within or around the drill string. This overprint wasusually removed easily by AF demagnetization of 15- to 20-mT fieldstrength (Fig. 3). Most samples had mean destructive fields (MDFs)of only about 2 to 4 mT, indicating a low coercivity component in themagnetization and explaining the propensity of the samples to acquirean overprint, but some samples had somewhat higher MDFs, in therange of 10 to 15 mT. Most samples also displayed stable behaviorduring AF demagnetization (Fig. 3) and yielded ChRMs with CSDvalues of a few degrees or less (Appendix B).

Only about 15% of the samples displayed unreliable behavior.These unreliable samples were not uniformly distributed; often theywere found preferentially in the deeper APC cores. In Hole 840C,Cores 135-840C-9H, -10H, and -12H were particularly affected. InHole 841 A, Cores 135-841A-4H and -5H yielded excellent results,whereas Cores 135-841A-6H and -8H gave poor results. This trendmay be a result of some process, such as anoxic dissolution (e.g.,Karlin and Levi, 1985), that destroyed the magnetization with pro-gressive burial.

In initial shipboard studies, a trend of declination values rotatingdowncore by 2.7° to 5.5° per meter was noted in whole-core measure-ments made of Hole 841A APC cores (Parson, Hawkins, Allan, etal., 1992). Similar trends have been noted by other paleomagnetistsworking on ODP APC cores and had been attributed by some torotation of the core within the liner. Discrete samples from the Hole841A cores did not show the same trends in properly isolated ChRMdirections (Fig. 4). Because shipboard whole-core paleomagneticmeasurements were restricted to AF demagnetization levels of 15 mTor less, this result suggests that the apparent rotation was caused byan incompletely removed overprint magnetization.

Because of the large orientation errors, it was impossible toperform a reversal test by comparing normal and reversed polaritysample directions between cores. This necessitated comparing thedirections from samples within a single core. Only one core, 135-841A-4H, had a sufficient number of normal and reversed samplesfor a reliable comparison. The normal and reversed samples from thiscore gave mean directions with overlapping 95% confidence circlesthat show the two means are not significantly different (Fig. 5). Thus,the Core 135-841A-4H data pass a reversal test, implying that theChRM directions are reliable and free of overprint. Because thesamples from the other APC cores are similar in lithology and mag-netic behavior, it is likely that their directions are also reliable.

APC Paleomagnetic Poles

Precision parameters (k) and α 9 5 values for core-mean paleomag-netic poles from the APC cores ranged from 5.5 to 199.8 and 4.8° to28.3°, respectively, indicating a large variability in the precision foreach pole (Table 1). Most of the poles are relatively well determined,with α 9 5 values of 20° or less. However, three poles were determinedwith only three samples each (Cores 135-840C-2H and -10H, Section135-840C-5H-2), owing to low recovery, so they had relatively largeoc95 values. In addition, Cores 135-841A-6H and -8H gave poles withlarge α 9 5 values because of high sample direction scatter, resultingfrom less-stable magnetizations.

APC cores from Hole 841A contained dipping beds with averagedips of 6.7° to 26.0° (Appendix A). Three different orientation meth-ods were tried on these samples (no dip correction, average dipcorrection, individual layer correction; see Appendixes B and C), andthe best results, defined by the least scatter indicated by the largestFisher precision parameter, k, were found with different methods on

767

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

Table 2. Mean paleomagnetic poles.

Pole

ABCDEFGHIJKLM

Location

Latitude(°N)

-69.0-53.7-63.0-62.9-69.6-70.2-37.0-28.5-77.8-38.7-28.8-57.7-76.2

Longitude(°E)

112.2104.0100.7100.699.190.4

106.4110.595.2

109.3113.0104.093.0

N

101111111111062419262219275457

95% confidence ellipse

Major

(°)

14.26.87.37.6

10.09.5

11.711.47.2

12.812.67.56.3

Minor

(°)

2.91.81.91.81.93.03.02.93.85.14.61.92.0

Azimuth

o352

7664

1213155

101358

(°)

Var31.934.235.645.721.122.626.115.725.229.824 120.9

σ/(°)

Var10.510.910.511.08.87.48.1

10.011.712.27 68.8

O

62.165.566.266.266.269.266.865.566.764.163.265 067.3

δ

o2.31.41.51.41.52.42.42.33.04.13.61 51.6

D

(°)

23.039.829.529.621.921.158.067.913.356.768.235 414.9

ΔD

o12.56.06.46.68.78.1

10.210.06.3

11.411.36.55.4

R

(°)

20.4

15.552.462.3

7.751.162.626 2

5.7

ΔΛ

(°)

12.6

8.310.410.26.6

11.511.4675.7

x2

1.51.21.21.31.21.51.41.31.31.21.21.41.4

Age

(Ma)

4.0

8.68.68.68.68.68.6

I S !15.3

Notes: Letters representing poles are as follows: A = Holes 840C and 841A APC mean pole; B = Hole 841B, all data, no bed-dip corrections; C =Hole 841B, all data, individual bed-dip corrections; D = Hole 841B, all data, FMS average bed-dip corrections; E = Hole 841B, all data, FMSaverage dip direction alignment; F = Hole 841B, late Miocene normal data, no bed-dip corrections; G = Hole 841B, late Miocene reversed data,no bed-dip corrections; H = Hole 841B, late Miocene normal outlier + reversed data, no bed-dip corrections; I = Hole 841B, late Miocene normaldata, individual bed-dip corrections; J = Hole 841B, late Miocene reversed data, individual bed-dip corrections; K = Hole 841B, late Miocenenormal outlier + reversed data, individual bed-dip corrections; L = Hole 84 IB, early Miocene data, no bed-dip corrections; M = Hole 841B, earlyMiocene data, individual bed-dip corrections. N = number of independent samples. "Major", "Minor", and "Azimuth" = major and minor semi-axislengths, and azimuth of major axis clockwise from north; σD, σ, = Univariate standard deviations of declination and inclination; δc, δ = correctedpaleocolatitude (Cox and Gordon, 1984) and 95% confidence limit; D, AD = mean declination and 95% confidence limit; R, ΔR = declinationanomaly and 95% confidence limit; x2 = reduced chi-square statistic with 2iV-2 degrees of freedom.

135-840C-2H-CC, 18-20 cm

(0-60 mη

1 32 mA/m

DNRM

135-841A-5H-3, 40-42 cm

(0-βOmT)

• Horizontalβ Vertical N | U p

• Horizontal w D o w n

o Vertical

J/Jo I

0 -

L J O -

V33.8 mA/m

20 40AF Field (mη

60 20 40AF Field (mη

135-840C-7H-2, 22-24 cm(0-θOmη

E. Up

400 mA/m

• Horizontalo Vertical

πNRM

135-841A-8H-3, 70-72 cm

(0-60 mη

N, Up4S0mA/π

ΠNBM

-IS

W.Down

11

J/Jo "

0 -

l JO

\

- 600.6 mA/m

• Horizontalo Vertical

S, Down

60 60

Figure 3. Stepwise alternating-field (AF) demagnetization behavior of four representative samples from oriented APC cores Holes 840C and 841 A. Three plots

are shown for each sample: (left) orthogonal vector end-point plot; (upper right) vector end-points on equal-area plot (open symbols denote point on upper

hemisphere; solid symbols, lower hemisphere); (lower right) normalized magnetization intensity vs. demagnetization field strength. Magnetic directions are shown

in sample coordinates.

768

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

Corel 35-841A-4H Corel 35-841A-5H

Inclination (°)

-90 -60 -30 30 60 90-90

Declination (°)

0 90 180

Inclination (°)

270 •90-60-30 30 60 90

Declination (°)

90 180 270 36027

Figure 4. Inclination and declination trends in Cores 135-841A-4H and -5H. Shipboard whole-core (small symbols) and land-laboratory discrete sample (largesymbols) inclination and declination data are plotted vs. depth (mbsf). Whole-core data are measurements after 15-mT AF demagnetization; discrete sample datawere AF demagnetized to isolate characteristic remanence direction. In declination plots, discrete sample normal-polarity declinations have been inverted (opensymbols) to accentuate trends. Dashed lines are declination trends determined from whole-core measurements (see Parson, Hawkins, Allan, et al., 1992). Whole-coretrends are not evident in the more thoroughly demagnetized discrete sample data.

180

270

Figure 5. Reversal test, Core 135-841A-4H. Normal and inverted reversed-polarity VGPs are compared. Open circles = reversed-polarity VGPs. Solidcircles = normal-polarity VGPs. Open and solid squares = mean VGPs forreversed and normal-polarity samples. Large circles with stipple = 095 circlesfor mean pole positions. Map is an equal-area plot of the Southern Hemisphere.

different cores. Because the differences in the amount of scatteramong methods were usually small and probably not significant, wesimply used the core-mean magnetic directions from the dip correc-tion method that gave the best results for each core (Table 1).

The core-mean paleomagnetic poles (Table 1) display consider-able scatter, most of it trending along the 110° meridian (Fig. 6). This

is simply the effect of the large variance in declinations caused by thelarge core-azimuth orientation error (So = 18.6°). This orientationerror is probably larger than it might ordinarily be owing to a problempeculiar to the Hole 840C APC cores. Those cores typically containedplugs of nannofossil ooze, with lengths of a few tens of centimetersto a few meters, separated by coarse, unconsolidated volcanic ash(Parson, Hawkins, Allan, et al., 1992). During handling on deck, itwas not uncommon to see water and ash sloshing about in some ofthese cores. As a result, it was possible for the small plugs of ooze tobreak free from the liner and rotate. A good example is the distinctlydifferent average declinations (1.9° and 89.7°) given by two differentsections of Core 135-840C-5H (Table 1).

Because of the considerable scatter in paleomagnetic declinationscaused by orientation errors, we used the outlier test to reduce thescatter of averaged VGPs. Those from Core 135-840C-6H and Section135-840C-5H-1 were found to be outside the 09 5 circle and were dis-carded from subsequent calculations. The mean paleopole calculatedfrom the 10 remaining core-means is located at -69.0°N, 112.2°E. Ithas a 95% confidence ellipse with a minor semi-axis length of 2.9° anda major semi-axis of 14.2° oriented along 352°. The reduced %2 statistichas a value of 1.5 with 18 degrees of freedom, a value within 95%confidence limits (0.52 and 1.60), indicating the estimated error valuespropagated to determine the error ellipse are consistent with the ob-served data scatter.

Hole 841B Sample Magnetic Properties

Sediments recovered from Hole 84IB range from clay to sand-stone and are mainly volcaniclastic in nature. Most of the paleomag-netic samples taken for this study are from the finer grained clay andclayey siltstone sections because fine-grained sediments often givebetter paleomagnetic results than coarser grained sediments. Like theAPC core sediments described above, those from Hole 84IB arestrongly magnetic because of volcanogenic components. NRM values

769

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

180

160 -

Figure 6. Paleomagnetic poles from Hole 840C and Hole 841A APC cores.Each pole location is shown by a small, filled circle and is surrounded by itsα 9 5 circle. Mean APC pole is shown by an open square and is surrounded bystippled 95% confidence ellipse. Grayed poles and α 9 5 circles were consideredoutliers and rejected from mean-pole calculation (see text). Solid square -location of drill sites. Map is an equal-area plot of the Southern Hemisphere.

for the oriented samples from Hole 84IB range from 4.10 to 1205.78mAra"1, but values of 50 to 200 mAm~1 are typical (Appendix D).

IRM acquisition curves for Hole 84IB samples generally displaysaturation in low applied fields (Fig. 7), implying that titanomagnetiteis the most probable type of magnetic grain in the sediments. Like theAPC core samples, those from Hole 84IB were overprinted with anupward-directed magnetization, probably resulting from the drill string.It was usually possible to remove this overprint with AF demagnetiza-tion at field strengths of 20-40 mT (Fig. 8). The behavior of mostsamples during AF demagnetization appeared to be reliable, with theremoval of an overprint leaving segments of univectorial decay onorthogonal vector plots (Fig. 8). Accordingly, CSD values for theoriented Hole 84IB samples are mostly <5°, indeed 64% are <2°(Appendix D), indicating well-determined magnetization vectors. Asmall percentage (<5%) of samples behaved erratically during AFdemagnetization or gave no stable ChRM direction; those with CSDvalues in excess of 12° were not included in subsequent analyses.Notably, the best behaved samples, with low CSD values, are often inadjacent cores or core sections (Appendix D), implying that broad-scale changes in sediment composition affect the magnetic stability.

Hole 841B Paleomagnetic Poles

Despite uncertainties in correlating the FMS logs and cores, theorientation of the samples worked well. Most sample VGPs are in theSouthern Hemisphere in a streak from the South Pole to the equatoralong about 110°E (Fig. 9). Only about 10% of the VGPs appear tobe outliers. Half of these, six samples in all, appeared to have beeninverted, that is, they were the only "reversed polarity" samples inotherwise totally normally polarized sections (see Appendix D). Theinversions of these samples were corrected so that their VGPs couldbe used in subsequent analyses.

As described above, we tried four methods for making azimuthand bed-dip corrections to examine the effects of the reorientationprocedures. The four methods were (1) azimuth reorientation without

0.0 0.2Magnetizing field (T)

Figure 7. Acquisition of isothermal remanent magnetization (IRM) by Hole

84IB samples.

bed-dip corrections, (2) individual reorientation and bed-dip correc-tions, (3) correction to the average FMS dip and dip direction, and (4)reorientation to the average FMS dip direction but with individualbed-dip corrections (see Appendix E). The mean pole positions cal-culated from the four different treatments are little different (Fig. 9and Table 2), the maximum distance between them being only 16.4°.The VGP scatter is greatest for Method 4 and least for Method 1, forwhich no bed-dip corrections were made. This is not only evident inthe standard deviations of the declination and inclination values (Table2) but is also clear visually (Figs. 9A and 9D). The fact that the sampleVGPs cluster most tightly without bedding corrections could indicatethat the samples acquired their magnetizations after they were tiltedand that if a fold test were possible for these strata, they would fail topass it; however, the difference in scatter is not statistically significant,so such an interpretation might not be correct.

When early Miocene VGPs were analyzed separately, we foundthat they did not seem to show as much declination scatter as the entireset of VGPs (Fig. 10). In this analysis, the least scatter occurred forthe bed-dip corrected VGPs, which suggests that the scatter statisticsare dependent on the subdivision of the data, so conclusions fromscatter observations must be drawn carefully. Furthermore, a signifi-cant difference was evident in the mean pole positions calculated withand without bed-dip corrections: uncorrected and corrected VGPsgave poles at -57.7°N, 104.0°E and -76.2°N, 93.0°E, respectively.The bed-dip corrected pole is much closer to both the geographic poleand the appropriate point on the reference polar wander path (Tables2-3). It implies a declination of 14.9° ± 5.4° (95% confidence limit),which yields a declination anomaly (observed declination minusreference declination) of only 5.7° ± 5.7°. On the other hand, theuncorrected data imply a declination of 35.4° ± 6.5° and a declinationanomaly of 26.2° ± 6.7° (Table 2). Tectonic conclusions made fromthese two poles are drastically different; one implies that the TongaArc has rotated about 26°, but the other suggests little rotation.

The late Miocene data behave differently than the older data. Thescatter appears to be greater, similar to the total data set, and polescalculated with and without bed-dip corrections are indistinguishable(Fig. 11 and Table 2), probably because the late Miocene sedimentsare tilted less. In addition, reversed polarity samples are present in thelate Miocene subset, unlike the older subset, and they give a signifi-cantly different pole position than the normal polarity data. Thereversed VGPs are mainly located at low latitudes (Fig. 11) and ifanalyzed by themselves give a pole location near -37° to -39° (Table2). If the reversed VGPs are removed, the remaining normal VGPsscatter from the South Pole to the equator, but with a definite cluster-ing near the geographic pole. This cluster is in the same location asthat of the early Miocene poles, so we used the early Miocene mean

770

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

135-841B-12R-2,104-106 cm

DNRM^

E Ito150mA/m ^

Wr

S, Down . Horizontalβ Vertical

135-841B-18R-1, 105-107 cm(0-βOmT)

•Horizontalβ Vertical

S, Down

135-841B-23R-4, 36-38 cm(0-100 ml)

135-841B-37R-4, 89-91 cm(0-βOmT)

E.Up

0 20 40 60 SO 100*FFWd|mη

πNRM

150mA/rn •

W.Down

Figure 8. Stepwise AF demagnetization behavior of four representative oriented samples from Hole 841B. Plot conventions as in Figure 3.

Table 3. Reference poles.

Age(Ma)

Pole location

Latitude

PN)

Australia APWP:2.9

12.026.0

-83.2-77.2-68.4

Hotspot APWP:2468

101214161820

-88.7-87.3-86.0-84.6-83.3-82.3-81.2-80.2-79.2-78.2

Longitude

PE)

103.6123.5118.7

124.3124.6124.9125.1125.4124.9124.5124.3124.2124.1

N

143020

6.24.24.8

k

42.540.747.8

Notes: Australian APWP data from Idnurm (1985), andHotspot APWP data from Gordon and Jurdy (1986). N= number of independent samples, OLQ5 = radius of coneof 95% confidence about mean pole, and k = precisionparameter.

pole (M, Table 2) and θ 9 5 to divide the late Miocene normal VGPsinto two groups. Most VGPs fall within the θ 9 5 circle and give a poleposition indistinguishable from the early Miocene corrected data(Table 2). The remainder are consistent with the reversed VGPs (Fig.11). The failure of these samples to pass a reversal test implies thatan overprint has not been removed. The observations that most of the

normal VGPs give a pole similar to the older, early Miocene data andthat they are smeared from the mean reversed pole to this polesuggests that most of the normal VGPs record an overprint. Thereversed polarity samples and a few normal polarity samples may notbe overprinted and therefore yield VGPs at low latitude.

DISCUSSION

Our main purpose in undertaking this study was to quantify theamount of rotation of the Tonga Arc and to constrain the timing of theLau Basin opening. To do this, we proposed determining paleomag-netic vectors from the arc and comparing them to expected directions.Choosing a reference direction is made more difficult by the complextectonics of the southwest Pacific; however, most investigators haveconcluded that the Lau Ridge has remained part of the Indo- AustralianPlate since the formation of the South Fiji Basin in Oligocene time(Watts et al., 1977; Malahoff, Feden, and Fleming, 1982; Taylor andKarner, 1983; Kroenke, 1984). Thus, we should be able to compare theTonga Arc magnetic vectors with those from the Indo-Australian Plate.

Unfortunately, the apparent polar wander path (APWP) for theIndo-Australian Plate is not well constrained, particularly for lateTertiary time. Many of the data used to locate the Australian APWPare undated (Embleton, 1981; Idnurm, 1985; Musgrave, 1989). Onlya few reliable, dated paleomagnetic poles are available (Table 3). A2.9-Ma pole from the Newer volcanics and Werriko limestone islocated a distance of 6.8° from the South Pole. Poles from theGlenample formation/Port Campbell limestone and Point Addis lime-stone are 12.8° and 21.6° from the pole and have ages of 12 and 16Ma, respectively. All three of these poles lie along the 120° meridian(Fig. 12) and suggest rapid Tertiary polar wander (Idnurm, 1985).

771

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

180 180

270

Figure 9. VGPs and mean poles of all oriented samples, Hole 84IB. A. VGPs calculated by reorienting beds or faults within sample to individual FMS features,but with no bed-dip corrections. B. VGPs calculated by reorienting beds or faults with individual FMS feature azimuths, with dip-correction from dip of samplebeds. C. VGPs calculated using core-average dip and dip-direction from FMS logs for azimuth orientation and dip-correction. D. VGPs calculated by reorientingsample bedding to core-average FMS bed-dip direction and making dip-correction from dip of beds in sample. Mean pole positions shown by open square andsurrounded by stippled 95% confidence ellipse. Solid square = location of drill site. All maps are equal-area plots of the Southern Hemisphere.

Another source of reference poles is to use a model of the drift ofthe Indo-Australian Plate relative to the hotspots to backtrack theSouth Pole and construct a predicted APWP. We used the recent platemotion model of Gordon and Jurdy (1986) for these calculations(Table 3), but most models (e.g., Duncan, 1981) have similar rotationparameters for the late Tertiary and give similar results.

The hotspot APWP shows far less rapid polar wander, accumulatingonly 11.8° of distance from the South Pole in 20 Ma (Fig. 12). Thedifference in polar wander rates between hotspot and paleomagneticstudies may result in part from the small number of dated poles(Musgrave, 1989) and in part because of long-term nondipole compo-nents in the geomagnetic field (Idnurm, 1986). The nondipole field biasdepends on latitude and age, but typically it is about 5° or less for thelate Tertiary (Livermore et al, 1983; Idnurm, 1986). Furthermore,models of the paleofield geometry indicate that the nondipole fieldsare probably axi-symmetric, which means that they should primarilyaffect paleomagnetic inclination. Nondipole field-biased paleomagneticdata from the Australian continent may indicate more apparent polarwander because the APWP trends toward the continent (Fig. 12). On

the other hand, the Lau Basin is in a direction nearly perpendicular tothe APWP trend, so nondipole fields will not strongly affect the separa-tion of Leg 135 paleopoles and the South Pole. Because the AustralianAPWP is not well constrained and may be biased by nondipole fields,which themselves may not be modeled well, we prefer to compare theLeg 135 paleomagnetic poles to the hotspot APWP.

APC Pole

Interpretation of the APC pole seems relatively straightforward.Most of the data comes from Hole 840C, which contains no evidenceto suggest that these sediments are tilted or metamorph°se<l Hole841A samples may have been tilted, but the corrections have littleeffect on the mean pole position. Likewise, no evidence was foundthat the sediments cored with the APC in Hole 841A were metamor-phosed. Furthermore, the APC core sediments pass a reversal test. Allof these factors imply that the APC mean is a reliable pole.

Oriented samples from Holes 840C and 841A all appear to rangein age from early to late Pliocene. Biostratigraphic markers indicate

772

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

270

270

Figure 10. VGPs and mean poles for early Miocene oriented samples, Hole84IB. A. VGPs calculated by reorienting beds or faults in samples to FMSfeature azimuths, but with no bed-dip corrections. B. Same as Figure 10A, butwith dip-corrections made from sample bed dips. Open symbols = outlierVGPs. Other conventions as in Figure 9.

that Cores 135-840C-2H through -4H were deposited during the timerepresented by calcareous nannoplankton Zone CN12, centered onabout 2.7 Ma, whereas Cores 135-840C-5H through -10H were depos-ited during Subzone CNlOb, centered on 4.8 Ma (Parson, Hawkins,Allan, et al., 1992). All of the Hole 840C paleomagnetic data arenormally magnetized (Appendix B), but this is consistent with thecorrelation of the Gauss normal polarity epoch (Subchron 2A, 2.48-3.40 Ma) to the lower part of Zone CN12 and the Thvera polarity event(a subchron of Chron 3,4.57^.77 Ma) to Subchron CNlOb (Berggrenet al, 1985). The dating of the cores from Hole 841A are less certainbecause the sediments are barren, but magnetic stratigraphy implies anearly Pliocene age (Parson, Hawkins, Allan, et al., 1992). Consideringthe distribution of ages, we take 4 Ma for the age of the APC pole.

The paleodeclination for the APC pole is 23.0° ± 12.5°, that forthe reference pole is 2.6° ± 1.8°, and the implied declination anomalyis 20.4° ± 12.6° (Fig. 13). To derive error limits, we used 2° for the95% error in predicted pole position, the major semi-axis length ofthe APC pole 95% confidence ellipse, and followed the methods ofBeck (1980) and Demarest (1983). The former error estimate seemedreasonable given that hotspot volcanic tracks can usually be locatedwithin a degree of distance; the latter is appropriate because the major

90 270

270

Figure 11. VGPs and mean poles for late Miocene oriented samples, Hole84IB. A. VGPs calculated by reorienting beds or faults in samples to FMSfeature azimuths, but with no bed-dip corrections. B. Same as Figure 11 A, butwith bed-dip corrections from sample bed dips. Open symbols = VGPs fromreversed-polarity samples. Plus sign, open square, and open diamond = polepositions for high-latitude normal, reversed-polarity, and low-latitude normal-polarity VGPs, respectively (see text and Table 2). Other conventions as inFigure 9.

semi-axis of the APC pole confidence ellipse trends toward thereference pole (Fig. 12). The APC core declination anomaly is indis-tinguishable from the 17° of clockwise rotation expected from thetrends of the Tonga and Lau ridges. Additionally, the paleolatitudeindicated by the APC pole, -27.9° ± 2.3°, compares well to theexpected paleolatitude, -24.0° ±1.6°, suggesting that the rotation wasmainly about a vertical axis. These observations imply that the LauRidge and Tonga Arc have separated like the blades of a pair ofscissors, rotating about a nearby hinge (pole of rotation). Furthermore,the youngest, oriented APC cores are of Subchron 2A age and theirpaleopoles are indistinguishable from the paleopoles of older APCcores (Table 1). Although confidence in this observation must be lowbecause of the orientation errors and small number of oriented cores,it implies that rotation began at the time of Subchron 2A or later. Thisis consistent with the oldest magnetic anomalies at the edges of theLau Basin being identified as Subchron 2A(Weissel, 1977; Malahoff,Feden, and Fleming, 1982) rather than as Subchron 3 A as suggestedby some investigators (Parson, Hawkins, Allan, et al., 1992).

773

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

N

Figure 12. Comparison of paleomagnetic poles from Tonga Arc with those fromIndo-Australian Plate. Filled circles = mean pole positions from Tonga Arc,with letters identifying poles (Table 2). Open circles = paleomagnetic polesfrom Australia (Idnurm, 1985; Table 3), labeled by age (in Ma). Small squaresconnected by a line with the South Pole = apparent polar wander path for theIndo-Australian Plate predicted from the relative motion of the plate vs.hotspots (Gordon and Jurdy, 1980); squares at 2-m.y. intervals. Outlines ofAustralia and Antarctica, in their present positions, are shown for reference.Map is an equal-area plot of the Southern Hemisphere.

Hole 841B Poles

In contrast to the APC cores, interpretation of the Hole 84IB datais complicated by uncertainties about overprints and corrections fortectonic tilting. The late Miocene VGPs failed a reversal test, and theirdistribution implies a normal polarity overprint. Such an overprint is apossibility because shipboard studies showed that within 100 m be-neath the fault zone at 449 mbsf sediments had been clearly affectedby low-temperature hydrothermal metamorphism (Parson, Hawkins,Allan, et al., 1992). This zone includes Cores 135-841B-31R through-41R, from which 90%-95% of the early Miocene samples were taken.Moreover, all of these samples are of normal polarity, which results,perhaps, because of remagnetization of the entire zone within a singlepolarity chron. Although shipboard descriptions suggest that the over-lying sediments were not as clearly affected by metamorphism, roughlyhalf of the VGPs from Cores 135-841B-2R to -29R have directionsconsistent with the early Miocene data, implying that hydrothermalmetamorphism may have affected much of the sediment column.

The reversed polarity samples suggest that some of the sedimentsfrom the late Miocene section may have escaped remagnetization. Thelate Miocene sediments contain calcareous nannoplankton from Sub-zone CN8b (Parson, Hawkins, Allan, et al., 1992), the center of whichis at about 8.6 Ma (Berggren et al, 1985). A declination anomaly of51.1° ± 11.5° was found by comparing the reversed polarity pole (J,Table 2) made with structural corrections, which has a declination of56.7° ± 11.4°, with the expected declination of 5.6° ± 1.8° (Fig. 12).The pole calculated without structural corrections (G, Table 2) is notsignificantly different, so it yields a similar result. As was true for theAPC cores, these data show little inclination anomaly: pole J impliesa paleolatitude of -25.9° ± 4.1°, nearly the same as the expected paleo-latitude, -26.2° ± 1.6°. Although these observations once again implyrotation about a vertical axis, the declination anomaly is nearly threetimes the expected 17° angle of deflection. Either this pole position isincorrect or local tectonics affected the Tonga Arc at the site. Large

Figure 13. Comparison of paleodeclinations with expected declination values,Sites 840 and 841. Paleodeclination values were determined from paleomagneticpoles (Table 2); expected declination values were calculated from synthetic ap-parent polar wander path derived from motion of the Indo-Australian Platerelative to the hotspots (Table 3). Plot at top shows the APC core mean decli-nation; plot in middle shows late Miocene declinations from Hole 841B; plot atbottom shows early Miocene declinations from Hole 84IB. Arrows = the refer-ence declination, R, and paleodeclinations (labels correspond to poles in Table2 and Fig. 12). Hachured and stippled sectors show 95% confidence regions.

local rotations within island arcs are not without precedent and mayresult from tectonic collisions (e.g., McCabe, 1984; Keating and Helsley,1985). Because Site 841 is located on the forearc of the Tonga Ridge,it may have been particularly susceptible to tectonic deformation.

Drawing a tectonic conclusion from the early Miocene pole iscomplicated by the addition of a large dependence of pole position onbed-dip corrections (Fig. 12). Cores 135-841B-32R through-41R havecalcareous nannoplankton indicative of biostratigraphic Zone CN4,which has a central age of 15.3 Ma (Berggren et al., 1985). The ap-propriate reference declination is 9.2° ± 1.8°, so poles without (L) andwith (M) bed-dip corrections give anomalies of 26.2° ± 6.7° and 5.7°± 5.7°, respectively. Both poles give similar paleolatitude values,-25.0° ± 1.5° (L) and -22.7° ± 1.6° (M), which are not very differentfrom the predicted paleolatitude of -21.1° ± 1.6°, so again the differ-ence is primarily in declination.

Two possible hypotheses can be formulated from these observa-tions. On the one hand, if the pole calculated without bed-dip correc-tions (L; Fig. 12) is correct, it indicates 26.2° of rotation, which isapproximately the same amount as the APC pole. These sedimentscould have been remagnetized by a hydrothermal event at the timethe Tonga Arc began to separate from the Lau Ridge. If this occurredduring Subchron 2A, the normal polarity of the overprint would beexplained. An alternate hypothesis is that the pole calculated fromsamples corrected for bed-dips is better. It implies a declination thatis virtually indistinguishable from the expected direction, as if thesediments were remagnetized recently. Remagnetization during the

774

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

Brunhes polarity epoch (since 0.73 Ma) could account for the normalpolarity, but if correct, this implies that the stratal dips mostly devel-oped recently. One significant tectonic event during The Brunhesperiod is the subduction of parts of the Louisville Ridge beneath thearc. At Site 841, this would have occurred about 0.5-1.0 Ma (MacLeod,this volume). Without additional data, we cannot decide with certaintywhich hypothesis is best.

CONCLUSIONS

Analysis of paleomagnetic data from oriented APC cores fromHoles 840C and 841A and samples from Hole 84IB, oriented relativeto the FMS wireline logs, yielded paleopoles from close to the SouthPole to the equator, trending along the 110° meridian. The scatter inpole latitudes apparently results from several factors: the AustralianAPWP trends along this meridian, core-orientation error causes scat-ter in this direction, and Tonga Arc rotation has moved the poles inthe same direction. In addition, structural corrections for bed dips,many of which trend perpendicular to the Tonga Trench, also movepaleomagnetic poles along this meridian.

The scatter of APC core poles implies a mean orientation error of18.6° (standard deviation). This value is probably inflated by a prob-lem with the lithology of Hole 840C, which caused the possibility ofcore rotation within the core liner. The scatter of sample declinationsreoriented relative to the FMS logs was variable and depended on thegroupings of samples by age and polarity. Declination-mean standarddeviations ranged from 15.7° to 31.9°, with higher values such as thelatter probably including a significant component of tectonic rotationand apparent polar wander.

The mean APC pole, with an age of 4 Ma, is located at -69.0°N,112.2°E, with a 95% error ellipse having a minor semi-axis of 2.9°and a major semi-axis of 14.2° with an azimuth of 352°. This polegives a paleodeclination of 23.0° ± 12.5°. When compared with anAPWP predicted from plate-hotspot motion, it implies a declinationanomaly of 20.8° ± 12.6°. Because the paleolatitude is nearly the sameas the expected value, this anomaly implies that the sites have rotatedclockwise around a vertical axis. This anomaly and sense of rotationare those expected if the Tonga Arc rotated away from the Lau Ridgeas a more or less rigid body. The youngest cores used to calculate thispole (Cores 135-840C-2H through -4H) all have declinations indis-tinguishable from the rest, despite their Gauss Epoch age (Subchron2A, 2.48-3.40 Ma). This observation suggests that most of the tec-tonic rotation has occurred since these sediments were deposited. Thelarge mean scatter of APC pole declinations, however, indicates thatthis conclusion must be taken with a degree of caution.

Interpretation of paleopoles from Site 841 was complicated by thevertical tectonics and overprints. Many of the sediments were hydro-thermally altered. Because normal and reversed samples yielded sig-nificantly different pole positions, it appears that many are overprinted.A pole calculated from late Miocene (age 8.6 Ma) reversed-polaritydata, which may be relatively free of overprint, is located at -38.7°N,109.3°E and has a 95% confidence ellipse with a minor semi-axis of5.1° and a major semi-axis of 12.8° with an azimuth of 10°. This polehas a paleodeclination of 56.7° ± 11.4°, giving a declination anomalyof 51.1 ° ± 11.5°. As before, no significant anomaly occurs in paleolati-tude, implying a vertical axis rotation. However, this rotation is muchgreater than that expected from the geometry of the Tonga Arc. EarlyMiocene paleopoles are highly dependent on structural corrections. Apole calculated from data not corrected for bed dips is located at-57.7°N, 104.0°E, and has an error ellipse with semi-axes of 1.9° and7.5°, with a major axis azimuth of 5°. A pole calculated from the samedata corrected for bed dips is located much closer to the geographicpole, at -76.2°N, 93.0°E, and has an error ellipse with semi-axes of2.0° and 6.3°, with the latter having an azimuth of 8°. The two polesyield declination anomalies of 26.2° ± 6.7° and 5.7° ± 5.7°, respec-tively. One could be interpreted to indicate an amount of rotationsimilar to that of the APC core pole, the other, to suggest little rotation.

ACKNOWLEDGMENTS

We thank Fred Vine and Simon Allerton for constructive reviews.The senior author thanks JOI/USSAC for project support. This isTexas A&M Geodynamics Research Institute Contribution No. 90.

REFERENCES*

Auzende, J.-M., Lafoy, Y, and Marsset, B., 1988. Recent geodynamic evolu-tion of the north Fiji Basin (southwest Pacific). Geology, 16:925-929.

Beck, M.E., Jr., 1980. Paleomagnetic record of plate-margin tectonic processesalong the western edge ofNorth America. J. Geophys. Res., 85:7115-7131.

Berggren, W.A., Kent, D.V., Flynn, J.J., and Van Couvering, J.A., 1985.Cenozoic geochronology. Geol. Soc. Am. Bull, 96:1407-1418.

Cox, A., and Gordon, R.G., 1984. Paleolatitudes determined from paleomag-netic data from vertical cores. Rev. Geophys. Space Phys., 22:42-72.

Demarest, H.H., Jr., 1983. Error analysis for the determination of tectonicrotation from paleomagnetic data. J. Geophys. Res., 88:4321-4328.

Duncan, R.A., 1981. Hotspots in the southern oceans—an absolute frame ofreference for motion of the Gondwana continents. In Solomon, S.C., Vander Voo, R., and Chinnery, M.A. (Eds.), Quantitative Methods of AssessingPlate Motions. Tectonophysics, 74:29^1-2.

Embleton, B.J.J., 1981. A review of the Paleomagnetism of Australia andAntarctica. In McElhinny, M.W., and Valencio, D.A. (Eds.), Paleorecon-struction of the Continents. Am. Geophys. Union, Geodyn. Ser., 2:77-92.

Falvey, D.A., 1978. Analysis of paleomagnetic data from the New Hebrides.Bull. Aust. Soc. Explor. Geophys., 9:117-123.

Fisher, R.A., 1953. Dispersion on a sphere. Proc. R. Soc. London A, 217:295-305.

Gordon, R.G., and Cox, A., 1980. Calculating paleomagnetic poles for oceanicplates. Geophys. J. R. Astron. Soc, 63:619-640.

Gordon, R.G., and Jurdy, D.M., 1986. Cenozoic global plate motions. J.Geophys. Res., 91:12389-12406.

Hamburger, M.W., and Isaacs, B.L., 1988. Diffuse backarc basin deformationin the southwestern Pacific. Nature, 332:599-604.

Hawkins, J.W., 1974. Geology of the Lau Basin, a marginal sea behind theTonga Arc. In Burke, C, and Drake, C. (Eds.), The Geology of ContinentalMargins: Berlin (Springer-Verlag), 505-520.

Idnurm, M., 1985. Late Mesozoic and Cenozoic palaeomagnetism of Austra-lia—I. A redetermined apparent polar wander path. Geophys. J. R. Astron.Soc, 83:399-418.

, 1986. Late Mesozoic and Cenozoic Paleomagnetism of Australia—III. Bias-corrected pole paths for Australia, Antarctica and India. Geophys.J. R. Astron. Soc, 86:277-287.

James, A., and Falvey, D.A., 1978. Analysis of palaeomagnetic data from VitiLevu, Fiji. Bull. Aust. Soc. Explor. Geophys., 9:115-117.

Karig, D.E., 1970. Ridges and basins of the Tonga-Kermadec island arcsystem. J. Geophys. Res., 75:239-254.

Karlin, R., and Levi, S., 1985. Geochemical and sedimentological control ofthe magnetic properties of hemipelagic sediments. J. Geophys. Res.,90:10373-10392.

Keating, B.H., and Helsley, C.E., 1985. Implications of island arc rotations tothe studies of marginal terranes. J. Geodyn., 2:159-181.

Kroenke, L., 1984. Cenozoic Development of the Southwest Pacific. U.N.Econ. Soc. Comm. Asia Pac, Suva, Fiji. CCOP/SOPAC, Tech. Bull., 6.

Livermore, R.A., Vine, FJ., and Smith, A.G., 1983. Plate motions and thegeomagnetic field. I. Quaternary and late Tertiary. Geophys. J. R. Astron.Soc, 73:153-171.

MacLeod, C.J., Parson, L.M., Sager, W.W., and ODP Leg 135 Scientific Party,1992. Identification of tectonic rotations in boreholes by the integration ofcore information with Formation MicroScanner and Borehole Televiewerimages. In Hurst, A., Griffiths, CM., and Worthington, RF (Eds.) Geo-logical Applications of Wireline Logs II. Geol. Soc. Spec. Publ. London,65:235-246.

Malahoff, A., Feden, R.H., and Fleming, H.S., 1982. Magnetic anomalies andtectonic fabric of marginal basins north of New Zealand. J. Geophys. Res.,87:4109^125.

Malahoff, A., Hammond, S.R., Naughton, J., Keeling, D.L., and Richmond,R.N., 1982. Geophysical evidence for post-Miocene rotation of the island

Abbreviations for names of organizations and publication titles in ODP reference listsfollow the style given in Chemical Abstracts Service Source Index (published byAmerican Chemical Society).

775

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

of Viti Levu, Fiji, and its relationship to tectonic development of the NorthFiji Basin. Earth Planet. Sci. Lett., 57:398-414.

Malahoff, A., Kroenke, L.W., Cherkis, N., and Brozena, J., in press. Magneticand tectonic fabric of the North Fiji Basin and Lau Basin. In Kroenke, L.,and Eade, J.V. (Eds.), Basin Formation, Ridge Crest Processes, andMetallogenesis in the North Fiji Basin. Circum-Pac. Counc. EnergyMiner. Resour., Earth Sci. Ser.

McCabe, R., 1984. Implications of paleomagnetic data on the collision relatedbending of island arcs. Tectonics, 3:409-428.

McFadden, RL., 1980. Determination of the angle in a Fisher distributionwhich will be exceeded with a given probability. Geophys. J. R. Astron.Soc, 60:391-396.

Musgrave, R.J., 1986. Palaeomagnetism and tectonics of the eastern Solomonislands [Ph.D. dissert.]. Univ. of Sydney, Sydney, Australia.

, 1989. A weighted least-squares fit of the Australian apparent polarwander path for the last 100 Myr. Geophys. J. R. Astron. Soc, 96:231-243.

-, 1990. Paleomagnetism and tectonics of Malaita, Solomon Islands.Tectonics, 9:735-759.

Parson, L., Hawkins, J., Allan, J., et al., 1992. Proc. ODP, Init. Repts., 135:College Station, TX (Ocean Drilling Program).

Parson, L.M., Pearce, J.A., Murton, B.J., Hodkinson, R.A., Bloomer, S.,Ernewein, M., Huggett, Q.J., Miller, S., Johnson, L., Rodda, P., and Helu,

S., 1990. Role of ridge jumps and ridge propagation in the tectonic evo-lution of the Lau back-arc basin, southwest Pacific. Geology, 18:470-473.

Sclater, J.G., Hawkins, J.W., Mammerickx, J., and Chase, CG., 1972. Crustalextension between the Tonga and Lau Ridges: petrologic and geophysicalevidence. Geol. Soc. Am. Bull, 83:505-518.

Tarling, D.H., 1967. The palaeomagnetism of some rock samples from VitiLevu, Fiji. N. Z. J. Geol. Geophys., 10:1235-1247.

Taylor, B., and Karner, G.D., 1983. On the evolution of marginal basins. Rev.Geophys. Space Phys., 21:1727-1741.

Watts, A.B., Weissel, J.K., and Larson, R.L., 1977. Sea-floor spreading inmarginal basins of the western Pacific. Tectonophysics, 37:167-181.

Weissel, J.K., 1977. Evolution of the Lau Basin by the growth of small plates.In Talwani, M., and Pitman, W C , III (Eds.), Island Arcs, Deep SeaTrenches and Back-Arc Basins. Am. Geophys. Union, Maurice Ewing Ser.,1:429^36.

Date of initial receipt: 1 October 1992Date of acceptance: 25 March 1993Ms 135SR-119

APPENDIX ACore-mean Bed-dip Directions, Site 841

Core number

135-841A-4H5H6HSH

135-841B-2R3R4R5R6R7R9R12R13R15R16R

Above 308.85 mbsf16R

Below 308.85 mbsf17R

Above 315.67 mbsf17R

Below 315.67 mbsf18R19R20R22R24R29R34R35R36R37R3 SR39R40R41R

Above 550.0 mbsf41R

Below 550.0 mbsf42R43R44R

Dip(°D)

10.513.926.0

6.7

20.311.4

8.18.2

10.76.66.25.8

11.510.2

7.8

8.7

6.0

11.312.924.313.3

8.43.85.2

21.039.227.727.025.422.223.3

25.9

14.815.331.239.6

Azimuth(°E)

21.8135.3

71.175.9

78.482.062.492.448.454.257.9

290.6327.3

53.8

57.4

328.3

336.2

59.4345.6292.3275.5103.5221.8183.5200.0

79.1128.9127.9117.0128.2146.4

146.9

81.288.198.596.2

k

181.37.26.2

10.1

19.915.227.233.511.825.112.8

5.83.2

20.3

221.5

16.4

12.6

21.39.7

10.121.010.719.521.1

30.329.432.861.521.9

38.5

11.211.610.065.2

oc95

5.529.738.451.7

35.317.312.811.519.97.59.5

15.223.4

6.9

9.5

12.2

25.3

19.113.435.8

7.113.335.634.2

13.67.7

10.53.56.0

6.5

23.126.613.813.3

N

6653

37777

1721151924

13

11

5

515

4

143:146

148

2929

15

65

144

Notes: Hole 841A dips oriented with multishot tool; Hole 841B dipsoriented with FMS log. " D i p " and "Azimuth" = average dip magni-tude and direction, k = precision parameter, α ^ = radius of 9 5 %confidence circle, N = number of dip and azimuth measurementsused to calculate mean.

776

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

APPENDIX BAPC Core Paleomagnetic Data

Core, section,interval (cm)

135-840C-2H-1, 27-292H-1, 48-502H-CC, 7-92H-CC, 18-203H-1, 10-123H-1, 21-233H-1, 30-323H-1,49-513H-1, 103-1053H-1, 109-1113H-1, 114-1163H-1, 140-1423H-2, 105-1073H-2, 110-1123H-2, 120-1224H-1, 10-124H-2, 93-954H-2, 133-1354H-3, 10-124H-3, 31-334H-3, 50-524H-3, 60-624H-3, 73-754H-3, 97-994H-4, 43-454H-4, 60-625H-1,67-695H-1, 80-825H-1, 91-935H-1, 102-1045H-1, 110-1125H-1, 128-1305H-2, 130-1325H-2, 142-1445H-2, 146-1486H-3, 71-736H-4, 56-586H-4, 80-826H-4, 74-766H-4, 86-886H-4, 103-1056H-4, 107-1096H-4, 128-1307H-2, 10-127H-2, 14-167H-2, 22-247H-2, 56-587H-2, 64-667H-2, 90-927H-2, 108-1107H-2, 134-1367H-3, 18-207H-3, 20-227H-3, 30-327H-3, 95-9710H-1, 54-561 OH-1,62-6410H-1, 63-65

135-841A-4H-1, 120-1224H-2, 30-324H-2, 50-524H-2, 82-844H-2, 124-1264H-3, 30-324H-3, 80-824H-3, 120-1224H-4, 38-^04H-4, 55-574H-4, 65-674H-4, 80-824H-4, 95-974H-4, 120-1224H-5, 43^54H-5, 60-624H-5, 75-774H-5, 100-1024H-6, 40^25H-1,40-425H-1,75-775H-2, 40-425H-2, 80-825H-2, 120-1225H-3, 10-12

Depth(mbsf)

47.7747.9848.0948.2057.1057.2157.3057.4958.0358.0958.1458.4059.5559.6059.7066.6068.9369.3369.6069.8170.0070.1070.2370.4771.4371.60

124.67124.80124.91125.02125.10125.28126.80126.92126.96137.21138.56138.80138.74138.86139.03139.07139.28144.60144.64144.72145.06145.14145.40145.58145.84146.18146.20146.30146.95172.04172.12172.13

28.7029.3029.5029.8230.2430.8031.3031.7032.3832.5532.6532.8032.9533.2033.9334.1034.2534.5035.4037.4037.7538.4539.3039.7040.10

AF step(mT)

10-4030-5020-5015-4015-4030-4025-4020-4020-4025^010-2525-5020-4020-3520-5015-3030-5020-5020-5020-3015-2520-3520-3520-4015-2525-3520-5025^025-4025^020-5020-4020-5015-3025-4010-2515-3020-5025-4025-4020-3020-5015-5025-4020-3525-5025-5025-4025-5025^025-5025-5030-4025-5030-4015-3015-40

25

20-5030-5030-4020-4020-5020-4020-4020-4020-4010-257-20

20-5020-4020-4030-5020-5030-5020-5020-5015-4020-5020-4020-5020-5020-50

N

6355534544444454355334443354344454444544356444444444343451

432343333444433434454444

Jo ,(mArrT1)

25.0645.1615.4533.79

421.29172.47141.7741.69

117.5469.3381.15

193.44208.04113.67500.89

58.6919.7437.55

114.1325.3838.2219.9429.0429.3137.9618.84

184.89128.94251.12140.48232.74283.93379.00516.70355.5244.0979.0468.23

102.24188.2474.44

159.1278.69

630.24475.09600.62460.62133.91357.55316.65894.71632.85354.50450.51184.0646.1643.6836.64

351.02110.5598.77

106.39106.9496.53

117.26163.77145.35145.50127.15154.1493.25

113.47100.1140.51

118.9091.5081.5748.8019.5526.3625.9251.6117.39

Wo(%)

0.240.050.080.140.040.020.110.190.030.050.150.030.030.050.040.160.140.090.110.200.130.120.080.180.060.110.020.050.040.030.020.030.020.020.010.070.170.070.100.080.070.040.170.020.030.030.030.140.050.040.030.020.040.020.050.090.100.07

0.020.010.020.040.030.050.030.030.040.110.100.050.060.050.020.200.030.090.100.500.330.450.300.31

Dec(°E)

43.070.852.422.624.148.533.647.9

358.824.616.671.333.228.134.05.4

11.923.433.917.546.726.826.241.953.528.484.385.992.187.098.290.20.76.5

358.7345.7321.2327.9324.3327.5335.8318.3334.661.154.437.140.626.737.971.052.625.711.421.836.8

352.1353.7359.6

14.5172.2196.8198.7182.0193.918.8

356.70.4

33.317.1

356.218.8

351.4220.1186.4346.9192.4184.9226.3213.3222.6212.8219.7224.1

IncCD)

^0 .9-41.2-35.1-30.6*-62.4-64.7-54.1-40.8-69.2-37.2-46.2-56.2-57.8-57.6-47.4-37.4-46.8-33.6-38.4-42.1-36.9-33.6-39.1-26.3-51.7-50.5-51.0-48.1-41.7-48.1^2 .2-51.8-31.4-52.6-53.4-58.2^5 .6-47.9-44.7^4 .3-35.8^6 .7-53.2-53.0-53.2-50.8-55.9^6 .5-56.3-48.7^9 .6-42.6-51.4-52.4-42.5^48.9-28.8-30.4

-50.641.3R

22.6R

27.7R

43.0R

25.1R

-46.1-50.4-48.0^6 .5-50.9-27.4-38.5-29.7

28.8R

38.8R

-43.846.1R

35.4R

38.0R

32.1R

29.3R

25.2R

32.3R

46.6R

CSD

o

1.82.86.21.11.91.50.70.65.63.22.63.31.81.81.11.95.34.41.84.92.87.25.56.84.22.94.71.00.81.44.43.02.42.93.96.22.72.01.42.34.32.00.52.52.81.31.60.60.90.72.51.22.83.93.64.05.72.5

3.26.22.93.25.13.34.05.92.12.11.32.52.22.0

10.40.64.31.91.53.11.31.21.40.63.0

777

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

Appendix B (continued).

Core, section,interval (cm)

5H-3, 40-425H-3, 44-165H-3, 83-855H-4, 40-425H-4, 69-715H-4, 118-1205H-5, 85-875H-6, 37-395H-6, 80-825H-6, 120-1226H-1,40-426H-2,40-426H-3, 84-866H-4,43^56H-4, 81-836H-4, 122-1246H-5, 40-428H-1, 22-248H-1, 28-308H-1,80-828H-1, 142-1448H-2, 40-428H-2, 80-828H-2, 120-1228H-3,40-428H-3, 70-728H-3, 80-828H-4, 40-42

Depth(mbsf)

40.4040.4440.8341.4542.1942.6843.8544.8745.3045.7046.9048.4050.3451.4351.8152.2252.9065.7265.7866.3066.9267.4067.8068.2068.9069.2069.3070.40

AF step(mT)

15-M)20-5020-5020-4020-5020-5020-5020-5020-5020-5020-5030-6050-6040-6030-5020-5020-5030-5040-6030-6020-5030-6030-6040-6040-6020-5030-6040-60

N

5444444444442334433444433443

(mAπT1)

38.4435.5750.1941.7540.0428.4230.6982.35

114.75142.35248.01330.47261.41387.32179.49210.86220.29432.91385.00581.80433.42611.73303.02355.98110.87405.45127.39725.16

•Wo(%)

0.410.370.470.220.260.280.170.040.050.050.020.020.020.040.030.030.020.010.040.040.060.020.040.070.070.060.070.01

Dec(°E)

237.1216.7226.9232.8210.9223.6217.9

61.5164.1158.553.3

204.2232.0

16.4351.8

90.9331.3226.0162.9163.9226.4162.9256.3187.6182.8243.5184.4146.0

Inc(°D)

51.5R

5 0 • 3 P

48.8R

39.3R

39.3R

45.0R

29.4R

-51.6-69.8*-54.3*-52.4

16.0R

22.2R

-61.5-54.7-66.3-57.9-52.6*

62.7R

38.9R

45.8R

70.1R

43.8R

35.9R

13.2R

21.4R

27.1R

-54.2*

CSD

(°)

1.41.10.62.01.31.74.26.56.44.36.93.44.73.38.96.4

10.33.72.02.03.42.23.01.31.32.82.27.4

Notes: AF step = demagnetization minimum and maximum alternating-field strengths used in determining characteristicremanent magnetization (ChRM); Jo = natural remanent magnetization (NRM) intensity, in I0"3 Am"1; J/Jo =normalized magnetization intensity at maximum demagnetization used in ChRM determination; "Dec" and "Inc" =paleomagnetic declination and inclination, in sample coordinates; CSD = circular standard deviation of mean direction.Superscript R = reversed polarity sample, and asterisk (*) = vector not used in mean pole calculations. N as in AppendixA.

778

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

APPENDIX CBedding-corrected Paleomagnetic Data, Hole 841A

Core, section,interval (cm)

135-841A-4H-1, 120-1224H-2, 30-324H-2, 50-524H-2, 82-844H-2, 124-1264H-3, 30-324H-3, 80-824H-3, 120-1224H-4, 38^04H-4, 55-574H-4, 65-674H-4, 80-824H-4, 95-974H-4, 120-1224H-5, 43-454H-5, 60-624H-5, 75-774H-5, 100-1024H-6, 40-425H-1, 40-425H-1, 75-775H-2, 40-425H-2, 80-825H-2, 120-1225H-3, 10-125H-3, 40-425H-3, 44-465H-3, 83-855H-4, 40-425H-4, 69-715H-4, 118-1205H-5, 85-875H-6, 37-395H-6, 80-825H-6, 120-1226H-1,40-426H-2, 40-426H-3, 84-866H-4, 43^56H-4, 81-836H-4, 122-1246H-5, 40-426H-1,22-246H-1, 28-306H-1, 80-826H-1, 142-1448H-2,40-428H-2, 80-828H-2, 120-1228H-3, 40-428H-3, 70-728H-3, 80-828H-4, 40-42

Dip(°D)

2.02.02.02.08.08.08.08.8

13.013.013.013.013.013.06.02.02.02.0

15.015.015.010.010.010.013.013.013.011.09.09.0

10.020.020.020.020.020.026.021.021.021.021.021.010.010.010.04.04.04.04.03.03.03.08.0

Individual layer

Azimuth(°E)

269.0269.0269.0269.0

19.012.012.012.022.022.022.022.022.026.026.0

224.0224.0224.026.026.026.04.04.0

137.0148.0148.0148.0216.0280.0280.0

83.0105.0105.0105.0105.0105.0105.0121.0121.0121.0121.0121.0248.0248.0248.0258.0258.0258.0258.0359.062.062.0

157.0

Dec(°E)

16.8174.0197.6199.7179.3194.120.0

353.0352.737.315.0

352.218.0

345.9221.0187.3348.5193.4179.3232.2215.1226.9215.7213.7212.1221.4204.1225.1237.4217.2215.8205.6

25.0222.5189.125.0

196.0222.4348.2336.3

30.0321.1219.4180.9171.6228.4174.0256.4190.2182.9243.5183.0143.2

Inc(°D)

-50.041.522.027.050.633.1

-54.0-58.8-59.8-59.2-63.8-38.9-51.5-40.1

34.637.2

^12.744.449.151.847.036.933.830.542.149.544.238.032.935.652.335.3

-62.9-70.5-61.9-60.9

18.428.1

-51.0-39.1-78.4-38.8-61.7

60.337.242.470.139.834.516.224.428.7

-62.0

Average bedding

Dec(°E)

12.2166.4196.3198.3177.6193.018.0

349.5354.7

36.315.6

353.218.2

347.6222.4183.5339.6190.0182.1215.6205.5215.1207.0211.4210.0219.3201.6211.4221.3201.0210.2210.643.4

203.0170.524.2

194.9225.1321.7316.4185.4300.3229.7150.3158.6222.4145.4256.4182.8181.2242.8181.0155.3

Inc(°D)

-61.050.233.138.252.735.5

-56.6-59.6-57.6-56.7-61.4-36.8-49.0-38.6

38.748.9

-52.056.445.436.928.327.721.529.944.652.246.447.339.734.743.026.7

-53.4-79.6-66.5-75.2

32.446.3

-65.5-51.3-81.4^16.2^ 6 . 7

61.638.451.568.750.538.115.127.929.0

-56.0

Notes: Individual layer = corrections made for individual layers; "Dip" and "Azimuth" = dip magnitude anddip direction; "Dec" and "Inc" = corrected paleomagnetic declination and inclination; "Dec" and "Inc"for average bedding, = paleomagnetic declination and inclination corrected using average value forcore-dip and dip direction.

779

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

APPENDIX DPaleomagnetic Data, Hole 841B

Core, section,interval (cm)

135-841B-2R-1, 116-1182R-2, 63-652R-2, 101-1032R-2, 126-1283R-1, 16-183R-1,24-263R_2, 42^43R-2, 54-564R-1,67-694R_2, 49-514R-2, 55-575R-1, 15-175R-1, 37-396R-1, 71-737R-1,23-257R-2,12-149R-1,59-619R-1, 65-679R-2, 91-939R-2, 99-10112R-2, 104-10612R-2, 114-11613R-4, 3-513R-4, 55-5715R-1,73-7515R-1, 104-10615R-1, 113-11515R-2, 42-4416R-1,78-8016R-1,95-9716R-1,98-10016R-2, 124-12616R-2, 138-14016R-3, 29-3116R-3, 53-5516R-5, 37-3916R-5, 61-6316R-5, 84-8616R-5, 92-9417R-2, 19-2117R-2, 43^517R-2, 68-7017R-2, 83-8517R-3, 135-13717R-4, 49-5118R-1, 105-10718R-2, 126-12819R-1.2-420R-2,12-1422R-3, 47-4922R-3, 68-7024R-1,96-9829R-2, 26-2832R-1, 102-10432R-1, 128-13032R-1, 136-13832R-2, 7-933R-1, 11-1333R-1, 74-7633R-1, 136-13833R-2, 70-7234R-2, 3-534R-2, 15-1735R-1, 107-10935R-2, 114-11635R-3, 124-12636R-1,56-5836R-3, 66-6837R-1, 76-7837R-2, 6-837R-2, 33-3537R-2, 35-3737R-2, 106-10837R-2, 113-11537R-3,48-5037R-3,111-11337R-3, 114-11637R-3, 119-12137R-3, 133-13537R-4, 14_1637R.4, 33_3537R-4, 55-5737R-4, 89-9138R-1,27-29

Depth(mbsf)

170.96171.93172.31172.56179.56179.64180.95181.07189.77191.09191.15198.85199.07209.11218.33219.72237.99238.05239.81239.89268.84268.94280.13280.65295.63295.94296.03296.82305.38305.55305.58307.34307.48307.89308.13310.97311.21311.44311.52315.49315.73315.98316.13318.15318.79324.55326.26333.22344.10364.87365.08381.66430.86459.12459.38459.46459.61467.91468.54469.16469.98479.01479.13488.17489.76491.38497.36500.46507.16507.96508.23508.25508.96509.03509.88510.51510.54510.59510.73511.04511.23511.45511.79516.37

AFstep(mT)

15^4015-4015-4015^4015-4015^015^015-4015-4015^4030-6015-4015^015-5015^015-3015-4015^020-5020-4015^015^4015-4015^030-5020-4020-5015-4015-4020-4015-4015^010-2515-2515-2515-2515-5015-4015^015-3040-6020-3015-5015-3015^015^015-4015^015^015-3015-3015^4015-5015-5015-5015-6015-5015-5015-5015-5015-5015-5015-5015-4015-5015-4015-5015-5015^020-5020-4020-4020-5030-5020-5020-5020-5015-5020-5020-5020-5020-5015-5020-50

N

5

555555554555545543555535455

454333655533644555544566666666666565666433434446444464

Jo ,(mAm~1)

223.64222.76317.58710.10569.79464.50654.96797.83292.68118.82123.03311.64526.94

1205.78720.44333.79153.92117.75104.25116.96180.93381.89178.32188.97282.59

88.33181.48107.8362.1490.9622.7659.6363.9048.3267.4184.6459.0471.5672.9255.9462.9324.2737.0497.5975.47

145.51473.91

96.5968.35

137.30134.05589.55290.98185.77307.30165.11330.9942.0544.3530.6723.3074.8561.48

123.1146.804.10

23.11323.5128.59

572.3632.3570.95

112.3148.3564.10

118.07102.94179.2096.58

135.99178.78159.02262.20244.24

Wo(%)

0.310.420.350.430430400.370.430.350.400.140.320.330.350.570.560.350.360.280.440.300.470.370.320.050.250.090.220.160.100.410.220.270.740.450.320.120.210.270.480.120.820.170.730.420.320.460.350.420.370.430.310.130.230.320.140.340.230.260.220.270.220.230.230.270.190.240.330.230.410.310.430.270.320.250.220.270.200.280.240.330.230.400.34

Dec(°E)

30.0355.0153.3161.1

7.0108.9222.4227.4247.5108.0310.366.9

179.573.277.0

302.1237.0141.9350.7256.7

99.9103.6315.6316.3279.1122.2160.9110.2153.5355.0352.9207.7300.745.542.1

352.8333.5234.592.9

181.4186.1288.8236.4180.1224.3289.2212.5236.4

60.0201.7200.348.117.1

168.90.5

337.915.043.2

323.3182.6207.3240.2194.9217.9172.150.30.1

268.0140.0155.349.847.9

218.7327.4199.7189.8198.0162.5189.6178.2181.1206.5171.6

10.7

IncCD)

-41.8-47.1^ 9 . 2-37.1-30.5-35.7-31.2-35.8-42.0-45.1^5 .7-26.6-40.5-39.6-22.6-34.4-58.7-62.5

57.8R

-66.9-55.3-56.6-57.2-41.8-10.9

57.4R

44.4R

^8 .9^5 .7

45.5R

39.8R

31.9*-25.1

42.5*31.0*

^ 2 . 0-15.8

24.1R

-23.434.3R

21.5R

38.9R

51.5R

34.8R

35.0R

44.2R

39.1R

30.5R

40.3R

21.5*42.0R

47.3R

^8 .7-45.1-38.9-50.5^44.7-44.1-41.6-37.9-37.3-84.5-77.9-31.1-28.9-29.9-14.8^ 7 . 0-34.2-49.7

42.4*44.0*

-33.635.3*

-42.2^5 .8^ 6 . 0^ 5 . 4-52.0-50.9-42.4-AOA-55.4-19.8

CSD(°)

0.80.40.70.10.30.40.10.30.40.73.60.70.50.60.40.21.21.12.11.10.80.31.30.62.93.13.32.43.82.95.94.52.11.51.82.73.83.02.11.34.13.03.50.62.70.80.91.91.11.81.41.21.11.30.81.70.80.90.50.90.71.11.32.70.78.71.01.02.73.82.50.51.82.61.00.70.81.21.31.21.21.30.40.5

780

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

Appendix D (continued).

Core, section,interval (cm)

38R-1,45-4738R-1, 57-5938R-2, 23-2538R-2, 103-10538R-2, 121-12338R-2,129-13138R-2,134-13638R-2, 145-14738R-3, 8-1038R-3, 34-3638R-4, 21-2339R-1, 24-2639R-1, 55-5739R-1,71-7339R-1, 118-12039R-2, 128-13039R-2, 137-13940R-1, 50-5240R-1,86-8840R-1, 97-9940R-1, 120-12240R-2,49-5140R-3, 11-1340R-3, 113-11540R-3, 131-13340R-4, 31-3340R-4, 56-5841R-1,40-4241R-2, 29-3141R-3, 24-2642R-4,125-12743R-1,45^743R-1, 130-13244R-2,16-1844R-2, 9-11

Depth(mbsf)

516.55516.67517.83518.63518.81518.89518.94519.05519.18519.44520.81525.94526.25526.41526.88528.48528.57535.80536.16536.27536.50537.22538.34539.43539.64540.11540.29545.40546.79548.24560.05564.45565.30575.22594.49

AF step(mT)

20-4015-5020-5020-5020-5020-5020-5015-5015-5020-5015-5020-4020-5020-5015-5020-5020-5015-5030-5020-4015-4015^015-2510^010-3020-5010-3015-2510-2510-2550-010-2515^010-2515^0

N

46444446646344644633557434534444545

(mAm~1)

225.58101.87135.87199.1747.9835.0034.6343.28

282.42147.12193.0732.84

141.5094.38

108.0093.9684.20

302.61188.11173.5756.89

110.90246.30

84.2526.6134.5644.90

196.14154.36260.21

55.5845.8615.57

147.5274.42

J/Jo

(%)

0.470.340.310.370.320.320.350.240.250.290.270.460.340.340.220.250.280.260.060.030.140.200.630.090.160.050.270.140.090.150.080.300.980.210.07

Dec(°E)

5.412.3

309.611.510.5

202.09.7

358.1337.7332.0

1.4190.0168.3198.0194.7197.7221.3171.6167.6310.5187.9189.9196.3181.9257.1186.4234.2219.7332.1202.6311.759.6

352.60.6

32.4

Inc(°D)

-18.4-22.0-10.0^2 .9-41.2-35.1^ 0 . 9^8 .8-36.7-41.8-54.9-4S.\^ 3 . 9-38.9^2 .7-53.1^ 9 . 4-20.8^ 6 . 0-40.5-57.9-39.8-34.7-32.3-21.7^ 2 . 5-58.3-54.8-56.4-57.2-41.8-56.9

28.1*-65.7-38.0

CSD(°)

0.90.52.50.90.70.40.52.30.61.00.51.90.70.40.81.30.60.55.78.74.12.50.46.38.1

11.82.72.13.61.73.63.43.23.42.3

Notes: AF step = demagnetization minimum and maximum alternating field strengths used in determining characteristicremanent magnetization (ChRM); Jo = natural remanent magnetization (NRM) intensity, in I0"3 Am"1; J/Jo = normalizedmagnetization intensity at maximum demagnetization used in ChRM determination; "Dec" and "Inc" = paleomagneticdeclination and inclination, in sample coordinates; CSD = circular standard deviation of mean direction. Superscript R= polarity sample, and asterisk (*) = sample inverted in handling. N as in Appendix A.

781

W.W. SAGER, C.J. MACLEOD, N. ABRAHAMSEN

APPENDIX EReoriented Virtual Geomagnetic Pole Data, Hole 841B

Core, section,interval (cm)

135-841B-2R-1, 116-1182R-2, 63-652R-2, 101-1032R-2, 126-1283R-1, 16-183R-1, 24-263R_2,42-443R-2, 54-564R-1,67-694R-2,49-514R-2, 55-575R-1, 15-175R-1,37-396R-1, 52-546R-1, 71-737R-1, 23-257R-2,12-149R-1, 59-619R-1,65-679R-2,91-939R-2, 99-10112R-2, 104-10612R-2, 114-11613R-4, 3-513R-4, 55-5715R-1,73-7515R-1, 104-10615R-1, 113-11515R-2, 42-4416R-1,78-8016R-1,95-9716R-1, 98-10016R-2, 124-12616R-2, 138-14016R-3, 29-3116R-3, 53-5516R-5, 37-3916R-5, 61-6316R-5, 84-8616R-5, 92-9417R-2, 19-2117R-2, 43^4517R-2, 68-7017R-2, 83-8517R-3, 135-13717R-4, 49-5118R-1, 105-10718R-2, 126-12819R-L2-420R-2, 12-1422R-3, 47-4922R-3, 68-7024R-1,96-9829R-2, 26-2832R-1, 102-10432R-1, 128-13032R-1, 136-13832R-2, 7-933R-1, 11-1333R-1, 74-7633R-1, 136-13833R-2, 70-7234R-2, 3-534R-2, 15-1735R-1, 107-10935R-2, 114-11635R-3, 124-12636R-1, 56-5836R-3, 66-6837R-1,76-7837R-2, 6-837R-2, 33-3537R-2, 35-3737R-2, 106-10837R-2, 113-11537R-3, 48-5037R-3, 111-11337R-3, 114-11637R-3, 119-12137R-3, 133-13537R-4, 14-16

Dip(°D)

7.011.03.04.09.0

14.010.012.011.04.09.05.08.0

21.04.0

10.05.0

17.S8.96.4

10.87.07.07.02.0

10.09.07.8

11.015.04.5

15.07.05.07.87.0

11.09.0

10.018.09.0

12.04.07.08.06.08.06.09.0

17.013.013.023.0

6.020.017.017.015.015.013.024.09.09.0

18.026.023.032.025.020.027.116.518.818.825.126.318.730.030.026.029.019.0

Bedding

Azimuth

105.0107.0105.0154.0214.0123.0302.0301.096.0

323.0179.090.0

203.027.048.053.0

323.0206.1206.5218.6

33.663.063.0

331.0298.0306.8302.0320.2135.0222.0222.0222.0

8.011.058.098.0

322.0305.0294.0

24.0221.0287.0124.090.0

256.0342.090.0

329.0305.0118.072.072.085.039.0

107.0121.0121.0136.0136.0226.0134.0131.0159.0149.0307.0351.0149.094.0

136.0266.1283.7328.4328.4237.0261.8285.1271.1272.0270.0270.0272.9

Sampleazimuth

(°E)

78.0150.0128.0339.0176.0123.0302.0301.0326.0

17.0154.090.0

203.0334.0169.053.0

323.0206.1206.5218.6

33.6105.079.029.064.0

306.8320.0320.2135.0222.0222.0222.0

8.011.0

241.098.0

322.0305.0294.0

24.0221.0228.0124.090.0

256.0342.0

90.0329.0305.0118.072.072.0

175.039.010.0

154.0154.0171.0149.033.0

331.018.07.0

178.0307.0120.0149.094.0

175.0266.1283.7328.4328.4237.0261.8285.1271.1272.0270.0270.0272.9

FMSazimuth

(°E)

66.0189.0352.0189.0173.0101.095.095.077.0

5.05.0

83.078.058.0

113.052.047.057.957.957.947.022.068.0

196.0196.053.863.053.870.085.085.085.071.071.0

201.058.0

348.0348.0315.0303.0243.0344.076.076.0

346.0346.028.0

7.0329.0277.0126.0126.0

9.0199.0185.0172.0172.0185.0184.0184.0184.0184.0190.0

5.0110.0286.0103.0133.0245.0136.0127.9127.9127.9127.9127.9127.9127.9127.9136.0127.9127.9

Orientationtype

F/MF/MF/MF/MF/MF/MB/MB/MF/MF/MF/MB/MB/MF/MF/MB/MB/MB/AB/AB/AB/AF/MF/MF/MF/MB/AB/MB/AB/MB/MB/I

B/MB/MB/MF/MB/MB/MB/MB/MB/MB/MF/MB/MB/MB/MB/MB/MB/MB/MB/MB/MB/MF/MB/MF/MF/MF/MF/MF/MF/MF/MF/MF/MF/MB/MF/MB/MF/MF/MB/MB/AB/AB/AB/AB/AB/AB/AB/AB/MB/AB/A

Dec(°E)

18.034.017.311.14.0

86.915.421.4

358.596.0

161.359.954.587.917.276.026.188.8

353.3190.0281.0

16.992.6

122.688.326.1

243.2254.5

45.216.5

218.0215.9270.7

0.75.52.1

18.816.5

255.511.9

203.4302.1240.8222.4270.1228.3227.2250.5260.4219.0255.7254.3242.1177.1343.9

18.5355.929.078.2

114.335.613.363.221.920.9

338.14.3

39.1338.0

9.9359.529.327.4

109.613.542.546.653.928.547.533.2

Inc(°D)

-41.8-47.1^ 9 . 2-37.1-30.5-35.7-31.2-35.8^12.0^ 5 . 1^ 5 . 7-26.6-40.5-32.3-39.6-22.6-34.4-58.7-62.5

57.8-66.9-55.3-56.6-57.2-41.8-10.9

57.444.4

^ 8 . 9-45.7

45.539.831.9

-25.1-42.5-31.1-42.0-15.8

24.1-23.4

34.321.538.951.534.835.044.239.130.540.321.542.047.3

-48.7^ 5 . 1-38.9-50.5^44.7^14.1-41.6-37.9-37.3-84.5-77.9-31.1-28.9-29.9-14.8-47.0-34.2-49.7-42A^ 4 . 0-33.6-35.3^42.2^15.8-46.0^ 5 . 4-52.0-50.9

VGPA

Latitude(°N)

-73.5-59.1-73.2-79.4-82.1-10.5-74.2-69.8-88.4-5.336.4

-32.9-40.2

-8.7-74.1-17.4-65.3-15.5-68.8-72.7-24.5-70.8-11.9

8.5-10.8-59.2-34.8-23.2-49.3-74.6-55.5-57.0

-6.2-79.8-84.8-83.1-72.8-68.0-18.1-74.2-67.7

23.7-34.3-51.7

-7.4^ 5 . 0-473-25.7-15.0-54.3-17.4-22.9-34.5

36.9-75.1-72.9-81.3-63.6-20.0

10.6-57.1-77.4-27.9-44.4-69.3-67.9-81.7^19.3-69.6-79.7-82.8-63.3-65.0

9.6-76.9-51.3^ 7 . 9-41.5-64.0-47.4-59.5

Longitude(°E)

100.8111.2122.082.533.9

111.973.388.8

241.5121.5163.996.8

106.0110.194.9

100.389.3

131.5198.1157.6229.2139.0130.3142.4116.463.3

125.8114.3113.9112.4108.8100.7110.9

8.6109.921.7

101.553.33101.052.287.4

119.5106.2117.6112.499.1

108.3109.0106.4102.199.6

112.3114.2181.6258.6

93.7208.6107.2115.1126.898.285.8

173.7172.880.0

291.734.478.2

256.370.2

188.2103.2105.9119.680.2

105.1110.1111.4108.5118.2117.8

Dec(°E)

11.721.913.89.46.2

83.49.2

12.5349.298.7

157.758.851.2

102.415.478.224.7

116.4335.9181.3268.5

8.499.1

126.989.124.8

243.3257.8

38.00.5

214.3203.9272.4358.4

10.55.3

25.318.1

258.820.4

206.7301.2239.9216.6275.2232.3230.4255.2264.9228.7250.8243.1222.5174.1331.4

6.6343.1

16.663.3

104.919.57.89.0

350.65.9

332.9349.3

30.6332.8

0.2348.046.745.999.434.325.716.721.79.1

12.2112.3

Inc(°D)

-41.8^ 9 . 2^ 9 . 2-33.9-22.6-49.2-32.5-38.3-39.7-42.3-54.2-31.2^ 7 . 7-^0.3^43.0-31.7-39.1-71.8-65.1

61.7-58.2-59.8-62.5-50.9-40.1-19.7

66.451.6

-58.6-49.2

48.448.538.5

-26.7-37.5-26.9-51.1-23.7

18.8-28.7

27.29.9

42.857.132.537.651.741.626.930.029.449.167.0

-54.2-32.3-33.5-39.1-38.7-50.5-35.8-31.0-32.7-79.0-63.0-28.0-51.1-20.7-27.8-28.1-16.3-38.1^ 2 . 4-A4A-56.8

41.9^ 0 . 9^ 2 . 3^15.9-33.2-47.6^45.8

VGPB

Latitude(°N)

-79.3-69.4

76.0-80.0-77.0-16.9-79.7-78.3-80.0-2.228.2

-34.8^44.0

1.5-75.9-17.2-67.2

-6.1-59.4-70.4-13.3-71.3

-9.815.2-9.6

-62.9-34.7-22.1-54.2-83.3-58.8-67.8

-6.4-80.6-80.0-79.6-66.1-69.6-14.0-69.2-63.3

26.1-35.8-55.6

-2.3-41.8-45.0-22.0-10.2^13.8-23.5-33.9-47.8

31.7-62.6-82.0-74.3-74.6-33.9

5.0-70.5-80.8-AA.2-67.5-79.8-64.6-73.8-59.9-63.0-75.0-78.7^17.5^48.4

-7.0-58.7-66.4-74.7-70.0-80.0-77.8-78.2

Longitude(°E)

101.0118.4126.367.932.7

120.863.088.3

277.2120.5164.0112.9112.9120.6104.5106.496.5

155.0217.8181.9237.8164.5138.4139.8115.569.7

139.0121.6130.2181.0113.3115.8115.6355.1

83.134.4

120.965.399.375.780.0

113.2109.4128.0113.0102.7117.8112.3105.995.0

102.4116.2144.4179.0281.6

56.7275.9

92.3117.7118.878.358.9

180.2202.5

39.1249.2324.0

83.7287.9

5.5282.3106.2108.3132.5103.2100.0102.3110.664.5

123.3116.4

PALEOMAGNETIC CONSTRAINTS ON TONGA ARC ROTATION

Appendix E (continued).

Core, section,interval (cm)

37R-4, 33-3537R-4, 55-5737R-4, 89-913 8R-1,27-2938R-1, 45-473 8R-1,57-5938R-2, 23-2538R-2, 103-10538R-2, 121-12338R-2, 129-13138R-2, 134-13638R-2, 145-14738R-3, 8-1038R-3, 34-3638R-4, 21-2339R-1,24-2639R-1, 55-5739R-1, 71-7339R-1, 118-12039R-2, 128-13039R-2, 137-13940R-1,50-5240R-1, 86-8840R-1,97-9940R-1, 120-12240R-2, 49-5140R-3, 11-1340R-3, 113-11540R-3, 131-13340R-4, 31-3340R-4, 56-5841R-1, 40-4241R-2, 29-3141R-3, 24-2642R-4, 125-12743R-1,45-4743R-1, 130-13244R-2, 16-1844R-2,9-l l

Dip(°D)

16.116.527.221.123.630.516.520.217.016.512.019.923.317.718.713.020.025.029.024.025.011.025.026.015.029.034.032.026.035.031.033.033.037.012.014.013.045.053.0

Bedding

Azimuth

263.1283.7247.4

95.2116.9100.0103.7109.6117.7114.690.0

107.099.4

127.9105.1276.0261.0261.0298.0263.0284.0225.0257.0290.0284.0259.0262.0260.0290.0267.0283.0323.0116.0313.0329.0262.0217.0123.064.0

Sampleazimuth

(°E)

263.1283.7247.4

95.2116.9100.0103.7109.6117.7114.690.0

107.099.4

127.9105.190.0

261.0261.0298.0263.0284.0225.0257.0290.0284.0259.0262.0260.0290.0267.0283.0323.0116.0313.0329.0262.0342.0123.064.0

FMSazimuth

(°E)

127.9127.9134.0124.0117.0115.0117.0117.0117.0117.0124.0117.0134.0134.0134.0120.0120.0120.0120.0120.0120.0134.0126.0126.0125.0124.0124.0126.0126.0126.0126.0163.0161.0147.083.0

217.015.093.0

105.0

Orientationtype

B/AB/AB/MB/AB/MB/IB/AB/AB/AB/AB/MB/AB/MB/I

B/MB/AB/AB/AB/AB/AB/AB/MB/MB/AB/MB/MB/MB/AB/AB/AB/MB/MB/MB/MB/MB/MF/MB/MB/M

Dec(°E)

45.950.758.439.55.5

27.3322.9

18.99.8

204.443.7

8.112.3

338.130.348.234.764.724.162.664.880.636.3

166.228.354.657.767.8

112.865.776.559.717.136.665.714.625.6

330.673.4

Inc(°D)

-42.4^ 0 . 4-55.4-19.8-18.4-22.0-40.0^ 2 . 9-41.2-35.1^-0.9-48.8-36.7-41.8-54.9-48.1^ 3 . 9-38.9-42.7-53.1-49.4-20.8^16.0-40.5-57.9-39.8-34.7-32.3-21.7^-2.5-58.3-54.8-56.4-57.2-41.8-56.9-28.1-65.7-38.0

VGPA

Latitude(°N)

-48.2^ 3 . 7-38.5-50.1-75.1-61.4-56.0-72.7-81.0

41.1-50.1-80.4-78.3-70.0-61.1^16.7-58.4-30.8-68.0-34.8-32.5-12.8-57.1

41.6-61.5-40.0-36.4-26.8

15.8-30.6-24.6-37.4-70.0-55.6-30.4-71.1

44.2-56.1-22.9

Longitude(°E)

106.0104.9123.082.226.374.7

268.2103.898.5

215.9103.8137.882.8

268.2126.4113.1106.2107.2103.3120.5116.8101.0109.5167.6133.6105.3101.9103.3114.7110.5128.5122.3141.4128.2110.0146.5220.3220.9109.1

Dec(°E)

31.036.219.631.70.2

15.3318.6

2.4357.6215.9

33.1350.8

1.6333.7

9.027.816.941.9

3.627.731.476.612.0

183.16.5

27.630.446.4

101.934.616.028.2

2.54.7

61.320.323.7

296.53.0

Inc(°D)

—42.6^ 2 . 0-52.7-20.4

-8.6-20.0-24.9-37.1-34.4-34.2^ 1 . 8-39.4-22.5-25.3^ 7 . 0-47.4-39.5-45.7-30.0-56.2-54.1-27.1-40.5-64.5-53.1-43.6^H.O-32.7^ 2 . 8-38.4-64.8-37.8-27.1-33.9-53.1-43.7-40.8-29.2-65.0

VGPB

Latitude(°N)

-61.7-56.9-70.1-57.2-70.9-70.4^19.5-86.6-85.0

35.2-59.8-81.5-78.3-63.2-80.6-64.5-74.4-52.1-82.0-62.7-60.5-17.7-78.9

20.2-78.2-64.8-62.2-46.4

0.3-58.1-63.3-63.9-80.7-83.5-35.9-71.4-68.3-30.1-66.2

Longitude(°E)

103.8103.7129.077.1

5.355.5

282.045.3

337.6227.5102.9279.7

12.4292.0127.5112.394.5

109.430.4

130.5124.3103.195.9

187.0157.2105.2101.296.4

122.198.5

159.895.519.948.1

120.3105.699.4

269.9179.6

Note: Bedding = dip (in degrees below horizontal) and dip direction (positive clockwise) in sample coordinates (i.e., relative to core liner); Sample azimuth = azimuth (positive clockwise)of feature used for reorientation (bedding plane or fault-dip direction), in core coordinates; FMS azimuth = feature azimuth from Formation MicroScanner log as follows: B = beddingplane, F = fault, M = matched to feature in FMS log, A = average dip direction, I = inferred from adjacent piece. VGP A, "Dec" and "Inc", "Latitude", "Longitude" for VGPA =paleomagnetic declination and inclination, paleomagnetic pole latitude and longitude, from sample magnetic vectors reoriented with individual picks of FMS log feature azimuth(where available) but with no correction for bed dip. VGP B = VGP A, but with correction for sample bed-dip.

783