Exploration Geophysics - Volume 9, Issue 3, 1978

New finite poles have been calculated for the Tasman Sea based on both pre-existing magnetic data and recently released Australian Bureau of Mineral Resources magnetic data. The pole positions are more southerly than previously calculated; the stage pole positions imply that strike slip motion in the northern Tasman Sea was concurrent with accretion in the southern Tasman Sea during the early stages of opening from anomalies 33 to 32. Progressive northward movement of the stage poles from anomaly 32 time reflects the increasing development of spreading ridges and corresponding reduction in strike slip motion in the northern Tasman Sea until accretion was active throughout the entire basin from anomaly 30 time until its cessation at anomaly 24 time. Many of the characteristic features of the southeastern Australian margin, including its steep slope, its narrow width and scarcity of sediments, all difficult to reconcile with previously postulated rifted models, are more consistent with the present model of margin development.

Although the normal mode of rift margin development involves a long period of taphrogenesis prior to separation of the continental blocks, the search for evidence of this stage in the history of the Coral Sea is largely fruitless. What evidence is available is regarded as ambiguous. We further investigate this problem by magnetic modelling of the continent-ocean boundary and by examining the subsidence history of the Queensland Plateau and find little to support the idea of Cretaceous taphrogenesis.

The Papuan Ultramafic Belt (Figure 1A) is one of the best preserved peridotite-gabbro-basalt complexes in the world. The Belt is considered to be an over-thrust sheet of oceanic crust and mantle with a thicker than normal oceanic crust.

Earlier workers have described the overthrust as resulting from north-south compression produced by the northwards movement of the Australian Plate, and a northeasterly dip is indicated by:

• (a)   the sequence of rock types; ultramafic rocks at the base overlain to the northeast by gabbroic rocks which are in turn overlain by ocean basalts.
• (b)   a gravity high offset northeast of the main outcrop of ultramafic rock, indicates the presence of dense rock at shallow depth to the northeast.

The age of formation of the complex is regarded as Jurassic or Cretaceous, and overthrusting occurred some time before the Miocene when shallow water tuffs and agglomerates were deposited on the Belt.

The Belt is cut by one major left-lateral fault and several smaller ones. Reversal of the movement along these faults (Figure IB) suggests that the Belt was originally aligned north-south. Theoretical cross-sections of the Belt in its present configuration were constructed by assuming that it originally dipped at a shallow angle to the east and was subsequently sheared along northwest-trending strike-slip faults. The computed gravity and magnetic anomalies over these theoretical cross sections match the observed anomalies closely. Only small changes in the model are required to produce an acceptable final fit. The good agreement between the theoretical and computed cross sections support the idea that the Belt when originally emplaced had a north-south strike.

The oceanic basalt layer was assumed to produce the pronounced magnetic anomaly which is associated with the Belt. Magnetic models suggest that the inclination of remanent magnetism is about -60°, much steeper than the present field inclination of-31°. The conclusion from the study is that the Papuan Ultramafic Belt was emplaced by over-thrusting from the east rather than the north, and that formation and emplacement occurred some 30’ south of the present position.

Late Cainozoic volcanoes in the Papua New Guinea Highlands overlie cratonic crust but yet produce arc-type volcanics. The data from the study of these supposedly anomalous igneous rocks suggest that a mantle magma source which has been chemically modified during subduction and which has passed through rapid and pronounced changes in tectonic setting may later on be the source of magmas produced during a favourable but non-arc tectonic regime.

Changes in the local magnetic field associated with volcanic thermal activity on Kadovar Island, Papua New Guinea, were measured over a one-year period, 1976/77. The magnetic field changes were highly localised near vents and were quite shallow. They were most likely produced by thermal demagnetisation rather than through piezomagnetic effects.

The Sunda volcanic arc provides a good example of variation in the geochemistry of lavas across an island arc. In addition to the well-known correlation between K20/Si02 ratios and depths to Benioff Zone in Pleistocene-RecentJavas of Java, there are well-defined relationships for ‘incompatible’ elements (Rb, Cs, Ba) and light rare earth elements. Estimated primary magma compositions for individual volcanic centres of Java indicate a progressive change in the conditions of primary basaltic magma production across the arc. The proportion of partial melting of peridotitle upper mantle appears to decrease from about 25% for olivine tholeiite magmas associated with tholeiitic lava series, to 5 – 10% for basanite primary magmas of high-K caic-alkaline lava series. The corresponding depths of final equilibration of magmas with crystalline residues increase from about 30 km to 60 km. These variations in conditions of magma production are probably superimposed upon an increase in the content of K and incompatible elements in the magma sources. The distinctive potassium-rich, strongly undersaturated Pleistocene lavas of northern Java probably originated from magmas produced at greater depth, in mantle enriched in phlogopite. Preliminary liquidus phase relationships for a primitive leucite basanite indicate a maximum depth of mantle origin of about 80 km. Such a composition could not be derived by melting of a Iherzoiite mantle in the presence of H20 alone, and the probable presence of C02 in the mantle is indicated.

An asymmetrical rise, 450 km long, on the Bismarck Sea floor extends between Manus Island and Willaumez Peninsula on the north central coast of New Britain (Figure 1). Northeast of the steeper flank of the rise is the Manus Basin, 2500 m deep, which is deeper and apparently younger than the New Guinea Basin southwest of the rise. The Willaumez-Manus Rise is coincident with a Bouguer gravity trough; Wiiicox (1977) calculated its crust to be 21-25 km thick, in contrast to values of mainly 19-21 km for the two adjacent basins.

The Bismarck sea floor is thought to be a marginal basin (e.g. Karig, 1973). An anomalous feature, however, is the pronounced east-west zone of earthquakes that extends across the sea, and cuts the Willaumez-Manus Rise. The earthquake zone has been regarded as a left-lateral strike-slip plate boundary (e.g. Johnson and Molnar, 1972), but the rise is not obviously displaced by the boundary, and more recent studies have shown that sea-floor spreading is probably taking place along parts of the earthquake zone, particularly in the Manus Basin (Connelly, 1975, 1976; Taylor, 1975).

Quaternary volcanic rocks form the islands in the St Andrew Strait area and the Witu islands at the northwestern and southeastern ends of the rise, respectively. Two possibilities for the origin of the rise may be discussed with regard to the compositions of these volcanic rocks. One possibility is that the rise is the trace of a mantle melting spot. Basaltic rocks on one of the St Andrew Strait islands have compositions similar in many respects to those of oceanic islands, but the most abundant rock type in the area is alkali-rich rhyolite which appears to be the product of partial melting of basaltic crust, implying high geothermal gradients (Johnson et al., 1978). Although St Andrew Strait may overlie an intra-plate melting spot, the Willaumez-Manus Rise is unlikely to represent its trace across the Bismarck Sea floor. There is, for example, no evidence along the rise crest for a chain of volcanoes that are progressively younger northwestwards; and if the Manus Basin is younger than the New Guinea Basin the presence of a melting-spot trace at the southwestern margin of the Manus Basin would be highly coincidental.

Many basaltic rocks in the Witu islands are petrologically similar to the rocks that are found in marginal basins formed by back-arc sea-floor spreading (Johnson and Arculus, 1978). If these rocks are representative of the submarine volcanic rocks that may exist elsewhere along the rise crest, then the rise may be an extinct spreading axis. But this interpretation is also unlikely, because, together with andesites, dacites, and rhyolites, the basalts of the Witu Islands are consistent with progressive changes in chemical composition of rocks from New Britain volcanoes, and these changes are apparently related to different depths to the Benioff Zone that dips northwards beneath the New Britain island arc (Johnson, 1977). Volcanism in the Witu Islands may therefore be dependent upon mantle upwelling above a downgoing slab rather than at a spreading axis. Moreover, a spreading-axis interpretation does not account for the apparent difference in age between the two basins, or for the asymmetry of the rise, and fracture zones orthogonal to the postulated spreading axis have not been identified.

The Willaumuz-Manus Rise does not seem to be primarily a volcanic feature. Its origin is more likely to be tectonic – perhaps uplift of old lithosphere at the edge of the New Guinea Basin near its contact with hotter and younger lithosphere beneath the Manus Basin (Johnson et al. in prep.).

Some of the data from seismic studies, gravity, and the distribution of Quaternary surfaces and deposits in the Solomon Islands are consistent with convergent subduction at the margin of the Pacific and Australian — Solomon Sea Plates, that is, the convergent Benioff Zones meet at considerable depth and are overlain by a small wedge-shaped plate composed of lithosphere. Such convergent subduction probably occurred in the Solomons during much of Late Oiigocene through Quaternary time.

A brecclated alnoite volcanic centre in central-north Malaita is of deep-seated origin and contains xeno-lithic suites that normally characterise kimberlites in continental cratonic settings. There is a wide range of xenolith types. The unique occurrence of garnet Iherzolite (unmodified mantle material) xenoliths is especially exceptional in a supposed island arc situation or oceanic submarine plateau. A pyroxene geotherm calculated from the garnet-bearing nodules indicates a steeper gradient than those constructed for continental kimberlites. A high-temperature inflection, defined by discrete bronzite and subcalcic diopside nodules, is suggestive of a lithosphere thickness of 110 km. These considerations, the age of the Malaita pipe, the crustal nature and thickness of the Ontong Java Plateau (of which Malaita is an edge), its suggested low heat flow values and its geological history support the notion that the Plateau is continental, possibly even a ‘proto-continent’ as suggested by Kroenke (1974).

In eastern Choiseul, Solomon Islands, the mineralogy and structure of basement lavas, grading from pillowed, massive, cleaved to hornblende schist, show similarities with parts of the basement of Guadalcanal and Malaita, especially in their Early Tertiary developmental history. These similarities throw doubt on the frequently proposed idea that Malaita and NE Santa Isabel (as parts of the Ontong Java Plateau) arose in a distant region far removed from the other Solomon Islands.

Basal pelagites deposited within the sequence of basal basic pillowed lavas are found in river sections on northeastern Santa Isabel. These deep-water carbonates are tough and lightly metamorphosed, and have to be studied in thin-section. Thin section determination of the plank tic foraminifera indicate the presence of pelagites ranging in age from Late Paleocene through to Early Oligocene; younger tuffaceous sediments lie above. The lavas, pelagites and other sediments can be correlated with Malaitan sequences to the east and south-east and support the notion that northeastern Isabel is also a part of the central south-west flank of the Ontong Java Plateau. The ages of the pelagites are especially significant to the basal ages obtained in DSDP Sites 288 and 289: it appears that the plateau was created by spreading and that it is more youthful from north to south, that is, it youngs to the south.

Recent palaeomagnetic field sampling of Lower Pliocene sediments and lavas on Viti Leva, Fiji has produced seven new stable sites. These data are comparable with results from nine sites in the same formations previously published by Tarling (1967). Analysis suggests 21° anticlockwise rotation in 4 to 4½ m.y. This is interpreted as active rotation of the whole Fiji Platform, with microplate motion commencing at the close of the Miocene (6 m.y. BP).

Palaeomagnetic samples have been collected on eight islands of the New Hebrides island arc. Data are presented for 31 stable sites out of 46 with ages ranging from Pleistocene to early Late Miocene. These data show 30° clockwise rotation of the arc commencing 6 m.y. BP. Synthetic polar wander paths corresponding to microplate rotation may be computed from known major plate polar wander paths. Thus a physiographically reasonable reconstruction of the New Hebrides plate is shown to have a polar wander path which compares well with the observed data. A reconstruction of a pre-Late Miocene double arc, consisting of the Solomons, New Hebrides- Vityaz, Fiji and Lau-Tonga island arcs, is possible which supports interarc geological correlations and which suggests Plio-Pleistocene growth of the North Fiji Basin byr-r-rtriple junction development concomitant with the development of the Lau Basin.

Lower to Middle Miocene sediments on Maewo contain clasts derived from the ‘Vitiaz arc’, a former tholeiitic volcanic belt similar in age, lithology and geochemistry to Upper Eocene-Middle Miocene rocks on Viti-Levu, Fiji. The configuration of Outer Melanesia from the New Hebrides through Fiji to Tonga-Lau in the Early to Middle Miocene was a double arc compising a frontal arc (Vitiaz arc — Viti Levu — Tonga Ridge) and a rear arc (Western Belt-Lau Ridge) couple. Tectonism in the Middle Miocene brought about inter-arc rifting between the Vitiaz arc-Western Belt couple and the anti-clockwise rotation of Viti Levu into a rear arc position; further fragmentation and the development of marginal basins had also affected the Tonga-Lau couple by the Early Pliocene. Mio-Pliocene lavas of the New Hebrides Eastern Belt were erupted in the Vitiaz arc-Western Belt inter-arc basin in response to further rifting and crustal thinning associated with the developing proto-North Fiji Basin. Expansion of this basin was accompanied by subduction reversal from the eastern to western margins of the New Hebrides Ridge which then migrated south-westwards to its present position. The present configuration of the New Hebrides Ridge is atypical of that usually attributed to SW Pacific arc systems.

The results of various recent geophysical studies suggest that the southern New Hebrides Basin and the western South Fiji Basin were once a single marginal basin. The basin originated above an ancient subduction of the Indian Plate under the Pacific Plate, active during Eocene-Oligocene. The ‘fossil’ subduction zone stretched along the western side of the Loyalty Rise and the Three Kings Rise both of which would have then belonged to the same volcanic arc.

The Pacific Plate is being consumed under the Lau-Tonga Ridge down a subduction zone dipping at 45°. The trench inner wall has ultramafic rocks exposed, possibly within the accretionary wedge or prism. The fore-arc basin has a pre-Late Eocene basement overlain by 3-5 km of Late Eocene to Recent sediments. Plio-Pleistocene volcanics occur at the rear of the present fore-arc zone. The Lau basin was formed by rifting apart of the former Lau-Tonga Ridge between the fore-arc and the volcanic arc in the Early Pliocene, the Lau Ridge (the old volcanic arc) is capped by a line of volcanic piles certainly as old as Miocene but probably extending back to the Late Eocene. The size of the accumulated pile increases towards the north. The western flank of the Lau Ridge was built by a volcanic accumulation of over 3 km thickness on oceanic crust of the South Fiji Basin formed during the Ofigocene, This sea floor postdates the inception of arc volcanism. In the discussion it is suggested that (a) the region has had high heat flow since the volcanism started and most of the magma has come from the mantle wedge overlying the downgoing slab, (b) the volcanism was initiated in the Tonga region over a steeply dipping Benioff Zone, (c) the locus of volcanism moved rearwards during the Late Eocene as the dip of the Benioff Zone became gentler, (d) the crust of the region reached its present thickness by (i) addition of sediment to the Tonga Ridge after the first volcanic phase and volcanic activity on the Lau Ridge and (ii) intrusions into the deeper crust, (e) the crustal thickening by both processes increased to the north as subduction rates increased away from the Pacific-India plate poles of relative rotation.

Palaeomagnetic studies in the New Zealand region have lagged behind those in eastern Australia, due mainly to the more complex problems posed by tectonic rotations (about both horizontal and vertical axes) associated with the late Cainozoic Kaikoura orogenic movements. These tectonic movements affect principally a 50 km-wide belt either side of the present Indian-Pacific plate boundary (Alpine Fault in part). Within this tectonically disturbed belt, Cretaceous-Cainozoic volcanic rocks are generally unsuitable for palaeomagnetic work because of varying degrees of magnetic instability and uncertainty in estimation of tilt corrections.

By concentrating on undisturbed or gently tilted volcanic sequences away from the Alpine Fault some progress has been made over the past five years. At the Chatham Islands on the eastern end of the Chatham Rise, Late Cretaceous (70-80 m.y. BP), Eocene-Oligocene (35-40 m.y.) and latest Miocene-Pliocene (5 m.y.) alkaline basaltic vo/canics lie flat and have provided data for an apparent Polar Wander Path for the eastern New Zealand region (part of Pacific Plate) over the past 75 m.y. (Grindley et. al., 1977). Closer to the Alpine Fault at Oamaru in North Otago, Late Eocene-Early Oligocene basalts have provided good N.R.M. data (Coombs and Hatherton, 1959) from which a pole position has been calculated which compares closely with the age-equivalent pole position from the Chatham Islands. At Mt Somers in Mid-Canterbury only 75 km from the Alpine Fault, an excellent 95 m.y. pole position has recently been obtained from early Late Cretaceous caic-alka/ine andesitic to rhyolitic lavas, tuffs and ignimbrites (Oliver, 1977; Oliver et al., in press., and compared with the apparent Polar the New Zealand Polar Wander Path provided no rotation is allowed between Mt Somers and the Chatham Islands.

West of the Alpine Fault, few reliable results have been obtained due to strong tectonism and related magnetic instability. In the Buller gorge area of North Westland, early Late Cretaceous (80-90 m.y.) basalts, iamprophyres and acid tuffs have recently yielded promising results comparable with those at Mt Somers which are still being investigated.

Younger rocks in the New Zealand region have yielded pole positions indistinguishable from those of the present day. The Upper Miocene Dunedin Volcanics have provided N.R.M. data from which a pole position close to the present axial dipole can be inferred (Coombs and Hatherton, 1959). The Upper Miocene Akaroa Volcanics and the Stoddart Formation of the Lyttleton Volcano have yielded results statistically different from the present axial dipole (Evans, 1970) but in the case of the Stoddart Formation at least, this can only be regarded as a virtual geomagnetic pole, secular variation not being eliminated. Ignimbrites and andesites of the Taupo Volcanic Zone, Auckland and Taranaki all less than 1 m.y. old have yielded a pole within 8° of the present pole with a circle of confidence of T (Cox, 1969).

All the reliable New Zealandpalaeomagnetic data are shown in Table 1 using the conventions of McElhinny (1973). The significant palaeopoJe positions are plo tted on Figure 1 (after Oliver et al., in prep.).and compared with the apparent Polar Wander Path for Eastern Australia (McElhinny, 1973). The New Zealand Polar Wander Path lies some 1000-1500 km east of the Australian Polar Wander Path, the differences corresponding to the finite rotational displacements across the Indian-Pacific and other plate boundaries (such as the Tasman spreading) over the past 90 m.y. Finite poles for these rotations lie on great circles bisecting the lines joining age-equivalent positions on the separate Polar wander paths (Figure 1).

Magnetic anomalies over the Campbell Plateau show up a linear belt of positive anomalies ex tending for 900 km with an approximate east-west trend. This belt is identified as part of the Stokes Magnetic Anomaly System which extends throughout New Zealand and is associated with rocks of the New Zealand Geosyncline. The system defines a trans-current fault with a dextral displacement of 330 km which can be traced from the SW tip of the Campbell Plateau to the Bounty Trougt’ in the north. Restoration of this Campbell Fault requires the Bounty Trough to be a rift feature, prooably associated with Indian-Pacific plates’ movements during the Miocene or Oligocene. The e

Site 208 from Leg 21 of the Deep Sea Drilling Project con tains a well preserved and almost complete sequence of Pliocene and Pleistocene planktic foraminifera and calcareous nannoplankton. The upper 42.5 m of sediment has been analysed at 50 cm intervals. Several major cold episodes are recognised. The strongest occur at 39 m, 27 m and about 11m. The first of these may represent the onset of major continental glaciation in the Late Pliocene, about 3.0 m.y. ago. Oceanic palaeo-temperatures were found to vary by at least 8°C, implying a latitudinal water mass movement of about 10°. The biostratigraphy has strong affinities with Tropical zonations, except that two species whose extinctions are used elsewhere as datum planes were found to occur here much higher in the sequence. They were Globorotalia multicamerata and Globigerinoides fistulosus. Four coiling reversals were found in Pulleniatina obiiquiloculata in the Pleistocene and were apparently unrelated to climate. Dates for these reversals agree closely with those found elsewhere. They may be useful datum points for Early Pleistocene correlation.