MARINE GEOLOGICAL INVESTIGATIONS IN THE SOUTHWEST PACIFIC AND ADJACENT AREAS
COMMITTEE FOR CO-ORDINATION OF JOINT PROSPECTING FOR MINERAL RESOURCES IN SOUTH PACIFIC OFFSHORE AREAS (CCOP/SOPAC)
G. P. GLASBY AND H. R. KATZ
(Editors)
PAPERS PRESENTED AT THE I.D.O.E. WORKSHOP SUVA, FIJI, 1-6 SEPTEMBER, 1975
1976
LIST OF CONTRIBUTORS
Andrews, J. E. Department of Oceanography University of Hawaii Honolulu, Hawaii '96822, U.S.A. Archer, A. A. Institute of Geological Sciences Exhibition Road London SW7 2DE, ENGLAND Burns, R. G. Department of Earth and Planetary Sciences Massachusetts Institute of Technology Cambridge Massachusetts 02139, U.S.A. Coleman, P. J. Geology Department University of Western Australia Nedlands Western Australia 6009, AUSTRALIA Cronan, D. S. Applied Geochemistry Research Group Imperial College of Science and Technology London SW7 2BP, ENGLAND Davies, H. L. Geological Survey P.O. Box 778 Port Moresby PAPUA NEW GUINEA Friedrich, G. H. Technische Hochschule Aachen 51 Aachen, WEST GERMANY Glasby, G. P. New Zealand Oceanographic Institute Department of Scientific and Industrial Research P.O. Box 12-346, Wellington, NEW ZEALAND Goodell, H. G. Department of Environmental Sciences University of Virginia Charlottesville Virginia 27903, U.S.A. Hawkins, J. W. Geological Research Division Scripps Institution of Oceanography University of California La Jolla California 92093, U.S.A. Katz, H. R. New Zealand Geological Survey Department of Scientific and Industrial Research P.O. Box 30-368 Lower Hutt, NEW ZEALAND
Melguen, M. Centre Oceanologique de Bretagne B.P. 337, Brest 29273, FRANCE Meylan, M. A. Offshore Exploration Shell Oil Company P.O. Box 60124 New Orleans Louisianna 70139, U.S.A. and Department of Oceanography University of Hawaii Honolulu, Hawaii 96822, U.S.A. Mizuno, A. Geological Survey of Japan 135 Hisamoto Kawasaki-shi, JAPAN Monzier, M. O.R.S.T.O.M. B.P. A5 Noumea Cedex, NEW CALEDONIA Moorby, S. A. Applied Geochemistry Research Group Imperial College of Science and Technology London SW7 2BP, ENGLAND Moritani, T. Geological Survey of Japan 135 Hisamoto Kawasaki-shi, JAPAN Mutter, J. C. Bureau of Mineral Resources P.O. Box 378 A.C.T. 2601, AUSTRALIA Pautot, G. Centre Oceanologique de Bretagne B.P. 337 Brest 29273, FRANCE Taylor, G. R. Department of Geology University of New England Armidale New South Wales 2351, AUSTRALIA
FOREWORD
During the period 1-6 September, 1975, a workshop was held in Suva, Fiji, under the auspices of the Committee for Co-ordination of Joint Prospecting for Mineral Resources in the South Pacific Offshore Areas (CCOP/SOP AC) to consider in detail two topics: (a) "Regional and specific studies of the structure of the oceanic-island arc-continental margins of the south-west Pacific", and (b) "Occurrence, mode of formation, and environmental factors of manganese nodule deposits, hot brines, and metalliferous muds in the south-west Pacific." Sixty participants from 16 countries took part in the workshop. The first two days of the workshop were devoted to the presentation of twenty lectures by invited experts in order to give a detailed background understanding of the various factors relevant to each topic. The remainder of the workshop was then devoted to discussion sessions in which key projects for further substantial work on the structure and mineral resources of the South Pacific were identified and evaluated. For this purpose, delegates were divided into three committees dealing with 1. Marginal and interarc basins, 2. Convergent margins, and 3. Manganese nodules/metalliferous sediments. These sessions resulted in a comprehensive set of proposals for future research in the South Pacific. The resulting workshop document was unanimously adopted at the 4th Session of CCOP/SOPAC held in Honiara, British Solomon Islands, during 8-17 September, 1975, and is published as part of the Proceedings of the 4th Session of CCOP/SOPAC and as the Intergovernmental Oceanographic Commission Workshop Report No.6. In this volume, individual background papers presented at the workshop are reproduced. All the papers dealing with manganese nodules and related topics presented at the workshop are included here, but only four of the structural papers and one abstract were available for publication in this volume. In addition, a number of papers not presented at the workshop were contributed for inclusion in this volume. These are marked with an asterisk in the list of contents. As editors of this volume, we believe that this workshop was of considerable value in setting guidelines for future work in the South Pacific and are optimistic that it will serve not only as a focus but also as a stimulus for future work in this region. Finally, we would like to thank the International Oceanographic Commission and the International Decade of Ocean Exploration who funded the majority of the delegates to this workshop and the organising committee of Dr L. W. Kroenke, Dr G. P. Glasby, Professor P. J. Coleman and Mr R. N. Richmond for helping make this conference a success. The Australian government generously donated funds for the printing of this volume.
G. P. G. and H. R'. K.
TABLE OF CONTENTS Page 1 Aspects of the Geological Setting of Manganese Nodule Deposits - J. E. Andrews
14 Metallogenesis of Transition Elements at Convergent Plate Margins in the South-west Pacific - R. G. Burns 21 Prospects for the Exploitation of Manganese Nodules: The Main Technical, Economic and Legal Problems - A. A. Archer 39 Manganese Micronodules in Deep-sea Sediments and their Relation to Manganese Nodule Fields - G. H. Friedrich
54 Deep Bottom Currents, Sedimentary Hiatuses and Polymetallic Nodules _ G. Pautot and M. Melguen * 62 Some Results of Surveys for Manganese Nodule Deposits in the Pacific Ocean by the Geological Survey of Japan - A. Mizuno and T. Moritani 80 Tertiary Manganese Mineralisation in Japan - T. Moritani 83 Styles of Mineralisation in the Solomon Islands - A Review - G. R. Taylor 92 A Comparison of the Morphology and Mineralogy of Manganese Nodules from the Southwestern Pacific Basin and Northeastern Equatorial Pacific - M. A. Meylan 99 Chemical Composition of Manganese Nodules from the Pacific-Antarctic Ocean, Drake Passage and Scotia Sea: Relation to Ferromanganese Oxide Mineralogy and Nucleus Type - M. A. Meylan and H. G. Goodell* 118 Preliminary Results of Renewed Investigations on Manganese Nodules and Encrustations in the Indian Ocean - D. S. Cronan and S. A. Moorby 124 Manganese Nodules and Encrustations in the Vicinity of New Caledonia and the Loyalty Islands - M. Monzier* 129 The Relation Between Basin Evolution and Marginal Plateau Subsidence in the Coral Sea - J. C. Mutter 133 Papua New Guinea: Geology and Tectonics (Abstract) - H. L. Davies
134 A Re-Evaluation of the Solomon Islands as an Arc System - P. J. Coleman
141 Tectonic Setting and Petrology of Samoa-Tonga-Fiji Region - J. W. Hawkins 153 Sediments and Tectonic History of the Tonga Ridge, and the Problem of the Lau Basin - H. R. Katz
ASPECTS OF THE GEOLOGIC SETTING OF MANGANESE NODULE DEPOSITS
J. E. ANDREWS
The factors controlling the distribution of The tectonic elements which form the manganese nodules in the world oceans may setting of Pacific nodule deposits fall into four be seen most simply as: 1) the supply of general catagories: the East Pacific Rise _ constituent materials (source and rate): 2) the active spreading centre; abyssal hills _ the modes of supply or incorporation: and 3) the dominant (or "average") basin the processes which interfere with the first topography; fracture zones _ the. signature two. This paper will deal with the patterning of sea floor spreading history and zones of or geologic setting of nodule deposits, and the crustal weakness; and volcanic features _ relationship of these patterns to the more seamounts, guyots, and volcanic islands. complex questions in an area which falls Abyssal hills cover some 85% of the Pacific mainly within category three. basin floor and are the "normal" site of From a geological standpoint, the con- nodule occurrances. Three types of abyssal ditions under which nodules have been, and hills have been described in the literature. are being, formed derive from plate tectonics Two of these are generally linear with relief considerations. That is to say that the of 50-250 m. They are oriented parallel to the proximity of a deposit site to volcanic centres, crest of the mid ocean ridge at which the the depth of the sea floor (and therefore lithospheric plate was generated and are in relationship to calcium carbonate fact primary topography of the second layer compensation depth), the latitude or palaeo- of the oceanic crust (layer 1 being sedi- latitude, sedimentation in terms of both types mentary). and rates of deposition,. and the evolution of The primary hills form at the ridge crest as the structures underlying the deposits - linear volcanic structures during the abyssal hills, fracture zones, and volcanic generation of the plate (Luyendyk, 1969). A features - are all directly related to the second group of hills forms as the lithosphere generation of the sea floor (lithosphere) at moves down the ridge flank, undergoing ridge crests and its subsequent motions. This thermal contraction and subsidence. These is basically the generation of crustal material hills are also linear and parallel to ridge crest and its migration down the ridge flank. This (Luyendyk, 1969). Fig. 1 (after Luyendyk, motion exposes the sea floor and the 1969) illustrates such a set of hills as mapped developing nodule deposit to a variety of by the S.1.0. Deep Tow system. Such sys- oceanic environments. terns are necessary for accurate definition of From the oceanographic standpoint, one slopes and dimensions at these scales. Fig. 2 must consider the surface currents and the documents the existence of this pattern on a bottom currents (benthic boundary layer) regional basis from a sidescan sonar and their respective evolutions. From these evaluation of topographic patterns (Andrews are derived changes in biogenic productivity, et al., 1975). A third category of abyssal hills both in type and distribution, the zonation of occur in the vicinity of fracture zones as the this productivity, and the consequent flux of result of stresses imposed on the plate by particulate matter to the sea floor. The on- motion of the fracture zone (Andrews, going evolution of the ocean basins deter- 1971). This motion arises from the motion of mines access of these phenomena to different plates on a sphere during the sea floor regions within the basins. spreading process. Motions are rotations of
1
the north island of New Zealand. and tectonic evolution of the zones exerts a , Seamounts, guyots, and volcanic islands considerable influence on nodule develop are distinctive features of great relief, and ment. This appears to be also true for a group have had obvious periods of volcanic activity of fracture zones which are less well defined. - often postdating the formation of the Where the original transform fault at the seafloor in the region. They occur singly and ridge crest had little offset of the ridge crest, in groups, and present a unique environment the resulting fracture zone will show little or due to the depth ranges, availability of no relief in many cases. The only evidence volcanic materials, and the accelerating effect existing may appear as a magnetic signature, of the topography on oceanic currents. Ferro- a zone of disturbed basement, or a series of manganese crusts or pavements (generally oblique (type three) abyssal hills (e.g. cobalt enriched) are normally found instead Handschumacher & Andrews, 1975). of nodule fields in these areas. Without a specific survey this type of zone The East pacific Rise is somewhat dif- will seldom be clearly defined, and as a ferent in that the area is geologically young, consequence its effects on the local nodule with continuing volcanism as a result of population will not be recognised. lithosphere generation. Being well above the It is pertinent here to comment on two carbonate compensation depth, pelagic questions which will ultimately playa major carbonate sediments accumulate as distance role in the definition of the actual influence of from the ridge increases. The basal sediment seafloor and oceanographic history on nodule layer has been shown to be generally fer- growth and origin. These are the questions ruginous clays - products of hydrothermal nodule growth rate and nodule diagenesis. In activity associated with the on-going the case of growth rate, the preponderance of volcanism. Ferromanganese deposits appear data supports slow rates of oxide accumuto form rather rapidly in this environment, lation - ca. 1 mm/l06 yrs (Ku, in press). and crust or pavement type deposits are However, some much faster rates have been common, encrusting lava flows. reported. These are generally for shallow Fracture zones are linear features which water deposits - near volcanic islands or originate as transform faults offsetting a ridge seamounts (Lalou et al., 1973; Morgenstein, crest. The major Pacific fracture zones show 1973), for material from active ridge crests rugged topography and a marked change in (Burnett and Morgenstein, in press), and for depth of the sea floor across the zone. Relief rare material from the deep basin - possibly of the fracture zones becomes less with associated with fracture zone activity increasing distance from the spreading ridge. (Burnett and Morgenstein, in press). In the Thermal contraction of the lithosphere with case of nodules from the vicinity of the aging brings seafloor depths towards a Mendocino fracture zone rates of 40 to 81 common base level. As mentioned above, it mm/106 yrs are reported. With data such as appears that the fracture zones may act as these in mind - a suggestion that volcanic zones of volcanic activity and sources of influences may greatly enhance the rates for supply to the nodule environment. This is growth - it is possible to suggest that it is the due to continuing strike-slip motion on the average growth rate which is slow, and that fracture zones to accomodate the increases in instantaneous rates are high and intermittant. linear spreading rates which occur at This then might also be reflected by the increasing distance form the pole of relative zoned structure of nodules. If the structure is motion. The major fracture zones are well a direct function of this intermittant input, defined and nodules from these areas then all nodules in a region (size unspecified) generally show cobalt enrichment. In two should show similar internal structure. From cases, this is uniquely on the "down thrown" a geologic point of view each of these possi- side of the zone (Andrews, 1971, and bilities has a rather different effect on a dis- unpublished date from the Murray Fracture cussion of nodule genesis. If rates are fast Zone). It is clear that the on-going volcanic then we must expect that all existing nodules 5
are a product of post-Pliocence growth, and hydroxides. These then reorder through time therefore that they are a reflection of to a-Mn02, birnisite, and todorokite. In turn, oceanographic conditions essentially as they the lattice structure of todorokite as exist today. interpreted by Burns (1975) would provide On the other hand, slow growth rates the most stable sites for copper and nickel would indicate that the nodules have ions, and therefore as the todorokite devel- performed the feat of continually climbing to oped so would its ability to absorb Cu and Ni. the top of a relatively more rapidly accumu- In this case nodule enrichment would be a lating pile of sediment - in itself not an function of nodule aging as well as of the obstacle to credibility. At the same time the events supplying the metals. That is, metal nodules have been exposed to a continually supply is necessary, but not sufficient for changing environment of depth, enrichment to occur. Here nodule structure sedimentation rate, sediment mineralogy, provides no clue to the environment of volcanism, and particulate flux from the sea growth, and reflects only the sum of events surface. What we read in the nodules as bulk that each nodule was capable of absorbing properties are then the average of all of these (literally and figuratively). changing inputs. We may therefore look to It should be reiterated that at present our the internal structure as a growth history. knowledge of nodule distribution and relation The same is true if growth rates are fast and to geologic setting is such that the patterns of intermittant, although here the input events bulk properties are in agreement with the might also be expected to appear well defined present marine geologic environment. in the stratigraphic column. The events The gross pattern then is rather simple. In would have to trigger, as well as promote, the North and South Pacific, nodules occur in growth. The inclusion of sediments in layers all areas of low to moderate sedimentation within nodules lends attraction to this idea, rates - predominately on red clays and and work to examine the isotopic com- silicious oozes. Deposits are densest in the position of this material may assist in North Pacific where sedimentation rates are unravelling the problem. lowest and average sea floor depths are The interesting matter here is that, by and greatest. In the South Pacific, deposits are large, the recognised gross patterns of the most abundant at or below the carbonate distribution of nodule bulk properties fit compensation depth on the flanks of the East within the framework of present oceano- Pacific Rise. Deposits are also concentrated graphic conditions. That is to say, on the within fracture zones where depths are basis of bulk measurements, there is no greater, local volcanism more frequent, and strong reason to suspect anything other than often where there exists access for the flow of a post-Pliocene history for nodules now Antarctic Bottom Water. Nodule Deposits on the seafloor. Definitely this is in conflict are denser in the Western South Pacific with radiometric age determinations (Ku, in (Glasby and Lawrence, 1974; Meylan et al., press). 1975), except where continental sedi- An alternative solution, which raises its mentation interferes as near New Zealand. In own problems, is the possibility of nodule the Antarctic, dense nodule fields are found diagenesis. That is, that internal structure is beneath the Antarctic circumpolar current. the result of internal re-ordering. Here, no These rest on a major erosional specific inter-nodule stratigraphic patterns unconformity which is a result of the current would be expected to occur and structure activity. In the Western North Pacific, man- would have no direct relation to growth. ganese crusts and pavements are common on Burns et al. (1975) and Meylan (1975) have the numerous seamounts, island slopes, and proposed, from microscale and bulk regional guyots. In general, all of these deposits show studies respectively, that the iron and low values of included copper and nickel manganese in nodules are initially deposited (Horn et al., 1973). Nodules showing as disordered iron and manganese oxy- enrichment in these elements lie principally
6
along the east-west band between the Clarion diagenesis leading to enrichment. In the first and Clipperton fracture zones in the case, no consideration of plate motions is equatorial Northeast Pacific. The band needed as it will have no effect. In the second appears to extend into the western portion of case, no consideration of plate motions is the basin, but rough topography has limited needed as only the most recent environment commercial interest and sampling efforts is reflected. A third possibility, of course, is have therefore been limited. The influence of that the patterns of oceanic productivity and increasing sedimentation rate is clear. The particulate flux are not the mechanism occurrence ofCu-Ni enriched nodules along a involved in nodule enrichment and we must band coinciding with the north-equatorial look to the underlying crust for clues to the productivity band also looks reasonable. The enrichment of this one specific band. This is association with productivity does not, how- the least satisfying of the possibilities. Testing ever, extend to the areas of the Antarctic di- these questions will require continued vergence. The abundant nodules there are regional sampling and specific sampling of relatively poor in copper and nickel (Meylan individual environments in great detail. At and Goodell, this voL). The denisty of these present the data base in both number and deposits may be influenced by the rapid types of samples is inadequate. bottom currents in the region. It appears that The local or small-scale variability in this association of current flow and nodule nodule deposits is large. Often bulk patterns occurrences can be extended to areas such as may vary to the extremes of the values for a the Samoan Passage, and even to hill slopes given region over distances of 1 km. or less. of the equatorial zone exposed to prominent These variations may be related to local currents. Most likely, this reflects the effect structure and its influence on the of inhibiting sedimentation by transport or mechanisms supplying the materials. Under- erosion. It must be noted, however, that a water photographic sequences routinely show major erosional unconformity in the South- these patterns physically and a set has been west Pacific shows no trace of present or selected from a station in the copper-nickel previous nodule occurrences. This is the enriched belt to demonstrate the patterns. Eocene-Oligocene unconformity traced by Camera station number 20f cruise Kana the Deep Sea Drilling Project (Burns, Keoki Mn 7201 was taken moving down Andrews et al., 1973; Kennett et al., 1972). slope on the south facing slope of a 250 m. This unconformity appears to represent the high abyssal hill at 130 12' N 1400 48' W. introduction of Antarctic Bottom Water into The station covers a distance of 4.8 km., the Pacific basin. The fact that no nodules moving downslope from a depth of 4850 to occur on this surface suggests either that 5000 m. Photographs in Figs. 6 and 7 are none was forming in the Oligocene or that numbered by frame number from the 500 they were later remobilised by pore water frame spool. At a time interval of 12 seconds circulation. the distance between frames should average So, in terms of regional patterns, nodule 10 m. This is an average, however, and distribution fits best with present oceano- several frame sequences provide nearly full graphic conditions - with a large number of overlap coyerage (e.g., frames 70 and 71 in question areas. Reconstruction of these Fig. 6). patterns to allow for lithospheric plate Near the hill crest, nodule populations are motions is not immediately satisfactory, small and rock fragments are sometimes seen particularly the southward shift required for protruding from the sediments. Scour moats the equatorial belt. If this is a function of around these are evidence of current action equatorial productivity, then enrichment (e.g., frames 3 and 4). On the slopes should drop off toward the centre of the outcroppings of basalt (normally coated with nodules. This trend is not found. ferromanganese oxides) are seen, together This tends to leave us with the position of with dense nodule fields. In many instances either rapid nodule growth or nodule individual lava flows are identifiable (frames
7
nodules which were collected from the speculated previously, we may label the seafloor on or adjacent to seamounts and results of average geologic and oceanographic fracture zones. Clusters Band C show a influences, as well as those of local or specific more scattered correlation - cluster B influences, and that the two are producing showing some correlation to the "chemical" two distinct populations of nodules (and a factor and cluster C showing some large indistinct group where they overlap). At correlation to the "physcial" factor. These the same time, we may not specify their samples (and those of the undefined group) nature from these data - only that they are are from areas of hills which may be related equally important in development of nodule to fracture zones of low relief or small deposits. volcanic centres. These are areas such as the "Clapperton" fracture zone between the ACKNOWLEDGEMENTS Clarion and Clipperton fracture zones, and The State of Hawaii Marine Affairs are more clearly hypothesised than mapped. Coordinator, the I.D.O.E. - N.S.F. Thus, the geologic influences on these Manganese Nodule Project, and the samples cannot be clearly defined at present. Alexander von Humboldt Foundation are What is clear here is that, as has' been thanked for their support of this study.
REFERENCES
ANDREWS, J. E. 1971: Abyssal hills as evidence of ANDREWS, 1. E., CRAIG, 1. and HARDY, W.A. transcurrent faulting on North Pacific Fracture Zones., 1974: Project Swathmap: First results. Hawaii Inst. Bull. Geol. Soc. Am. 82: 463-70. Geophys. Rep. 74-4: 14 pp + 11 Figs. ANDREWS, 1. E. and MEYLAN, M. A. 1972: Results BURNS, R., ANDREWS, 1. et al. 1973: Initial Reports of bottom photography: Kana Keoki cruise Manganese of the Deep Sea Drilling Project, Vol. 21. U.S. Govt. '72. Hawaii Inst. Geophys. Rep. 72-23: 83-111. Printing Office, Washington, D.C. ANDREWS, 1. E. et al. 1974: Ferromanganese deposits BURNS, R. and BURNS, V. 1975: Crystal structure of of the ocean floor: cruise report Mn-74-01 R/V Moana marine manganese Minerals. Geol. Soc. Am. Annual Wave. Hawaii Inst. Geophys. Rep. 74-9: 194 pp. Meeting Abstracts.
12
METALLOGENESIS OF TRANSITION ELEMENTS AT CONVERGENT PLATE MARGINS IN THE SOUTHWEST PACIFIC
ROGER G. BURNS
Abstract
Oceanic lithosphere contains significant amounts of transition metals, which occur in basaltic rocks and pillow lavas, in overlying pelagic sediments and ferromanganese crusts and nodules, and in underlying cumulates and upper mantle rocks. These transition elements become remobilised and fractionated during partial fusion at depths along subduction zones, leading to the formation of metalliferous deposits in island arc and cordilleran belts delineating convergent plates. The metallogenesis and fractionation patterns of the transition elements may be explained by crystal field theory . The southwest Pacific is a region of Tertiary and present-day convergence. In early Tertiary times, the Pacific plate was subducted beneath the Australian plate, leading to the formation of porphyry copper deposits in calc-alkaline igneous rocks and of nickel laterite deposits in highly weathered obducted ophiolite suites along the island arc chains. Subsequently, the subduction zone has changed polarity, so that the Australian plate is now being over-ridden by the Pacific plate. This is producing manganese mineralisation in many areas, which may be indicative of porphyry copper deposits at greater depths along the island arc belts of the SW Pacific.
INTRODUCTION 1975). We are now just beginning to In assessing the mineral potential of a understand the origin of certain metal region such as the southwest Pacific, it would deposits within continents and at their be desirable first to be able to interpret the margins. It seems appropriate, therefore, to origin of documented world-wide ore de- review some of the concepts of plate tectonics posits and then to predict the location of new theory which have a bearing on the genesis of ores in the region of interest on the basis of ore deposits, to mention some of the well-established geological principles. This is successes and complexities of the concepts, the aim of the present paper. and to relate the processes to known In recent years, the theory of plate examples and possible future deposits likely tectonics has had a profound impact in to be found in the oceanic islands and the geology; data from diverse branches of the seafloor of the SW Pacific. earth sciences - seismic, magnetic, oceano- graphic, tectonic, petrologic, geochemical, Plate Tectonics - The New Global Tectonics mineralogical, palaentologic - have been Theory brought to bear on developing a unified The essential features of plate tectonics, explanation for many features of the earth's the new theory of global tectonics, are that surface that were formerly considered as the earth's surface is divided into six larger independent phenomena. Plate tectonics has and several smaller lithospheric plates which also had an enormous impact on economic are generated (accreted) at oceanic ridges by geology: a large body of literature on uprise of subcrustal or mantle-derived relationships between metal deposits and basaltic magma, move laterally (sea floor plate tectonics has appeared during the past spreading) away from the oceanic ridges, and three years (for example, Mitchell and are destroyed or consumed along subduction Garson, 1972; Sillitoe, 1912a, b; Sawkins, zones when plates converge and one over- 1972; Guild, 1972, 1974; Rona, 1973; rides the other. Confrontation of an oceanic Mitchell and Bell, 1973; Hurley, 1975; Titley, plate with a continent, such as the Pacific
14
Plate with South America, gives rise to a polarity of subduction zones change over Cordilleran or mountain belt parallel to the geologic time. The Benioff zones, island arcs, margin. Convergence of two oceanic plates, and trench systems grow outwards as new such as the Pacific and Australian or Asian crustal material is formed on the earth's plates, gives rise to an island arc system. This surface by volcanic. activity above the characterises most of the Western Pacific, subduction zone. At the same time, the including much of the present day SW Pacific inclination of the Benioff zone may change, region. A third type of collision results at and reverse in polarity. This has happened, in transform margins when two plates slide past fact, near the Australian continent in the one another without accretion or consump- past. We shall have more to say about this tion. These produce huge transcurrent fault aspect later when the nickel deposits on New systems such as the Alpine Fault in New Caledonia are discussed. Changing position Zealand and the San Andreas Fault system of the Benioff zone also affects the location of in California. porphyry copper deposits and manganese A whole new jargon has arisen as plate mineralisation, which also is discussed later. tectonics theory has developed. Many of ' these concepts and type examples are derived Composition of Oceanic Lithosphere Born at from the SW Pacific area (for example, Isacks an Oceanic Rise et al., 1969; Karig, 1971; Karig and Let us examine more closely the Mammerickx, 1972; Oliver et aI., 1973; composition of material accreted at oceanic Macdonald et al., 1973; Coleman, 1975). rises which becomes oceanic lithosphere Thus, in addition to geological features doomed to destruction along subduction mentioned already, there are spreading zones. The most obvious material which is centres, triple junctions, trenches, outer arcs, seen or sampled at the crests of mid-oceanic arc-trench gaps, marginal basins, third arcs, rises is basaltic rock. The basaltic magma obducted lithosphere, ophiolites, and Benioff formed by partial melting of the upper zones which are currently active subduction mantle pours out onto the ocean floor as zones highlighted by intense earthquake pillow lavas. As seafloor spreading occurs and generation throughout the dip of the the new oceanic crust moves away from the downgoing plate. We shall discuss some of oceanic rise, ferromanganese crusts are these concepts later. Two important features deposited as well as a variety of sediment to note are: first, island arcs and their types. Particularly noteworthy for our associated deep trenches are generally discussion of subduction processes in the SW concave towards the side of the over-riding Pacific is that, in certain areas of the seafloor, plate. This is a geometric constraint of metalliferous sediments are deposited. In the colliding shells on the surface of a sphere. vicinity of the East Pacific Rise, for example, The Tonga-Kermadec trench system with its (where oceanic lithosphere is born and is predominantly linear trend is unique in this later subducted along the Tonga-Kermadec respect. Along this trench system, the Pacific Trench and Peru-Chile Trench systems), plate moving west is subducted beneath the sediments enriched in Mn, Fe, Ni, Cu, Zn, Australian plate. However, the New Co and Pb are deposited, as shown by Hebrides, Torres, and Solomon trench extensive work in the Bauer Deep located systems portray the Australian plate moving between the East Pacific Rise and the rapidly north and currently being subducted Galapagos Rise, a fossil spreading centre (for beneath the Pacific plate to the north. As a example, Dymond et 01., 1973; Cronan and result, the tectonics of the SW Pacific, Garrett, 1973; McMurtry and Burnett, 1975; particularly beneath the Fiji plateau, is Bonatti, 1975). It is generally believed that complicated by these oppositely facing such metalliferous sediments are formed by subduction zones between the converging circulating hydrothermal fluids, perhaps Australian and Pacific plates. involving seawater as it seeps through the Secondly, the position, inclination, and volcanic rocks formed at mid-oceanic rises.
15
One example is the stratiform iron- the top downwards, ophiolite suites (oceanic manganese deposits presently forming at lithosphere) consist of thin sequences of Matupi Harbour, New Britain, which are ferromanganese sediments, chert, argillite, thought to originate from seawater circulating and pelagic limestone. This corresponds to down through volcanic rocks and bringing layer 1 of the oceanic crust. It is underlain by dissolved Fe and Mn to the surface predominantly extrusive tholeiitic basalt, (Ferguson and Lambert, 1972). Many areas some hyaloclasite and alkali basalt, which of present day deposition of metalliferous commonly display pillow structures and sediments near oceanic rises are associated conversion to spilite. These basaltic rocks with high heat flow. The Fijian plateau is comprise layer 2 of the oceanic crust. Layer 3 believed to have formed by spreading from below the layer 2 basalts consist of grabbo the Nova Rise which strikes approximately and trondhjemite. Frequently, the lower part north-south along longitude 174°E, and also of layer 2 and the upper part of layer 3 are appears to have anomalous heat flow intruded by diabase dikes. Layers 1, 2 and 3 (Luyendyk et al., 1974). However, the of the oceanic crust overly partly ser- sediments on the Fijian Plateau have been pentinized dunite, harzburgite, lherzolite, and poorly studied. One wonders whether garnet peridotite of the upper mantle. From a metalliferous sediments, analogous to those metallogenesis viewpoint, important metal- found in the Bauer Deep, might also be found bearing assemblages in these igneous rocks associated with the Nova Rise or other ridges are: (1) massive sulphide deposits containing in the SW Pacific. mostly pyrite but also some copper-bearing As the oceanic plate moves further away sulphides such as chalcopyrite, chalcocite, from the spreading centre, it sinks to greater and covellite, which are often associated with depths through isostatic adjustment, and the layer 2 basalts; (2) podiform chromite different types of pelagic sediments are deposits occur as cumulates in layer 3 rocks; deposited on top of the metalliferous and (3) the olivines of underlying upper sediments overlying the basaltic rocks. mantle rocks contain significant Manganese nodules, also containing rela- concentrations of nickel. Manganese is an tively high abundances of Mn, Fe, Cu, Ni, important accessory element throughout the and Co, form at the sediment-seawater basic and ultrabasic rock sequences. interface. This sequence of pelagic sediments Some of the most intensively studied (referred to as layer 1) and underlying basalt ophiolite complexes are those occurring in (layer 2) transports significant concentrations Cyprus (the Troodos MassiO, Turkey, of transition metals towards the island arc- Greece, Iran, the Himalayas, Papua, Celebes, trench systems to be later remobilised during Phillipines, and the Californian coast range. geochemical processes along subduction Rarely is the whole ophiolite suite exposed at zones. anyone place, however. In the SW Pacific, an Further information on the nature and ophiolite complex has been identified in the composition of the ocean lithospheric plates Buena Vista Islands of the Florida Group in has been deduced from ophiolite suites the Solomon Islands (Taylor, 1974). It (Gass, 1968; Moores and Vine, 1971; Sillitoe, contains segments of layers 1 and 2 of the 1972a). Ophiolites are n6w generally oceanic crust, including massive sulphide regarded as slices of oceanic crust and upper deposits bearing Cu sulphide minerals and mantle which are tectonically emplaced overlying ironstone and manganiferous (obducted) on island arcs in orogenic belts, sediments. The Florida Islands are con- and are believed to mark locations of ancient sidered to be the eroded remnants of oceanic sites of lithospheric plate consumption. The crust once marginal to an island arc-trench obduction process is analogous to stripping of system active in Miocene times with a a wood chip (ophiolite suite) off a block of subduction zone dipping to the south-west wood (oceanic lithosphere) as it scrapes (c.f. the nearby present day Solomon Trench under a chisel (island arc-trench zone). From dips north-east). In the Tongan Islands, there
16 is evidence to suggest that the topmost part current views on the petrological evolution of 'of an underlying ophiolite complex is exposed calc-alkaline magmas and island arc systems. on the island of Eua (Ewart and Bryan, Ringwood (1974) recently surveyed some of 1972). the common petrological and geochemical The assemblage of rock-types constituting features of island arcs. He suggested that the the ocean lithosphere plate, which is basic rocks of layers 2 and 3 are converted to modelled on the ophiolite suite - pelagic quartz eclogite when the oceanic crust is sediments enriched in Mn, Fe, Cu, Ni, and subducted to depths of 100-150 km. A high Co; layer 2 basalts containing significant partial pressure of water (PH20) is amounts of Cu; layer 3 plutonic rocks maintained by the dehydration of serpen- containing Cr minerals; and underlying tinite derived from the ultramafic rocks in ultramafics containing Ni - is transported as the ocean lithospheric plate. As the the ocean lithospheric plate to convergent temperature in the oceanic crust rises above plate margins. During consumption of the 750°C, partial melting of the quartz eclogite plate along subduction zones, the metals are occurs, leading to the development of remobilised and have the potential to form rhyolitic magmas, characterised by high ore deposits. KINa ratios, and by high abundances of "incompatible elements" including Cu and The Formation of Porphyry Copper Deposits Mn, and volatiles such as HC!, HF, C02, Perhaps the best-known correlation and P205. The rhyolitic liquids are postulated between metallogenesis and plate tectonics to react with mantle material (pyrolite) involves porphyry copper deposits. These overlying the subducted plate, converting the ores consist of wned deposits of dis- pyrolite to garnet pyroxenite and producing seminated copper andlor molybdenum sul- calc-alkaline magmas which rise to the phides with associated hydrothermally earth's surface to form island arc systems altered host rock several thousand metres and Cordilleran mountain chains. Extending thick lying at high elevations in either Ringwood's arguments to metalliferous mountain belts or island arcs, or in regions deposits, it is probable that the "incompatible eroded to moderate depths. They typically elements" are transported with the volatiles occur with calc-alkaline rocks, such as in hydrothermal fluids. At lower tempera- andesites and diorites. Several geologists have tures in aqueous environments, hydrolysis suggested that porphyry copper deposits occurs as a response to cooling, reaction with occur consistently over active or recently wall-rock, and interaction with the lower active subduction zones throughout the temperature and diluted salinity ground Pacific, Caribbean, and Alpine orogenic belts water system. This leads to the deposition of (for example, Mitchell .and Garson, 1972). extremely insoluble sulphide minerals of Cu As was mentioned earlier in the description found in porphyry copper deposits. of the ophiolite complex in the Florida According to this model for the formation Group, which lies southeast of Bougainville of andesitic island arcs and of porphyry Island, the present day subduction zone copper deposits, partial fusion of the ocean associated with the Solomon trench system lithospheric plate occurs at relatively constant now dips to the northeast. One wonders depths. Provided the subduction zone whether further detailed mapping of remains constant (position and inclination) Bougainville Island and the Florida Group and the oceanic crust contains significant will reveal that the copper deposits in the two amounts of copper, there exists the necessary areas are actually identical, being related to ingredients (source and processes) to enrich either an ophiolite suite or to a porphyry the copper into ore grade deposits near the copper system, but not both. earth's surface. However, migration of the Since porphyry copper deposits are always subduction zone is likely to produce more associated with andesites, diorites, or diffuse distributions of copper, particularly granodiorites, it is worthwhile summarising around island arc systems.
17
Manganese Mineralisation in Porphyry earth's surface, what happens to the nickel, Copper Systems especially quantities contained in layer 1 A review of the global distributions of Cu pelagic sediments? and Mn ore deposits reveals that there is a Theories of chemical bonding (e.g. crystal close space-time relationship between field theory, Burns, 1970) predict that NP+, porphyry copper deposits, manganese ore because of its high crystal field stabilisation deposits, and subduction zones, leading to the energy (CFSE) in octahedral coordination in suggestion that volcanogenic processes minerals, will remain behind in the refractory generating copper systems also produce phases resulting from partial fusion of the manganese ore deposits (Thonis and Burns, oceanic lithosphere and become enriched in 1975). Copper itself need not be present for the upper mantle (Burns, 1973, 1975, 1976). the porphyry copper system to exist and to The nickel deposits that are found at bear manganese. continental margins are generally derived The source and transportation processes of from obducted ophiolite complexes which manganese from the subducted, partially have been hydrothermally altered, exten- molten, oceanic crust through to the sively eroded, and deeply weathered. Under overlying andesitic island arc system are these conditions, the ultramafic (upper probably very similar to copper. However, mantle) rocks of ophiolite suites are manganese is transported to higher levels in serpentinised, and, in equatorial regions, the system and may either be trapped near weathered to nickeliferous laterite deposits. the surface of a volcanic pile or escape out of Again, the enrichment of Ni in these deposits the system into submarine pyroclastic reflects the high CFSE ofNP+, which makes sediments to be deposited as manganiferous if difficult to remove the NP+ cations from carbonates or silicates. Under oxidising octahedral sites in the oxide mineral conditions, the manganese is oxidised to very structures into permeating aqueous solutions. insoluble Mn (IV) oxide minerals such as Nickel-rich serpentinized peridotites and pryolusite, nsutite, todorokite, and psilo- laterites occur at New Caledonia, Cuba, melane in eugeosynclinal environments. The Solomon Islands, Phillipines, and Indonesia, manganese deposit found at Forari on Efate for example (Boldt, 1967). Perhaps the best Island in the New Hebrides, as well as many known of these are the nickel ores of New Fijian occurrences, probably formed from Caledonia. Here the nickel deposits are seawater leaching of the manganese from derived from highly weathered serpentinised submarine andesitic pyroclastic sediments peridotites. Emplacement of the peridotite (Warden, 1970). Here, manganese probably took place during late Eocene _ mineralisation occurs at the contact between early Oligocene when the subduction zone limestones and volcanic rocks, and above and dipped to the southwest and oceanic litho- within impermeable clay horizons. sphere from the northeast was obducted onto island arcs to the southwest (Brothers and Factors Controlling the Formation of Nickel Blake, 1973; Dubois et al., 1974; Rodgers, Deposits 1975). Intense erosion of the peridotite took As mentioned earlier, the oceanic crust place during early Miocene, and weathering and overlying pelagic sediments are enriched led to peneplanation. After peneplanation, in several transition elements, particularly New Caledonia was further uplifted during Mn, Fe, Cu, and Ni in ferromanganese the Miocene and the peneplaned peridotites nodules and associated sediments. Nickel is more deeply eroded. This has resulted in the also contained in ultramafic rocks underlying high concentration of nickel left behind in the oceanic crust. A question that arises is: if zones of mutiple enrichment above the Mn, Cu (and Fe) are liberated by partial fractured and weathered peridotite. fusion of oceanic lithosphere along a The nickel-bearing laterites found on Santa subduction zone and transported hydro- Isabel, San Jorge, and Chioseul in the British thermally to island arc environments at the Solomon Islands are also derived from
18
weathering of underlying ultrabasic rocks, southwest. The manganese mineralisation ,which may be part of an ophiolite complex. currently occurring at the surface of the New The ophiolite suite occurring on the Buena Hebrides and Fijian Islands may herald the Vista Islands, Florida Group in the genesis of porphyry copper systems at greater Solomons, has probably not been sufficiently depths, which have resulted from reversed weathered to enrich nickel to ore grade polarity of the subduction zone as the deposits. Australian plate under-rides the Pacific plate to the northeast. Along the Tonga-Kermadec CONCLUSIONS trench, there is high potential for porphyry The preceding discussion leads one to the copper and manganese mineralisation to following conclusions. The Solomon Islands occur above the over-ridden Pacific plate to have the greatest potential for lateritic nickel the west. This accounts for the observed deposits formed from weathering of copper mineralisations found in the North serpentinised peridotites in ophiolites and for Auckland and Coromandel peninsula areas porphyry copper deposits, which were of New Zealand, and for the prediction of formed when the Pacific plate was being mineralisation along the Tonga and Colville subducted beneath the Australian plate to the Ridges.
PROSPECTS FOR THE EXPLOITATION OF MANGANESE NODULES: THE MAIN TECHNICAL, ECONOMIC AND LEGAL PROBLEMS
A.A. ARCHER
With increasing concern being expressed smaller potential "mine sites". A mine site about the ability of traditional sources to may be defined as an area in which there are continue to meet an exponentially growing sufficient nodules with a high enough grade demand for mineral-based raw materials, it is and density (or abundance, the weight of not surprising that in recent years a new nodules per unit area) to sustain a source, manganese nodules, has attracted commercially viable mining operation. As increasing attention. It has been implied that both grade and abundance vary within short the mineral resources of the deep-sea bed are distances, the full evaluation necessary to virtually inexhaustible, perhaps even growing identify a mine site involves examination at faster than the demand for the contained close intervals to provide results which are metals. The pace at which the exploitation of sufficiently accurate to be the basis of a manganese nodules is likely to proceed and mining venture. Obstacles, small and large the importance it will assume depend scale topographical relief and unfavourable principally on: engineering properties of the underlying (a) the development of new technologies, sediment, all of which could inhibit the use of (b) the economics of nodule mining, an otherwise suitable site, must also be including the costs of winning metals recorded. New or refined techniques are from this source compared with the necessary to determine some of these range of costs applicable to mining on properties to the required degree of accuracy. land and the wider economics of the Among the problems that have to be solved world metal market, is the satisfactory correlation of nodule (c) the regime under which the resources population, as determined with television or of the sea bed and ocean floor beyond photographic cameras, with density. the limits of national jurisdiction will Although the detail supporting the Deepsea be exploited, Ventures' claim to a mine site suggests that (d) the extent of the resources. adequate techniques are now available, more recently the difficulties and the need to DEVELOPMENT OF THE develop cheaper methods have been TECHNOLOGY emphasised (Amann, 1975). Many technological problems associated Although much has been written on deep- with the three main activities, the sea mining, that is nodule recovery systems, identification of a mine site, mining and little detailed information is available on the processing, must be solved before progress of industrial research and commercial exploitation becomes possible. development towards meeting the challenge The first step is prospecting to provide a of designing equipment to collect and lift better understanding of the distribution and nodules from depths of more than 4,000 m. characteristics of "prime areas", that is areas in the open ocean. There is little doubt that in at least part of which there are abundant this is now the main technological obstacle to nodules with significantly higher grades than be overcome. All three of the widely elsewhere. Increasingly detailed prospecting discussed systems, airlift, hydraulic pumping of prime areas is necessary to identify the and continuous line bucket dredging, are
21
apparently the subject of continuing research, Secretariat, the economics of the exploitation the cost of which is the principal element in of manganese nodules had had little the investment in research and development attention; recent papers include those by programmes by the major groups concerned. authors associated with the Deepsea Even if the other critical constraints are Ventures (Rothstein and Kaufman, 1973), eliminated, it seems likely that at least five Kennecott (Moncrieff and Smale-Adams, years will be needed for the further 1974) and AMR (Boin and Muller, 1975) development of mining systems and the consortia. The subject, of which only a brief construction of full-scale operational account is appropriate here, can be equipment. subdivided into operational and market There is also a substantial literature on the economics. extractive metallurgy of nodules, the third The following assumptions seem to be main technological activity. Physical tentatively agreed at least in relation to separation is ruled out and of the alternatives hydraulic dredging systems: a hydrometallurgical route is most likely to be 1. Economies of scale, particularly in the used. The details of the process used will processing plant, will be important. A depend in part on what metals are to be typical operation is likely to recover recovered (alternatively the metals that are between 1 and at least 3 million tonnes available for marketing will depend upon the of nodules a year (dry weight). route chosen). Thus, the Deepsea Ventures 2. The operation should have a life of at consortium is likely to produce manganese least 20 years. whereas the Kennecott process leaves this 3. For at least the first generation of metal in the residue. Metallurgical recovery mining equipment, mine sites will be of about 90% of the copper and nickel is needed with: widely predicted; the Deepsea Ventures (a) An abundance of not less than 10 kg process may lead to the recovery of as much of wet nodules/m2 (i.e. 10,000 as 96% of the nickel and 94% of the copper, tonnes/km2).1 (The dry weight of as well as 93% of the manganese and 96% of nodules after heating to 110°C is at the cobalt (Rothstein and Kaufman, 1973). least 30% less than the weight of A recent comprehensive review (Agarwal et drained wet nodules.) al., 1975) concludes that "It is likely that one (b) Nodules with an average grade of 1.2- or more of these processes is now ready for 1.3% nickel and 1-1.1 % copper. commercialisation requiring only limited (These may be accompanied by about further effort on scale-up of equipment. 0.25% cobalt, 27.5% manganese and Certainly, processing will not be the perhaps 0.05% molybdenum.) bottleneck in achieving commercialisation at Alternatively, the average grade may the end of the decade." An the processes be expressed as about 2.25-2.4% being considered will involve the combined Ni and Cu. It has also been construction of large plant with access to the suggested that 2% combined Ni, Cu necessary supply of reagents and energy and, and Co is the "lower economic limit" perhaps, a market for chemical by-products. (Flipse et al., 1973), or cut-off grade. Unless manganese is recovered, each 3 4. After account is taken of the proportion million tonne a year operation will involve of the mine site that cannot be mined the disposal of over 2.9 million tonnes of (because of physical obstructions and tailings a year. There is little or no prospect of grade or abundance below the required any of the processing being carried out at sea. level), the proportion that will be swept and the proportion of the nodules in the ECONOMIC FACTORS Until 1973 when the first of a series of 10 kg/m2 ~ equivalent to 2.05 Ib/ft2; the difference , between thiS abundance (used throughout thiS paper) papers was prepared by the Secretary-General and the often quoted abundance of 2 Ib/ft1 is negligible of the United Nations and the UNCT AD compared with other errors.
22
increased demand for other metals; for present position under customary inter- example, some is associated with national law is that coastal states have sover- nickeliferous laterites. Without production eign right over the natural prolongation of from nodules, therefore, there could be a gap their land territory into and under the sea. between supply and demand by 1985. Thus under the Convention and customary law there are resources beyond the limits of Manganese national jurisdiction. On the suggested scale of nodule mining Although the 1958 Convention on the activity, about 750,000 tonnes of manganese High Seas, which enshrines the long would be available to the market from three 1 established principles. of the freedom of the million tonne a year operations by 1985, high seas, makes no express reference to the when it would amount to only 8% of resources on the sea bed, it is clear that States demand. Thereafter the number of ventures have the right under existing law to exploit recovering manganese is likely to depend on these resources but not to acquire exclusive competition between metallurgical processes rights to do so in any particular area. The for nodules and the development of outlets provisions include that no state may assume for manganese. At present it is not clear in sovereignty over areas of the high seas and which form the manganese would be that reasonable regard must be paid to other marketed, but the product, whether it be users which applies to such activities as the manganese metal, ferro-manganese or oxide, laying of submarine cables and pipe-lines. It would have to compete with the same was against this background that an Ad Hoc product from other sources (and not with Sea Bed Committee was established by the manganese ore) for its markets. The General Assembly of the United Nations in manganese content of ores now mined is up 1967, following an initiative by the Maltese to twice as high as in nodules which, government. The work was continued by the additionally, contain metals deleterious to Sea Bed Committee from 1968 until 1973, steel making. Manganese from nodules must during which time its mandate was amended therefore bear the additional processing costs to allow it to become, in effect, a preparatory and this, together with the captive nature of committee for a conference on all aspects of much of the current manganese ore the law of the sea. In 1970, the General production, suggests that at lower Assembly adopted a Declaration of Prin- consumption growth rates it might well not ciples, proposed by the Sea Bed Committee. be economic to extract manganese from These included that "The sea-bed and ocean nodules for use in ferro-manganese floor, and the subsoil thereof, beyond the production. limits of national jurisdiction· (hereinafter referred to as the area), as well as the LEGAL ISSUES resources of the area, are the common The extent of coastal state jurisdiction over heritage of mankind" and that "The explo- the resources on the sea bed was not defined ration of the area and the exploitation of its unambiguously in the Continental Shelf resources shall be carried out for the benefit Convention, concluded in Geneva in 1958. It of mankind as a whole, irrespective of the provided that coastal states should have geographical location of States, whether landsovereign rights for the purpose of exploring locked or coastal, and taking· into particular and exploiting the resources of the adjacent consideration the interests and needs of sea bed to a depth of 200 m. or beyond to developing countries". where the depth of the superajacent waters The Law of the Sea Conference has had admits of exploitation. Although it could be two substantive sessions, in Caracas in 1974 argued that this allows coastal states to and Geneva in 1975, and is due to meet again extend jurisdiction as technology develops in New York in March, 1976. Of its three until it impinged on another state's shelf, this main committees, the First Committee is interpretation is not generally accepted. The concerned only with formulating new
26
international law for the international sea bed THE RESERVES AND RESOURCES area. In the Second Committee, which The recoverable metals in the nodule includes all the questions of limits in its deposits that are economically workable in mandate, there is widespread support for the the current, locally prevailing economic proposal that a 200 mile zone should be circumstances will be reserves. Those in established within which coastal states would deposits with a lower grade or abundance have jurisdiction over both the living and that, so far as can be foreseen, are likely to be mineral resources, with the further proposal workable at some time in the future, together that jurisdiction over sea-bed resources with reserves are resources. Thus the should extend to the edge of the continental specification for a mine site will change with margin where this is more than 200 miles economic circumstances including predicted from coastal baselines (Archer and Beazley, metal prices and changes in costs: costs are 1975). related, among other things, to the stage of It is widely accepted that an international development of mining, transport and authority will be established to supervise the processing technology. mineral resources in the sea-bed area beyond Information about the average grade of the limits of national jurisdiction. Various nodules in whole oceans is analogous to data regimes to ensure that these are exploited for on the crustal abundance of metals and can the benefit of mankind as a whole have been be related to the economic concepts of discussed during the many meetings. "reserves" and "resources" only with Proposals have ranged from a regime by caution. The use of these terms in a general which the authority would issue exclusive sense can be misleading, as for example in licenses to states or to companies and would the recent conclusion that "Reserves are benefit financially by way of royalties or other currently estimated to be of the order 1.5 arrangements,' to providing the authority trillion tons" (e.g., assuming these to be short with the right to exploit the area exclusively, tons, 10 kg/m2 over 136 million km2) "in the by establishing its own mining enterprise. In Pacific Ocean alone. Unlike many resources, Geneva this year, the possibility was ferromanganese nodules are currently discussed that the Convention should forming. The rate of formation in the Pacific provide that the authority should enter into Ocean has been estimated at 6 to 10 million either contractual or equity joint ventures tons per year. Most of the nodules are not with states and other entities. It is impossible economically mineable. Consequently, the within the compass of this paper even to mineable reserves are probably on the order summarise the debates or to discuss the of 10 billion to 500 billion tons" (Mielke, "single negotiating text" that has been 1975) (e.g. 10 kg/m2 over 0.9 million to 45 prepared by the chairman of the First million km2). Committee as a basis for further work at the next session. Potential Reserves Security of investment is an essential Present evidence suggests that deposits prerequisite to the further research and with an average of about 2.25-2.4% combined development needed to solve the remaining Ni and Cu and not less than 2% Ni + Cu + completely new and difficult technological Co, occurring with a minimum abundance of problems before a practical deep-sea mining about 10 kg/m2, are likely to be economically system is available. Major investment workable as soon as production can begin and decisions will have to be taken in the next may, therefore, be described as potential few months if mining is to start by 1980. It is reserves. For convenience, the areas in which clear that the costs of full-scale mining, these occur may be described as "first transport and processing systems will be very generation mine sites", that is sites to which high. The financial resources will be available the economics of first generation mining and only if an adequate return on the investment processing equipment will apply. is ensured. An attempt to estimate the potential 27 reserves of metals in nodules (or the crusts at these stations. However, there is corresponding number of mine sites) is very little published data relating to the severely hampered by the paucity of the extent to which acceptable grade and evidence and will certainly not be better than abundance coincide so that this critical at the "inferred" level of accuracy. information must be inferred. Information . about grade and how it is Nevertheless, these compilations provide distributed has been published in a some indication of the total area suitable for convenient form as part of the U.S. National first generation mines. The relevant Science Foundation's IDOE programme information on grade frqm the 530 stations is (Frazer and Arrhenius, 1972; Horn et al., summarised in Table 6 and its distribution at 1973a). However, the analyses (of varying intervals of 0.25% combined Ni and Cu is reliability) refer to only 530 stations for the shown in Fig. 1. Hopefully, any conclusions 362 million km2 or so of the ocean floor, that drawn from these data may not is there is an average of one per 683,000 km2• be vitiated by the inclusion of inaccurate The sampling pattern suggests that it is not analyses, as smaller errors may be within the random and substantial areas are unexplored. 0.25% intervals and there may be enough On the other hand, there is increasing samples to ensure that larger errors are self- understanding of the environment in which canceling. It has been shown by Healing and "mine-grade" nodules form, which enables Archer (1975) that, if grades below 0.25% are considerable nreas to be eliminated. An excluded, in the Atlantic, Indian and South indication of the worldwide variation in Pacific Oceans there is a geometrical increase nodule abundance has been provided by in the cumulative number of samples for unit Ewing and others (Ewing et al., 1971), in the decreases in average grade, following the form of maps showing the sites of bottom distribution suggested by Lasky (1950) and photographs and the proportion of the sea that a similar relationship applies to cut-off floor covered by manganese nodules and grade. The distribution of grade in the North 28
variables, the area in which acceptable grade This basis for estimating reserves is open and abundance coincide will be reduced to to similar criticism to that offered above about 3 million km2, as in only 13% is the about consideration of the ocean wide grade likely to be above the assumed cut-off average grade, as ocean wide average of 1.76%. The corresponding estimate is 2.19 abundance has little relevance to the million km2 if an average nodule diameter of identification of mine sites. An alternative 1.00 inch is assumed. By interpolation in the method is to attempt to estimate the extent data derived from Ewing et al., the average of prime areas. abundance is 18.9 wet kg/m2 over the 3 It is clear by inspection of the maps (which million km2 with a minimum abundance of are similar and can be directly compared) about 10 wet kg/m2, or 14.7 kg/m2 for showing the distribution of grade (Frazer and nodules with an average diameter of 1.0 in. Arrhenius, 1972; Horn et al., 1973a) and
A different population distribution is If it is assumed that the proportion, about -shown by Flipse et al. for "a typical nodule 2/3 of the Clarion-Clipperton prime area with deposit", which may refer to part of a nodules with an average grade that is potential mine site rather than of a prime acceptable applies to all prime areas and that area. On the basis of this distribution as the proportion, about l/3, of the western half much as about 60% of the area may be of the Clarion-Clipperton prime area with an occupied by a population of at least 35%, acceptable nodule abundance can also be corresponding to 1.45 inch nodules with an applied generally (or if the proportion in abundance of 10 wet kg/m2, but if the which acceptable grade and abundance average diameter is 1 inch this abundance is appear to coincide in the latter area applies to represented by a nodule population of at least all prime areas) then as a rough 50%, which is present in only 13% of the approximation the total potential first area. generation mine site area would be only Finally, the proportion of the prime areas about % of the total prime area, that is about that may be regarded as first generation mine 1.3 million km2 in the suggested 6.5 million site will be determined by the proportion of km2• the area in which acceptable grade and Although the estimates obtained by these abundance coincide. Schatz (1971) published two methods are both subject to a wide this data for a very small area, about 330 margin of error, it may be significant that km2, described as not too unlike many they are the same order of magnitude. The examined by Kennecott Exploration Inc. The effect of the choice of some variables on the nodules have an abundance of more than 10 possible number of first generation mine sites wet kg/m2 and contain an average of more that might be identified' within these than 2.25% Ni and Cu at 14 (42%) of the 33 estimated total areas is shown in Table 9. stations. The high average copper and nickel Not only are these estimates uncertain content (2.46% combined Ni and Cu) of the because of the inadequacy of the data used, nodules suggests that this may also be part of but also because of the uncertainty attached a potential mine site, in which case it is not to the assumptions that must be made. Thus representative of a prime area. the specification assumed for a commercially Figures 2 and 5. of Schultze-Westrum viable mine site is based on estimates of the (1973) can be combined to show that the costs associated with first generation mines, population at the 17 stations from which future prices and recoveries, all of which may potentially mine-grade nodules were re- be proved wrong. For example, if a covered is as follows: minimum abundance of 15 wet kg/m2 is found to be necessary, the total mine site area Population No. of stations might be reduced by about one-third, for 2 Less than 20% 7 (41%) example from 1.3 million km to 867,000 2 20 to 50% 8 (47%) km • Alternatively the whole of the 15% of More than 50% 2 (12%) the ocean floor o~ which nodules may be present (Table 7) would meet the This is consistent with the view that grade requirement for an average, rather than a and abundance are independent variables. By minimum, abundance of 10 kg/m2: the in terpolation, only 6 (21 %) of the 29 samples corresponding first generation mine area from this area have both. the specified would be about 7 million km2 (362 x 15% x abundance and grade. A similar proportion is 13%). The latter illustration clearly identifies obtained by multiplying the factors for grade the caution needed in applying this method, (59%) and abundance (30%), Interpolation as it is clear that the average abundance of all indicates that the average abundance is 18.3 nodules on the ocean floor has no direct kg/m2 in the area that satisfies both criteria relevance to their economic potential. For (or 14.5 kg/m2 if the average diameter is 1.0 this reason the "prime area" approach may in). be more useful, that is the total area of first
33
generation mine site may be nearer 1.3 than nodules: metallurgical losses are likely to be 3.0 million km2• similar and neglible compared with other These estimates do not take account of the errors in the estimates, need also to meet the required specification for the topography, which would reduce the Resources area. These estimates further assume that the As is the case with potential reserves, an nodules with acceptable grade and abundance attempt to estimate resources is attended by are so distributed as to allow more or less difficulties which impose severe limitations continuous mine site areas to be identified. on the confidence that can be placed in the The method used to derive the higher result. It follows from the definition of estimate of total first generation mine site resources that the grade and abundance that area (3 million km2) leads to the inclusion of will be economically exploitable in the future small widely dispersed areas which while must be predicted. These are particularly carrying the required grade and abundance, difficult to forecast as this involves, among might present such mining problems as to other factors, consideration of the costs likely disqualify them from consideration as first to be associated with the future development generation mine sites (although they might of currently untried technology. The fol- be regarded as sub-marginal resources). In lowing discussion is, therefore, highly specu- this context, the apparently most promising lative. area, between the Clarion and Clipperton The US Geological Survey (Brobst and Fracture Zones, may contain between 23 and Pratt, 1973) emphasises that the quantities of 37 of the sites (depending on the estimated metal likely to be recovered will be limited by total area required). the distribution of suitable grade and The average combined nickel and copper abundance and by the environmental setting, grade is 2.22% in deposits with a cut-off grade and comments on the wide range· of of 1.76% (Healing and Archer, 1975). The estimates of resources that have been made. inferred potential reserves of recoverable According to the US Bureau of Mines (1975) metals from the same two total mine site nodules contain several orders of magnitude areas, assuming an average grade of 1.22% more than 63.5 million' tonnes of nickel (the Ni, 1.00% Cu, 0.25% Co and 27.5% Mn, estimated world resources on land in deposits 22.5% mining recovery and 90% metallurgical averaging 1 %) and about 360 million tonnes recovery are suggested in Table 10. These can of copper (to which the corresponding be compared with the estimates of identified amount of nickel would be about 420 million world reserves on land (Table 11): although tonnes). In a recent report by the National these refer to quantities of contained metal in Academy of Sciences (Mineral Resources situ, mining recovery on land is much closer and the Environment, 1975), the recoverable to 100% than it is likely ever to be for copper has been estimated by a "geological
34 control approach" to be at least 450 million abundance is ignored, 30% of the total nickel .tonnes in the Clarion-Clipperton area and and copper resources (in deposits with at perhaps 1000 million tonnes in all areas (it least 0.25% combined) occur in deposits with should be noted that these estimates seem to a cut-off of 1.76% increasing to 62% if the be subject to an arithmetical error so that cut-off is halved. The corresponding increase these figures might be 4.5 and 10 million in the tonnage of nodules is from 13% to tonnes respectively): by an "area of influence 43%. approach" the copper in nodules is estimated However, the effects of decreasing the as 300 or 490 million tonnes. On the other required abundance and of recovering a hand, Mero (1972) has suggested that the substantially larger proportion of the nodules nodules in the Pacific Ocean contain 14.7, as mining techniques improve must also be 7.9, 5.2 and 358 billion tonnes of nickel, considered. The data on nodule population copper, cobalt and manganese respectively, distribution (Table 7) provided by Ewing et most of which is regarded as potentially al. 1971) may be used to indicate the available. proportion of the ocean floor that may be No better than hazardous guesses must be occupied by nodules with an abundance of at made to serve as guides to the lower limit of least 5 kg/m2• Again by interpolation, and grade and abundance that will be exploitable assuming an average diameter of 1.45 in, this in the future. Half the cut-off grade assumed may be about 60% (that is the proportion of for potential reserves, that is 0.88% combined the stations at which the population is at least Ni and Cu, might not be an unreasonable 17.5%) of the 15% of ocean floor on which guess of the future cut-off. Similarly, half the nodules may be present, corresponding to an minimum abundance, that is 5 kg/m2, might area of about 32.6 million km2 (362 million x be assumed. 15% x 60%). About 43% (Fig. 2) of this area, As a consequence of the Lasky relation- that is about 14.0 million km2, may have ship between grade and (inferred) tonnage nodules with more than 0.88% combined Ni there is a very large increase in the tonnage and Cu. If the average abundance is 15.6 wet only when the combined nickel and copper kg/m2, this area would contain about 218,400 content falls below about 1%, so that in all wet or 153,000 dry million tonnes of nodules. oceans (including the North Pacific if the Improvements in the proportion of the Clarion-Clipperton area is excluded) more nodules that are recovered are likely to be than half the samples have less than 0.88% one of the major factors leading to deposits combined Ni and Cu. The effect of the bi- with lower grade and abundance being modal distribution, notably in the Clarion- worked. Bearing in mind the problems Clipperton area, is that in the North Pacific as . involved in recovering nodules from oceanic a whole there are slightly less samples depths and that the topography may prevent between 0.88 and 1.75% than there are with some deposits ever being worked, it may not more than 1.75%. be unreasonable to postulate that the If the sampling is random with respect to recovery efficiency will not improve beyond all manganese nodule deposits and therefore about three times that assumed for reserves to the distribution of their abundance that is, 67.5%. It may be reasonable to (although neither can be demonstrated), the suppose, however, that this will be achieved number of samples can be taken to be only in the distant future and that the proportional to the tonnage of nodules. average recovery achieved by "later Considering the criterion of grade alone, generation" mining systems will be, say, 45%, therefore, its distribution suggests that the so that the total recoverable nodules would quantity of the metals in deposits with cut- then be about 69,000 million tonnes. It is offs between 1.76% and 0.88% is probably the perhaps possible that a considerable same order of magnitude as the potential proportion of the 69,000 million tonnes is so reserves (Le. above a cut-off of 1.76%). This widely dispersed as to be unworkable, even is confirmed by Fig. 2 which shows that, if eventually. Furthermore, at lower grades and
35
MANGANESE MICRONODULES IN DEEP-SEA SEDIMENTS AND THEIR RELATION TO MANGANESE NODULE FIELDS
GUNTHER H. FRIEDRICH
Abstract Microscopic and microprobe studies and chemical investigations have been carried out on ferromanganese concretions and micronodules separated from different facies of deep-sea sediments. The external features as seen with the scanning electron microscope and the internal features of single micronodules are similar to those described for macronodules. Zonal chemical variations occur within micronodules, and sympathetic variations of Mn, Ni and Cu were noted. In a deep-sea sediment core, the distribution of micronodules (size> 20 JJ-) varies with depth: From 20 mg micronodules per kg "Siliceous Ooze" at a depth of 0.5 m. up to 460 mg/kg at depths of 4 to 8 m. A decrease of micronodules was noted at depths of 8 to 10 m. It was also noted that Mn, Ni and Cu are enriched in the smaller size classes (from 0.63 to 20 JJ-) of this sediment core as found by studying the element distribution in the hydrogeneous phase of different size classes. In general, areas containing micronodules with high Ni and Cu contents also contain macronodules with more than 2.5% Ni + Cu. Core samples from the "Calcareous Ooze Area" where no macronodules are found, contain very small amounts of micronodules: From 2 to 8 mg micronodules/kg sediment at a depth of 1 to 3 m. and about 25 mg/kg at depths from 3 to 5 m. as a maximum. The Ni + Cu content of micronodules separated from "Calcareous Ooze" is surprisingly high: Ni varies from about 0.5 to 0.9% in different zones of the micronodules and Cu from 0.6 to 1.9%. Fe varies from 1.8 to 6.5% and Mn from 21.0 to 31 %, also in different zones of one single micronodule.
INTRODUCTION sediment facies: red clay, siliceous ooze, and In our previous studies which are part of calcareous ooze (Beiersdorf and Wolfart, the German Research Programme on Deep- 1974). In one of our previous studies; it was Sea Manganese Nodule Deposits, we have noted that manganese micronodules were concentrated on the investigation of found in most of the deep-sea sediment manganese macronodules collected from samples, and preliminary data were different areas south and east of Hawaii presented on the distribution of Mn, Fe, Ni (V ALDIVIA expeditions V A 04, OS, 08 in and Cu (Friedrich and Schmitz-Wiechowski, 1972, 1973 and 1974, respectively), between 1975). The same zonal distributions and the Clarion Fracture Zone in the North and geochemical correlations between Mn, Ni the Clipperton Fracture Zone in the South. It and Cu have been found in micronodules as has been shown that the chemical described in macronodules (Burns and composition of the manganese nodules are Ftirstenau, 1966; Friedrich et al., 1969). The not uniform within nodule fields of 40 km x importance of further detailed studies on the 40 km and that the areas around sea mounts distribution of micronodules within the contain nodules with definitely higher sequence of underlying deep-sea sediments contents of Fe and lower contents of Cu and below manganese nodule fields was indicated Ni than areas with smooth topography and their possible role in the genetic (Friedrich and Pltiger, 1974). Based on a interpretation of nodule growth was pointed detailed bathymetric and seismic reflection out. In his paper on manganese nodules as survey (Durbaum and Schluter, 1974), deep- indicators of long-term variations in sea floor sea sediment samples were taken at defined environment, Sorem (1973) wrote "in short, positions within the three different types of the chances are good that manganese nodules 39 can tell us much more about the marine gravity box corer 4, 6, 12 m. long (Kogler, environment, past and present, but we need 1974). A description of the core material and to know a great deal more about nodules a final interpretation will be given by Kogler before we can understand their language". In after completion of laboratory work in Kiel. this paper, I want to present a small The sedimentological-biostratigraphical contribution to the whole story on the investigations of most sediments used for our microscale, discussing the physical features studies have been carried out by Beiersdorf and the chemical composition of and Wolfart (1974) on board of R.V. ferromanganese concretions and micro- Valdivia. Based on preceding geophysical nodules found in different depths of selected investigations (Durbaum and Schluter, deep-sea sediments. Work, including 1974), four target areas were chosen where microprobe studies on single micronodules good correlations between different seismic from additional sample sites and quantitative facies and corresponding sedimentary facies atomic absorption analysis, is continuing on had been obtained (Fig. 1). In all sediments, the material. ferromanganese concretions or micronodules have been found with different GEOLOGICAL SETTING AND concentrations in certain sediment layers. In DESCRIPTION OF THE DEEP-SEA general, they occur as single grains or SEDIMENTS AND THE micronodules with shell-like structure similar MICRONODULES to the structure of most macronodules (Fig. 2). The size vanes from some microns to Undisturbed deep-sea sediments were about 200 MI-'. Larger "micronodules" from sampled in different areas in the Central 0.2 mm up to 2 mm) are uncommon, so that Pacific in depths· between 5000 and 5500 m. the "size distribution of the micronodules (Schultze- Westrum, 1973; Beiersdorf and makes them, as a population, distinct from Wolfart, 1974; Kogler, 1974) using a piston nodules which have grown larger than - 1 corer, box grab sampler (Kastengreifer) 20 x cm" (Greenslate et al., 1973). In some 30 x 40 cm. and box corer (Kastenlot) 30 x sediments the ferromanganese concretions 30 cm. In addition, three cores were taken are forming rod-like or worm-like, tubular from definite targets by use of a heavy features (Fig. 3).
40
layering and zonal textural variations as Microprobe Studies on Selected Micronodules known from macronodules. Organic and The distribution of Fe, Mn, Ni and Cu in inorganic detritus commonly occurs within individual micronodules is studied here by the micronodules or on nodule surface. microprobe analysis (Fig. 9) because no other Ancient growth surfaces are often covered by presently available analytical technique can small globule-like particles which may have a be used due to the very small size of the colloidal or a bacteriological origin (Fig. 6). micronodules. The data may be used only in Some ferromanganese concretions consist a qualified sense because of the well known of numerous idiomorphic non-ferro- difficulties of microprobe analysis of porous manganese oxide minerals (orthoclase) in a ferromanganese oxide material. groundmass of ferromanganese material Our preliminary investigations have shown (Fig. 8). that the same zonal chemical variations occur 45
was found in the sediment size class 20 to 50 smooth topography covered by "Siliceous 1-'_ Finer sediment material is very rare in Ooze" containing macronodules with higher this foraminiferal sediment. Mn, Ni and Cu contents and lower Fe content are characterised by higher amounts SUMMARY AND CONCLUSIONS of micronodule material in the sediment (e.g. The internal and external features of core KH 70). At present no comparison can micronodules are generally similar to those of be made between micronodule distribution in macronodules. A distinct correlation between sediments and seafloor concentrations of micronodule growth and the formation of macronodules because bottom photographs macronodules is therefore assumed. Small are not available from our sample sites. globule-like particles found on ancient growth Future detailed studies in Area 2 of V A 08-1 surfaces of micronodules (Fig. 5) may have a expedition are being planned for 1976; this colloidal or bacteriological origin. will include investigations of the relation The geochemical data of deep-sea between the micronodule distribution in ferromanganese concretions and micro- sediments and the distribution patterns of nodules obtained by microprobe studies and manganese macronodule fields. selective chemical analysis and their The occurrence of small amounts of distribution patterns in deep-sea sediment micronodule material with considerably high cores show distinct relations between the amounts of Mn, Ni and Cu in the occurrence and composition of macronodules "Calcareous Ooze" where no macronodules and the ferromanganese material within the have been found is in good agreement with sediments. Manganese nodule fields of the data given by Greenslate et al. (1973) for economic interest occur in areas of "Siliceous the rate of biogenic deposition of copper and Ooze" which contain considerable amounts nickel in the equatorial high-productivity of micronodule material. In general, the top zone, where the abundance of macronodules layers of the sediment core contain less than "is determined partly by the drowning effect 20 mg micronodule material per kg wet exerted by the total rate of accumulation. of sediment. An increase of the micronodule sediments" (p. 55), which is very high in content has been observed in depth of about areas covered by "Calcareous Ooze". 4 to 8 m. The amount of micronodule A schematic representation of the matter material> 20 I-' decreases at depths of 8 to of occurrence of transition metals in material 10 m. and the hydrogeneous manganese of different marine environment and the increases in the size classes 0.63 to 20 I-' at distribution of nickel in sea water, deep-sea this depth, probably due to the selective sediment and ferromanganese materials from decomposition of ferromanganese material. the area V A 05-1, Central Pacific is shown in Unfortunately, no sediment samples have Fig. 13. Sea water taken near the sea floor been available for our studies from greater contains nickel in the range of 0.022 ppm in depths. the colloidal phase and 0.002 ppm Ni have The variation in chemical composition of been found in the soluble phase of near manganese macronodules within a seafloor water and near sea surface water manganese nodule field of 40 km x 40 km, (Zuleta, 1975). Manganese macronodules depending on the deep-seafloor topography and micronodules contain similar amounts of may be reflected by the distribution of about 1.5% Ni, and 700 ppm Ni were found micronodules in the deep-sea sediment. in the hydrogeneous sediment phase in the Areas around submarine hills, containing 20 to 50 I-' size class. The average Ni content macronodules with lower Mn, Ni and Cu of "Siliceous Ooze" sediments is about 150 contents and higher Fe content are ppm. The new observations, together with characterised by lower amounts of the data discussed by other authors, suggest micronodule material in the sediment (e.g., the following conclusions regarding the origin core KH 20). Deep-seafloor areas with of macronodule fields:
51
DEEP BOTTOM CURRENTS, SEDIMENTARY HIATUSES AND POLYMETALLIC NODULES
G. PAUTOT AND M. MELGUEN
Abstract
Recent studies undertaken by the Centre National pour I'Exploitation des Oceans have focused some attention on thfJ environment of nickel- and copper-bearing iron-manganese nodules in the Tuamotu Archipelago region of the South Pacific. The results indicate that the distribution of the nodules is closely linked to that of deep currents in the area. We attempt to show that these currents are associated mainly with the Antarctic Bottom Water and we attribute the episodical character of nodule formation in this' region to major pulsations of the Antarctic Bottom Water. A synthesis of what are considered to be primary factors in the nodules formation are incorporated into a tentative model which is presented as a guide for future exploration for polymetallic nodules.
INTRODUCTION 2. To establish in this zone the Numerous works have underlined the relationship existing between erosion local importance of bottom currents on structures, sedimentary hiatuses and sedimentation in the Pacific (Johnson et al., the passage of the Antarctic Bottom 1970; Johnson, 1972), the presence of Current; outcrops of pre-Quaternary sediments and of 3. To illustrate the role played by the sedimentary hiatuses (Bramlette, 1961; Antarctic Bottom Current on the Riedel and Funnell, 1964; Moore, 1970; genesis of polymetallic nodules; Tracey et al., 1970; Hays et al., 1972), the 4. To illustrate in the light of recent relationship existing between bottom research carried out in the South Pacific topography, sedimentary facies, submarine (Glasby, 1972, 1973b; Glasby et al., volcanism and the distribution of 1974; Hollister et al., 1974) and, polymetallic nodules (Menard, 1964; Mero, especially, from our own observations 1965; Barnes, 1967; Horn et al., 1972; (Pautot and Melguen, 1975), a "model" Glasby 1973a; Friedrich and Pliiger, 1974; for the formation of polymetallic Beiersdorf and Wolf art, 1974; Andrews, nodules rich in nickel and copper. This 1974). However, most of these works deal "model", integrating the principal almost exclusively with the North Pacific and factors controlling the formation of only rarely emphasise the relationship nodules of economic interest, is existing between the structural and proposed as a "guide for exploration". sedimentological situation and the distribution of polymetallic nodules. Further, I. Proposals for a traverse of the Antarctic these works consider polymetallic nodules in Bottom Current in the Tuamotu-Marquesas a global fashion, without dissociating those Islands (South Pacific) nodules whose high contents of nickel and The Tuamotu Archipelago, bounded in copper justifies an economic interest. The the north by the fracture zone of the aims of the present article are as follows: Marquesas, is interpreted (Pautot, 1975) as a 1.To propose a course of the Antarctic volcanic structure comparable to an Bottom Current to the latitude of the undeveloped oceanic ridge. The Marquesas Tuamotu Archipelago and the Islands Archipelago is composed of a modern Marquesas Islands. volcanic mass (Plio-Pleistocene) (Pautot and
54
Dupont, 1974). The region studied is situated bordering the volcanic reliefs (Fig. 2, I) show between the Tuamotu Archipelago and the a superposition of greatly differing facies and Marquesas Islands (Fig. 1). This zone has a stratigraphic gap (for example: sample KS been the subject of detailed bathymetric and 13, Fig. 2, III). At the top of the cores a few structural research (Monti and Pautot, 1974; ems. of chocolate-brown mud ("Red clay" or Pautot and Dupont, 1974; Pautot, 1975). mud) is Quaternary in age. Below this and The course of. the Antarctic Bottom separated by a very sharp contact is a white Current in the South Pacific has been Upper Oligocene calcareous ooze. described, at least partially, by Wooster and The origin of this sedimentary hiatus, with Volkman (1960), and more recently by a stratigraphic gap of almost 30 million years, Hollister et al. (1974). The Antarctic Bottom has been discussed in detail (Pautot and Current enters the Pacific south of Tasmania, Melguen, 1975). The Antarctic Bottom at the latitude level of the Pacific ridge, and Current appears to be the major factor flows in fracture zones through channels or controlling the presence of this gap, by its basins at a depth greater than 4,000 m. From influence on the "abyssal fractionation" of the Tonga-Kermadec trench, it flows into the carbonates (Berger, 1970: Berger and Samoan Basin and then into the North Winterer, 1974) and on sediment erosion. Pacific through the Samoan Passage. The upper limit of the Antarctic Bottom Nevertheless all the Antarctic Bottom Water Current corresponds, in the South Atlantic reaching the Samoan Basin do not flow (Melguen and Thiede, 1974) as in through the Samoan Passage. A second the Pacific (Berger, 1974), with the water passage exists. It was brought to light on the depth of the calcite lysocline. Towards 4,000- new bathymetric chart by Mammerickx et al. 4,200 m. water depth, calcareous sediments (1974) and named the Aitutaki Passage and more particularly planktonic foraminifera (Pautot and Melguen, 1975). are subject to a drastic increase in the degree It seems that the Antarctic Bottom Water of dissolution. Because of its relationship with masses flowing through the Aitutaki Passage the presence of the Antarctic Bottom Water reach the Tuamotu-Marquesas zone. In masses, this water depth (4,000-4,200 m.) is effect, some hydrological measurements also called "hydrographic lysocline" (Berger, taken during the Transpac I cruise suggest 1974). The passage along the core KS 13 the presence of a bottom current along the (Fig. 2) from a white calcareous ooze to a north flank of the Touamotou Archipelago. brown mud much poorer in carbonate reflects The Antarctic Bottom Current flows towards a great fluctuation of the hydrographic the east along Marquesas fracture zone lysocline at the end of the Oligocene. (Pautot and Dupont, 1974), and reaches the Furthermore, it seems that the fluctuations Tiki Basin where it is probably stopped by the of the lysocline (during Oligocene, upper ridge. The Antarctic Bottom Water then Miocene, Pliocene, and Quaternary) (Berger, flows towards the NW, overflowing into the 1973, 1974)is linked to the pulsations of the Tapu Basin (Fig. 1). Antarctic Bottom Water masses during the major Cenozoic glaciations (Watkins and II. Antarctic Bottom Current, erosional Kennett, 1971; Weisell and Hayes, 1972). In structures and sedimentary Hiatuses effect, at the end of the Cenozoic, the The evidence given of the Antarctic intensity of the Antarctic Bottom Current Bottom Current flowing along the north flank fluctuated in response to worldwide climatic of the Touamotou Archipelago explains the variations and particularly to the ml\ior presence of erosional structures (for oscillations of glacial and interglacial periods example: channel of erosion, Fig. 2, I and II) (Hayes et al., 1969; Kennett, 1970). The and of sedimentary hiatuses (for example: corrosive flow of the Antarctic Bottom core KS 13, Fig. 2, III). The cores sampled a Current has played a role of primary water depth of about 4,200 m. in the basins importance in the presence of sequences of
56
of primary importance for the reconstruction considered oceanic region: palaeo-bathymetry of the palaeocirculation of the deep water depending notably on the relative mobility of masses and, consequently, for the exploration ocean plates, on subsidence, on tectonics, on of polymetallic nodules. major oceanic transgressions and regressions.
III. Antarctic Bottom Current, sedimentary IV. Proposal for a "model" of formation and hiatuses and polymetallic nodules distribution of polymetallic nodules rich in The analysis of results (put at our disposal nickel and copper by our colleagues from the Nodules Project) It is well known at present that from the nodules-prospecting cruises in the polymetallic nodules form around a nucleus South Pacific leads to the following consisting of volcanic debris, of indurated observations: clay, of fish teeth or of any other organic or (i) the ocean floor situated at depths less inorganic debris exposed on a well- than 4,000 m. is poor in nodules and the oxygenated sea floor with a low sedi- nodules encountered there are economically mentation rate (deep basins far away from uninteresting. continents) and situated near active volcanic (ii) the ocean floor rich in economically zones (mid-oceanic ridge, volcanic chains). interesting nodules is found at water depths The bottom topography also controls the greater than 4,000 m. (for example: the Tiki nature and the distribution of nodules and Tapu Basins), that is, below the (Menard, 1964; Barnes, 1967; Friedrich and hydrographic lysocline, which itself, as Pluger, 1974) i.e. nodules rich in nickel and previously indicated, marks the upper limit of copper are concentrated in the basins, while the Antarctic Bottom Water masses, and also nodules rich in cobalt are more frequent in coincides with the distribution of higher regions. In spite of considerable sedimentary hiatuses. Similar observations number of dating studies undertaken on have been made in the South Atlantic in the manganese nodules (isotopes, potassium- Vema Channel, where indurated beds rich in argon, biostratigraphy of nuclei of organic manganese nodules have been encountered. origin), the rate of formation of the nodules They are present in the zone of the Antarctic seems to remain an open question. Therefore Bottom Current passage, at a depth of more no discussion on the subject is presented in than 4,200 m. They are associated with non- this article. A particular emphasis is put on calcareous sedimentary facies (abyssal clay or the episodic character of the nodules "red clay" showing at least locally a Miocene- formation. Lower Pleistocene sedimentary hiatus The cortex of manganese nodules, (Melguen and Thiede, 1947a, 1947b; Pautot consisting of rings of growth, strongly and Melguen, 1975). suggests an episodic formation. Conse- If, however, the present hydrographic quently, this reflects the presence of episodic lysocline delimits potentially rich zones of factors in the environment of formation of polymetallic nodules on the ocean floor, it is the manganese nodules, factors such as important to note that it has varied during volcanism and bottom currents. The the geological periods, correlating to the proposed model (Fig. 3) is very largely based pulsations of Antarctic Water masses. Thus, on these two factors. in the Vema Channel, for example, it has The role of volcanic activity in the genesis undergone fluctuations in depth of 800 m. or of nodules was first studied at the level of the more in the Miocene and Pleistocene axis of the East-Pacific Rise (Transpac (Melguen and Thiede, 1975). Therefore all cruise): a zone where transition metals can polymetallic nodules exploration must take be released in an ionic state in the sea water. into account not only the present depth,but But in the samples taken in the immediate also the palaeo-depths of the hydrographic proximity of the emissive zones of the ridge, lysocline and the palaeo-bathymetry of the we have found only fragments of
58
manganese-rich crust. Furthermore, it seems copper can not only be transported in an ionic 'that enrichment in nickel and copper must be state but also be fixed on biogenic fragments, looked for, not along the axis of an organic matter or shells. The fixation of these oceanic ridge but, above all, in the metals on calcareous micro-organisms neighbourhood of the fracture zones cutting certainly plays an important role in the across this ridge. In the fracture zones, in genesis of nodules. In the abyssal depths, at effect, the bottom water is enriched in the level of the hydrographic lysocline, cal- transition elements (enrichment confirmed careous organisms, abruptly dissolved by cor- by the results of the Geosecs programme). rosive cold water masses, free metallic ions, The enrichment of the water comes about which precipitate on a substrate: a fragment either by weathering of outcropping ultrabasic of volcanic rock, erratic rock fragment or rocks or from hydrothermal sources. indurated argile. The precipitation of At the outlet of the fracture zones the flow of transition metals seems to take place water enriched in metallic ions meets the preferably between the hydrographic cold bottom current. Low temperatures and lysocline and the carbonate compensation an oxygenated environment are favourable to depth or approximately between 4,100 and the precipitation of metallic ions and the 5,000 m. At greater water depths, their crystallisation of solid phases. Nickel and amounts diminish. The enrichment of
nodules in nickel and copper is a function not (ii) in an area with a thin sedimentary only of the availability of metallic ions in the cover (showing the absence of water and their rate of precipitation but also reworking of sediments, and of of the rate of precipitation of the manganese. turbidity currents); In a well oxygenated environment (i.e. zones (iii) in the proximity of an active oceanic with local acceleration of a bottom current: ridge, cut by fracture zones (a source channels, base of a great depth), the of metallic ions and a circulation area precipitation of manganese is rapid and the for waters enriched in metals towards dilution of transition metals by the the abyssal depths); manganese is important. Consequently, the (iv) not far from active volcanic relief (a manganese nodules which are the richest in source of metallic ions and of nuclei nickel and copper are found on sea floor of crystallisation); swept by bottom currents but more in zones (v) on the course of bottom currents, but of deceleration (and therefore less even more in a zone of deceleration oxygenated) then in those of acceleration of of these currents (a zone sufficiently these currents. The example of the Tiki and oxygenated to allow the precipitation Tapu basins is significant. of nickel and copper, but not favourable to a too rapid precipitation of nickel and copper, but not favourable to a too rapid precipitation of manganese); (vi) at a greater depth than that of the hydrographic lysocline (4,100-4,200 m., a zone of intensive dissolution of carbonates). In the context of nodule exploration, these Criteria may serve as a “guide”. This guide, However, must depend not only on Knowledge of the present oceanic Environment but also on that of the palaeo- Environment: palaeobathymetry, palaeo- Currents, and palaeolysocline.
deductions from our observations in the ACKNOWLEDGEMENTS South Pacific, we propose (Fig. 3) a model of We thank the crew of the CNEXO formation and distribution of "nickel- and "Nodules Project" for kindly giving us access copper-rich nodules", that is to say of nodules to their cruise results. We also thank those of of economic interest. In brief, in the South the geophysical-geological department of Pacific, a field of "nickel- and copper-rich C.O.B. who helped us, whether in the course nodules" will be found: of discussions or in putting their talents as (i) far from continents (in order to avoid draughtsmen at our disposal. This paper was dilution of metallic ions by a too rapid originally presented at the 9th International rate of sedimentation); Congress of Sedimentology, Nice, 1975.
REFERENCES ANDREWS, J. 1974: Recent investigations of the an Sedimenten aus dem zentralen Pazifischen Ozean. geological setting of manganese nodule deposits and Meerestechnik Mar. Technol. 5(6): 192-8. new approaches to the problem. lIe Colloque BERGER, W. H. 1970: Planktonic Foraminifera: International sur l'Exploitation des Oceans, Bordeaux, selective solution and the lysocline. Mar. Geol. 8: 111- Octobre 1974 .. 38. BARNES, S. S. 1967: Minor element composition of BERGER, W. H. 1974: Deep-sea sedimentation. In ferromanganese nodules. Sei., N. Y. 157: 63-5. Burke and Drake (Eds.), The Geology of Continental BEIERDORF, H. and WOLFART, F. 1974: Margins. Springer-Verlag, Stuttgart. p. 213-41. Sedimentologisch-biostratigraphische Untersuchungen BERGER, W. H. 1974: Deep-sea sedimentation Pp. 213-
60
SOME RESULTS OF SURVEYS FOR MANGANESE NODULE DEPOSITS IN THE PACIFIC OCEAN BY THE GEOLOGICAL SURVEY OF JAPAN
ATSUYUKI MIZUNO AND TOMOYUKI MORITANI
Abstract
In 1974, the Geological Survey of Japan began its systematic investigation of manganese nodules in the Central Pacific Basin on the new geological research vessel Hakurei Maru. The first cruise (GH74-5) was carried out over an eastern part area of the Basin (6°-10030'N, 164°30'-171030'W), and we report here the preliminary results on the occurrence of manganese nodule deposits, paying . particular consideration to their relationship to submarine topography and surficial and sub-bottom sedimentary facies. The surveyed area comprises a deep-sea basin at 5,000-5,400 m. defined to the north and east by the chain of seamounts and guyots of the Christmas Ridge. The deep-sea basin is divided roughly into two contrasting topographic features. The eastern part is characterised by flattened topography resulting from continuous deposition of turbidites; the meridian and western parts are characterised by gently rolling topography and the existence of a large number of deep-sea hills. Manganese nodules are almost lacking in the former flattened eastern area, whereas they are widely distributed in the latter rolling meridian and western parts. The population density of nodules varies from less than 1 kg/m2 to 26 kg/m2 and the higher density is found in the siliceous-calcareous ooze zone of rather small, flat basins surrounded by deep-sea hills. The density is closely related to the thickness of the transparent layer obtained by 3.5 kHz PDR profiling over the whole area. COJ1Sidering the various data of grab sampling, 3.5 kHz PDR profiling and to a lesser extent of deep-sea television and camera observations, the most promising mangane~ field in the present area seems to be confined to the north of the western sector of the area
INTRODUCTION preparatory studies and four subsequent Deep-sea manganese nodule deposits are cruises (Figs 1, 2). These investigations were now regarded as a feasible future. mineral carried out on chartered vessels putting resource and continuing efforts have been emphasis on the survey techniques and made by many countries of the world to relying mainly on the conventional dredge understand and clarify the nature and and bathymetry. distribution of the deposits. In April 1974, a geological research vessel In Japan, the Geological Survey of Japan Hakurei Maru (1,882 gross tons, 87 m. (GSJ) is playing a leading role in the length) equipped with various modern investigation of the manganese nodules and instruments was launched, and this is the main research organisation conducting contributed to enlarging the scale of study surveys on the deposits. Primarily, our and to raising the quality of surveys for investigations aim at basic scientific studies to manganese nodules. The Central Pacific clarify the occurrence, mineralogy and Basin was chosen as the main survey area for genesis of the deposits, together with studies the manganese nodule deposits rich in nickel on the prospecting methods. This will also and copper in the following year's studies. serve as the reconnaissance phase of the The cruise GH74-5 lasting 65 days on the prospecting. East Pacific Basin and cruise GH74-7 lasting The GSJ has conducted preliminary 35 days to the east of Okinawa Islands were investigations on manganese nodule deposits carried out as case studies on the manganese in the northwestern Pacific areas during the nodule deposits in the area near Japan (Fig. period 1969-1973, including the first one-year 1).
62
Here, we would like to report the western Pacific areas, and from 13 localities preliminary results of the cruise GH74-5 as by the cruise GH74-7, 1974 in the area east the first instance of a systematic survey, of Okinawa Islands (Figs 1, 2 and Table 1). based on Mizuno et al., (1975) and Mizuno The nodules occur mainly on the upper and Chujo (1975), and partly including the slopes of seamounts and in small basins results of other cruises (Takeda et al., 1974 between seamounts. Although detailed and other data concerned). examination and comparative studies of their mineralogy and chemistry have not yet been RESULTS OF THE INVESTIGATION made, these nodules from the northwestern CRUISES 1970-1973 AND Pacific areas generally show rather restricted G H7 4- 7, 1974 areal distribution and lack of persistency in ore amount of the deposits, higher Fe and Co Manganese nodules and crusts were and lower Ni and Cu contents and some dredged from 18 localities by the four cruises difference in the environment or condition of during the period of 1970-1973 in the north- manganese accumulation compared to the
topographically flattened eastern area and characterised by a remarkable change of ,part of the central-western areas. The deep- colour within them, showing a more or less sea clay shows a limited distribution in the mottled structure. From their occurrence, central-western area. most of such mottles seem to be due to The results of the sediment type analyses lebensspuren, and bioturbation is therefore demonstrate rather uniform vertical facies widely recognised throughout the surveyed distribution patterns through the uppermost area. Traces of activity of benthic organisms, 30 cm. columns of the grab samples, such as holes, tracks and pellets were although there are some exceptions such as observed on sea floor photographs (Figs 13, represented by the column at 81. 129, in 15). According to the analytical results, the which a poorly sorted sand layer with some change of colour probably results from the manganese nodules is found at the horizon of change in the content of stained radiolarian 20-25 cm. tests. However, the uniform facies in the According to some data of piston cores, the uppermost 30 cm of the columns are upper 5 m. sequence also shows no re-
The population density of nodules varies in The higher density tends to be found in , the nodule distributional area, ranging from siliceous-calcareous ooze/clay, and the lower less than 1 kg/m2 to a maximum of 26 kg/m2 density in the siliceous ooze/clay zones. (St. 143). Its distribution is irregular and th~ Considering the relation to submarine density changes abruptly even over short topography, the following can be deduced: distances. However, an area with density (1) the eastern area where nodules are absent more than 5 kg/m2 seems to exist rather corresponds to the wide flat plain with thick widely in the north of the western half of the sub-bottom turbidite layer. This suggests that area, particularly in the vicinity of St. 143. Kaufman's (1974) comment on the nodule The nodule distribution pattern is likely to occurrence and topography is also true in the be closely related to the type of sedimentary present area; (2) in the hilly central-western facies, topography and thickness of part of the area, the distribution of the higher transparent layer by 3.5 kHz PDR profiling. density nodules is restricted to the siliceous- Fig. 19 shows the relationship between the calcareous sediments in small flat basins population density and sedimentary facies. surrounded by deep-sea hills, and in the basin
extending along the boundary between the case, a roughly normal relationship is northern seamounts chain and hilly area. recognised. For example, a density of more It is suggested that there may exist some than 20 kglm2 is confined to the distribution relationship between the manganese nodule area of the transparent layer of 10-20 m. distribution and thickness of the transparent thickness; and (5) angular slab-like nodules layer by 3.5 kHz PDR profiling particularly tend to be occurred in the area with the on the western-central area. transparent layer of less than about 10 m. Figs 21 and 22 show some 3.5 kHz records thickness. and above-mentioned relationships. From Whether such a relationship is generally the results of the examination of thickness recognised or not in the environs of the changes of the transparent layer as shown in survey area remains a problem to be solved Fig. 21 and density changes of the nodules, in future studies but, if it is true over a large some interesting and probable relations are area, then 3.5 kHz PDR profiling will play an drawn as follows: (1) in general, the effective role in a quantitative survey for population density depends on the thickness nodule deposits, particularly during the of transparent layer with some rare earlier stages of prospecting. exceptions; (2) the transparent layer with a Data on the metal contents of the nodule thickness of more than about 30 m. relates to samples obtained here are not yet available. only very poor nodule occurrence with According to the summary of the previous density less than about 2 kglm2; (3) a density results of chemical analyses on nodules by of more than about 20 kg/m2 is related to the Horn et al. (1973), the nodules in the vicinity transparent layer with thickness less than of 81. 137 have the content of Mn 20.7%, Ni about 20 m.; (4) moreover, in the latter third 1.07%, Cu 1.22% and those near Sts 129, 144
have content ofMn 24.1%, Ni 1.11% and Cu Japan is in its preliminary stage. Both an 0.98%. improvement in the survey methods and the From the above discussions, the most accumulation of areal and detailed in- promising area for nodule deposits in the formation on the deposits in future survey surveyed area may be expected in the flat works is required in order to understand and deep-sea basin floor north of the western- clarify the nature and distribution of the central part of the area immediately south of manganese nodules as a feasible future the northern seamounts chain, where the mineral resource, and to contribute to the promising ore deposits are distributed in reconnaissance phase of prospecting. rather narrow belt prolongated in the direction of WoE, including Sts 143, 144 and 129 (220 km x 37-74 km.), although a ACKNOWLEDGEMENTS further detailed survey is necessary to Most part of this paper was cited from the delineate the exact extension of the deposits. manuscripts of the cruise GH74-5, 1974 and Another rather densely concentrated area of its preliminary summary (Mizuno and Chujo nodules with higher metal contents is found 1975; Mizuno et al. 1975), which were in the small flat basin near S1. 121, contributed by many colleagues of the surrounded by deep-sea hills, but the Geological Survey of Japan and the National promising deposits seem to be limited only Research Institute for Pollution and on the extent of about 37 km x 37 km. Resources. We are greatly indebted to J. Chujo, N. Yamakado, E. Honza, S. CONCLUSIONS Maruyama, M. Arita, K. Handa, K. Systematic and full-scaled investigations Ishibashi, Y. Kinoshita, F. Murakami, M. for manganese nodule deposits in the Central Nohara, K. Sawada, K. Tamaki and T. Pacific Basin area by the Geological Survey of Usami. 78
related to each other, and represent each took place along or related to boundary phase of mineralisation in the course of the fractures between them. evolution and differentiation of ascending or From the above-mentioned facts, the migrating hydrothermal solutions from depth palaeogeographic or depositional conditions to the sea floor. in which the bedded ore deposits have been In particular, the bedded manganese oxide formed seem to be those of a shallow sea deposits occurring in the sedimentary rocks near the margin of a sedimentary basin, and associated with jasper and argillised rocks where local depressions for trapping the man- or altered clays indicate their composite ganese compounds have also appeared on the nature and that they have been formed under uneven surface of the underlying rocks. the influence of both hydrothermal and The chemical environment of the bedded sedimentary factors as hydrothermal (or hot manganese deposits reflects high Eh or spring) sedimentary deposits. oxidising conditions, presumably of normal Stratigraphic horizons of the bedded sea water. manganese deposits are mostly in the In the regional polymetallic mineralisation, formations of the Nishikurosawa stage the manganese deposits are generally situated (middle Miocene), and only partly in that of along the outer or upper part of the main the Onnagawa and Funakawa stages (middle mineralised core zones with such deposits of to late Miocene). Au-Ag veins, Cu-Pb-Zn veins and Kuroko The age of predominant manganese deposits. This indicates that manganese mineralisation is therefore middle Miocene mineralisation took place in the oxidising (Nishikurosawa stage) and this roughly zone some distance from the reducing coincides with that of the Kuroko deposits. conditions of the core zones. The formations containing the bedded Although the manganese deposits are manganese deposits of the Nishikurosawa certainly related to the volcanic activity, its stage are, in most cases, of shallow sea type manner of association is very complicated. facies, and the structural features of the This is because the hydrothermal manganese districts are characterised by mineralisation generally takes place in the existence of dome and uplifted zones within waning stages of volcanic activity, with the geosynclinal basins, wider or local considerable lags in space and time after the unconformities or thinning of strata on the magmatic activity rather than in an uplifted part. These features reflect the phase independant hydrothermal activity. of the tectonic or geosynclinal history of the The problem of the source of manganese is Nishikurosawa stage, in which subsidence difficult to solve using only the data of and accompanying transgression prevailed regional geology, which indicates the almost over the entire Green Tuff Region. At possibility of both hydrothermal leaching of the same time, relative differential block metal components from the underlying movements of uplift and subsidence became volcanic rock formations or older basement apparent and characteristic volcanic activity rocks and the differention of the metal such as basalt eruption and mineralisation components from the deep parental magma.
REFERENCES
MORIT ANI, T. In press: Geology of the Miocene bedded manganese ore deposits in Japan. Report of Geological Survey of Japan.
82
STYLES OF MINERALISATION IN THE SOLOMON ISLANDS - A REVIEW
G. R. TAYLOR
Abstract
The distribution of vulcanicity and seismic activity in the Solomon Islands is seen as a consequence of fracturing in a tensional environment. Igneous rock types of the Solomons may be classified into two broad associations. The earliest island arc volcanics constitute an ophiolite suite and in it occur Cyprus type massive sulphide deposits and ferromanganese sediments. Younger mature island arc volcanics are highly differentiated and show calc-alkaline affinities. The whole is interpreted as a natural consequence of major rifting with Carey's Tethyan Shear System. The mode of occurrence, mineralogy and chemistry of manganiferous sediments indicate that there are two distinct categories. A demonstrably exhalative variety is often associated with massive sulphide deposits whilst the origin of pelagic manganese sediments is uncertain. The nature of mineralisation occurring within the mature island arc volcanics may be correlated with an evolution in the style of volcanism. As porphyry copper deposits are frequently ascribed to the redistribution of metals by thermally driven hot-water systems the chemistry of several such active systems are described. It is concluded that whilst such studies may explain present day exhalative phenomena their relevance to ophiolite associated. mineralisation is limited due to the different characteristics of the hydrothermal systems involved.
INTRODUCTION the available evidence which is indicative of a The geology of the Soiomon Islands is both regional tensional environment and the varied and complex. This paper presents a present author now believes that the genesis simplified view of the evolution of the island of the Solomon Islands may be explained arc and its integral rock types which without reference to subduction, obduction or facilitates the differentiation of two distinct downgoing lithospheric slabs. categories of mineralisation. The presentation The vulcanicity and seismicity is shown in of the regional geology relies heavily on the Fig. 1. The northern and southern zones of work of J. C. Grover, B. D. Hackman, P. J. most intense seismicity are concordant with Coleman, R. L. Stanton, R. B. Thompson the deepest development of the offshore and geologists of the Solomon Islands trenches which lie to the south of the island geological survey. chain. The seismic zones in these areas are close to vertical (Denham 1971); a feature A Synthesis of the Regional Geology which calls for severe flexure in any proposed "The anomalous nature of the Solomon down going plate. Intense Cenozoic Islands arc has been described by several Volcanism is confined to an area which is authors (Milson 1970, Denham 1971, relatively aseismic and is characterised by Hackman 1973). Recently Titley (1975) has minor shallow shocks (Westwood 1970). The explained porphyry copper mineralisation in NE-SW alignment of volcanic centres in the the Solomons in the terminology of the New New Georgia group, normal to the island arc Global Tectonics (Isacks et al. 1968). Titley axis was noted by Stanton and Bell (1969). believes the episodic nature of the This has been emphasised by the recent mineralisation to be related to variations in discovery of a major submarine volcano in the convergence rates of the Australian and the New Georgia sound and the recognition Pacific Plates in a compressional of the tensional environment within the environment. However, this is contrary to Vella Graben (G. R. Taylor, in press). A
83 graben like feature also affects N.W. intrusions of late Cretaceous to Miocene age. Guadalcanal. Major faults on Guadalcanal, It contains all known occurrences of cupri- Savo, Russel Is. and Vella Lavella parallel ferous massive sulphides and all significant this trend. The Coral Sea to the south of the deposits of exhalative manganiferous Solomons possesses a horst and graben sediments. These are the basement· morphology of extensional character (Ewing complexes of earlier authors which were et al. 1970) whilst sinistral displacement of given differing local names on each island. en-echelon horsts is seen on Rennell Island The second, mature island arc suite, ranges in (Taylor 1973). The oceanic crust in this age from the Pliocene to the present and in region is abnormally thin as is indicated by a composition from picritic basalts and basaltic mantle depth of 8 km. south of the New andesites to hornblende andesites of calc- Georgia Group (Furumoto et. al. 1970.). alkali affinity and is clearly a highly These features are all compatible with a differentiated series. This suite contains tensional environment affected by east-west several intrusive diorite stocks with as- shearing parallel to the axis of the island sociated porphyry copper mineralisation and chain. also minor high level epithermal gold It can be shown that the evolution of deposits. igneous activity through both space and time Major element analyses show the eruptive was accompanied by an evolution in the rocks of the ophiolite suite (Hackman 1971 styles of mineralisation. The igneous rocks of and 1973, Taylor unpublished data) to be low the Solomons may be divided into two major potassium tholeiites (K20, 0.17-0.75%). The associations. In practice there is probably a lavas are frequently pillowed and show gradation between the two. The oldest suite spilitic alteration and this is reflected in a high (Fig. 2) is an ophiolitic complex composed of variability of the total alkali content. island arc tholeiites, gabbroic and ultramafic Titanium is relatively high but variable 84
(Ti02, 0.57-2.10%) as is the silica content samples from Bougainville and New Georgia (44.6-59.2%). The volcanics on Hanesavo (Page and Johnson 1974) implies "derivation show a distinct iron enrichment trend. Rare from relatively homogeneous upper mantle earth abundances show a cholidritic pattern sources". (Jakes and Gill 1970). The suite is typical of Several authors (Coleman 1972, Hackman the island arc tholeiites as defined by Jakes 1973) have emphasised the importance of and Gill (1970) and the petrochemical Carey's interpretation of the regional variability relative to that of abyssal tholeiites structure. The Tethyan Torsion System is suggestive of considerable· near surface (Carey 1958), which includes the Solomons fractionation. Megashear, is an irrefutable feature of global The mature island arc suite shown in Fig. tectonics. Until such times as contrary 2 includes the New Georgia Suite and the evidence is presented it seems reasonable to Gallego Lavas of earlier authors (Stanton and analyse the structural evolution of the Bell 1969 and Hackman 1971). Published Solomons using Carey's approach. For the analyses by R. 1. Stanton (1969) indicate purpose of this analysis the two igneous that A120a and (Na20 + K20) are distinctly suites described above may be encompassed higher than found in the ophiolite suite. by two overlapping envelopes (Fig. 3), These There is a great range in silica contents from have been constructed by enclosing all the picritic lavas of New Georgia (Si02- known outcrops of each suite. Each envelope 44.99%) to the hornblende dacites of the is approximately rhombohedral in shape and Gallego lavas (Si02-67.96%). The rare earth is therefore, in Carey's terminology, a abundances (Jakes and Gill 1970) show rhombochasm. Grossly simplifying the main enrichment of the lighter elements (La and structural features, the oblique tensional Sm). An extremely narrow range in the fractures within the New Georgia strontium isotope ratios (Sr87/Sr86) of Rhombochasm, and the gravity folds of the
85
Malaita fold belt are seen to be the expected Mineralisation in the Ophiolite Suite consequences of sinistral shear in an approxi- Minor occurrences of several categories of mate east-west direction. The westerly mineralisation occur in the Ophiolite Suite. migration of the active shear zone is a natural Ultramafic and gabbroic intrusions are consequence of the generation of new crust frequently associated with minor nickel in the Florida Rhombochasm following the sulphides, disseminated pyrite/chalcopyrite, eruption of increasingly fractionated lavas podiform chromite and nickeliferous laterites. through the now annealed· fractures. This Cupriferous sulphide and manganese interpretation is merely a rationalisation of sediments of presumed exhalative origin are the mechanisms postulated by Coleman and the subject of this review and the distribution Hackman in 1974. There is increasing of these is shown in Fig. 4. This is assembled argument that island arc ophiolites represent from Geological Survey publications and primary mantle material erupted through contains no recent mining company data. tensional features (Stonley 1975, Miyashiro Although of particular interest the massive 1975, Thompson and Melson 1972) and this sulphide bodies are but a minor localised concept may have wider applicability than is facet of what is otherwise a monotonous pile presently suggested. of mafic volcanics. The ore bodies are Geophysical evidence (Rose et al. 1968) invariably small (less than 100,000 tons) but indicates that the Solomons has no root zone they contain high grades of copper, and more with the mantle depth being similar to that of rarely, zinc. They frequently occur within neighbouring oceanic areas. The degree of extensive regional fault zones which show fractionation of both igneous suites, but stringer pyrite mineralisation but no signifi- particularly the mature island arc suite, is cant amounts of other base metals (Fig. 4). therefore difficult to resolve, but it may be All the known occurrences are associated partly related to the extent of vertical with sedimentary deposits of manganese and development of the volcanic edifices. ironstones. The similarity of this lithological 86
The massive sulphide deposit consists Mineralisation and volcanic exhalative primarily of pyrite with subsidiary activity within the mature island arc suite chalcopyrite, chalcocite and covellite. The The mature island arc suite is characterised average copper content is above 5 percent by a marked change in the style of and other important accessories are zinc, gold mineralisation which is related to the change and silver. The massive sulphide bodies are in the style of volcanism. The monotonous underlain by a stockwork zone of highly mafic lavas of the ophiolite suite are probably altered silicified and argillaceous lavas, which the products of fissure type eruptions carry much disseminated sulphide. The whereas the calcalkaline volcanics of the overlying Hanesavo Beds are composed of mature island arc suite form interdigitating ochres, sinter, ironstone 'and manganese strato-volcano complexes. Unroofing of the lenses. A complicated distribution of base subvolcanic dioritic stocks has developed to and precious metals suggests that the beds differing stages according to the age of the are direct precipitates from volcanic complex. Associated porphyry copper exhalative solutions. Emplacement of mineralisation occurs at Panguna (Bougain- massive sulphide was probably preferentially ville), Fauro (Shortllind Islands), Limbo located above the erosion surface of the lower (New Georgia), Koloula, Poha and Hidden basaltic volcanics, where interbedded pillow Valley (Guadalcanal). High level, epithermal, breccias were exposed. The structural gold and copper mineralisation occurs at controls thereby imposed by the local Atamo (Bougainville), Gold Ridge stratigraphy may distinguish the mineralis- (Guadalcanal), on New Georgia, and on the ation from the otherwise similar Cyprus-type active peleean volcano of Savo Island. The category of massive sulphide mineralisation. later occurrence is interesting in that it The Hanesavo Beds show complex probably represents recent mineralisation of a variations in chemistry and mineralogy that pyrite/chalcopyrite/gypsum association being will enable an assessment to be made of the' effected by a vapour-dominated hydro- nature of the mineralising fluids and the thermal system (Taylor 1975, in press). The physico-chemical factors affecting sea floor porphyry copper category of mineralisation is precipitation. Analytical data is now available thought to be the result of a hot-water for thirteen trace elements. Interpretation of dominated hydrothermal system whose this and whole rock analysis is now in activity was contemporaneous with the progress. Systematic X-ray diffraction andesitic volcanism. The deposits offer analysis has revealed the occurrence of consilierable opportunity to reconstruct the significant concentrations of twenty-five physico-chemical nature of the mineralising different minerals. Fig. 5 shows the solutions by reference to alteration effects distribution of certain iron and manganese and fluid inclusions. However, the study of minerals. Open circles indicate trace amounts actual volcanic emanations has also proved to whilst the size of the filled circles is be informative. proportional to the approximate amount of The' Paraso thermal area is the latest each mineral. The implications of the relative feature along a linear progression of volcanic abundances' of -pyrolusite, ramsdelfite and activity within the Vella Graben which todorokite are not, as yet, fully understood. strikes northeast-southwest across Vella Other X-ray data shows that the ochreous Lavella Island. The thermal area has a sediments contain montmorillonite and natural heat flow of 36 x I03Kcals/sec saponite. Other ,phyllosilicates such as (Taylor; in press) which is of the same order ganophyllite, chlorite and nontronite also of magnitude as some geothermal ,power occur with a systematic distribution. Zincite, producing areas. Both acid-sulphate cuprite, gypsum, calcite and anatase add to mudpools and neutral hot springs occur. the complexity of this exhalative sediment. Chemical analyses of the mudpool water In many ways it is similar to the mineralised shows it to be typical of that found above a Red Sea deposits. vapour dominated hydro-thermal system
89
within a recently active volcanic area. The intrusions that occur in the Poha and Bonege hot springs however have lower sulphate and rivers show distinct petrochemical affinity higher chloride contents suggestive of a hot- with the Koloula diorite, and although of water dominated hydrothermal system probable Oligocene age may date from the which may have potential as a geothermal commencement of Tertiary to Recent power source. The hot springs contain . andesitic volcanism in Guadalcanal. The significant amounts of dissolved iron and analysis of water collected by B. D. Hackman manganese but the Ngokosole River into from a spring within the Bonege diorite which they drain contains no iron and shows the spring waters to be highly manganese at its mouth due to precipitation mineralised brine. of ferromanganese hydroxides. Determinations of the base metal contents The precipitation effect resulting when of this sample are as follows: Fe 0.50, Mn residual volcanic solutions enter an oxidising 0.07, Zn 0.50, Cu 0.00 mg/L. It may be environment can be seen around the argued that the occurrence together of saline dormant volcano of Simbo Island. The water spring and diorite pluton is entirely fortuitous of a salt water lagoon close to hot spring but the frequent association on Guadalcanal outlets has a pH of 2.5 and both iron and of saline springs with diorite plutons implies manganese in solution. A way from the that the brines eminate from an aquifer that springs the buffering effect of the sea water draws its heat from the original intrusion and causes a rise in pH and a corresponding not from some later feature. The moderate decrease in iron and manganese in solution. value of the derived subsurface temperature This correlates with an increase in these and the high concentration of calcium in the constituents in precipitated sediments col- solution would suggest that argillic alteration lected from the sea floor (Taylor, in press). is occurring in the underlying rocks. The Hackman (1971) listed 24 saline, warm evidence therefore suggests that low tem- and hot spring occurrences on the island of perature hydrothermal lateration can take Guadalcanal. There is an apparent increase in place during a period measured in tens of the temperature of the springs from east to millions of years after the emplacement of west with the hottest springs occurring the intrusion. within Pliocene to Recent andesitic lavas of the Gallego Volcanics. Potassium-argon age ACKNOWLEDGEMENTS determinations (Hackman, 1971) show the This paper was written and presented latter to be 6.4 ± .2 million years old. Dates during the tenure of a Conzinc Riotinto of of 7.8.± .8 m.y. and less than 1.6 m.y. have Australia Post-Graduate Research been obtained from different intrusive phases Scholarship for which the author would like of the Koloula mineralised diorite. Older to express his gratitude.
CONCLUSIONS
A simplified view of the geological associated mineralisation may be very evolution of the Solomons has facilitated a limited. Although both categories are twofold division of the major categories of probably effected by saline solutions the mineralisation. The study of present day relative roles of seawater, meteoric water and exhalative systems may have applicability in primary magmatic fluids have not been understanding porphyry copper and related determined. The origins of the primary mineral deposits but its relevance to ophiolite metals are likely to be another variable.
90
A COMPARISON OF THE MORPHOLOGY AND MINERALOGY OF MANGANESE NODULES FROM THE SOUTHWESTERN PACIFIC BASIN AND NORTHEASTERN EQUATORIAL PACIFIC
M. A. MEYLAN
Abstract
The nodule populations of the Southwestern Pacific Basin and Northeastern Equatorial Pacific differ in external morphology, internal structure, mineralogy and chemistry. It is tentatively suggested that these differences are at least partly the result of the two populations representing different stages of nodule development, the nodules from the S.W. Pacific Basin being "younger" than those of the N.E.Equatoriai Pacific. This paper primarily will consider the differences in morphology and mineralogy. While neither population is homogenous, most of the nodules of the S.W. Pacific Basin can be characterised as spheroidal, while in the N.E. Equatorial Pacific, discoidal and ellipsoidal shapes predominate. Whereas polynucleate nodules have been found at only a few stations in the S.W. Pacific Basin, they are relatively widespread in the N.E. Equatorial Pacific. Nodules from the S.W. Pacific Basin usually display granular surface textures, while N.E. Equatorial Pacific nodules show a greater variety of textures - rough and botryoidal textures appearing on most nodules, while many have a smooth surface. Nodules from both areas generally fail in the 0.5-10 cm size range, however, with the smaller nodules usually being spheroidal. ' Bulk X.R.D. analysis has demonstrated that regional differences also exist in ferromanganese oxide and detrital mineralogy of the nodules. Nodules from the N.E. Equatorial Pacific generally contain greater proportions of todorokite than those from the S.W. Pacific Basin, where delta-MnO2 appears to be the dominant manganese mineral. N'odf.1es from the S.W. Pacific Basin contain decreasing amounts of incorporated detrital quartz and feldspar as distance from the Antarctic continent 'increases, while nodules from the N.E. Equatorial Pacific contain small but relatively constant amounts of quartz and feldspar.
INTRODUCTION and N.E. Equatorial Pacific are related to Manganese nodules from the differences in mineralogy 'and internal Southwestern Pacific Basin and Northeastern structure. This paper considers primarily the Equatorial Pacific differ in morphology, mineralogical and morphological aspects, the internal structure, mineralogy and chemistry. latter often reflecting internal structure. Considering the nodule fields as potential ore The nodules used for this study were bodies, the most significant difference is in collected mostly during two cruises of the nodule chemistry. Nodules recently R.V. Moana Wave to the N.E. Equatorial recovered from the economically attractive Pacific (Mn-74-01 and Mn-74-02) and during zone of the N.E. Equatorial Pacific contain on a R.V. Tangaroa cruise to the S.W. Pacific the average more than 25% manganese, Basin. Additional samples were available in about 1.5% nickel, 1.2% copper, and 0.2% the N.E. Equatorial Pacific from a R.V. Kana cobalt (J. Mero, B. Dugolinsky, S. Margolis, Keoki cruise undertaken in 1972 and from and J. Craig, pers. comm.). These averages the R.V. Valdivia cruise V A-09 of 1974, contrast to those for nodules from the S.W. while in the S.W. Pacific Basin, V.S.N.S. Pacific Basin: about 17% Mn, 0.4% Ni, 0.2% Eltanin cruises provided supplementary Cu, and 0.4% Co (Glasby et al., 1975). It samples. The Mn-74-01, Mn-74-02 and VA- appears likely that the differences in nodule 09 cruise tracks are shown in Fig. 1; the composition between the S. W. Pacific Basin Tang,'o1oa station sites and nodule deposit
92
limits in the S.W. Pacific Basin are shown in Pacific and S.W. Pacific Basin is, however, Fig. 2. similar, being rolling abyssal hills and plains, While the significance in terms of nodule interspersed with seapeaks and seamounts. genesis is not clear, the N.E. Equatorial Depths are also generally similar, the nodules Pacific and S.W. Pacific Basin occurring mostly from 4,500-5,500 m., ferromanganese deposits occur in contrasting although the average nodule depth is structural and sedimentation regimes. The probably somewhat greater in the S.W. most extensive nodule fields in the N.E. Pacific Basin Equatorial Pacific occur on radiolarian clay and ooze between the Clarion and Clipperton Nodules in the N.E. Equatorial Pacific rest Fracture Zones (Horn, 1913a). The nodule upon a thin Quaternary sediment cover fields of the S. W. Pacific Basin occur on (probably mostly less than one metre thick) brown silty clay and clayey silt (Goodell et or on Tertiary sediment (Riedel and Funnell, al., 1971, Meylan et al., 1975), with the 1964; Theyer, pers. comm.), whereas the Tangaroa study area being east of the Tonga Quaternary sediments of the S.W. Pacific trench and just south of a region of volcanic Basin may be many metres thick seamounts. The small-scale structure of the (extrapolated from Goodell et al., 1971; nodule areas of both the N.E. Equatorial Burns, Andrews et al., 1973).
93
occur nodules with quite different Almost all nodules from both the S.W. 'morphologies. Pacific Basin and' N .E. Equatorial Pacific fall Nodules from the S.W. Pacific Basin into the 0.5-10 cm. size range. Tangaroa cruise have been examined by Concretions from the two study areas also Meyan et al. (1975) and Backer et al. (1976). show differences in surface texture. Those They reported a preponderance of spheroidal from the S.W. Pacific Basin generally possess shapes from the traverse southwest of granular surfaces, with gradations to Rarotonga, while south of Rarotonga, a microgranular and to microbotryoidal, often greater variety of shapes was encountered: with a cavernous aspect (for a discussion of discoidal, botryoidal, spheroidal with surface textural terms, see Meylan, 1974). equatorial "skirts", and more irregular forms Textural gradations frequently occur on in addition to many spheroidal concretions. larger individual nodules; on some the upper Botryoidal nodules (polynucleate nodules or surface is more coarsely granular than the nodules consisting of intergrown spheroidal lower surface, and on others the reverse is forms) were found only at four stations at true. the southern end of the traverse south of Nodules from the N .E. Equatorial Pacific Rarotonga. At several stations where the usually have rough (friable granular) or larger nodules deviated from a spheroidal botryoidal surfaces, with smooth surfaces form, the smaller nodules tended to be often being observed as well. The textures spheroidal. Many nodules of all sizes and are not as uniform as on nodules from the shapes displayed faceting i.e., had one or S.W. Pacific Basin, with larger specimens more flattened surfaces. frequently having a pronounced botryoidal Nodules from the N .E. Equatorial Pacific equatorial belt. Where lower surfaces are have been examined by Meylan (1974), rough, upper surfaces are likely to be among many others. Discoidal and ellipsoidal botryoidal, and where lower surfaces are shapes predominate and polynucleate botryoidal, the upper surface may approach a nodules are more frequently encountered smooth texture. than in the S.W. Pacific Basin, although they Probably one of the most significant seldom are the most abundant nodule form differences between nodules from the two in any particular sample. The frequency of areas of interest is the internal structure. occurrence of the different morphologic Concretions from the S. W. Pacific Basin classes of nodules collected during the Mn- Tangaroa collection consist of a thin (about 74-01 cruise is shown in Table 1, further 1-2 mm thick) coating of ferromanganese subdivided by size.
and N.E. Equatorial Pacific can be volcanic nucleus, but we must look at a wider rationalised if the nodules from the former geographic sampling of nodules to determine area are considered to be "younger" than if an increase in the Mn/Fe ratio and a those from the latter area. The apparently concurrent increasing predominance of thicker Quaternary sediment cover of the todorokite over delta-Mn02 are typical trends S.W. Pacific Basin, the deposition of which in the aging of a manganese nodule. might have restricted nodule growth until fairly recent times, along with the relative ACKNOWLEDGEMENTS thinness of the accreted ferromanganese Most of the support for this research has oxides, tends to support the idea that S.W. been provided by the Seabed Assessment Pacific Basin nodules are "younger". Since Program, IDOE-NSF, which also pro- the principal nucleus type of N .E. Equatorial vided a part of the travel support for Pacific nodules is a previously-existing nodule attendance at this meeting. Col1ection of fragment, most are obviously in at least a samples in the S.W. Pacific Basin was made second stage of growth. possible by an invitation from the New It is tempting also to attribute the Zealand Oceanographic Institute, D.S.I.R., to differences in chemistry and ferromanganese participate in the Tangaroa cruise with Dr. oxide mineralogy to differences in G. P. Glasby as cruise leader. Preparation of developmental stage, but existing data are the manuscript was facilitated by discussions insufficient to provide support for this with Drs J. E. Andrews and S. V. Margolis. hypothesis. Diagenetic changes are obviously Particular thanks is due the Mining Ventures an important aspect of the history of those Division of Shell Oil Company for supplying nodules which grow from an easily-altered the airfare needed to attend this meeting.
REFERENCES
BACKER, H., GLASBY, G. P. and MEYLAN, M. A. HORN, D. R., HORN, B. M. and DELACH, M. N. 1976: Manganese nodules from the Southwestern 1973b: Factors which control the distribution of Pacific Basin. N20I Oceanographic Field Rep. 6: ferromanganese nodules and proposed research 88 pp. vessel's track North Pacific. Tech. Rep. Off into BURNS, E. E., ANDREWS, J. E., et al., 1973: Initital Decade Ocean Explor. 8: 20 pp and 20 profiles. reports of the Deep Sea Drilling Project, vol. 21. U.S. MARGOLIS, S. V. and BURNS, R. G. 1976: Pacific Government Printing Office, Washington D.C. 931 pp. deep-sea manganese nodules: their distribution, GLASBY, G. P., BACKER, H. and MEYLAN, M. A. composition and origin. Ann. Rev. Earth Planet. Sci. 5. 1975: Metal contents of manganese nodules from the MEYLAN, M. A. 1974: Field description and Southwestern Pacific Basin Erzmeta/l 28: 340-2. classification of manganese nodules. Hawaii Inst. GOODELL, H. G., MEYLAN, M. A. and GRANT, B. Geophys. Rep. HIG-74-9: 158-68. 1971: Ferromanganese deposits of the South Pacific MEYLAN, M. A., BACKER, H. and GLASBY, G. P. Ocean, Drake Passage, and Scotia Sea. Antarctic Res. 1975: Manganese nodule investigations in the Ser. 15: 27-92. Southwestern Pacific Basin, 1974. N20I Oceanogr. HORN, D. R., HORN, B. M. and DELACH, M. N. Field Rep. 4: 24 pp. 1973a: Ocean manganese nodules metal values and RIEDEL, W. R. and FUNNELL, B. M. 1964: Tertiary mining sites. Tech. Rep. Off into Decade Ocean sediment cores and microfossils from the Pacific Ocean Explor. 4: 57 pp. floor. Q. JI geol. Soc. Lond. 120: 305-68.
98
CHEMICAL COMPOSITION OF MANGANESE NODULES FROM THE PACIFIC-ANTARCTIC OCEAN, DRAKE PASSAGE AND SCOTIA SEA: RELATION TO FERROMANGANESE OXIDE MINERALOGY AND NUCLEUS TYPE
M. A. MEYLAN AND H. G. GOODELL
Abstract
Manganese nodules from the Scotia Sea and Pacific Ocean s~tors of the Southern Ocean display regional differences in chemical and mineralogical composition. Mn/Fe and todorokiteI8-MnO. ratios decrease southeasterly from the Southwestern Pacific Basin to the Drake Passage-Scotia Sea. Relative to Mn, the greatest enrichments of Co, Ti, V, Zn, Mo and Sn occur in Drake Passage-Scotia Sea nodules, while Ni and Cu are most enriched in the Southwestern Pacific Basin. Manganese mineralogy is directly related to the Mn/Fe ratio, so that todorokite rather than S-MnO. predominates when high manganese vs. iron contents are found. Trace metal contents are also related to mineralogy, with high values of Ni, Cu, Zn and Sn, and low values of Co, Ti, V and Mo associated with high todorokite contents. Nodules having indistinct zeolitite/vesiculite nuclei tend to have Mn/Fe and todorokite/ll-MnO ratios significantly higher than average for the region in which they are found, whereas ferromanganese oxide fragments dredged from submarine rock surfaces usually have Mn/Fe and todorokite/ll-MnO ratios lower than regional average.
INTRODUCTION SAMPLE DISTRIBUTION The Scotia Sea and Pacific Ocean sectors of Ferromanganese accumulations of the the Southern Ocean were extensively study area are most abundant in two regions, sampled during cruises 5-27 of the U.S.N.S. one a belt along approximately 600S. Lat. Eltanin. Representative Fe-Mn concretionary beneath the Antarctic Circumpolar Current, material from most of the 83 sediment cores and the other an extensive zone in the and 122 dredges containing ferromanganese Southwestern Pacific Basin (Goodell et a/., nodules or crusts has been analysed for 1971). The locations of samples collected and composition and mineralogy. Statistical analysed from these two regions, as well as analyses of elemental relationships, trend those from other parts of the study area, are surfaces showing the geographic variation of shown in Fig. 1. All the samples were elemental contents, and maps of regional collected either by dredging (using devices of nodule mineralogy are contained in Goodell several sizes, including Blake and Menzies et at. (1971). The purpose of this paper is to trawls with coarse mesh nets, and rock present the results of the chemical analyses dredges with chain bags) or by piston coring (Appendix B), and compare nodule (inside diameter of plastic liner 5.6 cm.) composition with mineralogy and type of Samples selected for analysis were those nucleus after re-examining regional elemental nodules or crust fragments of sufficient size variations. The chemical data have to provide powdered material for bulk previously been incorporated in charts of chemical and mineralogical analysis element contents in nodules in the South Pacific (Glasby and Lawrence, 1974a-e; see ANALYTICAL PROCEDURE also Glasby, In press). Except for a number of nodules taken
99
from sediment cores which still retained divided into five regions: (1) Southwestern some moisture content, samples had been Pacific Basin, (2) Pacific-Antarctic Ridge air-dried for variable lengths of time prior to (East Pacific Rise or Albatross Cordillera), preparation. The nodules were first broken (3) Chile Rise, (4) Bellingshausen Basin open and the non-ferromanganese nucleus (Southeastern Pacific Basin), and (5) Drake material removed; in some cases the nucleus Passage-Scotia Sea (including Scotia Arc). material was prepared for analysis separately. Samples near the arbitrary boundaries of any The ferromanganese oxides (laminated two regions were considered to belong to crusts, at times with massive or irregularly both regions. structured subcrusts) were lightly ground The average metal contents of nodules with mortar and pestle and the incorporated from the five regions are listed in Table 1. sand or larger detritus picked out. The (Only those nodules used for both chemical powder was cloth screened to a size suitable and mineralogical analyses were included in for analysis. the calculations). Proceeding southeastward Chemical analyses were performed by the from the Southwestern Pacific Basin toward Materials Laboratory of the Newport News; the Drake Passage-Scotia Sea, it is' obvious Shipbuilding and Drydock Company. Fe, that the contents of manganese, nickel and Mn, Cu, Co, and Ni were analysed by X-ray copper show dramatic decreases, and fluorescence and Ba, Sr, Zr, Ti, Mo, Sn, and molybdenum a less pronounced one, while V by atomic absorption. The latter technique iron displays a significant increase. The was also used to attempt measurement of the highest cobalt and titanium contents are content of Cr, Ag, W, Hg, and Sc, but these found in the basins, while vanadium and zinc five elements were consistently below the occur in greatest abundance on the mid- detection threshold (Cr < 0.03%, Ag < ocean ridge system. Tin content, while 0.02%, W < 0.05%, Hg < 0.02%, Sc < highest in the Drake Passage-Scotia Sea, is 0.10%). also higher on the ridge system than in the Mineralogical analysis involved adjacent basins. Zirconium, strontium and interpretation of X-ray diffractograms. These barium do not appear to display coherent were prepared with a Norelco diffractometer regional trends. system with gas-proportional detector and Ferromanganese samples were further pulse-height analyser. Nodule powders categorised according to type of nucleation packed into an aluminium planchet were material in order to determine possible subjected to nlckei-flltered Cu-Ka radiation, relationships between growth site and nodule and scanned at a rate of 1 °29 per minute. chemistry/mineralogy. Average ferro- Diffractogram peak areas were calculated by manganese oxide metal contents for the multiplying the base width and peak half- various types of nucleus are listed in Table 2. height. Occasional checks on the Glacial erratic nuclei include gran,ule and reproducibility of the peak areas were made larger sizes of sedimentary, metamorphic, by multiple runs on a single powder sample. and non-oceanic igneous rock fragments, It must be pointed out that measurement of usually somewhat rounded. The unaltered the "peak" area of the Il-Mn02 reflection at volcanics category includes glass, palagonite, 2.45A is subjective because' diffractograms and basaltic rock fragments. Their alteration typically display a wide regi0p, elevated above products, often a buff-coloured mixture of background at this spacing which is phillipsite and montmorillonite, and composed of one or more broad peaks that frequently having a vesicular aspect, were may represent a Il-Mn02 reflection mixed placed in the zeolitite/vesiculite categories. with secondary todorokite reflections. The "indistinct" zeolitite/vesiculite nuclei were those where the volcanic alteration CHEMICAL COMPOSITION products had in turn been nearly completely In order to assess regional variations in replaced by ferromanganese oxides. The nodule composition, the study area was indiscernible nuclei were instances where no 101
indication of the original growth site was Pacific-Antarctic Ridge from the visible. Ferromanganese oxide accumulations Southwestern Pacific Basin to the Drake placed in the "broken from nucleation site" Passage-Scotia Sea, decreasing Mn/Fe ratios category were large fragments apparently are matched by decreasing relative contents dredged from extensive rock surfaces and of todorokite vs a-Mn02. At the same time, typically having a massive rather laminar Mn/Ni ratios increase. But Mn/Cu ratios cross-sectional structure; this type of material increase only to the Bellingshausen Basin, cannot properly be considered as being and then show lower values in the Drake nodular, but should be regarded as crusts. In Passage-Scotia Sea, indicating a relative two cases, the nucleus was obviously a enrichment of Cu vs. Mn. Relative fragment of an older nodule that had enrichments of cobalt and titanium occur in fractured on the seafloor (some of the both the Bellingshausen Basin and Drake "indiscernible" nuclei may well belong in this Passage-Scotia Sea, while the latter area also category). One nodule contained a shark's shows the lowest Mn/V, Mn/Zn, Mn/Mo, tooth at its centre. and Mn/Sn ratios. In the Southwestern Bearing in mind that there is a definite Pacific Basin, the greatest depletions of V, regional influence on nodule chemistry, the Zn, Mo, and Sn vs. manganese of any region variations in metal content of the oxide crust are found. with differences in nucleus type will only be Because manganese oxide mineralogy is mentioned briefly at this point. As seen in interrelated with Mn/Fe variations, the Table 2, nodules having an indistinct nucleus types with lower Mn/Fe ratios zeolitite/vesiculite nucleus have the highest display higher 9.7A./2.4SA. peak area ratios manganese, nickel, copper and zinc contents (Table 4). Nodules with indistinct (disregarding the single nodule with a shark's zeolitite/vesiculite nuclei display the highest tooth nucleus), and have the lowest iron, todorokite proportions and Mn/Fe ratios. cobalt and titanium contents. Conversely, Relative to manganese, they are also the ferromanganese oxide fragments broken most highly enriched in nickel and copper, from accretion sites have the highest iron, and depleted in cobalt, titanium, vanadium cobalt, titanium (and vanadium), and the and molybdenum (disregarding the single lowest nickel and copper contents. Most of shark's tooth' nodule). Ferromanganese the nodules with indistinct zeolitite/vesiculite accumulations with the greatest relative nuclei are from the Southwestern Pacific content of 8-Mn02 are the large fragments Basin, which partially accounts for the high broken from accretion sites. This type of manganese (and associated element) concretionary material also shows the percentages. Similarly, concretions grown on greatest depletion of nickel and copper glacial erratics have the lowest manganese· relative to manganese contents, and the contents, but most of these are from the greatest relative enrichments of cobalt, Drake Passage and Scotia Sea, a region noted titanium, vanadium and molybdenum. for low Mn contents in all nodule types. The obvious correlations between geographic region and Southern Ocean MINERALOGY nodule chemistry and mineralogy might Apart from the ubiquitous detrital minerals obscure correlations between nucleus type incorporated into Southern Ocean nodules, and composition/mineralogy, so an attempt the principal crystalline material is a mixture was made to remove the regional influence of todorokite and 8-Mn02, which on X-ray approximately. A proportional deviation diffractograms can be identified by peaks at from regional mean was calculated for each 9.7 A. and 2.45A. respectively. The relative sample ratio (9.7 A./2.45A., Mn/Fe, etc.), and proportion of these two minerals is closely the values grouped according to nucleus type related to the Mn/Fe ratio, as can be seen in so that an average for each type could be Table 3. Proceeding southeastward across the calculated (Table 5). A value of 1.0 indicates
102
that a particular nucleus type has on the Conversely, high proportions of todorokite average the same mineral or elemental ratio are accompanied by relative depletions of Co as all nodules from the geographic region in and Ti, and to a lesser degree, V and· Mo. which it is located. The relationships just stated seem to be most Examination of Table 5 leads us to believe clearly developed for 9.71\/2.45A peak area that the two nucleus types already identified ratios above about 1.2-1.3. as apparently unusual - the "indistinct zeolitite/vesiculite" and the "broken from DISCUSSION nucleation site" - do indeed possess the Regional variations in nodule chemistry most deviant mineralogiCal and com- are reasonably well documented. Mero positional characteristics. In fact, very few (1962, 1965) was the first to divide the significant variations in any other nucleus Pacific into compositional regions, noting type are evident from Table 5 (disregarding iron-rich, manganese-rich, nickel/copper-rich, the last two categories which only have a and cobalt-rich zones. Although limited by a total of three samples between them). The scarcity of samples from south of 400S. Lat. oxide crusts of nodules with glacial erratic in the Pacific, Mero placed the Southwestern nuclei may be somewhat enriched in V and Pacific Basin and the Pacific-Antarctic Ridge Sn relative to Mn, while Cu and Sn may be southeast of New Zealand in the high iron depleted in nodules seeded by unaltered category; likewise, samples from the Chile volcanic material. Rise were similarly categorised. Most .of the Plots of Mn/Fe, Mn/Ni, etc. vs. central part of the South Pacific was 9.71\12.45A peak area ratio for each sample considered to belong in the nickel/copper-rich were also made in order to check the zone. Refinement and modification of these composition vs. mineralogy relations more geographic trends has resulted from the work closely. The plots provide some evidence that of Skornyakova et al. (1962), Cronan and certain of the minor metals are preferentially Tooms (1969), Price and Calvert (1970), associated with todorokite, rather than with Cronan (1972), Skornyakova and Andrush8-Mn02 or an iron-rich phase. In general, chenko (1974), and Glasby (In press). The increasing 9.7A/2.45A peak area ratios are few samples available to Cronan and Tooms associated with lower Mn/Ni and Mn/Cu (1969) seemed to indicate that the Southeast ratios (and apparently Mn/Zn and Mn/Sn, Pacific nodules were nickel-rich and cobalt also), indicating a relative enrichment of Ni, poor. Cu, Zn and Sn when todorokite assumes A multitude of samples became available increasing abundance relative to 8-Mn02. from the Southern Ocean as a result of
104
Eltanin cruises; Goodell et al. (1971) utilised restricted to the east of 1300W. Long., while . nodules collected between 1962 and 1968. 8-Mn02 and 7A manganite, which they Regional variations in nodule chemistry for lumped together as birnessite, were most the Scotia Sea and Pacific Ocean sectors have abundant in the western Pacific. Glasby been presented in the form of trend surfaces (1972) believed that a-Mn02 was found by Goodell et al., who noted three typically in nodules from open ocean geochemical provinces. A lithophile (Fe) environments at all depths, whereas loA and suite of elements was enriched in concretions 7' A manganite were characteristic of nodules from the Pacific-Antarctic Ridge (600S., from continental borderlands. Meylan (1968) 1600W), a chalcophile (Mn) suite in the observed that todorokite (or lOA manganite) Southwestern Pacific Basin, and a mixed Fe- was the dominant Mn oxide mineral at all Mn suite was found on the Chile Rise-East depths throughout the Pacific-Antarctic Pacific Rise (400-500S.). region and that the content of todorokite The regional compositional trends relative to incorporated detrital minerals observed in this study support and extend generally increased away from the Antarctic those detected by previous workers. By continent. arbitrarily defining geographic regions of Evidence presented in this study and the interest, differences in composition (and accompanying one (Meylan, this publication) mineralogy) have been noted. The principal indicates that Mn/Fe and todorokite/a-Mn02 differences are concurrent southeastward ratios are directly related. A depth decreases in the Mn/Fe and todorokitela- dependency to this relationship is only Mn02 ratios. This may be attributable coincidental, Le., higher iron (and hence 8- primarily to two factors, one being the Mn02) contents would be expected on distance of the sampled portion of the volcanic seamounts. The factors that control Southwestern Pacific Basin from terrigenous iron and manganese precipitation on the sediment sources. Skornyakova et al. (1962) deep-sea floor appear to operate instead on a found that maximum Mn (and minor regional scale, and the relative availability of element) contents were to be expected where Fe and Mn to a depositional surface will in detrital accumulation rates were minimal. turn indicate the proportions of todorokite, li- The other probable factor is the activity of Mn02 and amorphous iron that form bottom currents. Goodell et al. (1971) used ferromanganese iron concretions. bottom photographs to delineate a belt of The mineral phases responsible for strong bottom currents extending incorporating each of the various trace metals approximately from the junction of the have not been definitely identified, although Pacific-Antarctic Ridge and Eltanin Fracture some understanding of cation locations now Zone, along the fracture zone, through the exists. Burns and Fuerstenau (1966), using Drake Passage and into the Scotia Sea. electron microprobe techniques, discovered Possibiy the bottom currents could serve to that Ni, Cu, Zn, K and Mg were enriched disperse iron from local volcanic sources, with manganese in loA manganite layers, preferentially increasing the rate of Fe and Co, Ca and Ti were enriched in layers deposition relative to Mn. with high Fe concentrations, the latter Qualitative differences in the manganese usually containing 8-Mn02. Barnes (1967) oxide mineralogy of Pacific nodules found at concluded that Co and Pb were enriched in 8- various depths and geographic locations have Mn02 relative to 10A-manganite, whereas also been noted. Barnes (1967) determined the latter mineral appeared to concentrate Cu that Mn-oxides are basically depth and NL Cronan (1967) observed similar Co, dependent, II-Mn02 tending to occur at Pb, Cu and Ni associations, and noted in shallower depths than either 7 A or lOA addition that Ti was enriched in the II-Mn02 manganite. Cronan and Tooms (1969) noted phase, whereas Mo, V, Ba and Cr did not a similar depth dependency, and also seem to show a preference for either phase. observed that todorokite-rich nodules were Southern Ocean nodules also display Cu and 105
Ni enrichment when high in todorokite, and Pacific-Antarctic Ridge, Chile Rise ' at the same time show Zn and Sn and Bellingshausen Basin samples. enrichment somewhat less clearly; nodules (2) Regional differences in trace element with relatively abundant ~-MnO. were seen composition also exist. Ni content is to be most likely to have greater proportions directly related to Mn, as is Cu (except of Co and Ti, and possibly V and Mo. in the Drake Passage-Scotia Sea, where Although beyond the scope of this study, it is relatively enriched despite low knowledge of accretion rates could provide a manganese contents). The greatest rationale for the chemical and mineralogical enrichments of Co, Ti, V, Zn, Mo and composition of Southern Ocean nodules. It Sn . vs. Mn are found in the Drake appears likely that accretion rates may reflect Passage-Scotia Sea. Nodules of the both proximity to element sources and the Southwestern' Pacific Basin show the influence of bottom currents. Thus the greatest depletions of V, Zn, Mo and presence of a nearby volcanic source of Sn relative to Mn. elements or enhanced bottom current activity (3) When classified according to type of could lead to increased rates of iron and/or nucleus, only two classes of manganese oxide precipitation. Scott et al. concretionary material show signifi- (1974) found that a manganese oxide crust cant differences from the from the median valley of the Mid-Atlantic compositional means of the regions in Ridge had an extremely high MnlFe ratio which they were found. Nodules with and a low trace metal content. The crust, of indistinct zeolitite/vesiculite nuclei probable hydrothermal origin, accumulated have significantly higher todorokitel _ two orders of magnitude faster than the MnO., MnlFe, Mn/Co, MnlTi, accepted normal rate for deep-sea ferro- Mn/V, and Mn/Mo ratios, and lower manganese deposits. Typical hydrothermal Mn/Ni and Mn/Cu ratios than Fe-Mn deposits have much more iron than regional averages. Concretionary manganese and are also low in trace metals material broken from submarine rock (Bonatti et al., 1972). The concretionary surfaces has todorokitela - MnO., material termed "broken from nucleation Mn/Co, MnlTi, MnN and MnlMo site" in the present study may well be of a ratios lower than regional averages, hydrothermal nature. Contrasted to this type and higher Mn/Ni and Mn/Cu ratios. of Fe-Mn deposit would be the nodules (4) In general, increased proportions of having indistinct zeolitite/vesiculite nuclei. todorokite relative to a-MnO. are Falling into the hydrogenous category of accompanied by higher contents of Ni, Bonatti et al. (1972), they have probably Cu, Zn and Sn relative to Mn, and accreted metals slowly from sea water, lower contents of Co, Ti, V and Mo. resulting in high Mn/Fe ratios, high (5) Fe-Mn oxide mineralogy is directly todorokite 'contents, and enhanced related to the availability of iron and accumulation of Ni and Cu in particular. manganese. The proportion of todoro- kite increases as the Mn/Fe ratio increases, whereas a -MnO. is CONCLUSIONS associated with iron-rich phases. (1) Southern Ocean nodules display (6) Trace metal contents are probably regional differences in Fe-Mn controlled by rates of accumulation as chemistry and mineralogical well as by the nature of the Fe-Mn composition. The Southwestern Pacific oxide minerals formed. Basin samples have the highest, (7) Southern Ocean nodule distribution Mn/Fe and todorokitela-MnO. ratios, and composition appear to be related while values for these ratios are lowest to the combined influence of in the Drake Passage-Scotia Sea. volcanism and the Antarctic Intermediate values characterise Circumpolar Current.
106
PRELIMINARY RESULTS OF RENEWED INVESTIGATIONS ON MANGANESE NODULES AND ENCRUSTATIONS IN THE INDIAN OCEAN
D. S. CRONAN AND S. A. MOORBY
Abstract Chemical analyses of previously unanalysed nodules and encrustations from more than 60 sites in the Indian Ocean have shown the maximum concentrations of the potentially economically valuable elements Mn, Ni, and Cu to be found in the Central Indian Basin, where nodules are usually more abundant than elsewhere. High concentrations of these elements also occur in some of the smaller basins of the western Indian Ocean. By contrast, Fe and Co are highest in the westem Indian Ocean, Fe in the vicinity of the Mid-Indian Ocean Ridge, and Co in some other elevated areas.
INTRODUCTION AND PREVIOUS occurrence. In this presentation we give some WORK preliminary results of our studies. Manganese nodules were first discovered in the Indian Ocean during the voyage of DISTRIBUTION H.M.S. Challenger (1872-76). Further The distribution of nodules in the Indian samples were collected sporadically by Ocean is known in much less detail than in various research vessels until about 1950, the Pacific. This is not only because there has after which more detailed investigations by been much more sampling in the former the U.K., U.S.A. and U.S.S.R. resulted in the than in the latter, but also because there have collection of the samples which form the been few attempts at bottom photography basis of much of our present knowledge of over wide areas in the Indian Ocean. Detailed Indian Ocean nodules and encrustations. The studies by Laughton (1963), Glasby et at. geochemical results of the Russian in- (1974) and Vincent (1972), for example, vestigations have been summarised by have shown variable and occasionally dense Bezrukov and Andrushchenko (1974), who concentrations of nodules in parts of the extended the conclusions reached by Cronan western Indian Ocean, usually in areas of and Tooms (1969) on the basis of material dissected topography, but nodule distribution largely collected by U.K. and U.S. ships over most of the sea floor is imperfectly during the fifties and early sixties. known. Since these initial studies many more On an ocean wide basis, the limited data samples have become available which enable available indicate that the distribution of a more detailed assessment of the economic nodules in the Indian Ocean appears to be potential of Indian Ocean nodules to be related to gross sedimentation rates and made. These samples, together with data bottom topography. Bezrukov and previously collected, are being used at the Andrushchenko (1972) collected together AGRG to construct a detailed regional most pre-existing data and found nodules to geochemistry of Indian Ocean nodules. This be most abundant in the basins far from land, research aims in part to delineate areas and particularly on either side of the 900E containing nodules of possible ore grade in Ridge. the Indian Ocean, and to explain their This is likely to be due to the low
118
sedimentation rates in such areas as a result themselves may promote manganese of their remoteness from land, leading to a oxidation and precipitation. Encrustations lack of burial of nodule nuclei and from the Carlsberg Ridge have been protonodules at the sediment surface. described by Cronan and Tooms (1967), and Concentrations of nodules also occur locally Glasby et al. (1974), and were found to be in other basin areas where sedimentation similar in composition to associated nodules. rates are low such as in parts of the southern More recently R.R.S. Shackleton dredged a Indian Ocean, and the basins of the western range of seamounts on the equator north of Indian Ocean. However, the basins in the the Seychelles, and obtained ferromanganese northern Indian Ocean which receive detritus oxide encrustations in almost every haul. from Asia, contain very few nodules. Both Leaving compositional considerations the Arabian Sea and the Bay of Bengal are aside, in order to be mineable a nodule turbidite floored, and nodules are very rare deposit must ideally be both extensive and under such conditions of rapid sedi- dense, and occur in an area of limited mentation. Furthermore, recent results from topographic contrasts. This rules out the a cruise of R.R.S. Shackleton in the north- elevated areas such as the mid-ocean ridges western Indian Ocean have shown that and seamounts, and probably also precludes nodules are very sparse in the equatorial those nodule covered areas with rock outcrop zone. This is probably a result of the as a result of strong bottom current scour. Of widespread carbonate sedimentation in this the remaining nodule bearing areas in the area. Indian Ocean, the sediment floored basins of Nodules Occur in abundance in some areas the central Indian Ocean would appear to be of the Indian Ocean where sedimentation is the most suitable for nodule accumulation in inhibited by strong current action. There are sufficient density to warrant exploitation. many nodules under parts of the Antarctic Unfortunately, the remoteness of these areas Circumpolar Current for this reason, and also has resulted in them being amongst the least in the vicinity of Madagascar. explored of the whole ocean. Nodules have In addition to sedimentation rates, bottom been found at many stations in the central topography affects nodule distribution in the Indian Ocean, but the sampling methods that Indian Ocean. Bezrukov and Andrushchenko were used have, for the most part, not (1972) have reported high nodule enabled an assessment of nodule abundance concentrations at the foot of some submarine to be made. This needs to be rectified as from ridges, sometimes in the vicinity of major the compositional point of view, these fault zones. This may be a result of an nodules are amongst the most economically abundance of volcanic nuclei, in these attractive in the whole ocean. probably tectonically active areas, suitable for encrustation by manganese and iron oxides. GEOCHEMISTRY Some more elevated areas such as the Chemical analyses for Mn, Fe, Ni, Co, Cu, Carlsberg Ridge also contain limited fields of Zn, Ca and Al have been conducted on densely concentrated nodules (Laughton, samples comprising material from both 1963), although the rapid carbonate closely spaced sites in some small areas, and sedimentation typical of such elevated areas from widely spaced sites throughout the most probably inhibits widespread nodule whole ocean. They therefore facilitate the formation in them. interpretation of both local and regional In most elevated areas, manganese variations in the composition of Indian encrustations rather than nodules are the Ocean nodules (Fig. 1). In this preliminary principal form of ferromanganese oxide presentation the new data are discussed on deposit. This is probably because the rock their own. In our final synthesis of the outcrops in such areas form a suitable surface regional geochemistry of Indian Ocean on which the ferromanganese oxides can nodules they will be considered together with accrete. Indeed, the alteration of the rocks all previously published data on Indian Ocean
120
nodules in order to arrive at a more detailed area. It is somewhat lower to the east of the analysis. 900E Ridge, averaging about 0.5 to 1 %. In order to examine regional variations in Similar values occur sporadically in the the concentration of Mn, Fe, Ni, Cu, Co and western Indian Ocean, but values are Zn throughout the Indian Ocean, the values generally lowest overall in this area of these elements in the samples analysed particularly in the continental margin and have been charted individually. also in the vicinity of the Mid-Indian Ocean Manganese reaches its highest Ridge. concentrations, about 35%, in the continental Copper tends to follow Ni fairly closely, margin area of the south-western Indian and is also highest in the central Indian Ocean, off the coast of South Africa. In this Ocean where, as is the case with Ni, values in regard, Indian Ocean nodules are similar to excess of 1 % are common. Like Ni it is lower those from the Pacific, where highest Mn to the east of the 900E Ridge, and lowest concentrations also occur in continental overall in the western Indian Ocean where margin environments. High Mn values of less than 0.1 % are frequently concentrations, often in excess of 25%, occur encountered on parts of the Mid- Indian in the central Indian Ocean between the Mid- Ocean Ridge and in the continental margin. Indian Ocean Ridge and the 900E Ridge. Zinc tends to follow both Ni and Cu, being Somewhat lower concentrations, 15-25%, highest in the central Indian Ocean where occur to the east of the 900E Ridge. Other values over 0.15% are common. It is lower than near the continental margin, manganese and somewhat more variable in abundance to concentrations are relatively low in nodules the east of the 900E Ridge and over much of over much of the western Indian Ocean, the western Indian Ocean. However, some often less than 15%, and are lowest overall in unusually high values of Zn also occur in the the vicinity of the Mid-Indian Ocean Ridge. Mozambique Channel off S.W. Madagascar The distribution of iron in the samples which are unrelated to enrichments of any of analysed is more or less the opposite of that the other elements normally concentrated in of manganese, although there are notable zinc rich nodules. exceptions. It is highest in the western Indian Cobalt varies in a different manner to the Ocean, both in an isolated sample from the other elements of potential economic value. continental margin in which it reaches over It is very low in the central Indian Ocean, 30%, and overall in the region of the Mid- often less than 0.1 %, and only slightly higher Indian Ocean Ridge where it is usually over (0.1-0.3%) to the east of the 900E Ridge. 15% and often over 20%. With regard to the Cobalt is variable in the western Indian latter, Fe is behaving similarly to its Ocean with low values in the basins and on behaviour on the Mid-Atlantic Ridge parts of the Mid-Indian Ocean Ridge, but is (Cronan 1975). The high Fe value of the higher overall than in the east, especially on continental margin sample coupled with high some isolated topographic elevations. In this values of Mnin the same general area further regard, Co is acting in a manner similar to its illustrates just how compositionally variable behaviour in the Pacific and Atlantic Oceans nodules can be in continental margin (Cronan & Tooms,1969, Cronan 1975), environments. Iron values are very low, where it tends to be negatively correlated often in the region of 5-10%, in the central with increasing depth. There are also some Indian Ocean, and only a little higher to the quite high Co values (up to 0.67%) off South east of the 900E Ridge, where they range Africa, but these are associated with low between 5 and 15%. values (down to 0.04%) in the same general Nickel follows manganese, other than in area, illustrating the variability of Co in the the continental margins, and is highest in the Western Indian Ocean. central Indian Ocean between the Mid-Indian The regional variations of Mn, Fe, Ni, Co, Ocean Ridge and the 900E Ridge. Nickel Cu and Zn in Indian Ocean nodules shown concentrations commonly exceed 1 % in this by the present data are similar to, but extend
12l
and amplify, the conclusions on the regional was quite fortuitous, more detailed surveying geochemistry of Indian Ocean nodules would probably result in the discovery of reached by Cronan and Tooms (1969) and even higher grade nodules. Bezrukov and Andrushchenko (1974). They are also similar in many respects to data on SUMMARY the regional variability of Pacific nodules, and Manganese nodules in the Indian Ocean to a lesser extent, those from the Atlantic. are most abundant in the basins of minimal Nodules from all three oceans show the sedimentation far removed from land, and highest values of the economically important their maximum concentration of the elements Ni and Cu in the deep basins of economically attractive elements Mn, Ni and minimal sedimentation. The distribution of Cu, are similar to the maximum iron seems to be influenced by volcanic concentrations found in the Pacific. The activity, and that of Co by depth. The regional variations of Mn, Fe, Ni, Co, Cu and continental margins of the ocean appear to Zn have been charted. Manganese, nickel, contain deposits generally low in minor copper and zinc tend to be correlated with elements, but of variable Mn and Fe each other and are enriched in the central contents. The causes of such variations have Indian Ocean, with somewhat lower values been discussed by Mero (1965) Cronan and of these elements to the east of the 900E Tooms (1969), Price and Calvert (1970), Ridge. Values of these elements are much Glasby (1970) Horn et al. (1972) and lower overall in the western Indian Ocean Cronan (1974, 1975, in press), and further than in the east, and apart from some high discussion on this topic is beyond the scope Mn and Zn values, are also low in the of this preliminary presentation. continental margin off Africa. Iron and cobalt From the economic point of view, the behave in a different manner. Iron is highest most promising area to undertake further overall in the western Indian Ocean with exploration for manganese nodules on the sporadic high values in the continental basis of their composition in the Indian margin area, and lower in the central and Ocean, is the central Indian Basin, although eastern Indian Ocean. Cobalt is also higher in some of the smaller basins in the the western Indian Ocean than in the east, northwestern Indian Ocean might also be but its distribution is influenced by promising in this regard. These areas have topographic variations in a way that the other received only a fraction of the attention of elements are not. The results suggest that the areas of potential economic value in the central Indian Basin is the most likely area Pacific, yet the very limited sampling that has for the location of nodules of economic grade, been conducted has resulted in the discovery with the basins of the north-western Indian of nodules of potential ore grade. Unless this Ocean being of lesser potential.
REFERENCES
BEZRUKOV, P. L. ANDRUSHCHENKO, P. F. 1972: ferromanganese oxide deposits from the Atlantic Iron-Manganese nodules from the Indian Ocean. Izv. Ocean. J. geophyy. Res. 80: 3831-37. Akad. Nauk. SSSR, Ser Geol .• 7: 3-20. CRONAN, D. S. In press: Manganese nodules: BEZRUKOV, P. L., ANDRUSHCHENKO, P. F. 1974: distribution and geochemistry. In Glasby, G. P. (Ed.), Geochemistry of Iron-manganese nodules from the Marine Manganese Deposits, Elsevier, Amsterdam. Indian Ocean. Int. Geol. Rev. 16: 1044-61. CRONAN, D. S., TOOMS, 1. S. 1967: Geochemistry of CRONAN, D. S. (1974): Authigenic minerals in deep- manganese nodules from the N.W. Indian ocean. sea sediments. pp. 491-525 in Goldberg, E. D. (Ed.) Deep-Sea Res. 14: 239-49. The Sea, vol. 5: Wiley Interscience, New York. CRONAN, D. S., TOOMS, 1. S. 1969: The geochemistry CRONAN, D. S. (1975): Manganese nodules and other of manganese nodules and associated pelagic deposits
122
THE RELATION BETWEEN BASIN EVOLUTION AND MARGINAL PLATEAU SUBSIDENCE IN THE CORAL SEA
JOHN C. MUTTER
Marginal plateaux such as the Queensland with the acoustic zonation found on seismic Plateau may be thought of as structures reflection profiles. The Eocene/Oligocene caught within the transition zone between hiatus in DSDP 209 corresponds well with a continental and oceanic crust (Mutter, 1975). marked unconformity on the reflection They are continental structures which have profiles. The unconformity is often angular been modified by the effects of tectonism and is widespread on the Plateau. The lowest associated with the formation of an ocean acoustic unit is a well stratified, structurally basin. The structure of such a pleateau owes disturbed sequence which pinches out against as much to a parental relationship with the basement highs, the tops of which have been continent adjoining it on one side, as it does levelled by erosion. It represents the products to a tectonic relationship with the ocean basin of erosion of the basement highs and adjoining it on the opposite side. The history correlates with the basal rock unit in DSDP of development of a marginal plateau is thus 209. the manifestation of continent/ocean The geological history of the Queensland interaction at a newly developing continental Plateau can be deduced from the distribution margin. through time of the different litho-acoustic The Queensland Plateau (Fig. 1) lies units. The history consists of three major between northern Queensland on its western episodes: side, and the Coral Sea Basin on its eastern (1) Uplift of the basement, erosion of side. It was covered by an extensive and basement highs, and deposition of the systematic multisensor geophysical survey in erosion products in shallow water. 1971 (Mutter, 1974) and was tested by Uplift probably occurred during the DSDP drilling (hole 209) in the same year Upper Cretaceous and erosion (Burns, Andrews et al., 1973). Drilling continued through to and including revealed three basic lithologic units: Middle Eocene. (1) A basal unit consisting of shallow- (2) Differential subsidence accompanied water, largely terrigenous sediments by faulting, in which the outer edge of deposited during Middle Eocene and the Plateau subsided with respect to an unknown time interval before this the inner. The shallow-water sequence (drilling did not penetrate to was deformed, and, as terrigenous basement). source areas diminished, a hemipelagic (2) A hemipelagic rock unit deposited sequence developed. This occurred during Upper Eocene. This is during the Upper Eocene to Oligocene. separated by a depositional hiatus (3) Uniform subsidence; source areas from were lost and wholly pelagic sediments (3) Pelagic sediment deposited from blanketed the Plateau. This occurred Upper Oligocene onward. from Upper Oligocene time onwards. This lithologic zonation has been matched The graben-troughs which form the
129
explained by invoking rapid basin evolution: Lithospheric cooling and contraction in the so rapid that the characteristic rift valley basin after spreading ceased would heighten structures and sediments did not have time this effect. to develop. Beforehand, a protracted uplift Thus the continental margin of northern stage occurred. To explain this it is necessary Queensland appears to have formed by a to postulate a warm thermal anomaly, or modified type of Atlantic margin more likely a hotter anomaly located some development, the key to which is rapid distance from the Coral Sea Basin. This evolution following some time after the thermal anomaly must have been capable of action of a remote thermal anomaly located causing uplift of the crust without inducing outside the Coral Sea Basin. metamorphism at its base. If the influence of this anomaly was later replaced by an intense but short-lived anomaly located in the Coral ACKNOWLEDGEMENTS Sea Basin and associated with its opening, This paper represents a condensation of then the observed evolutionary scheme some parts of a larger work by the author presents no difficulties. Subsidence of the currently in press as a BMR Bulletin (Mutter, Plateau will follow basin evolution as the 1976). It is published with permission of the intense anomaly would promote thermal Director of the Bureau of Mineral Resources, metamorphism and crustal subsidence. Canberra, A.C.T., Australia.
REFERENCES
BURNS, R. E., ANDREWS, 1. E. et al. 1973: Initial MUTTER, J. C. 1974: Geophysical results from the Reports of the Deep Sea Drilling Project, Vol. 21. U.S. Coral Sea: Continental margins survey report. Bur. Government Printing Office, Washington, D.C. 931 Miner. Resour. Aust. Rec. 1974/116. pp. MUTTER, J. C. 1975: Basin evolution and marginal F ALVEY, D. A. 1974: The development of continental plateau subsidence in the Coral Sea. Bull. Aust. Soc. margins in plate tectonic theory. APEA J. 14(1): 95· Explor. Geophys. 6(2/3). 106. MUTTER, J. C. 1976: The Queensland Plateau. Bur. KARIG, D. E. 1971: Origin and development of Miner. Resour. Aust. Bulletin 179 (in press). marginal basins in the western Pacific. J. geophys. Res. 76: 2542·61.
132
A RE-EVALUATION OF THE SOLOMON ISLANDS AS AN ARC SYSTEM
PATRICK J. COLEMAN
Abstract
The Solomon chain is made up of a part of the Ontong Java Plateau on the north-east flank, a central Early Tertiary tholeiitic province, and a Plio-Pleistocene volcanic pile on the west-southwest flank. Evidence that the Early Tertiary province arose as an arc with NE-facing polarity is ambiguous and limited especially by lack of refined geochemical data If low-angle obduction of the Plateau over the central province could be proven, then this would be strong indirect evidence of an old subduction zone, but the relation between these two can just as well be interpreted as one of high angle strike-slip. The young volcanics are geochemically peculiar and not a typical young arc assemblage. The seismicity is puzzling and does not define a subduction zone. In terms of juvenile arc development the absence of volcanics east of western Guadalcanal is altogether anomalous. Given the fact of the Line Islands and notion of volcanic accumulations along fracture zones and leaky transforms, then the Solomons can be speculatively interpreted as an old oceanic fracture system that has been incorporated into an essentially transform boundary between India and Pacific plates.
The Solomon Islands are a composite angles to the axis of the Solomon chain. This geological entity made up of three distinct set defines the gap between the Solomon or areas and sequences (Coleman, 1970, Bougainville Trench to the north-west and Hackman, 1973) (Fig. 1). The Pacific the San Cristobal Trench to the south-east Province occupies the northeastern flank and and includes the W oodlark Basin supposed is typified by the island of Malaita; it is an spreading ridge (Luyendyk et al., 1973). uplifted, overthrust part of the southern edge These young volcanics are geochemically a of the Ontong Java Plateau to the north difficult group. On Bougainville, high- (Kroenke, 1972). Malaita consists of ocean- alumina, low-titania andesites are dominant floor lava overlain by deep-water pelagite (Blake and Miezitis, 1967), whereas in the oozes of Albian/Cenomanian to Pliocene age New Georgia group (Stanton and Bell, 1969) (Deventer and Postuma, 1973) which there are olivine basalts, feldsparphyric become increasingly contaminated with basalts and hornblende basaltic andesites terrigenous detritus higher in the sequence. (roughly in that order of abundance) and This admixture is considered to mark the equivalent pyroclastics in which the collision between the proto-Solomons and the dominant clasts are olivine basalts; there are Plateau in the Late Miocene. On the other dykes of picrite basalt and a stock of flank (Volcanic Province), there is a pile of pyroxene andesite. Geochemically, this young lavas, intrusives and volcaniclastics, unusual but essentially tholeiitic assemblage, probably less than 4 m.y. old, with well- although comparatively low in titania, has preserved cones and several active volcanoes. affinities as much with mid-oceanic igneous The Volcanic Province is boomerang-shaped rocks as with those of arcs. A suggestion of a and stretches westwards from western genetic link between the young crust and Guadalcanal to Bougainville. The central spreading ridge of the Woodlark Basin and part, the new Georgia group, lies adjacent to a these New Georgia volcanics is in order, southern set of major horsts and graben, with granted their youth and the fact that there is, trend SW-NE, more or less at right and has been, a strong westerly component
134
in the India-Pacific and Solomon-Pacific plate is taphrogenic,with deep wacke-filled graben motions. The petrologically variable diorite- resulting in steep local gravity gradients tonalite stocks of Bougainville and western (Coleman and Day, 1965); the Aola Trough Guadalcanal are significant features of the in eastern Guadalcanal is estimated to Volcanic Province; the now-famous Panguna contain over 5,000 m. of Mio-Pliocene porphyry-copper mine in central Bougainville sediment. This taphrogenesis was a is located in one of these. They range in age continuous if episodic process throughout the from about 3.5 m.y. to 1.5 m.y. BP, time of accumulation of the sediment pile; The islands of the Central Province horst blocks were torn down to fill include Bougainville, Choiseul, southwestern intermediate graben and these in turn were Santa Isabel, Nggela, Guadalcanal and, uplifted and their sediment caps recycled to possibly, San Cristobal. These were once fill new graben. The sequences are marked by linear; their present roughly en echelon intense lateral and vertical changes in arrangement is most probably a fairly recent lithofacies. dislocation and an expression of an increasing The boundary between the Central and vector of convergence between India and Pacific Provinces is marked by high-angle Pacific plates. The basement rocks are fractures with consequent ultramafic pods dominantly tholeiitic with minor calcalkaline such as those along the Korigole Thrust content (Hackman, 1971). In southeastern on Santa Isabel (Stanton, 1961). High- Guadalcanal and eastern San Cristobal, the temperature albite-epidote-amphibolite younger part includes deep-water, schists, probably metamorphosed in the mid- manganese-bearing pelagites, caught up with Eocene, are spatially and possibly genetically the lavas and lightly metamorphosed, which related to this boundary. There are also large contain keeled Globorotalia (akin to areas of serpentinised harzburgites on Truncorotaloides) and Chiloguembelina Guadalcanal, related to large fracture systems suggestive of an Eocene-Early Oligocene age. and probably emplaced in the Late Oligocene; The bulk of the igneous rocks is smaller bodies of ultramafics are found on metamorphosed to zeolite and greenschist San Cristobal. Kroenke (1972) suggests that facies and this makes difficult the task of the boundary ultramafics, together with the assessing its geochemistry. On the basis of higher crustal lavas and pelagites, such as less than a dozen analyses of the Guadalcanal those on Malaita, may constitute a tectonic Mbirao Group, Hackman tentatively suggests flake, the result of low-angle obduction over that the basaltic rocks of this group have the central province of the southern edge of features consistent with both oceanic and the Ontong Java Plateau. A high-angle circumoceanic derivation. A full-scale project collision is a possibility, however, following of analysis of the basement rocks of the convergent transform motion of the Plateau, central islands is urgently needed. with consequent emplacement of ultramafics. Shallow-water, late Eocene biogenic foram- Geophysical results complement geological algal limestones are found in the studies. Bouguer anomaly patterns confirm island chains to the east and west, but so far the horst and graben structure, the presence have not been found in the Solomons. This is of ultramafic bodies, and the presence of a puzzling gap. The familiarly ubiquitous highs of older basement rocks adjacent to the Melanesian cover of volcanic calcwackes and severely downfaulted" southern margins of biogenic limestones of Late Oligocene each of the central islands (Grover, 1968; through mid-Miocene is, however, wide- Laudon, 1968; Rose et al., 1968). Most of spread. Volcaniclastics dominate the rest of the islands appear to be isostatically the succession, although foram-algal compensated. Local magnetic anomaly limestones are fairly common, especially patterns reflect the presence of ultramafics towards the top. The coral content in these and younger intrusives. The regional limestones becomes marked only after the magnetics (Woollard et al., 1967) suggest a Miocene. Central province regional structure strong east-west grain, subparallel to and
135
partly delineating such features as the that supposedly marks an old trench, a Woodlark Rise and Pocklington Trough to postulated belt of high heatflow south of and the south. Marine seismic refraction studies parallel to the chain and said to be a lingering (Furumoto et aI., 1970) suggest a linear after-effect of subduction (Halunen and Von block-like character for the Solomons, with Herzen, 1973); concomitant with this, the crustal thickness varying from 15 to 20 km. suggestion of hanging remnant slabs of The crust beneath the Ontong Java Plateau to lithosphere beneath the chain (Denham, the north is thicker, up to 35 km., to the 1975); and the supposed obduction of the south beneath the Solomon Sea it is thinner, Ontong Java Plateau over the Solomon block. about 10-12 km. High heatflow values have Cumulatively, this evidence is indirect been recorded to the south-west, especially in and ambiguous. The supposed old trench sits the Woodlark Basin and Solomon Sea, by squarely on the thick Plateau; it might just as Halunen and Von Herzen (1973) and by well be the trace of a fracture system. There MacDonald et al. (1973). Seismicity is are not enough heatflow measurements to highest at each end of the Solomon chain, justify the notion of a belt and, in any case, although the New Georgia group has its fair the higher western ones can be attributed to share of earthquakes with hypocentres down spreading in the Woodlark Basin. The to 40 km. In profile, the hypocentres lie question of hanging slabs remains open and within a diffuse, near-vertical wedge, apex is part of the general problem of explaining down, on the southwestern flank; Santa the pattern of Solomon seismicity. The Isabel and Malaita are nearly a seismic. There consumption of crust down a south-dipping is a suggestion of a gap between 200 and 400 subduction zone with consequent obductive km., and no record of a hypocentre deeper collision of the Plateau and the Central than 550 km. (Curtis, 1973, Ripper, 1975). Province is a notion which begs the question The foregoing preamble summarises the as to whether or not there was indeed a essential geological data (necessarily, it does subduction zone: as said before, the collision not include fine but pertinent points of characteristics can also be explained by a detail). With this as a base, we can evaluate prior change in India-Pacific relative plate the Solomons as an arc system. It has been motions so that what was an active described as such by Coleman (1970, with transform boundary segment becomes refs), as an arc which has undergone polarity converted to a zone with a convergence reversal by Karig and Mammerickx (1972) vector; the old transform could be and Halunen and Von Herzen (1973), and as represented now by the large fractures which part of a large arc complex (Solomons - separate Central and Pacific provinces. The New Hebrides - Fiji, Tonga) in models presence of high-temperature amphibolite proposed by Packham (1973) and Falvey schists on the northeastern side, but of lower (1975). Considered as an arc system, its grade metamorphics on the other side, and Pacific Province has to be excised: this is a the comparative dearth of calcalkaline rocks, Late Tertiary tectonic addition. The Volcanic do not support a north-east-facing polarity. Province is also latter-day, but may be Although sparse, the geochemical evidence genetically a part of the system. The Central on the basement volcanics does not support, Province is the essential part. Granted this, unequivocally, their arc derivation. and operating within today's model of arc It seems premature to speak of the formation (Karig, 1974; Coleman, 1975), it is Solomons as an arc which has undergone difficult to nominate the position of a polarity reversal over the last 10 m.y. The subduction zone whether on northeastern or most that can be said is that the Solomons southwestern flank, which would have led to are now behaving, and have been over that its development. The features which favour a period, as a newly initiated arc system. The northeastern, south-dipping subduction zone present-day tectonic features, however, are include: a set of poorly defined troughs, up to not straightforward with respect even to 4,000 m. deep, on the northeastern flank, juvenile arc development. The Bougainville
137
and San Cristobal trenches are not reconstruction of this part of the Pacific necessarily subduction signatures; their would be useful, but the attempt soon shows relation to the near-vertical, diffuse pattern of that whether or not the Solomons arose as an hypocentres, extending to depth greater than arc system is itself a crucial element in the 500 km., is quite' uncertain. To speak of a reconstruction. A full-scale kinematic "Benioff zone" would be a misnomer and approach is not possible just now for lack of this applies even to the Bougainville area data (Coleman and Packham, in press). where at first there appears to be a relation If the Solomon chain did not develop as an between shallow, intermediate and deep arc system, in what way did it develop? In earthquakes: there appear to be seismic simplistic model terms, there are two discontinuities in this area (Denham, 1975). speculative answers. The first is that the The pattern of these earthquakes is not central province arose from the build-up of challenged, but whether they can be said to volcanic piles along a great fracture system define a steeply north-dipping slab of within the Pacific plate. Analogues are lithosphere. The high positive gravity presented by the Line Islands (Jackson, anomalies on the southwestern flank of the Schlanger et al., 1974), the Louisville Ridge Solomons block are also anomalous. If the and even the Ninety East Ridge (Hekinian, young volcanics are the by-product of north- 1974). The notion that deep fractures dipping subduction, then there are problems (including 'leaky' transforms) may produce here, too. The lavas from the Bougainville great volumes of extrusives and intrusives is volcanoes are dominantly andesitic (Blake an old one and is being presently further and Miezitis, 1967), as also are those of investigated as part of the study of oceanic western Guadalcanal, and so are distinct fracture systems (e.g. Van Andel, 1971; from those of the New Georgia group (and, Thompson and Melson, 1972; Cochran, incidentally, from the exotic lavas of the 1973; Harrison and Ball, 1973; Hekinian and Lihir chain (Johnson et al., in press) which Aumento, 1973; Udintsev et al., 1974; Vogt extends north-west, in line with and Johnson, 1975). This proto-Solomon Bougainville). The most active volcanoes in system then experienced tectonic uplift and the New Georgia group are submarine and was incorporated as part of an essentially located further south than the older, transform boundary between India and presumed extinct ones; this implies a Pacific plates. This change possibly came narrowing, not a widening of the arc trench about as a consequence of an abrupt change gap (see Dickinson, 1973). The possibly in direction of motion of the Pacific plate at extinct but well-preserved volcanoes on about 40 m.y. BP (Clague et al., 1975, with Choiseul are difficult to relate to either a refs) together with accelerated creation of northern or southern subduction zone. new sea floor between Australia and Finally, if the San Cristobal Trench marks a Antarctica. The older islands reached sea subduction zone, then the depths of level in the Late Oligocene. The subsequent hypocentres provoke the question as to why history, including further uplift, collision with there are no young volcanic by-products east the Ontong Java Plateau, extrusion of the of western Guadalcanal; the youngest young volcanics and dislocation of each volcanic terrigenous sediments on San island block is part of the continuing Cristobal are mid-Miocene. interaction, essentially sinistral strike-slip, A partial explanation for these anomalies between India and Pacific plates. and ambiguities may lie in the fact that India- The second possibility, a rather more Pacific interplate motion has had a larger remote one, is that the central province horizontal than vertical component, roughly represents the edge of the Ontong Java parallel to the axis of the Solomons for some Plateau, which has a history extending back millions of years past, but the problem they beyond the Aptian, Early Cretaceous occasion is thereby highlighted rather than (Kroenke, 1972; Packham and Andrews, solved. To solve it, an historical 1975). This would involve movement from
138 the east in the transform direction between untestable at the present time, but India and Pacific plates with convergence and nevertheless worth considering. Whether it shearing collision by the Pliocene. A difficulty arose as an arc, or along a fracture system, it here is that there is a real stratigraphical, is now disrupted, locked into the Ontong lithological and structural break between Java Plateau and so part of the Pacific plate. Pacific and Central Provinces. As well, the The Solomons may be the site of juvenile arc Solomons would have to represent the young, development but if so they remain thin edge of the Plateau; assuming that such anomalous. The present"day seismicity is plateaux develop about spreading centres peculiar and difficult to relate to a subduction then one is faced with the problem of locating model, although an alternative explanation is its "southern" half, and of explaining why it not part of any of today's earth models. This is that the plateau remains submarine. is a major problem. Another centres on the The first speculation is preferred. It is lack of refined geochemical data. Overall, the difficult to test against actual examples Solomon group illustrates the dangers of because oceanic fracture systems have been stretching simple models to cover complex relatively little studied. If one checks through situations. the literature referenced above (Van Andel, 1971 etc.) one can find gross similarities to ACKNOWLEDGEMENTS Solomons features, from morphology to the I thank Loren Kroenke, Hawaii Institute of petrography of dredged samples, but a Geophysics, for helpful discussion and meaningful comparison is not possible. A argument, and especially for his organisation better target would be the Line Islands, the of the CCOP/SOPAC-IOC Workshop at age of which ranges back into the Cretaceous Suva, Fiji, which gave me the opportunity to (Jackson, Schlanger et al., 1974), but not deliver the views expressed above. A travel enough is known about this group. grant from the Intergovernmental In summary, the Solomon chain is Oceanographic Commission (IOC) made it anomalous as an arc system. That it could possible for me to attend the conference: I have arisen as a fracture system is an idea am most grateful for this help.
REFERENCES
BLAKE, D. H. and MIEZITIS, Y. 1967: Geology of Melanesian Borderlands and India-Pacific plates' Bougainville and Buka Islands, New Guinea. Bur. boundary. Earth Sci. Rev. Miner. Resour. Geol. Geophys. Aust., Bull. 93: 56 pp. CURTIS, J. W. 1973: The spatial seismicity of Papua CLAGUE, D. A., DALRYMPLE, G. B. and New Guinea and the Solomon Islands. J. geol. Soc. MOBERLEY,R. 1975: Petrography and K-Ar ages of Aust. 20: 1-20. dredged volcanic rocks from the Western Hawaiian DENHAM, D. 1975: Distribution of underthrust Ridge and the southern Emperor Seamount Chain. lithospheric slabs and focal mechanisms - Papua New Bull. geol. Soc. Am. 86: 991-8. Guinea and Solomon Islands region. Bull. Aust. Soc. COCHRAN, J. R. 1973: Gravity and magnetic Explor. Geophysicists 6: 78-9. investigations in the Guiana Basin, Western Central DEVENTER, J. V AN and POSTUMA, J. A. 1973: Atlantic. Bull. geol. Soc. Am. 84: 3249-68. Early Cenomanian to Pliocene deep-marine sediments COLEMAN, P. J. 1970: Geology of the Solomon and from North Malaita, Solomon Islands. J. geol. Soc. New Hebrides Islands, as part of the Melanesian Re- Aust. 20: 145-50. entrant. Paci! Sci. 24: 289-314. DICKINSON, W. R. 1973: Widths of modem arc-trench COLEMAN, P. J. 1975: On Island Arcs. Earth Sci. Rev. gaps proportional to past duration of igneous activity in 11: 47-80. associated magmatic arcs. J. geophys. Res. 78: 3376·89. COLEMAN, P. J. and DAY, A. A. 1965: Petroleum FALVEY, D. A. 1975: Arc reversal and a tectonic model possibilities and marked gravity anomalies in north- for the North Fiji basin. Bull. Aust. Soc. Explor. central Guadalcanal. Rec. geol. Surv. Brit. Solomon Is. Geophysicists 6: 47·9. 2 (1959-1962): 112-9. FURUMOTO, A. S., HUSSONG, D. M., CAMPBELL, COLEMAN, P. J. and PACKHAM, G. H. in press: The J. F., SUTTON, G. H., MALAHOFF, A., ROSE, J. C.
139
TECTONIC SETTING AND PETROLOGY OF SAMOA- TONGA-FIJI REGION
JAMES W. HAWKINS
Abstract
The region bounded on the east and west by the Tonga and New Hebrides Trenches respectively and lying between the Samoan Island chain on the north and Fiji Islands on the south, includes a complex tectonic province which acts as a trench-trench transform in the broad scheme of plate tectonics. In detail, the region is far more complicated than would be expected if it were simply a zone or left lateral slip. East of Fiji, there are at least three distinct tectonic-petrologic regimes: (1) The andesitic-dacitic Tonga Island arc which presumably represents volcanism related to fractional melting of the Pacific Plate in the Tonga subduction zone. (2) The Lau Basin, a dilational marginal basin floored by oceanic ridge-type tholeiite. This basalt owes its origin to fractional melting of rising diapirs of mantle material in the region between the Tonga and Lau Ridges. (3) The linear volcanic chain of the Samoan Islands. The chain is characterised by alkalic and strongly silica- undersaturated lavas. The orientation of the linear chain, and the chemistry of the magma forming the islands, appear to be an effect of plate bending and rupture as the Pacific Plate is deformed at the north end of the Tonga Trench. Magma chemistry suggests that the source of erupted material may be close to the base of the lithosphere which, under Samoa, may be on the order of 80 to 90 km. deep. The crustal geology and geometry of tectonic features north and west of the Fiji Islands is equally complex. Seamounts showing the alkalic chemical characteristics of the Samoan Chain are found at least as far to the west as Hazel Holme Bank. The deep area (> 3.5 km. deep) which marks the westtrending part of the Tonga Trench may be traced more or less continuously to the Vitiaz Trench. There are no data in this region to support the postulated existence of the "Hazel Holme Fracture Zone", Reconnaissance studies of the Fiji Plateau have shown that there is at least one area (near 15°30'S, 173°E) in which high heat flow, bathymetry, magnetic anomaly data and the presence of MORB-type tholeiitic rocks indicate the existence of a spreading centre oriented in a generally north-south direction. The location of this spreading centre, with respect to the New Hebrides Trench, indicates a tectonic setting distinctly different from that of the Lau Basin.
INTRODUCTION end of the Tonga Trench. The "Hunter The region between the Tonga and New Fracture Zone" lies on this trend as does a Hebrides Trenches (Fig. 1) includes several zone of shallow focus (< 100 km.) different tectonic - petrologic regimes. earthquakes, (Barazangi and Dorman, 1969) Areas of crustal dilation, crustal convergence and offer support for the assumption that the and large scale transform faulting have been trend is a major plate boundary. Another recognised or postulated. In addition, there zone of shallow earthquakes and a linear are isolated fragments of thick, and probably array of seamounts runs from the north end old, crust; linear oceanic deeps which may be of the TongaTrench westward at about 14°S. abandoned convergence zones; and linear This trend is also marked by several deep volcanic chains. When considered in the areas (3.5 to 4.5 km.) which form a more or framework of large-scale plate tectonic less continuous trough between the Vitiaz theory, the region between the two trenches and Tonga Trenches. Thus the transform must contain a trench-trench transform fault. fault "boundary" between the Pacific and A likely orientation for this transform fault Australia Plates may actually be distributed would 'be from the south end of the New over a broad zone and, as will be shown, this Hebrides Trench northeasterly to the north zone apparently includes dilational features as
141
well as the dominant left lateral slip faults. Trench, Lau Basin, Samoan Island Chain, Locally, it is probable that there is oblique Melanesian Border Plateau and the North convergence on the transforms. Fiji Plateau. The interpretations are based The term "New Hebrides-Samoan largely on data collected by the author at sea Lineament" (NHSL) will be used informally and on the Samoan Islands but also draw on in this paper to refer to the linear array of the published data of other workers. In seamounts and deeps in the regions bounded particular, the extensive studies on islands of by the New Hebrides Trench on the west, the Tonga Arc by Bryan et al. (1972) and the east end of the Samoan Island Chain on Ewart et al. (1973), in the Fiji and Lau the east and extending from 12°S to 15°S. Islands by Gill (1970), and in the Samoan The term is not intended to describe a and Wallis Islands by Stearns (1944, 1945), specific tectonic process or feature but rather MacDonald (1944, 1945), Kear and Wood it is to focus attention on a major bathymetric (1959), Stice (1968) and Natland (1974, feature of the oceanic crust. This region 1975) will be used. includes the Melanesian Border Plateau, a name proposed by Fairbridge and Stewart Tectonic Setting and Bathymetry (1960) to describe the high standing ridge The major morphologic and tectonic capped by seamounts and islands which features of the area are shown in Fig. 1. The trends east-west near latitude 14°S. Pacific seafloor, east of the Tonga Trench, is This paper will summarise petrologic and about 5.5 km. deep. The depth, the apparent geochemical data for four different tectonic trends of recognised and dated magnetic settings or regimes in the Samoa-Tonga-Fiji anomalies (Winterer et at., 1974) and sparse region. The whole region, and specifically the DSDP data indicate an age on the order of NHSL, has been the site of active tectonism 100 m.y. or more for the seafloor beneath the and plutonic-volcanic activity through Samoan Chain, for the lithosphere being Cenozoic time and is still active. Magma subducted at the Tonga Trench and for the leaks at zones of crustal convergence, crustal lithosphere north of the NHSL. This dilation and lateral slip or transform faulting lithosphere must be at least 80 to 100 km. form distinctive igneous rock suites. The thick based on age-thickness estimates chemistry of these igneous rocks offers (Parker and Oldenburg, 1973). The major insights to thermal gradients, lithosphere plate boundaries are defined by geophysical thickness, and helps to identify the style of features such as the west dipping Benioff lithosphere deformation on the postulated Zone of the Tonga Trench, which reaches plate boundaries. depths of 700 km.; the east-dipping New The purpose of this discussion is to show Hebrides Trench Benioff Zone which reaches how the petrologic and tectonic data may be to a depth of 300 km and a belt of shallow used to frame working hypotheses which earthquakes trending southwesterly from the help to explain some of the geological north end of the Tonga Trench towards the complexities of the region along the NHSL. south end of the New Hebrides Trench. This The interpretations to be presented are based trend coincides, in part, with the Matthew- on the assumptions that experimental Hunter Fracture Zone. There are several petrologic data offer insights to the nature of areas of shallow earthquakes both on the Fiji the chemical and physical processes involved Plateau and in the Lau Basin which may in the generation and evolution of magmas mark sub-plate boundaries (Chase, 1971; and that the plate tectonics hypothesis is a Sclater et al., 1972): Barazangi and Isacks viable model for explaining the major (1971, fig. 1) recognised a zone of high tectonic elements of the Samoa-Tonga-Fiji seismic attenuation under the Lau Basin. area. Most of the data presented here are They attribute delays in arrival time of based on extensive surveys made on seismic waves from earthquakes in the Lau expeditions of the Scripps Institution of Basin to the presence of either high Oceanography in the areas of the Tonga temperature or partly melted mantle under
143
from the 1.5 to 2 km. deep area which forms been other continuous ridge systems, one the northern boundary of the basin. trending north-easterly and one, west- However, the geological-geophysical feature northwest. Donna Seamount and Zephyr known as the Tonga Trench must change its Shoal are bounded by very steep scarps. The character at the sharp change in direction relief on these scarps is 3 km. or more and near 15°S. There is no island arc on the they are presumed to have a fault origin on northern boundary of the Lau Basin and the basis of morphology and on sheared and there is no deep seismicity to indicate a partly recrystallised gabbro and greenstone Benioff Zone. One may assume therefore dredged from the scarps. These samples may that at present there is little or no component represent fragments of seismic layers 3 and 2. of underthrusting on the north. There is no Between 19°5 and 24°S, a discontinuous evidence that this margin ever was a zone of central ridge appears to show right lateral underthrusting with related volcanic activity offsets, with the ridge becoming more but since there are no petrologic data to indistinct as it goes south (Fig. 2 and indicate the nature of the seafloor in this area Hawkins 1974, fig. 1); south of 21°S the it cannot be ruled out as a possibility. bathymetric data are extremely sparse but Geochemical-petrologic data for the Fiji-Lau due to the narrowness of the basin it is Island Group lead to the interpretation that unlikely the morphology differs significantly they were formed by volcanism related to a from that shown. The deepest soundings in north to northeast facing subduction zone the Lau Basin (- 3.4 km.), and some of the prior to Pliocene time (Gill, 1970). It is likely most complex bathymetry, lies in the region that remnants of this island arc may be found between 17°30' and 19°00'S. It is an area of between the Tonga and Lau arcs, however moderate relief (2 km.) with numerous there is no indication of island arc volcanism seamounts and steep scarps but lacking major in the area today. lineated bathymetric features. The morphology of the Lau Basin is Tonga Arc
dominated by several linear ridges The Tonga Island Arc has been described presumably formed by linear magma leaks as by Bryan et al., (1972) and Ewart et al., well as by numerous seamounts interpreted (1973) who presented detailed petrologic data as being derived from point source magma establishing it as an andesitic-dacitic island leaks (Hawkins, 1974, 1976). The linear arc deriving its magma by fractional melting ridges form two major trends - in the subjacent Benioff Zone and northwesterly and north, with a less well subsequent fractional crystallisation. The developed north-easterly trend also apparent island arc is clearly the expression of a (see. Fig. 2). The Peggy Ridge is a long compressive tectonic regime related to litho- continuous feature running northwesterly for sphere subduction at the Tonga Trench; it is 300 .km. With about 1.3 km. of relief; it in marked contrast to the tholeiitic basalt dominates the bathymetry between 15’S and floored Lau Basin which is the expression of 17’S30’S and has led to speculation concerning lithosphere dilation. Fig.3 is a model cross- its possible role as a spreading centre (Chase, section of the Tonga Arc and Lau Basin and 1971}. The interpretation which seems most shows schematically the contrasts in tectonic consistent With the data IS that at present It style magma type and depth of origin of the acts as a transform fault between sub-plate magamas. in the Lau Basin (Sclater et al., 1972; Hawkins, 1974, 1976). There is a Samoan Chain discontinuous but parallel broad ridge with The Samoan Islands form a linear chain of less relief on the southwest side of Peggy volcanic islands extending for 550 km. from Ridge. On the northeast side are several large Rose Atoll on the east to Savai'i on the west. seamounts including Donna Seamount, Seamounts and shoals between Savai'i and Zephyr Shoal and Niua Fo'ou Island which Wallis Islands indicate that the linear are built on fragments of what may have seamount-island chain is approximately 1100
146
km. Bathymetric details are lacking because Tholeiitic basalts (Table 1) have also been of poor coverage but it appears that the dredged from a scarp on the Fiji Plateau near maximum reliefis on the order of 1.5 km. 180°, 13°30'S. There are not enough data to Heat flow on the Plateau is generally high indicate whether these rocks are from a (Sclater and Menard, 1967; MacDonald et spreading centre or from some other type of magma leak. a/., 1973, Hartzell, 1975). An east dipping Benioff Zone underlies the Several models have been proposed for the west side of the plateau adjacent to the New North Fiji Plateau; Chase (1971) postulated a Hebrides island arc and trench (Dubois et a/., complex set of spreading centres and 1973). In contrast to the Benioff Zone transform faults which in detail does not fit associated with the Tonga Trench, it only newer data but he did recognise it as an area extends down to a depth of about 300 km. of crustal dilation and that several sub- plates for most of its length under the plateau. must be involved. Luyendyk et a/., (1974) However, there is one cluster of 600 km. expanded on Chase's model and discussed deep earthquakes near 14°S and 170°E. the western part of the North Fiji Plateau. A Shallow earthquakes on the North Fiji more recent model (Falvey, 1975) proposes a Plateau « 50 km.) reported by Dubois et a/. less complicated geometry and proposes a (1973) appear to define active zones along progressive eastward displacement of the the trend of the proposed Hunter Fracture spreading centre with time, coupled with a Zone and the western end of the NHSL in progressive termination of spreading between the region traversed by the postulated Hazel successive fracture zones from north to Holme Fracture Zone (Chase, 1971). There south. I do not believe that the new data are four other areas of shallow seismicity on collected on SIO expedition EURYDICE in the plateau; one of them coincides with a 1974 are sufficient to propose a new model, possible spreading centre. First motion however, the data appear to rule out specific studies of two earthquakes near 15°30'S, parts of both Chase's and Falvey's models. 173°30'E allow an interpretation of a left- The proposed Hazel Holme Fracture Zone lateral strike-slip solution for one and a NW (Chase, 1971) does not appear on our seismic trending normal fault, NE side down, for the reflection profiler records east of 173°E and other (Hartzell, 1975). there is no good evidence for the proposed The shallow earthquakes near 15°30'S, north-south spreading centre (Nova rise) 173°E (Hartzell, 1975) are near an area which should be tied to it.The URYDICE surveyed on SIO expedition EURYDICE in data indicate that the central Fiji Plateau has 1974 which has bathymetric and geophysical shorter ridge segments offset by many more features suggesting a possible area of crustal fracture zones than Falvey's model proposes. spreading. Seismic reflection profiles indicate However, it appears likely that his the presence of rough topography having interpretation of the Fiji Plateau may be thinly sedimented tilted fault blocks. The more correct than previous interpretations heat flow values are in the range 8-10 H.F.U. and there are symmetric magnetic anomalies Petrology and Geochemistry which may be explained by spreading at 1-2 The details of petrology and geochemistry cm/yr half rate for the last 2-3 m.y. Rock for the major tectonic provinces described dredged from scarps near the presumed above have been presented elsewhere and spreading centre are tholeiitic (Table 1) and only a summary will be given here. The have chemical and mineralogic properties major point to be made from these data is indicating close similarity to mid-ocean ridge that they offer insight to depths of magma basalt-type tholeiites and to basalts of the Lau generation, to lithosphere thickness and to Basin. The presumed spreading centre trends the nature of source material. northerly and comprises several 30 to 40 km. 1. Hawkins (1974, 1976) has showed that long segments offset by 3.5 to 4 km..deep the Lau Basin basalts are identical to mid- linear features presumed to be fracture zones. ocean ridge basalt-type tholeiite in terms of
148
major and minor elements. Strontium isotope collapse and infilling - post-caldera intrusion ratios are slightly more radiogenic than of differentiated magmas - erosion - post- average mid-ocean ridge basalt but they are erosion eruption of strongly undersaturated comparable (0.7032-0.7035) and overlap the rocks. The alkalic basalts may have come mid-ocean ridge basalt range. REE content from depths of 50 to 60 km. but the origin of normalised to chondrites show a mid-ocean the undersaturated rocks probably involves ridge basalt-type pattern for least fractionated small amounts of melting of the mantle at samples (Hawkins, 1976) and l\ 80 values are pressures (depths) which would be near the typical of ocean ridge basalts in least altered presumed base of the lithosphere (80-90 samples (Pineau et al., 1976). Table 1 km.) under Samoa (Hawkins and Natland, includes data for least altered Lau Basin 1975). Table 1 also includes data for a basalts and for altered samples. An origin by seamount in the Samoan chain which fractional melting of a rising peridotite diapir appears to have had a history similar to that seems to explain the data best. Separation of of the large Samoan Islands. A more detailed melt from residual solid phases at depths on discussion of these rocks is in MacDonald the order of 30 km. is proposed for these (1944) and Hawkins and Natland (1975). magmas and their most likely source is in the Fig. 4 shows a plot of alkalis vs. SiOz for mantle beneath the basin. Counterflow, Samoan chain samples and compares them to probably caused by viscous drag on the Low samples from the Hawaiian linear volcanic Velocity Layer by the descending Pacific chain and to Atlantic Ocean seamounts. Plate, is postulated as a mechanism to 4. Data for Fiji Plateau rocks are very. replenish the supply of magma and may help scarce but samples dredged from the move the plates apart. Fig. 3 shows a model presumed spreading centre have 48-51 % SiOz, to explain the possible origin of the basin 1.2-1.4% TiOz, .03-.16% KzO and .09-.11% according to this hypothesis. A more detailed PZ05; olivine microphenocrysts are FOsB and account of the basalts and their evolution is whole rock Mg/Mg+Feii is .47 to .57. They presented elsewhere (Hawkins, 1974, 1976). appear to be moderately fractionated tholeiitic 2. The Tonga Arc lavas seem to be typical lavas. Derivation from depths of about 30 of intermediate composition lavas presumed km. by fractional melting of a peridotitic to have formed by fractional crystallisation of mantle seems likely. They are close to Lau parental magma derived by high pressure, Basin basalts in major and minor element hydrous, fractional melting of basaltic chemistry. (eclogitic) rocks. Bryan et al., (1972) and The rocks dredged from a 1.1 km. high Ewart et al., (1973) have given detailed scarp at 180° and 13°30'S are tholeiitic basalts discussions of the arc and the petrology- similar to those described above. The average geochemistry of the rocks. data for the Fiji Plateau basalts given in Table 3. The rocks of the Samoan linear chain 1 are based on a composite of these samples (Table 1) are mainly alkalic basalt intruded and those from the presumed spreading by lesser amounts of phonolite or trachyte centre. All of the Fiji Plateau tholeiites have differentiates and capped by strongly mineralogic and chemical characteristics undersaturated rocks (MacDonald, 1944; which resemble marginal basin basalts such Stice, 1968; Natland, 1974; Hawkins and as those from the Lau Basin. However, the Natland, 1975). Basanites apd nephelinites Fiji Plateau samples appear to be more are the youngest erupted lavas on the chain. strongly fractionated than the average Lau They are considered to be significant as Basin basalt; they are quite similar to indicators of both depth of magma generation fractionated Lau Basin samples. and of lithosphere thickness based on The Hazel Holme Bank samples are experimental data bearing on their origin strongly fractionated alumina-rich trachy- (Bultitude and Green, 1971). The Samoan basalts; plagioclase phenocrysts are abundant. chain shows the classic volcanic island They plot in boundary for the field of alkali sequence of: shield building stage - caldera basalts on Fig. 4 and probably belong to the
149
Samoan Linear volcanic chain described sediment fill. The basin is asymmetric in above. cross-section with the thickest fill on the southern edge. The Paleo- Vitiaz Trench may New Hebrides - Samoa Lineament have been part of the north to northeast North of the Lau Basin, the Fiji Islands facing subduction zone-island arc complex and the North Fiji Plateau, there is an area of which Gill (1970) proposed on the basis of a deep water ranging from 3.5 to over 6 km. in petrologic-geochemical study of Fiji and Lau depth running in a generally west- Island volcanic rocks of late Tertiary age. northwesterly direction. The Vitiaz Trench is Falvey (1975) also recognised the likelihood at the northwestern end of this trench and that the Paleo- Vitiaz Trench was a major the east end is at the point of curvature of the subduction zone in mid-Tertiary time; he Tonga Trench. However, the deep water does proposed that the present en-echelon offsets not form a continuous trough but is broken of the deep water are secondary features due by a series of en-echelon steps; the trench to progressive termination of trench segments are separated by seamount clusters segments as the spreading centres on the Fiji or islands such as Wallis and Rotuma. Plateau migrated eastward on growth Winterer et al. (1974) modeled seafloor ages transform faults. north of the NHSL and showed the seafloor The abandoned trench appears to be aging to the southeast towards this deep serving at present as part of the left-lateral water feature which might therefore strike-slip regime which acts as a trench- represent an abandoned trench. Seismic trench transform. The young volcanic reflection profiles across this linear deep area features located on this lineament have (Hawkins, unpubl. data from EURYDICE petrologic and geochemical properties which expedition, 1974) show up to 1 second of indicate that they are part of the alkali basalt
150
- nephelinite magma series which forms the high standing areas of oceanic crust. Both Samoan chain. A possible explanation is that areas have actively spreading ridges and the islands such as Wallis, and seamounts seamounts from which tholeiitic basalt has such as Hazel Holme Bank, are magma leaks erupted to form new seafloor. The Lau Basin localised on the fractures which offset the separates an older inactive volcanic ridge Paleo- Vitiaz Trench. The NHSL is therefore (Lau Ridge) from the volcanically active a complex region - there are remnants of an Tonga Ridge. The "trench.trench trans. abandoned and partly filled trench, there are form", which separates the two trenches with young and perhaps still active alkalic basalt opposing dips, probably is distributed over seamounts, and there is active shallow more than one trace. Part of the required slip seismicity for which some left· lateral strike- probably is on the Hunter Fracture Zone, slip first motion determinations have been part is probably taken up along the informally made. named New Hebrides-Samoa Lineament. The linear volcanic chain of the Samoan SUMMARY Islands owes its origin and location to The Samoa-Tonga-Fiji region is complex complex lithosphere deformation at the north and the details of present or past tectonic and end of the Tonga Trench and Lau Basin. Part petrologic activity are certainly not well of the NHSL appears to follow an abandoned understood. At present it is possible to and now partly filled trench which formerly recognise some of the major geologic consumed oceanic lithosphere arriving from a elements of the region and a general model ridge system off to the northeast. may be presented. The Tonga and New Hebrides Trenches are crustal consumption ACKNOWLEDGEMENTS zones with adjacent island arcs which are This research was supported by grants sites of basaltic-andesitic to dacitic volcanism. from the National Science Foundation The Lau Basin and North Fiji Plateau are (USA).
SEDIMENTS AND TECTONIC HISTORY OF THE TONGA RIDGE, AND THE PROBLEM OF THE LAU BASIN
H. R. KATZ
Abstract
Previous studies have shown the Tonga Ridge to be a young volcanic island arc virtually devoid of sediments, which was formed by rifting and eastward migration of parts of an older arc system (the Lau Ridge) and/or the replacement of an original inclined seismic zone by a new one further east. In the course of these events, the Lau Basin would have been formed as new oceanic crust was being created in an area of extension between the Tonga and Lau Ridges. Based on new data from well sections which terminate at 5.500 ft (1685 m.) in the upper part of the Lower Miocene and detailed seismic sections which indicate a total sediment thickness of 3 km. or more, the Tonga Ridge is now seen to have had a long and complex history dating back at least to the Eocene. Sediments were to a great extent deposited in bathyal conditions, but important breaks occur in the Oligocene and Pliocene. Thick limestones were formed in the Late Eocene, while during the Miocene volcanic islands gave rise to deposition of local shallow-water carbonates and coarse slopebreccias, as well as mainly deeper-water fine-grained volcanoclastic sediments. Thick coralline limestone was deposited in Plio-Pleistocene and Recent times after the ridge as a whole had been uplifted. The Tonga Ridge is still undergoing comparatively strong differential block movements which are mainly, but not everywhere, upwards. Volcanic activity occurred during and possibly earlier than the Middle Eocene, in the Miocene, and in Plio-Pleistocene to Recent times. Recent andesitic-dacitic volcanism is concentrated further west of the area of Miocene volcanism; this contradicts the model of an eastward-migrating arc-subduction system. In addition, seismic sections across the Lau Basin suggest a tectonic history as long and as complex as in the Tonga and Lau Ridges. The two ridges could not therefore have originally been closely together. In many parts or the Lau Basin, a sedimentary series of pre-Late Miocene age is indicated; basement may thus be as old as Early Tertiary. Late Miocene to Plio-Pleistocene sediments in the area of DSDP hole 203 are up to 1,000 m. thick and may to a great extent consist of locally derived non-volcanic oozes; in addition, some intermittent ash layers were found in the 409 m. thick Plio-Pleistocene well section. While there is evidence in the Lau Basin for tectonic deformation in the pre-Late Miocene, strong faulting occurred during Pleistocene and Recent times; this was probably connected with regional uplift that occurred at this time in the Tonga and Lau Ridges. Associated with this faulting are intrusions and extrusions of igneous-volcanic bodies which locally penetrate older basement and sedimentary series, forming many high ridges and seamounts. Tholeiitic basalts dredged in the Lau Basin during Scripps Institution of Oceangraphy cruises are probably from these places and do not therefore represent original basement. The Lau Basin appears to be a structural basin formed by regional collapse between two belts of major uplift. Upwelling mantle material in this area of collapse has resulted in the formation of basaltic magmas which have pierced through to the surface, thus introducing new oceanic crustal material locally into an area of pre-existing continental affinities and structure. It is suggested that the entire region from New Zealand northwards to Tonga, Lau and Fiji, which from the Early Tertiary onwards has been a region of greater mobility, has on the whole experienced accretion to a quasi-continental structure; locally, however, as in the Lau Basin, recent foundering and oceanisation may cause its destruction and reversal to oceanic-type crust.
INTRODUCTION related to island arcs. Results of petrological In recent years much geophysical work has and petrochemical studies of volcanic rocks been carried out over the Tonga Ridge and seem to a great extent to support these Lau Basin regions and this has led to new concepts. Less work has been done regarding concepts regarding marginal oceanic basins as sedimentary rock formations, although these
153
can provide a greater and more direct insight evolution, it appears, must therefore be into the geologic evolution of an area; their interpreted somewhat differently. The data thickness distribution, facies and age are derived from an extensive marine relationships, and structural deformation are geophysical survey by Mobil Oil corporation, basic to the understanding of the tectonic from field studies by the author and detailed framework and history. geological and drill log reports by Shell now An excellent summary of available data on Open File in Nuku'alofa (Mulder and and a comprehensive discussion of the Nieuwenhuizen, 1971; Tonga Shell, 1972). geology of the Lau Basin have recently been given by Hawkins (1974 and this volume). In his interpretation, the Lau Basin is an area of THE SEDIMENTARY SEQUENCE crustal dilation related to the generation of OF THE TONGA RIDGE oceanic lithosphere. It is bounded in the west by the Lau Ridge, considered an extinct 1. Well logs of Kumifonua-l and -2 in andesitic island arc, and in the east by the still Tongatapu. active andesitic-dacitic volcanic arc of the Based on gravity and magnetic surveys on Tonga Ridge. According to Hawkins, the the island and tied to offshore' seismic sediment cover on both these ridges is thin to measurements, these wells were to test the absent (less than 0.01 sec.). Also in the Lau stratigraphic sequence and ascertain whether Basin, sediment cover is very thin, generally Eocene limestone above volcanic basement, less than 0.1 sec. penetration, with a few as found in outcrop on the island of 'Eua narrow ponds which reach 0.1 to 0.2 sec.; further east, is present in the central part of these ponds seem to be in fault-controlled the Tonga Ridge. Both wells showed that the depressions, or possibly behind seamounts sedimentary succession is much thicker than and ridges of volcanic origin that have acted anticipated and failed, within the available rig as dams to trap the sediments. Only along capacity, to reach the main objective. At total the edges of the basin parallel to the Lau and depth (T.D.) 5,525 and 5,529 ft respectively, Tonga Ridges were comparatively thick only the upper part of the Lower Miocene sedimentary wedges (up to 0.7 sec.) found. had been reached (Fig. 2; Tonga Shell, 1972). Hawkins' conclusion is that these sediments Below about 450 ft of Plio-Pleistocene on either side' of the basin were derived from reefal limestone and after a sharp break the volcanic areas, whereas some of the which suggests an unconformity or sedimentary filling of the central basin ponds disconformity, a somewhat monotonous was locally derived and has probably slumped alternation of coarser and finer detrital off the local volcanic ridges; in general, sediments rich in volcanic lithoclasts was however, the thin sediment cover of the traversed; the grains mainly comprise acidic central basin is also considered to be derived and basic andesite and basalt. An from the Lau and Tonga Ridges, having agglomerate cored at T.D. in Kumifonua-2 bypassed the sediment traps along the basin shows flow structures, and under the margins. The age of the basin is thought to be microscope a certain amount of autometa- Pliocene to Late Miocene, probably not older morphism or hydrothermal alteration is than mid-Miocene as suggested by the drill visible. This points to the occurrence in this log of DSDP hole 203; based on 20 dredge area of intra-Miocene (Lower Miocene) samples which were, however, all taken from volcanism which is also indicated from seamounts, ridges and steep scarps, it is observations on rock outcrops in the conduded that the basin is underlain by Nomuka-Mango islands further north. The young seafloor formed of tholeiitic basalt. age interval involved for the whole series This paper presents additional data on the (Lower Pliocene to Lower Miocene) and the sediment thickness and structure of the general lithological characteristics of the region, which show that the geologic situation volcanoclastic rocks are very similar to those is undoubtedly more complex; the tectonic of outcropping formations on 'Eua to the east
155
and on islands of the Nomuka Group to the locally also reefal, bioclastic and foraminiferal north. The open marine benthonic- limestones occur. planktonic faunal assemblages in the Miocene to Lower Pliocene sediments 3. Pliocene limestones of Vava 'u indicate depths of deposition between 50 and Unconformably below the Quaternary 200 m.; however, taking into account the raised terrace-forming coral limestones there volcanic character of the area, as well as the are tilted (up to nearly 20° dip) coral reef difficulty in assessing the amount of limestones below which is a thick sequence transported fossils, it seems highly probable of massively bedded detrital limestones and that a large part of these sediments has been local rather soft chalky bryozoan limestones, deposited in the bathyal zone. alternating with thinner Cycloclypeus packstones. The detrital limestones show 2. Comparison with Miocene outcrop sections large-scale foreset bedding, and contain on the Tongan platform clearly rolled coral colonies, thick-shelled On 'Eua the Miocene-Lower Pliocene gastropods and pelecypods: they are volcanoclastic sediments are mainly slightly interpreted as shallow water reef slope calcareous lithic sandstones with subordinate deposits. It has been speculated (Mulder and and lenticular siltstone or mudstone Nieuwenhuizen, 1971) that the whole of the intercalations: they all contain an abundant Vava'u main island could represent an old planktonic fauna (Orbulina). On the eastern drowned caldera during Pliocene time: its side of the island, bathyal submarine edges would be represented by the reef and channels filled with these sediments cut reef slope deposits now exposed in the through the Eocene limestones and into the outward cliffs, which are all open marine, underlying basal volcanics. shallow-water limestones with abundant On various islands of the Nomuka Group, larger foraminifera; the lagoonal limestones similar rocks which have ages of late Lower of Neiafu quarry would be part of a larger Miocene, early Middle, Middle and late lagoon occupying the central parts of the Middle Miocene, and late Upper Miocene, island. are represented by largely turbiditic detrital The total thickness of these older volcanoclastic and pyroclastic sediments, limestones at Vava'u is about 250 m.; while which show strong evidence of deep-water their age could not be more closely channelling. Evidence of contemporaneous determined than "Pliocene or younger", a volcanism is seen in several places. Pliocene age has generally been adopted. Globigerina mudstones, coarse slope breccias, and Middle Miocene reefal and shallow 4. Lower Tertiary sediments in the southern marine foraminiferal limestones apparently part of the Tongan platform, and total represent the remains of fringing reefs of sedimentary thickness former (volcanic) islands (Mulder and Overlying what is considered volcanic Nieuwenhuizen, 1971). An unconformity basement on 'Eua Island are 6-8 m. thick (a separates the Upper Miocene from raised thickness of 25 m. is mentioned from other Quaternary coral reefs. localities) conglomerates and tuffaceous In conclusion, a fairly wide area of the calcareous sandstones which are highly Tongan platform is covered by thick fossiliferous. They are followed above by sediments of Miocene age (ranging from massive hard limestones which are mostly Lower Miocene to Lower Pliocene); beneath micritic to bioclastic (mainly foraminiferal Tongatapu these are at least 1,500 m. thick. with some algal lime wackestones).,These are The area probably was dotted with partly 100-200 m. thick and have been deposited in active volcanic islands, between which the a protected environment, suggesting a period sediments were deposited in shallow- to of little or no volcanic activity and rather deep-water environments. They are mainly quiet tectonic conditions. The age of this coarse to fine grained volcanoclastics, but series is late Middle Eocene to possibly
157
lowermost Oligocene, but the upper surface is Fault-block tectonics combined with widely channelled and was obviously eroded subordinate fOlding are characteristic of the before deposition of the Lower Miocene entire segment which, north of New Zealand, mudstones and volcanoclastics. constitutes the 2,500 km. long Kermadec- The structure of 'Eua is essentially an Tonga Ridge (Katz, 1974). Apart from assymmetric faultblock anticline and offshore regional faults which are mainly meridional, seismic data show clearly its close connection there are many east-west, and northwest and with the rest of the Tongan platform. If the southwest-striking cross-faults on the Tongan correlation of a prominent reflector with the platform, where many of the relatively deep Eocene limestone is correct (Katz 1974, fig. and narrow channels between the main 11), the latter would be about 3 km. deep island groups are probably controlled by such below Tongatapu (cf. also Kroneke and faults. It can be seen on 'Eua that there has Tongilava, 1975). No certain basement has been intermittent tilting and warping of been distinguished on seismic records, while individual blocks since Eocene time; at the virtually conformable reflections are visible northern end of the platform, Vava'u down to over 2 seconds, and sometimes to 3 represents a block that has undergone very seconds below sea bottom, which gives a recent tilting to the south-southeast, which depth of perhaps 4-5 km. Sonobuoy velocities still continues as shown by drowned reefs in the Miocene volcanoclastics below about 1 between Vava'u and Ha'apai. North of km. depth are 2.5 to 3.5 km/sec, and 4.6 Vava'u there is some evidence that the km/sec for what is the inferred Eocene Tongan platform is terminated by a major limestone at 3.1 km. depth, while velocities at downthrow probably along a major west- the deepest reflections are 5-5.3 km./sec; northwest trending fault. these may, however, be from lavas of the Step-wise downfaulting seems to occur in volcanic basement, which in outcrop in one the east towards the Tonga Trench, but the place at least on the eastern side of 'Eua apparent lack of sediments along the lower Island are conformable below Eocene part of the slope above a widely developed sediments. terrace (the "trench slope break") and on To the north along the Tongan platform, this terrace itself and down the trench slope this basement (as indicated by the above makes it difficult to recognise th€( true velocities) becomes much shallower; this is structural situation. In the west a large also suggested by a corresponding decrease in normal fault, downthrown to the west about overall seismic penetration. 1500-2000 m., has been mapped over a Thus the total sedimentary thickness in distance of over 200 km. along the Tongan the southern part of the Tongan platform platform (Katz, 1974). where it comprises Eocene (to possibly Differential vertical movements thus have Oligocene), Miocene and Plio-Pleistocene affected the Tonga Ridge since at least the deposits, appears to have a minimum late Middle Eocene, i.e. for about 50 m.y. On thickness of 3 km., and perhaps as much as 'Eua, during the Eocene, the basal volcanics 4-5 km. Northwards the thickness had undergone considerable erosion undoubtedly diminishes, and seismic sections apparently due to uplift; submergence suggest that the Eocene may cut out near the followed during a rather more quiet period Nomuka island group, and the Miocene without volcanism. Although much of the underneath the Ha'apai group; around Oligocene is probably missing, it is not clear Vava'u, volcanic basement is probably whether the Upper Eocene limestones were directly overlain by Pliocene. actually uplifted above sea level prior to a period of renewed volcanism and tectonic STRUCTURE AND GEOLOGIC unrest during the Lower and Middle HISTOR Y OF THE TONG A RIDGE Miocene. Volcanic islands, possibly in a chain Cross sections show that the ridge lying some distance further west, developed structure as a whole is dominated by faulting. during this period; constant erosion of these
158
islands supplied thick volcanoclastic tectonic history since at least the Early sediments, often as turbidites and related Tertiary. high-energy pelagic deposits, to relatively As has been shown elsewhere (Katz, deep water between and adjacent to these 1974), the thick Sedimentary sequence of the islands, which were surrounded by shallow- Tonga Ridge is continuous southward along water carbonate rocks of fringing reefs. the Kermadec Ridge; across the deep This situation probably continued during sedimentary basin of the Raukumara Plain most of Miocene and also Pliocene times. A this sequence is connected with the East northward uplift, such as has occurred in Cape and East Coast region of the North post- Pliocene time around Vava'u, may have Island, New Zealand. started earlier on a regional scale. 'Eua, in the south, emerged during the Upper Pliocene, SEDIMENTS AND STRUCTURE IN and this was probably also the time when the THE LAU BASIN - DSDP SITE 203 Tongan platform came into being: seismic Fig. 3 shows part of a seismic section sections suggest that older sediments are across the southern Lau Basin. The welling sharply down thrown at the western of DSDP hole 203, which is located 11 km. boundary fault, while the newly created off this line, has been projected at right angles depression (Tofua Trough) was now being into the section. filled with thick sediments of younger, Very good and persistent reflections of a probably Plio-Pleistocene age. Also, volcanic conformable sequence are found in this and activity seems to have shifted further west adjacent sedimentary troughs to a depth of again; within the Tongan platform away 0.5 to 0.9 sec. below sea bottom. These from the western boundary fault, there is no sediments are considerably affected by evidence of contemporaneous volcanic faulting, as well as by draping and warping, eruption centres during this more recent and in some places they lap onto and wedge period. Stepwise uplift, evidenced by raised out against basement highs; also, an older coral reefs with well-developed reef terraces, and more strongly deformed sedimentary is responsible for the origin of most present- sequence occurs locally unconformably day islands. beneath the former. On the whole, the relatively thick A somewhat poor sonobuoy recording sediments of the ridge area indicate long- close to the site of DSDP hole 203 gave a lasting mobility with a general tendency to velocity of 2.5 km/sec at about 0.3 sec. below subsidence of the crust, which was therefore sea bottom. This may be slightly above the accreted and thickened to the quasi- total depth of hole 203 (409 m.). Sonic continental structure of today, as is shown by velocities measured on individual rock gravity and seismic wave velocities (Talwani fragments of indurated ashes and oozes from et al., 1961). This is in strong contrast to the cores 4 (230 m.) and 5 (400 m.) gave values area east of the Tonga Trench: here during ranging from 1.66 to 2.24 km/sec (Burns et nearly the same period of time (since the al. 1973); velocities of core 3 at 110 m. depth Oligocene) only 103 m. of abyssal clay were were 1.54-1.56 km/sec. Thus a conservative deposited on oceanic crust in a continuously average velocity across the maximum deep ocean environment, with layer 3 at sedimentary thickness shown in this section shallow depth underneath (Katz, 1974); west (Fig. 3, west of 0900 hours), which is 0.9 sec. of the trench the increased crustal instability (and taking into account increased resulted in pronounced differential vertical compaction with depth), would be about 2- movements, volcanicity and the deposition of 2.5 km/sec. This suggests a total thickness of an enormously thicker sedimentary sequence probably not less than 1,000 m. at much shallower water depths. The Tonga Hole 203 at 409 m. depth terminated in Trench therefore marks the site of an sediments of the Middle Pliocene about 3 important suture that has separated tectonic m.y. old (A. R. Edwards, pers. comm.). provinces of decidely different character and Extrapolating an average sedimentation rate
159
Tofua Trough (west of the Tonga Ridge) in al1 from high-lying ridges, fault-scarps and the northern part of which sediments are seamounts, represent, such younger over 1 second thick, are they probably to a intrusions and not the original, pre- great extent pyroclastic tuffs and derived sedimentary basement. Detailed seismic volcanoclastics. control around dredge stations (and for As mentioned above, the sedimentary selecting them in the first place), as well as sequence which, according to data from the age-dating of recovered basalts, should give a trough of DSDP hole 203, is of Late Miocene final answer to this question. to Plio-Pleistocene age, is in some places unconformably underlain by an older series of stronger deformation which also has GEOLOGY OF THE LAO RIDGE somewhat different seismic characteristics AND ITS COMPARISON WITH (e.g. Fig. 3 at 0700 and 0730 hours). It can THE TONGA RIDGE therefore be recognised where it is exposed locally at the surface, as it is clearly Karig (1970) suggested that the "third arc" distinguishable from the regularly straight (Lau-Colville ridge) and the "frontal arc" and paral1el reflections so typical of the (Tonga-Kermadec ridge) together formed an younger sequence. At the margins of the older frontal arc until the Early Pliocene; basin, this older sequence seems to continue increased island arc tectonism near the end into the structures of the Lau and Tonga of the Tertiary would have resulted in Ridges; particularly across the upthrow of the extensional rifting within this feature, the western boundary fault of the latter, it new frontal arc and trench position migrating appears to correlate with some part of the eastwards thereby creating new oceanic crust substratum mapped over the Tonga Ridge. ("interarc basin") between the two arcs. Its age which must be older than Late Hawkins (1974) essentially adheres to the Miocene may therefore go back to the Early same explanation of moving the seismic Tertiary. zone, trench and active arc which originally The basement in the Lau Basin may were at the Lau Ridge, to the east thereby therefore be Eocene or older, as on the Tonga creating an extensional situation and the Ridge. A pre-Late Miocene unconformity is generation of new oceanic crust. Of two also distinguishable on seismic sections. slightly different models presented by him, These sections indicate that tectonic he favours model B which does not require deformation undoubtedly occurred before splitting of the old volcanic arc (Lau Ridge) and during deposition of the older but postulates that the original inclined sedimentary sequence (Miocene to ?Eocene) seismic zone simply is replaced by a second which is on-lapping and thinning against seismic zone further east, which would result older basement highs, and is thickening in in the formation of a new volcanic (andesitic) fault-angle depressions which obviously were arc while counterflow above the seismic zone active at that time (cf. Fig. 3 at 0700 hours). is responsible for the opening of the basin The most conspicuous and widespread between the arcs. Contrary to Karig's and deformation in the Lau Basin is, however, Hawkins' model A theory, model B implies represented by young faulting of Pleistocene that the Tonga Ridge is younger than the Lau to Recent age, including considerable vertical Ridge. displacement and tilting of individual fault Obviously, the geology of the Lau Ridge blocks. Associated with this young faulting should give important clues to this question, there are probably igneous-volcanic and it is surprising therefore that neither of intrusions and piercement structures, which these authors adequately discussed previous form many high crests and mountains detailed work on the structure, stratigraphy throughout the Lau Basin. It is probable that and in particular the petrology of the volcanic the dredge samples mentioned and rock formations of the Lau Ridge. The investigated by Hawkins (1974), which are extensive work presented by Ladd and
161
Hoffmeister (1945) indicates that there were and Koro Mbasanga volcanic periods of the two periods of regional volcanism, separated Lau Ridge, or only one that correlates with by a period of non-volcanism during which the older Lau volcanism alone, is not known. the Futuna limestone was deposited; this In either case, Karig's and Hawkins' models limestone is assigned to the upper portion of both fail in one important aspect, in that the Lower Miocene. Emergence and erosion volcanism and volcanic island arcs have not followed the deposition of the Futuna moved east but, within Tonga at least, limestone and renewed volcanism took place migrated westwards from the Lower while the area was emergent. Submergence Miocene (perhaps even from the Eocene) to and deposition of the Ndalithoni limestone the Present; Hawkins' model B certainly does occurred in the Upper Miocyne or Pliocene. not hold true, since there is no evidence that According to Ladd and Hoffmeister the the Tonga Ridge, and the andesitic island arc age of the two volcanic periods is pre-Lower volcanism in Tonga as a whole are younger to Lower Miocene for the older Lau than the Lau Ridge. Both the Tonga and Lau volcanics, and Middle to Upper Miocene for Ridges, and their associated volcanism, date the younger Koro Mbasanga volcanics. The back to at least the Lower Miocene (Eocene Lau volcanics are described as andesitic in Tonga); and while on the Lau Ridge agglomerates, tuffs and some lavas, and the volcanism apparently ended in the Upper Koro Mbasanga volcanics as flows, Miocene, it has continued to the Present on agglomerates and tuffs most commonly of the Tonga Ridge - but has migrated olivine basalt. However, Alling (in: Ladd and westwards from the site of Miocene activity. Hoffmeister 1945, p. 191) states that From seismic sections (Fig. 1), it petrologically there occurs every degree of furthermore appears that the Lau Ridge is a gradation between the two, although the field complex and complicated structure with relationships do seem to confirm that rocks strong block-faulting and tilting of sedi- of andesitic composition, including andesitic mentary sequences that show a thickness of basalts were extruded before deposition of 0.5-1 seconds. They seem to be older than the Futuna limestone, whereas olivine basalts the less deformed Late Miocene to Plio- came afterwards. It appears therefore that the Pleistocene sequence which is seen on the Lau Ridge volcanism is complex and not flanks of the Lau Ridge, and in parts of the simply that of an andesitic island arc of Lau (and South Fiji) Basin. Also on the Lau intermediate to silicic composition and in Ridge, there are numerous intrusive and clear contrast with the basalts of the Lau extrusive igneous bodies recognisable on Basin, as is held by Sc1ater et al., (1972) and seismic sections; many are accompanied by Hawkins (1974). This, of course, may be true distinct magnetic anomalies. also of the Tonga Ridge, where the Miocene. In conclusion, the Lau Ridge appears to be volcanic rocks have not been described so far a structure of similar type and age to the to any detail. It is possible that they represent Tonga Ridge, apart from the Recent a similar change in volcanic character: andesitic-dacitic volcanism of the latter. They cuttings descriptions (Tonga Shell, 1972) both have experienced andesitic to basaltic from the volcanoclastic sequences in the two volcanism in Miocene and earlier times Kumifonua wells mention acidic and basic (including Eocene in Tonga), as well as andesite and basalt (whether, and to what quieter periods of non-volcanism and extent these Miocene volcanic rocks differ limestone deposition. The older sedimentary from the Recent ones which have been sequence of both ridges can be traced described as basaltic andesite, andesite and intermittently across the Lau Basin, where dacite - Bryan et al., (1972) - has to my there is evidence of a pre-Late Miocene knowledge not been established either). unconformity; on the two ridges, an Whether there was, on the Tonga Ridge, a unconformity occurs in the Pliocene, and an more extended and continuous period of older unconformity in Tonga is dated as Tertiary volcanism covering both the Lau Oligocene.
162
SUMMARY AND CONCLUSIONS occasional ash showers probably consists largely of nanno oozes. The present data clearly show that the Both the Tonga and the Lau Ridge are Tonga Ridge is a comparatively old feature structures which are strongly dominated by (for intra-Pacific regions) with a well-defined block-faulting. Intensive faulting also occurs geologic history dating back to the Eocene. within the Lau Basin; while some of the Sediments at least 3 km. thick, and possibly faults appear to be pre-Late Miocene, the age 4-5 km. thick, have been deposited, and of most of them is Pleistocene to Recent. volcanism occurred immediately before the Apparently associated with this young Middle to Late Eocene limestone deposition faulting are widespread intrusions and of 'Eua, during Early and Middle Miocene extrusions of igneous-volcanic bodies; times and possibly still later, and in Recent extrusive masses generally form pronounced time. Over 1,500 m. of Lower Miocene to highs, ridges and seamounts. It is probably Lower Pliocene sediments have been drilled from these places that samples of tholeiitic into beneath Tongatapu Island without basalts have been dredged (Sclater et al., reaching the base of the Miocene. Much of 1972; Hawkins, 1974); Hawkins' conclusion the area of the Tonga Ridge probably was that regionally the "Lau Basin is underlain by bathyal through most of the period, but young sea floor formed of tholeiitic basalt" locally horst-block faulting, uplift and erosion (1974; see also this volume) seems to be a occurred during and after the Late Eocene, generalisation not supported by sufficient while volcanic islands probably were built up evidence; in the writer's opinion these basalts in many places giving rise to shallow-water are a later addition to the overall basin carbonate deposition and coarse slope structure, the older basement and breccias in the Miocene. In Plio-Pleistocene sedimentary series of which they have locally and Recent time the Tongan platform was penetrated. To account for their accommo- strongly uplifted with individual blocks being dation, some extension must certainly have differentially tilted; thick coralline limestones taken place in the Lau Basin; just how much, were deposited during this time and have lo locally been uplifted as much as 180 m. above however, is difficult to say. The basin as such sea level (Vava’u). Recent volcanism has seems to be a structural basin that originally shifted westwards and is concentrated along formed by regional collapsed between the and west upfaulted Tonga and Lau Ridges, much in and west of a major boundary fault, which the same way as its southernmost extension, separates the Tongan platform from the Lau the Taupo Volcanic Zone or graben of the Basin. North Island, New Zealand, came into being. In some places the older (Eocene to However, because of much greater Miocene) sedimentary sequence of the attenuation of the crust in this northern Tonga Ridge, which is downfaulted to the region (Barazangi and Isacks, 1971), west, is seen on seismic sections to continue upwelling mantle material has resulted in the unconformably below younger sediments of formation of basaltic magmas which, without the Tofua Trough, but apparently also occurs being decidedly contaminated, have pierced intermittently across parts of the Lau Basin. through to the surface thus forming what has The age of the younger sediments in the Lau been called new oceanic crust. Basin probably is mainly Plio-Pleistocene, but The crust of the Lau Basin therefore is near DSDP hole 203 the base of the regarded as of mixed type, being originally corresponding sequence when extrapolated and basically quasi-continental as in the beyond the total depth of the well is found to Tonga and Lau Ridges, but having in its date back to about the middle Late Miocene. latest history received an admixture of newly The total thickness of this sequence is up to created oceanic crustal material. The tectonic 1,000 m; here in the central part of the basin process visualised here - and which is most it is mainly locally derived and not from the compatible with all the observed data _ is Lau and Tonga Ridges and apart from one of vertical collapse including extension,
163
and structural disrupture with widespread Lau Ridge, and possibly still further. intrusions; the process may lead to what Andesitic volcanism, on the Tonga Ridge, might be called "oceanisation". In the has shifted westwards from Tertiary to absence of gravity and seismic refraction data Recent times. If there is subduction below in the Lau Basin, the data of a single seismic the Tonga Ridge as suggested by the present- refraction station in the Havre Trough day seismic zone, and if the Tonga volcanism further south (Karig, 1970), which is the of Recent as well as Tertiary time is related to direct continuation of the Lau Basin, is of such subduction, then the seismic zone must particular interest here. Although the crust in have become less inclined with time, or the Havre Trough is thinner tMn under the migrated further to the west. It is suggested Colville and Kermadec Ridges, an that a westward migration may have intermediate layer several kilometres thick occurred in response to breakdown along the which occurs under the two ridges (5.1 and eastern edge of the Tonga continental 5.4 km/sec., respectively) also is present margin, which could have taken place in a under the Havre Trough between them (4.4 similar way to that found along the Pacific km/sec.); this corresponds to the 5.1 km/sec margin in South America (Katz, 1970, 1971). layer of comparable thickness underneath the Both westward migration and/or flattening of Tonga Ridge (Raitt et al., 1955). Thus a the seismic (subduction) zone could thus be similar crustal structure is found across the thought of as a measure of tectonic resistance entire region, basin and ridges alike; this brought against it by the opposing continental points to a comparable tectonic history. This plate: the Tonga continental plate, being crustal structure is clearly different, however, young and immature (not rigidly cratonised from the typically oceanic crustal structure of and relatively thin) is easily being the adjacent deep ocean basins of the Pacific overwhelmed and destroyed through tectonic and the South Fiji Basin (Karig, 1970, fig. fragmentation at its leading edge opposite the 10). It is concluded that the whole region Pacific oceanic plate. experienced crustal accretion from Eocene to In the Lau Basin, the locally observed Miocene times that has resulted in the build- occurrence of a sedimentary series uncon- up of a quasi-continental crust; through local formably below the Late Miocene and the collapse in the Lau Basin area from the Late implication of a pre-Miocene (perhaps as old Miocene and mainly Plio-Pleistocene as Eocene) basement indicate that the older onwards, through subsidence and structural basin history is comparable to that of the disrupturedue to extensional forces, and the adjacent ridges; thus the Lau and Tonga intrusion of basaltic magmas originating from Ridges have always been a considerable partial melting because of the upwelling of distance apart. Although more recently some hot mantle material, the crust has here extension has occurred in the Lau Basin, become of a mixed type and appears partly to probably in response to foundering between be reverting to an oceanic crust. upfaulted and outward-moving ridges which In this light, there is no support for the has resulted in a widening gap with partly idea that the Lau and Tonga Ridges would new oceanic crustal material being intruded, once have been together. Nor is there there was originally a continuous area of evidence for an eastward shifting of a seismic a probably solidly continental, or quasi- zone which originally would have been at the continental block; available data suggest that Lau Ridge, and later at what then became the it has embraced the whole region between Tonga Ridge. Volcanism on the Lau Ridge is New Zealand and the Kermadec-Tonga and not exclusively andesitic (commonly Colville-Lau ridges, and including Fiji. It is a postulated as the type of volcanism region of near-continental proportions developing above subduction zones) but is underlain by a continental-type-crust which mainly basaltic particularly in its later stages, has formed only in comparatively late while the geological history of the. Tonga geologic time, i.e. from the Early Tertiary Ridge dates at least as far back as that of the onwards, but which in places and before
164