The ocean ridges
Were the world’s oceans to be removed, a great system of mountain ranges, exceeding even the Himalayas in scale, would be revealed (Fig. 13.1). This mostly submerged system, extending for a total length of 65,000km throughout the oceans, was largely unknown until the introduction of electronic echo-sounding in 1923. Prior to this development, only sporadic soundings by conventional lead-line methods had been available.
Historical investigations
The presence of a ridge in the mid-Atlantic was well known by the late nineteenth century, and had been commented on by some of the early geologists, including Alfred Wegener, who incorporated it into his ideas of continental drift in 1912. The introduction of electronic echo-sounding enabled detailed surveys of the Mid-Atlantic and Carlsberg ridges to be completed by 1933. From the 1950s onwards, a comprehensive programme of oceanic surveying was undertaken by scientists from several oceanographic institutions using more advanced techniques including gravity, seismic, heat-flow and magnetic surveys, together with side-scan sonar imaging, to obtain a detailed picture of the ridge topography. In 1977, Bruce Heezen and Marie Tharp from Lamont-Doherty Observatory in Columbia University published a topographic map of the whole mid-ocean ridge system (Heezen and Tharp, 1977). Subsequent investigations have included visual imaging by towed cameras, manned submersibles and remote-controlled vehicles. Further information came from a series of international ocean-drilling programmes – the Deep-Sea Drilling Project (DSDP, 1968–83) and its successor the Ocean Drilling Program (ODP, 1985–2004), which provided rock samples from the various parts of the ridges.
Figure 13.1 The ocean ridge network from space. Oceanic raised features are shown in paler blue. Note that in addition to the mid-ocean ridges, which are typically broad and often poorly defined, there are many more sharply defined ridges such as the very obvious Ninety-east Ridge and the many island arcs. Ridges: AAR, Atlantic–Antarctic; ChR, Chile; CNR, Cocos–Nazca; CR, Carlsberg; EPR, East Pacific; MAR, Mid-Atlantic; NER, Ninety-east; PAR, Pacific-Antarctic; SEIR, SE Indian; SWIR, SW Indian; Compare with Figure 13.2. © Shutterstock, by McLek, from NASA image.
The role played by the mid-ocean ridges in plate-tectonic theory was described in chapter 3 (see Figs 3.7 and 3.8). To recap briefly, they are the sites where new oceanic crust is generated by the upwelling of mantle-derived melts, and are referred to as divergent (or constructive) plate boundaries, or as ocean-spreading centres. It is important to note that oceanic ridges are also formed along transform faults and submerged volcanic arcs, and many examples of these were described in the previous chapters; only those that are volcanically active spreading centres form part of the mid-ocean ridge network as shown in Figure 13.2.
The controlling mechanism of the spreading ridge is ultimately governed by the heat distribution in the uppermost mantle; warmer material underlying the ridges and cooler material at the trenches produce the downward force of slab pull at the trenches and the extensional force of ridge push at the ridges, and it is these forces that provide the motor that drives the movement of the plates. It is the extensional state of stress within the ridge that allows magma to access the crust and not, as might have been thought, the pressure of the rising magma. The ridge-push force is caused by the outwards gravitational pressure exerted by the excess mass of the ridge topography – the same mechanism that was discussed in the context of on-land mountain ranges such as the Himalayas, where the effects of the rising orogenic edifice are countered by gravitational gliding on the mountain flanks.
Ridges do not, in general, lie above uprising mantle columns, and their topography is not directly related to the mantle convection pattern, but is governed by the stress state in the strong upper crust. Ductile asthenospheric mantle is drawn upwards into an extensional region, and melting occurs because of the consequent release of pressure. The initiation of spreading systems, however, does seem to take place above mantle hotspots or plumes, often creating ‘triple junctions’ from which rifts spread outwards, as in the case of the Red Sea–Gulf of Aden–African Rift described in chapter 3 (see Fig. 3.9). Here the continental crust has first been deformed into a dome above the plume, which subsequently cracks into three rifts under the extensional stress.
As oceanic plates grow and their geometry changes, the initial trend of the ridge axis may become oblique to the spreading direction. In this case, the ridge generally adjusts by forming short transform offsets. Lengthy sections of ridge that are oblique to the spreading direction are uncommon (one example is the Reykjanes ridge south of Iceland, which is discussed below).
Figure 13.2 The ocean ridge network. AAR, American–Antarctic Ridge; CaR, Carlsberg Ridge; CIR, Central Indian Ridge; CNR, Cocos–Nazca Ridge; CR, Chile Rise; EPR, East Pacific Rise; ESR, East Scotia Ridge; GA, Gulf of Aden Rift; GaR, Gakkel Ridge; JFR, Juan de Fuca Ridge; KR, Kolbeinsey Ridge; LSR, Laptev Sea Rift; MAR, Mid-Atlantic Ridge; MR, Mohns Ridge; PAR, Pacific–Antarctic Ridge; RR, Reykjanes Ridge; RS, Red Sea Rift; SEIR, SE Indian Ridge; Sv, Svalbard; SWIR, SW Indian Ridge. Plates: Ar, Arabian; Aus, Australian; Ca, Caribbean; Co, Cocos; Ind, Indian, Na, Nazca, Ph Philippines; Sco, Scotia. The arrows give the direction of motion of each plate relative to the Antarctic Plate (regarded as stationary). After Vine & Hess, 1970.
Topography and structure of spreading ridges
In general terms, the spreading ridges are submarine mountain chains that can be between 1000km and as much as 4000km in width, and are elevated by between 1km and 2km from the surrounding ocean floor (Fig. 13.3). They typically incorporate a central rift valley, up to around 100km in width, along which the vulcanicity and associated seismic activity is concentrated. Individual ridges differ considerably in profile, depending on their spreading rate; faster-spreading ridges display a smoother topography with narrow, sharply-defined central rifts, whereas slower-spreading ridges are characterised by more rugged relief and wider central rifts (Fig. 13.4). The rifts are bounded by normal faults with displacements of several hundred metres. Central rifts are absent from very fast-spreading ridges.
Spreading rates of individual ridges are described as slower in the range 10–55mm/a, medium, 55–100mm/a, and faster over 100mm/a (note that 1mm/a=1km/Ma). It is often convenient to quote the ‘half-spreading’ rate, which is the velocity of an oceanic plate relative to its spreading axis, although the half-spreading rates on each side are not necessarily the same. Spreading rates on most ridges, including the Mid-Atlantic and Indian Ocean ridges, are slow, whereas those of the East Pacific Rise are much faster.
Figure 13.3 The Mid-Atlantic Ridge. Traverse from Eastern USA (Florida) to West Africa. The ridge is just over 1500km wide and occupies about one-third of the width of the Atlantic Ocean here, rising from a depth of about 6km to about 3km at the apex. After Searle, 2015.
Figure 13.4 Structure of ridge crests. A, East Pacific Rise at latitude 3°S; B, Mid-Atlantic Ridge at 37°N. A is fast-spreading (64mm/a) and B slow-spreading (23mm/a), F–F, zone of active fissures; faults in red; zone of recent volcanic eruptions in purple; sedimentary basins in yellow. The fast-spreading ridges have a much smoother topographic profile. After Searle, 2015.
The shape of the ocean ridges, like that of the on-land mountains, is governed ultimately by gravitational forces; because ridges are the sites of upwelling warm (and in part melted) mantle material, they are less dense than the surrounding cooler ocean floor, and the crust therefore expands to form a topographic bulge in order to attain gravitational equilibrium. However, compared with continental mountains, erosion plays a much reduced role in modifying their topography, which is governed essentially by a series of escarpments formed as a result of the extensional faulting, or by volcanic edifices created by escaping lava.
As the new crust created at the spreading centre cools, it moves sideways to accommodate fresh injections of new material at the ridge axis, and ultimately sinks to the general level of the deep-ocean floor. The time taken, and thus the distance achieved, by this process depends on the spreading rate, which explains why faster-spreading ridges are broader and vice versa.
Temperature and heat flow variation
Because the ocean ridges are in a state of gravitational equilibrium, the extra surface mass must be compensated by less dense underlying material. Ridges have a characteristically cuspate topographic profile in which the slope becomes shallower with distance from the ridge axis. The height varies inversely with distance, and thus with the age of the underlying crust. This model conforms with the pattern of changing heat flow (Fig. 13.5), which varies from several hundred mWm-2 (milliwatts per square metre) at the ridge axis to about 50mWm-2 on the flanks, indicating that the ridge topography matches the temperature distribution, which in turn determines the density of the underlying lithosphere.
Another way of illustrating this pattern is by the variation in the thickness of the lithosphere, the base of which is determined essentially by the temperature at which its properties change from brittle to ductile, usually taken to be the 750° isotherm. Thus defined, the lithosphere thickness varies from zero at the ridge axis to 9.5km in 1Ma-old crust, 19km in 4Ma-old crust, and 95km in 100Ma crust. The breadth of the ridge is determined by the cooling rate, which depends in turn on the spreading rate; for fast-spreading ridges like the East Pacific Rise, this can be as much as 4000km, whereas in slow-spreading ridges such as the Mid-Atlantic, the width is less than 1000km for most of its length.
Earthquakes
It was the distribution of earthquakes along the mid-ocean ridges that helped to establish the concept of plates, whose boundaries were defined by the lines along which current tectonic activity appeared to be concentrated (see chapter 3, Fig. 3.1). These earthquakes occur within what is termed the seismogenic zone, within which the crust is susceptible to brittle deformation, and this zone lies at progressively deeper levels as the crust becomes older and cooler. Earthquake foci on the spreading ridges are usually confined to the central axial zones, and are invariably shallow: on fast-spreading ridges, they typically occur at depths of less than 1km, whereas on slow-spreading ridges, they may occur at depths of 7–10km. However in old, cold lithosphere, earthquakes can occur at much greater depths.
Figure 13.5 Variation in heat flow across the Pacific Ocean floor. Red circles are observed data. Note that the heat flow declines steeply from the axis of the East Pacific Rise, then decreases gradually with increasing age across the ocean floor. Heat flow data in microcalories per cm² per sec. After Uyeda, 1978.
Earthquakes also occur along, and help to define, ridge offsets; these are the oblique or transverse sections, some of which correspond with transform faults that link adjoining segments of spreading ridge. Most earthquakes appear to be generated by extensional faulting in a tectonically active zone within a few tens of kilometres of the axis of spreading ridges, or by strike-slip displacements on the transform faults. Some transform fault zones have experienced compressional or extensional movements due to subsequent changes in plate geometry, and these produce uplifted topographic ridges or depressed trenches, respectively.
Vulcanicity
Active vulcanicity is concentrated in the axial region of the mid-ocean ridges (Fig. 13.3); in slow-spreading ridges this takes place in the central rift valley, often forming an elongate ridge a few kilometres wide above the currently active rift that marks the plate boundary. Eruptions on fast-spreading ridges typically create rather shallower ridges a few kilometres in height, which extend along the whole length of the actively spreading sector. Axial ridges generally exhibit a narrow valley along their crest, a few kilometres wide and several hundred metres high, situated above the active fissure; this is caused by gravitational collapse when the lava supply becomes exhausted.
About 20 eruptions occur somewhere along the ridge network every year – more frequently on the fast-spreading ridges. Individual lava flows vary from pillow, tube or ropy forms similar to those seen in near-shore underwater eruptions, to sheet-like shapes, depending on the local topography. Pillow shapes predominate on the slower-spreading ridges with steeper slopes, whereas sheet flows are more typical of the faster-spreading examples with higher flow rates.
Mid-ocean ridge magmas are invariably basaltic, produced by the melting of dry lherzolitic peridotite from the upper mantle, and either extruded as lavas or intruded into the crust as gabbroic sheets. Hydrothermal activity is common, particularly on fast-spreading ridges, and results in the formation of pipe-like vents that form cones or pillars rising from the sea floor. These are termed ‘black smokers’, and are a rich source of ore mineralisation because of the interaction between hot basalt and saline hydrothermal fluid. Magmas erupted from the volcanic islands – the ‘hotspot magmas’ – are typically more alkaline in composition, thought to indicate their derivation from a more enriched, higher-temperature mantle source.
There is insignificant sedimentary cover anywhere near the ridge axis owing to the slow rate of deposition in the deep-ocean environment: only fine siliceous or calcareous ooze mixed with small quantities of very fine clay is deposited there at rates of only a few millimetres per million years. However, significant accumulations of sediment may occur on the flanks, particularly of slow-spreading ridges, where the crust is over 10Ma old, as seen in Figure 13.4.
The mid-ocean ridge network
The active mid-ocean ridge network spans all the world’s oceans (see Figs 13.1, 13.2). The longest, and one of the best-studied, is the Mid-Atlantic Ridge. Major ridges also traverse the Pacific, Antarctic and Indian Oceans; the East Pacific Rise lies on the eastern side of the Pacific and continues south-westwards as the Pacific–Antarctic Ridge, which separates the Pacific Plate from Antarctica. Four separate ridge sectors are recognised in the Indian Ocean: the SW and SE Indian Ridges, the Central Indian Ridge and the Carlsberg Ridge. The short American–Antarctic Ridge separates the American Plate from the Antarctic Plate.
Short stretches of ridge occur on the margins of the smaller oceanic plates: the Scotia Ridge on the Scotia Plate; the Gulf of Aden Ridge (or Rift) along the Gulf of Aden, forming the south-western boundary of the Arabian Plate; the Chile Rise, between the Nazca and Antarctic plates (see Fig. 12.7); the Cocos–Nazca Ridge between the Cocos and Nazca Plates; and the Juan de Fuca Ridge between the Pacific and Juan de Fuca plates (see also Fig. 11.2). The northern part of the Mid-Atlantic Ridge itself is subdivided into a further four separate segments: the Reykjanes, Kolbeinsey, Mohns and Gakkel ridges. Some of the best-studied stretches of ridge are on the Mid-Atlantic, East Pacific, Nazca and Carlsberg ridges. The only place where the structures of the ocean ridge can be studied in detail on land is the island of Iceland, situated on the northern part of the Mid-Atlantic Ridge, where its unusual elevation is ascribed to a mantle hotspot located beneath the ridge.
The Atlantic
The Mid-Atlantic Ridge (MAR)
This is the longest of the mid-ocean ridges, extending for about 16,000km from its junction with the SW Indian Ridge in the Southern Ocean to the Siberian shores of the Arctic Ocean (Figs. 13.2, 13.6). It was the first of the mid-ocean ridges to be discovered, in the eighteenth century, and the first to be surveyed in any detail, in the 1970s, by manned submersible, side-scan sonar and other techniques.
The MAR is the ‘type example’ of the slow-spreading ridge, with a spreading rate varying from 41mm/a in the southern Atlantic to 18mm/a south of Iceland, and even lower in the Arctic. The topographic profile of a rugged ridge summit with a narrow central rift valley as shown in Figure 13.4B, which was established in these early studies, has been found to be typical of other parts of the MAR as well as of other slow-spreading ridges. The scale of these topographic features is considerable: the ridge as a whole is elevated by about 3000m from the surrounding ocean floor, and the central rift valley, which can be as much as 100km wide, is depressed by over 2000m from the flanking mountains. There is usually a ridge, up to 1km wide and over 100m high, in the centre of the rift, formed by the most recent volcanic eruption. Vertical extensional fractures occur on the floor of the rift valley, whereas normal faulting is characteristic of the flanking mountains (Fig. 13.4).
Figure 13.6 The Mid-Atlantic Ridge. Active ridges: KR, Kolbeinsey Ridge; RR, Reykjanes Ridge; MR, Mohns Ridge; SWIR, SW Indian Ridge. Hotspots: AH, Ascension; AzH, Azores; BH, Bouvet; CH, Canaries; CVH, Cape Verde; IH, Iceland; TCH, Tristan da Cunha. Inactive ridges and oceanic plateaux: AR, Aguilhas Ridge; IFR, Iceland–Faeroes Ridge; IP, Iceland Plateau; RFZ, Romanche Fracture Zone; RGR, Rio Grande Rise; RP, Rockall Plateau; WR, Walvis Ridge. Islands: GI, Gough; JM, Jan Mayan; SH, St Helena; TC, Tristan da Cunha. Other features: AzTF, Azores Transform Fault; BTJ, Bouvet Triple Junction; GTF, Gibraltar Transform Fault; SA, Scotia Arc. For the northern end of the Mid-Atlantic Ridge, see Figure 13.2. Based on topographic ocean-floor maps by National Oceanographic and Atmospheric Administration (USA).
The half-spreading rate on the western side of the ridge is less than that on the eastern side, at least along the central part of the MAR; that is, the American Plate is moving away from the ridge axis more slowly than either the African or the Eurasian Plates. Looked at another way, relative to the African continent, the ridge axis is moving west at the eastern half-spreading rate while the American continent is moving west faster, at the full spreading rate. The ridge is also moving west relative to both the Azores and Iceland hotspots, which are regarded as being in a (relatively) fixed global frame on timespans of at least 100Ma.
The South Atlantic sector
The southern sector of the MAR (Fig. 13.6) begins at the Bouvet Triple Junction, at 54°S, and keeps to the mid-line of the South Atlantic Ocean, accommodating to the double bend between North Africa and the Caribbean by means of numerous transform faults that offset the ridge, first in a sinistral, then in a dextral sense, while retaining its trend at right angles to the spreading direction, which varies from close to east–west near the Equator to WNW–ESE in the North Atlantic. The position of the major transform faults is inherited from the original shape of the break-up line when the continents split apart in the Mesozoic. Thus, for example, the Romanche Fracture Zone, which offsets the MAR sinistrally for a distance of 900km, helps to accommodate to the 90° bend in the Brazilian and West African coastlines.
Five islands lie on or near the ridge between Bouvet Island and the Azores: Bouvet Island itself at 54°S, Gough Island and Tristan da Cunha at around 40°S, St Helena at 16°S, and Ascension at 8°S. The Bouvet Triple Junction lies close to the Bouvet Hotspot. Gough and Tristan da Cunha are part of the same submerged plateau attributed to the Tristan da Cunha Hotspot, which occurs at the intersection of the MAR with the Walvis Ridge, a 3000km-long, rather sinuous feature striking north-eastwards towards the coast of Africa. This feature is thought to result from a deep-mantle plume, which was one of the foci of the line of break-up of Africa and South America in the Cretaceous. As the South Atlantic opened up, two hotspot trails were left, linking the present position of the hotspot to the original sites on the African and South American coasts. A similar ridge on the South American side, the Rio Grande Rise, strikes north-westwards towards Brazil. St Helena is a volcanic island which last erupted 7–10Ma ago and is no longer active. It lay on the ridge at that time but is now offset to the east. Ascension Island is actively volcanic and attributed to the Ascension Hotspot, which is offset about 80km to the east of the ridge axis.
These topographic swellings on or near the ridge are all attributed to mantle hotspots, which have provided enhanced supplies of magma locally into the ridge. The ridge elevation decreases away from these hotspot sites as the magma supply becomes less readily available.
The Cape Verde Islands and the Canaries
The Cape Verde Archipelago, at 17°N, 25°W, and the Canary Islands, at 28°N, 16°W both lie on elevated submarine plateaux and are attributed to hotspots that are located on the African Plate close to the African continental margin. Both groups still contain active volcanoes, but their vulcanicity appears to date back only to the Miocene. Neither has any current connection with the MAR, and both are thought to have remained approximately beneath their present position as the African Plate moved slowly eastwards across them.
The Azores Hotspot
At a latitude of around 39°N, there is a triple junction between the Mid-Atlantic Ridge and the Azores Transform Fault. The nine islands of the Azores lie on a submerged oceanic plateau formed by the Azores Hotspot, which has been volcanically active for the last 7Ma. The volcanic centre currently lies 100–200km east of the present position of the MAR, but it is likely that the ridge lay above the hotspot when it was formed, and subsequently migrated westwards away from it. The ridge itself is appreciably wider at this point. The Azores Transform Fault runs east from the Azores Plateau to enter the Mediterranean at Gibraltar, where it becomes the Gibraltar Transform Fault (see chapter 4 and Fig. 4.6).
The North Atlantic sector
North of the Azores Triple Junction, between latitudes 40°N and 50°N, there is a relatively short stretch of the Mid-Atlantic Ridge that trends NNE–SSW then, at a point halfway between southern Labrador and Ireland, veers off to the west, and is crossed by a number of “WNW–ESE transform faults. The ridge then follows a relatively straight path in a northeasterly direction to Iceland. This part of the MAR is known as the Reykjanes Ridge, and is notable because it is one of the few places on the network where the ridge is oblique to the spreading direction (which is still WNW–ESE).
The island of Iceland lies directly on the ridge and is the only place on the entire MAR where the geology of the ridge can be conveniently examined above ground. Because of its importance, it is discussed separately below.
Iceland to Svalbard
The section of the MAR lying between Iceland and the Svalbard Archipelago is divided into two; the southern part, known as the Kolbeinsey Ridge, continues the northeasterly trend of the Reykjanes Ridge, following a relatively straight path between Greenland and Norway until it reaches a point just north of latitude 70°N, near Jan Mayen Island, where it is displaced eastwards along a major transform fault. Jan Mayen itself lies on a fragment of continental crust left behind when Greenland and Norway separated during the Eocene. It hosts a currently active volcano, Beerenberg, which lies directly above the Mid-Atlantic spreading axis and last erupted in 1970 (Fig. 13.7).
The ridge is displaced eastwards on the transform fault for about 200km, whereupon it continues northeastwards as the Mohns Ridge, until it reaches a point south of Svalbard, where it is again displaced, this time to the west, for over 1000km, and continues across the Arctic Ocean, where it is known as the Gakkel Ridge (see Fig. 13.2).
Figure 13.7 Beerenberg, 2277m, on Jan Mayen Island: the world’s most northerly active volcano, last erupted in 1970. Shutterstock©P. Fabian.
This submerged ridge follows what is known topographically as the ‘Nansen Cordillera’ across the Arctic Ocean to the Laptev Sea in Siberia. Despite being obscured by thick permanent sea ice, evidence of ongoing volcanic activity along the ridge was obtained in 1999 from a nuclear submarine and has since been confirmed by scientists operating from ice-breaker vessels. The spreading rate along this ridge is very low, less than 10mm/a; it is classed as an ‘ultra-slow’ ridge, and it joins a convergent plate boundary that crosses eastern Siberia to form the eastern margin of the Eurasian Plate. The Siberian end of the ridge is termed the Laptev Sea Rift.
Iceland
The island of Iceland offers the ideal opportunity of examining the structures of a typical spreading ridge above sea level. The spreading axis (Fig. 13.8) comes on-shore from the Reykjanes Ridge and runs along the Reykjanes Peninsula, where it is represented by a narrow median rift trough, crossed by a pedestrian bridge with a sign indicating that the American Plate lies on one side and the Eurasian Plate on the other! The ground here is crossed by numerous deep fissures parallel to the spreading axis (Fig. 13.9A). Much of the ground surface here consists of lavas erupted so recently that erosion or vegetation has had little chance to disturb the original volcanic features; examples of ropy and blocky lava illustrate the typical morphology of MOR-type basalts (Fig. 13.9B, C).
Active volcanism is widespread on the island, and provides both geothermal power and district heating for most of the population. The word ‘geysir’ comes from the famous hot spring of that name near Reykjavik (Fig. 13.10), and volcanic eruptions are common, one of the most spectacular being the Eyjafjallajökull eruption in 2010, whose ash cloud grounded much of Europe’s air transport for days.
Figure 13.8 Iceland. Distribution of present-day volcanic activity: the whole island consists of volcanic material formed over the last 28Ma. The spreading axis approaching the island from the south, on the Reykjanes Ridge, is offset along a transform fault towards the hot spot, with its centre beneath the Vatnajökull glacier, west of Snaefell Mountain. In the north of the Island, the spreading axis is offset again along a second transform fault to continue north along the Kolbeinsey Ridge. The current volcanic activity on the island is concentrated above the hotspot and along the Reykjanes and Snaefelsnes Peninsulas. After Saemundsson, 1974; and Foulger & Anderson, 2005.
Figure 13.9 Morphology of Iceland volcanics. A. Open fissure in the axial rift zone. B. Blocky lava in the foreground with a cracked lava dome behind. C. Ropy lava.
Figure 13.10 Geyser. Steam plume from the hydrothermal vent at the type locality of Geysir, 40km NE of Reykjavik, SW Iceland.
The island of Iceland has long been thought to lie on a major hotspot, which has resulted in a concentration of volcanism that has enabled this section of the Mid-Atlantic Ridge to be elevated above sea level. Vulcanicity on the island has been continuous since at least 25Ma ago during the late Oligocene, and the currently active rift zone runs through the centre of the island, following a rather complex pattern (Fig. 13.8). This active rift, coinciding with a zone of high heat flow containing many active hydrothermal vents, extends from the Reykjanes Peninsula north-eastwards for 200km towards the centre of the island, where it is offset by 100km along a WNW–ESE transform fault zone towards the southeast. It then continues north-eastwards for a short distance to a point southwest of the mountain of Snaefell, where it turns northwards to continue to the north coast at the Raufarhöfn Peninsula. There, another transform fault, the Tjörnes Fracture Zone, offsets the spreading ridge 100km westwards to connect with the Kolbeinsey Ridge, which continues north-eastwards towards Svalbard. The main focus of volcanic activity, i.e. the ‘hotspot’, appears to lie beneath the Vatnajökull icecap and rifts seem to have propagated from here both south and northwards beyond the transform faults. Earthquakes associated with the rift zones are consistent with NW–SE extension at right angles to the main rift trend.
The axial rift in the north of the island is about 70km wide and contains flood basalts erupted over the last several hundred thousand years. The currently active belt contains N–S-oriented fissure swarms parallel to the rift margins, along with several individual volcanoes and calderas. The rift is situated within a wide zone where the older lavas have been uplifted into a broad ridge whose flanks are tilted towards the rift.
The island of Iceland sits on a large region of thickened crust that extends from Greenland in the west to the Faeroe Islands in the east (see Fig. 13.6). Beneath central Iceland, the crust is up to 40km thick and is thought to be entirely oceanic in nature. The original hypothesis attributing the enhanced vulcanicity of Iceland to a deep-seated mantle plume fails to explain some of its geochemical and structural characteristics, and the hotspot has more recently been attributed to a more shallow-rooted accumulation of hot mantle material, unconnected with the deeper mantle convective system.
The Indian Ocean
Three separate but linked ridges form the Indian Ocean ridge network: the Southwest Indian (SWIR), Southeast Indian (SEIR) and Carlsberg/Central Indian Ridges (Fig. 13.11). The SWIR and SEIR form the main part of the circum-Antarctic system, the other components being the American–Antarctic and Pacific–Antarctic ridges, discussed below. The Carlsberg Ridge is the best known of the Indian Ocean ridges, and was the first to be investigated in any detail.
The Carlsberg–Central Indian Ridge
This ridge was discovered by the Dana Expedition in 1928–1930 and named after the Carlsberg brewery, the expedition sponsor. Its southern end is the junction with the SW Indian Ridge at 20°S near Rodriguez Island, east of Mauritius. From there it extends north to 10°N near the entrance to the Gulf of Aden, where it connects with the Gulf of Aden and Red Sea rifts (see also Fig. 3.9). The ridge forms the boundary between the African Plate to the west and the Indian Plate to the east. It originated with the separation of India from Africa in latest Cretaceous time and propagated northwards until by the Eocene it had reached the Gulf of Aden, where it was influenced by the Afar hotspot, which initiated the Red Sea and African rift system.
The Carlsberg Ridge is a typical example of a slow-spreading ridge with an average spreading rate of 26mm/a. It is divided into two ‘first-order’ segments: the more northerly is at right angles to the NE–SW spreading direction, whereas the more southerly one (alternatively known as the Central Indian Ridge) is oriented N–S, at 45° to the spreading direction. Here, the spreading axis is segmented into numerous ‘second-order’ segments, tens of kilometres in length, which are separated by dextral offset zones that displace the axis by a few kilometres laterally. This geometry enables the spreading axis as a whole to maintain an overall N–S trend while individual segments are orthogonal to the spreading direction. Figure 13.12 illustrates typical along-axis and across-axis profiles obtained from the Carlsberg Ridge. Each segment has an axial valley with one or more volcanic ridges that decrease in height towards the bounding offset zones, which are small fault-bounded basins floored by sediments but devoid of volcanic material.
The Carlsberg Ridge is considered to be typical of slow-spreading ridges, in being characterised by short segment lengths and high topographic relief, compared with faster-spreading ridges, which have lower relief and more symmetrical rift flanks. The difference is ascribed to the greater availability of melts in the case of the faster-spreading examples.
Figure 13.11 Ocean ridges of the Indian Ocean. Spreading ridges and rifts: AAR, Atlantic–Antarctic Ridge; CIR, Central Indian Ridge; CR, Carlsberg Ridge; GAR, Gulf of Aden Rift; MAR, Mid-Atlantic Ridge; RSR, Red Sea Rift. Aseismic ridges: BR, Broken Ridge; ChR, Chagos Ridge; NER, Ninety-East Ridge; TR, Tasman Ridge; WR, Walvis Ridge. Hotspots: AH, Amsterdam–St Paul’s; B, Bouvet; Cr, Crozet; McQ, MacQuarie. Triple junctions: BTJ, Bouvet; McQ, Macquarie; RTJ, Rodriguez. Islands: B, Borneo; C, Cocos; Ch, Chagos; K, Kerguelen; M, Mauritius; Mg, Madagascar; Sey, Seychelles. Based on topographic ocean-floor maps by National Oceanographic and Atmospheric Administration (USA).
Figure 13.12 Morphology of the Carlsberg Ridge. A. Topographic profile along the ridge axis, showing the shape of the ridge segments. Note the c. 1000m height difference between the ridge crests and the offset basins. B. Topographic profile at right angles to the ridge axis showing the difference in height of c.500m between the western and eastern flanks. After Murton & Rona, 2015.
The Southwest Indian Ridge (SWIR)
This 7700km-long ridge (Fig. 13.13) is interesting both because of its very slow spreading rate and because it is extending parallel to its length at a faster rate than any of the other ridges. This is due to the fact that the triple points at each end of the ridge are receding from each other, as both the Atlantic and Indian oceans are expanding, and there is no compensating subduction taking place between them. The ridge forms the boundary between the African Plate to the north and the Antarctic Plate to the south, and has a spreading rate varying from 15mm/a (i.e. ultra-slow) to 30mm/a (slow). The ridge was initiated when Africa broke away from Antarctica during the late Cretaceous.
The western end of the ridge is joined to the Mid-Atlantic Ridge and the Atlantic–Antarctic Ridge just west of Bouvet Island, at 54°S, 1°W. The ridge here possesses a deep, wide axial valley, typical morphology of a slow-spreading ridge, and is offset by several NE–SW transform faults. From 10°E to 25°E, the ridge has an overall E–W trend, without transform faults. The western part is characterised by numerous orthogonal segments separated by oblique segments with variable orientations, which are devoid of vulcanicity (Fig. 13.14D). This section has an ultra-slow spreading rate. Between 16° and 25°E, there is a long section of the ridge that is perpendicular to the spreading direction.
From 27°E eastwards, the ridge is offset by numerous transform faults or fracture zones, several of which have very large offsets and are marked by troughs with depths of over 6km. The longest of these is the Andrew Bain Transform Fault with a sinistral displacement of 750km. These transform faults all trend NNE–SSW, parallel to the direction of divergence between the African and Antarctic Plates. Some of the major fracture zones (e.g. the Mozambique Ridge) extend to the continental margin and correspond with sharp breaks in the initial fracturing pattern of Gondwana. Much of this long stretch of the SWIR has quite irregular segmentation. The depth of the axial region here varies from as much as 4730m to 3050m near the Marion Hotspot at 36°E, where there has been an increased magma supply.
Figure 13.13 The SW Indian Ridge. AAR, American–Antarctic Ridge; BTJ, Bouvet Triple Junction; CIR, Central Indian Ridge; DH, Discovery Hotspot; Ma, Madagascar; MAR, Mid-Atlantic Ridge; MP, Madagascar Plateau; MR, Mozambique Ridge; MtR, Meteor Rise; MTF, Melville Transform Fault; RTJ, Rodriguez Triple Junction; SWIR, SW Indian Ridge. Symbols and colours as in Figure 13.6. Based on topographic ocean-floor maps by National Oceanographic and Atmospheric Administration (USA).
Figure 13.14 Types of discontinuity on spreading ridges. After Searle, 2015.
The easternmost section of the ridge, east of the Melville Transform Fault, is characterised by an ultra-slow spreading rate and is devoid of vulcanicity. The ridge flanks here are rounded and smooth. Divergence is achieved by extensional faulting parallel to the ridge axis rather than by magma injection. The ridge joins the Central Indian Ridge at the Rodriguez Triple Junction, at 70°E. The increase in length of the ridge, required by the increasing separation of the two ends, is achieved by the oblique sections of the ridge, which have experienced a component of NE–SW extension.
The Southeast Indian Ridge (SEIR)
This 6000km-long ridge marks the boundary between the Australian and Antarctic plates, and extends from the Rodriguez Triple Junction in the west to the Macquarie Triple Junction in the east (Fig. 13.15). The ridge was initiated in the Oligocene when Australia and Antarctica began to drift apart. The current average spreading rate is about 65mm/a. The ridge lies close to, and has been influenced by, three hotspots: Amsterdam–St Paul, Kerguelen and Balleny.
The Amsterdam–St Paul Hotspot is a large submerged plateau, situated around Longitude 80°E, 150x200km in extent and elevated 1–3km above the surrounding ocean floor. A wide ridge extends northeast from the plateau, containing a string of submarine volcanoes representing the trail of the hotspot across the Indian Ocean towards Broken Plateau.
The Kerguelen Hotspot is located further southwest, at 70°E. The Kerguelen Archipelago lies on an even larger submerged plateau about 1,250,000km2 in extent, situated within the Antarctic Plate about midway between the SEIR and Antarctica. It has been active volcanically since the early Cretaceous, since when a very large volume of igneous material has been accumulated. The hotspot dates back to the separation of Australia from Antarctica and has been split in two by the subsequent growth of the SEIR. The Australian half of the plateau is known as the Broken Plateau, or Broken Ridge, which extends for 1200km from the southern end of the Ninety-East Ridge towards the southwest corner of Australia. The Ninety-East Ridge extends from Broken Ridge north to the Bay of Bengal, and is considered to mark the track of the hotspot as the Indian Plate travelled north across it towards its present position. The combined Kerguelen–Broken Ridge topographic swell makes it one of the world’s largest hotspots, and it is attributed to a class of ‘large igneous provinces’ (LIPs) attributed to deep-mantle plumes.
About midway between the Kerguelen Hotspot and the Macquarie Triple Junction, around 120°E, the ridge narrows, and is characterised by unusually thin crust, rough topography and anomalously low heat flow. The ridge crest here is more than 2000m deeper than in the adjoining sectors. These properties are thought to indicate a region of mantle downwelling.
The detailed morphology of the SEIR is very similar to the other slow-spreading ridges, such as the Carlsberg and SWIR. A series of first-order segments are separated both by transform faults and by non-transform discontinuities, which are rift-like structures that extend the ridge obliquely to the spreading direction (see Fig. 13.14).
Figure 13.15 The SE Indian Ridge. AH, Amsterdam–St Paul’s Hotspot; BH, Balleny Hotspot; BR, Broken Ridge; CIR, Central Indian Ridge; ETP, East Tasman Plateau; KH, Kerguelen Hotspot; KP, Kerguelen Plateau; LHR, Lord Howe Ridge; MA, MacQuarie Arc; MTJ, MacQuarie Triple Junction; NER, Ninety-East Ridge; NR, Norfolk Ridge; NZ, New Zealand; PAR, Pacific–Antarctic Ridge; RTJ, Rodriguez Triple Junction; SEIR, SE Indian Ridge; SWIR, SW Indian Ridge; TR, Tasman Ridge. Symbols and colours as in Figure 13.6. Based on topographic ocean-floor maps by National Oceanographic and Atmospheric Administration (USA).
The easternmost stretch of the ridge is cut by several large transform faults with dextral offsets, which have the effect of changing the overall trend of the ridge from E–W to NW–SE as it approaches the Macquarie Triple Junction.
The Macquarie Triple Junction is the meeting point of the SEIR with the Pacific–Antarctic Ridge to the east and the Macquarie Arc, which forms the western boundary of the Pacific Plate. The triple junction is close to the Balleny Archipelago, which consists of a group of volcanic islands located around 163°E, 67°S. The islands lie on a NW–SE ridge that extends from the SEIR towards the East Tasman Plateau, southeast of Tasmania. They are attributed to the Balleny Hotspot, which is believed to date back to the separation of Australia from Antarctica.
The Pacific
The East Pacific Rise (EPR)
This ridge is the Pacific counterpart of the Mid-Atlantic Ridge, formerly running through the centre of the Pacific Ocean. However, the reconfiguration of the plates around the Pacific that occurred through the Mesozoic and Cenozoic Eras has resulted in the EPR being displaced eastwards relative to the surrounding continents, to the extent that the Pacific Plate, on the western side of the EPR, now occupies the greater part of the Pacific Ocean (Fig. 13.16) and the EPR has been partially over-ridden by the westward movement of the American Plate. The northern end of the EPR joins the Gulf of California Rift Zone, which separates the Baja California Peninsula from the North American Plate (see chapter 11, Fig. 11.7). This Rift links in turn with the southern end of the San Andreas Transform Fault, which forms the eastern boundary of the Pacific Plate further north.
The northern sector of the EPR runs from the Gulf of California southwards to near the Equator, where it is joined by the Cocos–Nazca Ridge. This sector forms the western boundary of the Cocos Plate, which is being subducted north-eastwards along the Mid-America Trench (see Fig. 12.1). From here, the EPR continues southwards to 35°S, where it meets the transform fault that connects with the Chile Rise, at the Easter Island Triple Junction. This sector forms the western boundary of the Nazca Plate, which is being subducted beneath the Peru–Chile Trench (see Fig. 12.7B).
Figure 13.16 The Pacific Ocean ridges. Ridges/rifts (R), transform faults (T) and fracture zones (FZ): CNR, Cocos–Nazca; EFZ, Eltanin; EPR, East Pacific (Rise); GCR, Gulf of California (Rift); GR, Galapagos (Rift); JFR, Juan Fernandez; MFZ, Mendano; MGC, Marshall-Gilbert Chain; NR, Nazca; MNR, Marcus–Necker; PAR, Pacific–Antarctic; QCT, Queen Charlotte; SAF, San Andreas; SEIR, SE Indian; SR, Shackleton; UFZ, Udintsev. Hotspots: EH, Easter; GH, Galapagos; HH, Hawaii; LH, Louisville; MDH, Macdonald; MH, Macquarie. Other tectonic features: CAC, Cook–Austral Chain; CoP, Cocos Plate; CTJ, Chile Triple Junction; ETJ, Easter Triple Junction; JFP, Juan de Fuca Plate; MAT, Middle America Trench; MQA, Macquarie Arc; MTJ, Macquarie Triple Junction; PCT, Peru–Chile Trench. Topographic features: Ai, Aitutaki Atoll; Ar, Arorae Atoll; BC, Baja California; Bi, Bikini Atoll; En, Eniwetok; Ki, Kiritimati (Christmas Island); KP, Kamchatka Peninsula; Ma, Marquesas; Ra, Rangiroa; Sk, Sakhalin; Sol, Solomon Islands; Tas, Tasman Plateau; Van, Vanuatu. Based on topographic ocean-floor maps by National Oceanographic and Atmospheric Administration (USA).
Most of the EPR is very broad, up to 4000km across, much smoother in profile than the Mid-Atlantic Ridge, and lacks a central rift valley (compare Fig. 13.4A with 13.4B). This difference is attributed to its much greater spreading rate; this reaches 150mm/a in the fastest-spreading sector, near Easter Island, but even further north, where it is around 60mm/a, it is appreciably faster than the MAR.
The detailed topography of the ridge has been explored in several places and is well illustrated in Figure 13.17, which represents a section of the ridge between latitudes 12° and 13°N. The summit is characterised by a shallow ridge where the current vulcanicity is concentrated. This axial ridge varies from about 8km wide and 250m high at 13°N to 15km wide and between 300m and 400m high at 20°S, on the superfast section of the ridge. The broad sectors, which have a roughly rectangular profile, are attributed to the presence of a shallow inflated magma chamber, whereas the narrower ridges, with a more triangular profile, appear to overlie regions where the magma chamber is deeper, or deflated.
Figure 13.17 Topography of the East Pacific Rise. Map of the seabed on part of the EPR, looking NNE along the ridge between latitudes 12° and 13°N. Depths range from white 2700m to deep blue 3250m. Note en-echelon spreading segments. Science Photo Library, © Dr Ken Macdonald.
The axial ridge along most of its length possesses a small narrow valley or trough less than 2km wide and about 100m deep. This is thought to indicate the presence of an empty fissure over which the ridge summit has collapsed. In some places the spreading segments are arranged en-echelon (Fig. 13.10E), as shown in Figure 13.17. Not all the vulcanicity is concentrated along the ridge. Occasional circular seamounts have been encountered some distance away from the ridge axis, and several hydrothermal vents have been explored. The morphology of the lava flows is very similar to those of the MAR.
The Galapagos Hotspot
The Galapagos Archipelago is a group of 18 main volcanic islands scattered across the Equator around longitude 90°W, situated on a submarine ridge that extends in an easterly direction towards the coast of Ecuador. The more westerly volcanoes are currently active and the vulcanicity becomes older eastwards, dating back to around 4Ma. The island chain is situated on the Nazca Plate, and is attributed to a hotspot over which the Nazca Plate has tracked, probably since the Miocene. The Galapagos Islands are situated east of the EPR but are linked to it via the Cocos–Nazca Ridge, which forms the boundary between the Cocos and Nazca Plates.
The Easter Island Hotspot
Easter Island consists of three extinct volcanoes forming the summit of a large volcanic mountain rising over 2000m from the ocean floor, and lies at the end of a submarine ridge containing numerous volcanic seamounts. This seamount chain extends for 2700km eastwards, then north-eastwards towards the Chilean coast, where it is currently subducting beneath the Chile Trench. The chain is attributed to the track of, first the Farallon Plate, and more recently the Nazca Plate, as they moved across the Hotspot.
The Pacific–Antarctic Ridge
The southern sector of the EPR, also known as the Pacific–Antarctic Ridge (PAR), runs from Easter Island, first southwards, then gradually bends south-westwards to where it connects with the SE Indian Ridge at the Macquarie Triple Junction (159°W, 55°S), south of New Zealand. This section of ridge forms the boundary between the Pacific Plate to the west and north and the Antarctic plate to the east and south.
At about 140°W, the PAR meets the Louisville Ridge (also known as the Eltanin Fracture Zone), which is a volcanic seamount chain extending for 4300km north-westwards to the Tonga–Kermadec Arc where it is being subducted beneath the Tonga Trench (see chapter 10). This ridge is regarded as the track of the Louisville Hotspot, which is believed now to lie close to the PAR but has left a trail of submerged volcanic seamounts as the Pacific Plate moved across it. Formerly much more vigorous, volcanic activity has decreased since the Oligocene.
Fracture Zones
The Pacific Ocean basin is traversed by a number of aseismic fracture zones. These are expressed topographically in the form of either ridges or rifts, or a combination of these features, and mark the course of formerly active transform faults, which in many cases date back to the Cretaceous Period. The more important of these, shown in Figure 13.16, from north to south, are: the Mendocino, Murray, Molokai, Clarion and Clipperton Fracture Zones, all of which extend westwards from the eastern margin of the Pacific Plate; the Mendanao and Easter Fracture Zones within the Nazca Plate; and the Eltanin–Louisville and Udintsev Fracture Zones, which extend north-westwards from transform faults on the Pacific–Antarctic Ridge. These structures are parallel to the relative movement direction of the plates that they cut. In the case of the North Pacific zones, this direction, which is roughly east–west, tracks the former movement of the Pacific Plate relative to the now-extinct Farallon Plate; the zones in the southern Pacific, which are markedly curved, indicate the current movement path of the Pacific Plate relative to the Antarctic Plate.
The Pacific seamount chains
The vast expanse of the central Pacific Ocean is also crossed by a number of long, relatively straight ridges, some of which have no obvious genetic connection either to the East Pacific Rise or to the volcanic arcs of the western Pacific rim (Fig. 13.16). From north to south, these include the Haiwaii–Emperor Chain, the Marcus–Necker Rise, the Line Islands, the Marshall–Gilbert Island Chain, the Tuamotu Island Chain, and the Cook–Austral Seamount Chain. All of these structures are submerged linear ridges with considerable topographic relief, dotted with small islands and coral atolls. Some of them are clearly hotspot trails, while others have a more enigmatic origin. The best known and most studied example is the Hawaii–Emperor Chain, which has become the type-example of a hotspot chain.
The Hawaii–Emperor Chain
This mostly submerged ridge extends for over 5800km from the Aleutian Trench in the north to the youngest active volcano, the Lo’ihi Seamount, 35km south of the main island of Hawaii. The ridge changes direction abruptly through 120° between the Emperor Seamount Chain and the Hawaiian Ridge. The Emperor seamounts range in age from 85Ma at the northern end to 39Ma at the southern, while the islands on the Hawaiian Ridge are between 28Ma and 7Ma old. The Hawaiian Archipelago itself is still volcanically active; there are five volcanoes on the main island, three of which are still active. The magmas on the main island are tholeiitic rather than alkali-basaltic, and atypical of non–MOR oceanic volcanoes.
The northward increase in age of the volcanoes, coupled with the increasing depth of the eroded seamount tops, led J. Tuzo Wilson in 1963 to propose that the chain represented the trail of a stationary hotspot as the Pacific Plate moved across it – the older, more northerly parts of the ridge becoming gradually deeper as the ridge cooled (Fig. 13.18). The abrupt change in trend was attributed to a change in the direction of travel of the Pacific Plate. Although the model of a stationary plume has been questioned more recently for the Hawaii–Emperor chain, it has acted as a template for explaining many of the other ocean ridges.
Figure 13.18 Formation of a hotspot trail. Schematic diagram to illustrate the creation of a hotspot trail by a plate moving across a fixed hotspot. Not to scale. After Wilson, 1963.
This structure is a 4500km-long ridge that extends from around Longitude 150°E in the west, to just south of the Hawaiian Ridge in the east, following an irregular path. The ridge is broad, up to 1000km in places, and elevated by 300–500m from the ocean floor. It contains many submerged peaks and several small islands, but differs from the other seamount chains in not following a NW–SE linear track, and its origin is unclear.
The Line Islands
These are a chain of atolls situated on volcanic seamounts that form a 4800km-long NW–SE ridge extending from around 20°N to well south of the Equator. They include Kiritimati, formerly known as Christmas Island. Volcanic ages range from mid-Cretaceous to late Eocene, becoming younger southwards. Although a hotspot origin has been proposed for the chain (see below), the age distribution does not appear to match that of a simple hotspot trail.
The Marshall–Gilbert Island chains
These two archipelagos form part of a largely submerged ridge that extends from Eniwetak Atoll, at 11°30’N, 162°20E to Arorae Atoll, at 2°38’S, 176°49’E, at the southern end of the Gilbert Chain. The Marshall Islands are scattered over a large area, whereas the Gilbert Islands form a more linear chain. Both groups of islands are composed of coral atolls situated on volcanic seamounts. Bikini Atoll, near the northern end of the Marshall Islands, achieved notoriety because of the atomic bomb testing carried out there by the USA between 1946 and 1958, which involved the removal and resettlement of the population. Recorded volcanic ages range from late Jurassic in the north to early Cretaceous in the south. The Marshall–Gilbert Ridge is approximately aligned with the Cook–Austral Chain further southeast and may represent the proximal part of the same hotspot track.
The Tuamotu Ridge
The Tuamotu Archipelago consists of nearly 80 islands and atolls situated on a broad submerged ridge aligned NW–SE. The large island of Rangiroa, at 15°07’S, 147°38’W is near the northern end of the chain. The ridge is aseismic and there are no recorded volcanic eruptions. The ridge is aligned with the Line Islands further northwest and both groups have been attributed to the Easter Hotspot, which lies on the East Pacific Rise further southeast.
The Cook–Austral Seamount Chain
This chain, consisting of 13 islands and atolls, extends for 2200km south-eastwards from the island of Aitutaki in the Cook Islands to the Macdonald Seamount, which is a currently active volcano located at 140°W, 30°S. Like the other NW–SE seamount chains, it is aligned parallel to the direction of Pacific Plate motion and is attributed to the Macdonald Hotspot.
