island arcs

island arcs Island arcs have long been recognized as tectonically active belts of intense seismic activity containing a chain or arc of active volcanoes. At the turn of the nineteenth century, W. J. Sollas drew attention to the correspondence of the arc-like form of the Aleutians/Alaskan Peninsula, the East Indies (Indonesia), and several mountain chains to a series of great circles, and C. Lapworth discussed the ‘Volcanic Girdle of the Pacific’ (better known as the Pacific ‘Ring of Fire’) as a continuous ‘septum’ separating ‘plates’ with different histories and thicknesses. The deepest parts of the oceans, the elongate deep-sea trenches, were located on the oceanward side of these arcs. As the nature of the ring of fire was examined, it was realized that a line, called the andesite line could be drawn around the Pacific outside which andesites occurred (named after their type area in the Andes) and inside which basalts predominated.

Little could, however, be done to elucidate the origin of island arcs until geophysical data were acquired. The first, in the 1930s, to tell us something about the nature of the internal structure of the oceans was the evidence obtained by the Dutch geophysicist F. A. Vening Meinesz, who showed that an extensive belt of large negative isostatic gravity anomalies ran parallel to the island arc. A surprising feature of this discovery was that the centre of this belt ran to the island side of the trench axis, not under the trench itself. it is therefore not related to the present topography but to some deeper root; and, since all surface evidence in trenches and island arcs shows that they are under tension, it was considered as proof that there must be some downward stress that was dragging down a mass of material of relatively low density. It gave rise to the concept of the tectogene, which involved the vertical downwarping of oceanic crust, either by compressional forces or the drag of converging convection currents. It was not, however, until 1949, when H. Benioff showed that earthquake epicentres became progressively deeper as one went from the ocean side of the trench to the volcanic arc, that the idea of a relatively simple steeply dipping thrust plane extending from near the trench to a depth of as much as 700 km was clearly established.

Thus, by the 1950s substantial geophysical data had been acquired around the Pacific, off Indonesia, and in the Caribbean to suggest that large slabs might be dragged down beneath island arcs along subduction zones, which are commonly known as Benioff zones.

It was nevertheless not until 1968 that the next significant advance was made. The establishment of the hypothesis of ocean-floor spreading in the 1960s had shown that new lithosphere was being continuously created (see mid-ocean ridges), and it was realized that unless the Earth was expanding (and at that time some geoscientists seriously argued in favour of this) an equal amount of lithosphere must be lost, and that this must happen at the subduction zones. As the slab of oceanic lithosphere goes down, it melts partially at about 150–200 km depth, giving birth to magmas that rise and are extruded in volcanoes located 150–200 km from the axis of the trench.

A subduction zone is identified by seismic foci, the seismic activity being concentrated on the upper surface of the downgoing slab of lithosphere. The seismic activity defines the ‘seismic plane’ of the subduction zone, which may be up to 20–30 km wide. Subduction zones dip mostly at angles between 30° and 70°, but individual subduction zones dip more steeply with depth. The dip of the slab is related inversely to the velocity of convergence at the trench, and is a function of the time since the initiation of subduction. Because the downgoing slab of lithosphere is heavier than the plastic asthenosphere below, it tends to sink passively; and the older the lithosphere, the steeper the dip.

It is important to remember that the term ‘island arc’ is commonly used as a synonym for ‘volcanic arc’, yet the two terms are not quite the same. Volcanic arcs include all volcanically active belts located above a subduction zone, whether they are situated as islands in the middle of oceans or on continents, as along the west coasts of Central and South America. Island arcs include only those separated from the land by a stretch of water.

There is therefore a spectrum of island-arc types. Some are truly intraoceanic, being situated entirely within the oceans (Marianas, New Hebrides, Solomons, and Tonga in the Pacific; the Antilles and Scotia arcs in the Atlantic). Others are separated from major continents by small ocean or marginal basins with a crust that is intermediate between continental and oceanic (Andaman islands, Banda, Japan, Kuril, and Sulawesi). The Aleutian arc passes laterally into a continental ‘Cordilleran-type’ fold belt on the North American continent. At the extreme end of the spectrum are those arcs built against continental crust, such as the Burmese and Sumatra/Java portions of the Burmese–Andaman–Indonesian arc, and finally the Andean chain, where the volcanic belt is located entirely within the continent and is not therefore an island arc. The age also varies. Some are very young: less than 10Ma (10 × 10⁶ years). Others are much older, with histories that go back at least to the Tertiary, if not to the Cretaceous.

The exposed magmatic island arc is only one of a number of features or tectonic zones that extend from the trench at the oceanward end to the marginal or back-arc basin on the continental side (Fig. 1).

Trenches are the deepest features of ocean basins, with depths ranging from 7000 to almost 11 000 m, the deepest being the Mariana and Tonga trenches. Most deep-sea trenches in the Pacific are floored by normal basaltic oceanic crust overlain by pelagic sediments and ash. This relatively thin sedimentary layer is easily subducted under the overriding plate. This simple subduction pattern may, however, be disturbed because, in some instances the trench may receive a substantial amount of terrigenous sediment derived from a continent. Although this feature is most important in the fill of deep-sea trenches immediately adjacent to continents (e.g. the Andean-type subduction zones of Central and South America), it may also be important where the trench is fed from one end, as off the Andaman islands. In addition, enormous volumes of sediment accumulate on ocean floor as deep-sea fans. The largest of these, the Bengal Fan, has a volume of 4 million km³, and this is being swept eastwards towards the Burmese–Andaman–Indonesian arc. The other complication is due to the fact that ocean floors are not simple, smooth features of a constant 5000–6000 m depth, but are composed of a variety of topographic features, such as oceanic plateaux, ridges, sea mounts, and guyots (flat-topped submarine mountains) that may rise thousands of metres above the ocean floor and have areas that range from a few square kilometres to thousands of square kilometres. When these irregularities reach the deep-sea trench and come up against the overriding plate they may either be subducted and incorporated into the downgoing slab or scraped off the oceanic plate and incorporated in the overlying wedge.

In this way an accretionary prism or wedge develops that has a ridge at the surface, known as the outer arc ridge or trench-slope break, behind which is the fore-arc basin. This prism comprises a thrust stack of scraped-off sediments from the trench, which forms by tectonic accretion. The sediments can be examined today in the Mentawai islands west of the main Indonesian islands and in Barbados, a low-lying non-volcanic island east of the Lesser Antillean volcanic chain. An important feature of the accretionary prism is that while each sedimentary succession gets younger upwards, there is a decrease in depositional and metamorphic age as one goes from the earliest-formed wedge on the volcanic arc side towards the toe on the ocean side, as successively younger slices of oceanic sediments and floor are scraped off. It is in fact this successive ‘underplating’ that progressively raises the older slices and causes the elevation of the fore-arc region. A striking feature of some accretionary prisms, recognized in Barbados for many years, has been the abundance of mud diapirs (see mud and sand volcanoes) formed by high-pressure fluids within the unconsolidated hemipelagic muds and biogenically generated methane that are carried down beneath the accretionary prism so that they become overpressured and rise through the prism. The surface of the accretionary prism is not a simple one. Because it was formed as a result of the successive accretion of individual slabs of crust, the surface displays a rugged and irregular sea-floor morphology governed by numerous tectonic ridges that form by folding and dislocation. As each successive thrust fault propagates oceanward through the wedge of uplifted trench sediments, a deformation front will separate the undisturbed trench environment from the disrupted accretionary prism, and this structural boundary migrates seaward with time. Simultaneously, however, older, higher thrusts may be reactivated. Thus this constantly reactivated complex and prism has innumerable small sedimentary basins known as trench-slope basins.

Between the volcanic arc and the accretionary prism is the fore-arc basin, which may be as much as 100 km wide. Typically such basins are filled by immature volcaniclastic sediments derived from the erosion of the volcanic arc. Sedimentation in fore-arc basins can, however, be quite complex and almost any type of sediment can be deposited. For example, in the very extensive fore-arc basin west of the Burma–Sumatra–Java volcanic arc and east of the Andaman– Nicobar–Mentawai islands outer arc, sediments range from fluvial–deltaic in the north, fed by the Burmese rivers, to deep-water turbidites and shelf sediments further south.

The volcanic suites of island arcs vary according to the thickness and composition of the overriding lithosphere. At true intraoceanic islands, mantle-derived magmas are relatively unimpeded in their ascent, resulting in eruption of very fluid, often aphyric (microcrystalline and without phenocrysts) tholeiitic basalts and basaltic andesites. As the arc develops, the thickening arc massif depresses the oceanic crustal layer. Once the crust has thickened to some 20 km, it may start to act as a density filter, and primary magma may become ponded in a series of interconnected high-level magma chambers. The upward ascent of magma, particularly under the centre of the arc, is a slow and fitful affair, and differentiation processes generate calc-alkaline magmas, which contain relatively large amounts of calcium (CaO) in relation to alkalis (Na₂O and K₂O), and more intermediate magmas such as andesite. In the Japanese arc it appears that erupted magmas become more alkaline away from the trench, apparently reflecting the increasing depth to the source of the magmas along the Benioff zone. Not all arcs follow this simple pattern, however, probably because of tectonic inhomogeneities in the overriding plate, variable depths of partial melting, and the type of lithosphere subducted.

Oceanic volcanic arcs are surrounded by large volcaniclastic aprons, kilometres thick, whose volume may far exceed that of the volcanoes. Although the foundation of the arc may be dominated by lavas, most of the apron consists of pyroclastic fragments generated by explosive volcanic activity in shallow water or on land, or reworked volcaniclastic rocks. As the submarine slopes of arc-related volcanoes are steep, there is great seismic activity and sedimentation is rapid resedimentation by slumping, sliding, and turbidity currents is common. Unfortunately, the sediments immediately adjacent to island arcs are not easily penetrated by deep-sea drilling, and we mostly know about such sediments from their uplifted equivalents on land, either in modern island arcs such as the New Hebrides and Japan or in their ancient equivalents such as the Ordovician of North Wales and in the English Lake District.

Some of the more interesting results that have come from the study of both modern (e.g. the Lesser Antilles, New Hebrides) and ancient examples have shown that:(1) the distribution of sediments around an arc is usually asymmetric, owing to the prevailing wind patterns, different arc slopes, and ocean currents (as in the Lesser Antilles where westerly winds blow most of the ash into the Atlantic);(2) as the island arc develops, older and more mafic volcaniclastics, including scoria, hyaloclastites (formed when hot lava comes into contact with water or wet sediment), and shards are succeeded by products of more differentiated magmas, including abundant pyroclastic material; dormancy of the island arc can result in blanketing of the volcaniclastic apron with silty epiclastic (redeposited) turbidites;(3) the position of the island or individual volcanoes can migrate, often oceanward towards the trench or along the arc;(4) as the volcano grows into shallow water and emerges, eruptions become more explosive, with ash dispersed over greater distances from the volcano;(5) sedimentary processes continually sort volcaniclastic fragments by grain size and density into a proximal coarse-grained facies of pillow breccias, debris-avalanche, and lahar (mudflow) deposits; a medial debris-flow facies; and a distal facies consisting of thin distal turbidites and fallout ashes; and(6) the rates of arc volcanism and back-arc spreading can vary with periods of continuous arc volcanism and reduced spreading favouring progradation of the volcaniclastic apron.As island arcs develop, enlarge, and become more mature, as in Japan and the North Island of New Zealand, terrestrial sediments and plants abound, and lagoons and lakes develop, especially within the caldera of the volcanoes. Uplift may reveal the plutonic core of the arc, to expose calc-alkaline, predominantly dioritic to granodioritic plutons. Because island arcs are zones of overall extension, rift valleys develop, some with a component of strike-slip, as in Japan or the Taupo rift of the North Island of New Zealand.

Although the type of mineralization found within an island arc varies according to its age, the characteristic class of ores are syngenetic massive sulphides, the so-called Kuroko ores named after their type locality in Japan. These contain pyritic zinc–lead–copper as well as silver and gold. They are closely associated with marine pyroclastic rhyolitic domes and calderas and were probably deposited from submarine brines in or on the flanks of local sedimentary basins.

A striking feature of the western Pacific Ocean is the enormous area covered by a large and complex pattern of basins that lie behind the volcanic arcs and are marginal to the continent. These marginal basins have been a source of controversy ever since it was realized that their crusts, while usually having thicknesses close to that of continental crust, have seismic velocities closer to those of oceanic crust. Most marginal basins are now known to be old ocean floor trapped behind an island arc and are recognized not only in the western Pacific but also in the Andaman sea behind the Burmese– Indonesian volcanic arc, and behind the Antillean and Scotia arcs. They range in age from very young back-arc basins that have developed within oceanic crust relatively recently (intraoceanic back-arc basins) to those mature basins adjacent to continents, such as the Japan Sea, which is inactive at present (continental back-arc basins).

We now know that these oceanic back-arc basins are formed by crustal extension producing first rifts and then new ocean crust by sea-floor spreading, similar to that which forms mid-ocean ridges (see mid-ocean ridges). The extensional origin of these basins is indicated by the presence of normal faults, high heat flow, and, in some instances, magnetic linear anomalies about a central rift. Back-arc spreading occurs particularly when older, colder lithosphere is being subducted and where the subduction zone dips steeply. Although there is no consensus as to the precise mechanism for the spreading, it is probably a consequence of several mechanisms, such as migration oceanward of the trench and volcanic arc over the subducted plate, or magma intrusion and upwelling due either to thermal upwelling of a magma diapir or to heat generated by a plume or by the subducted slab.

Harold G. Reading

Bibliography

Busby, C. J. and and Ingersoll, C. V. (1995) Tectonics of sedimentary basins. Blackwell Science, Cambridge, Massachusetts.
Fischer, R. V. and Smith, G. A. (eds) (1991) Sedimentation in volcanic settings. Special Publication No. 45, Society of Economic Paleontologists and Mineralogists, Tulsa.
Orton, G. J. (1996) Volcanic environments. In Reading, H. G. (ed) Sedimentary environments: processes, facies and stratigraphy, pp. 485–567. Blackwell Scientific Publications, Oxford.