Structure Of Subduction Zones From Earthquakes

Subduction zones exhibit intense seismic activity. A large number of events occur on a plane that dips on average at an angle of about 45° away from the underthrusting oceanic plate (Fig. 9.5). The plane is known as a Benioff (or Benioff-Wadati) zone, after

South North

South North

Subduction Isotherms
Figure 9.4 Gravity anomalies of an oceanic subduction zone (after Grow, 1973, with permission from the Geological Society of America).
Gravity Anomalies Subduction Zones

Figure 9.5 Vertical section perpendicular to the Tonga arc showing earthquake foci during 1965. Circles, foci projected from within 0-150 km north of the section; triangles, from 0-150 km south. Exaggerated topography (13: 1) above. Inset, enlargement of the region of deep earthquakes (redrawn from Isacks et al., 1969, with permission from the Geological Society of America).

Figure 9.5 Vertical section perpendicular to the Tonga arc showing earthquake foci during 1965. Circles, foci projected from within 0-150 km north of the section; triangles, from 0-150 km south. Exaggerated topography (13: 1) above. Inset, enlargement of the region of deep earthquakes (redrawn from Isacks et al., 1969, with permission from the Geological Society of America).

Tonga Trench___Rarotonga

Tonga Trench___Rarotonga

Tonga Arc Geology
Figure 9.6 Hypothetical section across the Tonga arc based on the attenuation of seismic waves (redrawn from Oliver & Isacks, 1967, by permission of the American Geophysical Union. Copyright © 1967 American Geophysical Union).

its discoverer(s), and earthquakes on it extend from near the surface, beneath the forearc region, down to a maximum depth of about 670 km. Figure 9.5 shows a section through the Tonga-Kermadec island arc system with earthquake foci projected on to a vertical plane parallel to the direction of underthrusting. The foci can be seen to occur at progressively greater depths with increasing distance from the site of underthrusting at the Tonga Trench. Further information on the nature of the Benioff zone was obtained from a study of the body wave amplitudes from deep earthquakes (Fig. 9.6). Seismic arrivals at the volcanic islands of the arc, such as Tonga, were found to be of far greater amplitude than those recorded to the front or rear of the arc at stations such as Raratonga and Fiji. The differences in amplitude are usually described quantitatively in terms of the Q-factor, the inverse of the specific attenuation factor, and in general the higher the Q-factor the stronger the rock. High Q travel paths give rise to little attenuation, and vice versa. Seismic waves traveling up the length of the seismic zone appear to pass through a region of high Q (about 1000), while those traveling to lateral recorders pass through a more normal region of low Q (about 150). The Benioff zone thus appears to define the top of a high Q zone about 100 km thick. The Benioff zone had originally been interpreted as a large thrust fault between different crustal provinces. The seismic data allowed a new interpretation to be made in terms of a high Q belt of Pacific lithosphere underthrust into the mantle. This interpretation was refined by Barazangi & Isacks (1971), by the use of a local seismometer network in the region of the Tonga arc (Fig. 9.7). In addition to the previous results, a zone of very high attenuation

(extremely low Q of about 50) was defined in the uppermost mantle above the downgoing slab in a region about 300 km wide, stretching between the active island arc (Tonga) and backarc ridge (Lau Ridge). This implies that the mantle beneath the backarc basin (Lau basin) is much weaker than elsewhere or that the lithosphere is considerably thinner. The data have important ramifications for the origin of backarc basins and will be considered in more detail in Section 9.10.

Detailed investigations of the region above the subducting lithosphere have also been carried out using seismic tomography (Section 2.1.8). Plate 9.1 (between pp. 244 and 245) shows a section through the Tonga arc in which the subducting slab is clearly defined by a region of relatively high P-wave velocity. Above this there is a region of low velocities, beneath the Lau basin (see also Section 9.10), corresponding to the region of extremely low Q in Fig. 9.7. The lowest velocities occur beneath the Tonga arc volcanoes.

The earthquake activity associated with the down-going slab occurs as a result of four distinct processes (Fig. 9.8). In region "a" earthquakes are generated in response to the bending of the lithosphere as it begins its descent. Bending, or downward flexure of the lithosphere, puts the upper surface of the plate into tension, and the normal faulting associated with this stress regime gives rise to the observed earthquakes, which occur to depths of up to 25 km (Christensen & Ruff, 1988).

Flexural bending of the lithosphere also gives rise to the topographic bulge present in the subducting plate on the oceanward side of the island arc. This regional rise of sea bed topography is located between 100200 km from the trench axis and has an amplitude of

Tonga Tonga

Tonga Tonga

Tonga Ridge Subduction Attenuation

1000 900 800 700 600 500 400 300 200 100 0 Distance from trench (km)

Figure 9.7 Schematic section across the Tonga arc showing the zone of very high seismic attenuation beneath the Lau backarc basin (redrawn from Barazangi & Isacks, 1971, by permission of the American Geophysical Union. Copyright © 1971 American Geophysical Union).

1000 900 800 700 600 500 400 300 200 100 0 Distance from trench (km)

Figure 9.7 Schematic section across the Tonga arc showing the zone of very high seismic attenuation beneath the Lau backarc basin (redrawn from Barazangi & Isacks, 1971, by permission of the American Geophysical Union. Copyright © 1971 American Geophysical Union).

i

■■■■■■■■■■■■■■■

i ■ mmMm

h'^-iT:?:::::::; Lithosphere I ■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■

100 -

Asthenosphere

200 -

AW

E

300 -

400 - ^.

w

500

Figure 9.8 Plate model of subduction zones; a, b, c, and d indicate regions of distinctive focal mechanisms.

several hundred metres. Simple beam theory predicts that the presence of this bulge is a consequence of the downward deflection of the subducting plate (Fig. 9.9). However, closer investigation of lithospheric behavior in this environment indicates that the flexure is not completely elastic, and must involve considerable plastic (permanent) deformation (Fig. 9.10) (Turcotte et al., 1978). Chapple & Forsyth (1979) deduced that the bending of a two layer elastic-perfectly plastic plate,

50 km thick, in which the upper 20 km are under tension and the lower 30 km under compression, fits most topographic profiles, and that the variations in these profiles are probably due to variations in the regional stress field.

Region "b" (Fig. 9.8) is characterized by earthquakes generated from thrust faulting along the contact between the overriding and underthrusting plates. Focal mechanism solutions for earthquakes associated with regions "a" and "b" of Fig. 9.8 are shown in Fig. 9.11,

Subduction Zone Earthquake Mariana
Figure 9.9 Downbending of an elastic or elastic-perfectly plastic plate at a subduction zone (redrawn from Turcotte et al., 1978, with permission from Elsevier).
Earthquake Marina Trench
Figure 9.10 Observed and theoretical profiles of lithosphere bending at a trench: (a) Mariana Trench, with an elastic lithosphere 29km thick; (b) Tonga Trench, better modeled by an elastic-perfectly plastic plate 32km thick (redrawn from Turcotte et al., 1978, with permission from Elsevier).
Earthquake Marina Trench
Figure 9.11 Focal mechanism solutions of earthquakes in the Aleutian arc, compressional quadrant shaded (redrawn from Stauder, 1968, by permission of the American Geophysical Union. Copyright © 1968 American Geophysical Union).

which represents the distribution of earthquake types around the Aleutian island arc (Stauder, 1968). The belt of earthquakes to the south of the islands is caused by normal faulting associated with the flexure of the top part of the Pacific Plate, which is underthrusting the Bering Sea in a northwesterly direction. The groups of earthquakes lying under or just to the south of the island chain are indicative of thrust faulting. The nodal planes dip steeply to the south and gently to the north. It is probable that the latter planes represent the fault planes, and that these earthquakes are generated by the relative movement between the Pacific and Bering Sea lithosphere. The single focal mechanism solution indicative of strike-slip movement is either on a sinistral strike-slip fault perpendicular to the island chain, as indicated on the diagram, or alternatively on a dextral strike-slip fault paralleling the island chain. In view of the oblique direction of underthrusting in this region, the latter interpretation is perhaps more likely to be correct (Section 5.3).

The earthquakes occurring in the Benioff zone in zone "c" (Fig. 9.8), at depths greater than the thickness of the lithosphere at the surface, are not generated by thrusting at the top of the descending plate, because the asthenosphere in contact with the plate is too weak to support the stresses necessary for extensive faulting. At these depths earthquakes occur as a result of the internal deformation of the relatively cold and hence strong descending slab of lithosphere. Hasegawa et al. (1978), making use of a local array of seismographs, identified two Benioff zones beneath the Japan arc that appear to merge down dip (Fig. 9.12). The arrival times of different seismic phases indicate that the upper of these zones corresponds to the crustal part of the descending slab, and the lower to the lithospheric mantle (Hasegawa et al., 1994).

Subsequently, double seismic zones, at depths between 70 and 200 km, have been documented in numerous well-studied subduction zones (Peacock, 2001), and it seems probable that they are a common feature of subduction zone seismicity. In some cases focal mechanism solutions for the upper zone earthquakes imply down-dip compression, and those for the lower zone earthquakes down-dip tension. This suggests that unbending of the downgoing plate may be important, the plate having suffered a certain amount of permanent, plastic deformation during its initial descent (Isacks & Barazangi, 1977). However the double seismic zones extend to depths well beyond the region of unbending of the downgoing plates. It is now thought that most of these earthquakes are triggered by meta-morphic reactions involving dehydration; those in the upper zone associated with the formation of eclogite (Kirby et al., 1996), and those in the lower zone with the dehydration of serpentinite (Meade & Jeanloz, 1991). It is suggested that dehydration reactions generate high pore pressures along pre-existing fault planes in the subducting oceanic lithosphere, producing earthquakes

What Causes Double Seismic Zones

Figure 9.12 Distribution of earthquakes beneath the northeastern Japan arc. Shaded line is probably the top of the descending lithosphere (redrawn from Hasegawa et al., 1978, with permission from Blackwell Publishing).

Figure 9.12 Distribution of earthquakes beneath the northeastern Japan arc. Shaded line is probably the top of the descending lithosphere (redrawn from Hasegawa et al., 1978, with permission from Blackwell Publishing).

Focal Mechaism Solution
Figure 9.13 Schematic focal mechanism solution distribution on a section perpendicular to an island arc. Inset shows alternative intermediate depth mechanism (redrawn from Isacks et al., 1969, with permission from the Geological Society of America).

by brittle failure. The process is termed dehydration embrittlement.

Peacock (2001), using a detailed thermal model for the subduction zone beneath northeast Japan, has shown that the lower seismic zone (Fig. 9.12) migrates across the isotherms, from approximately 800 to 400°C, as the focal depths increase from 70 to 180 km. If these temperatures and implied pressures are plotted on a P-T diagram, the pressure/temperature values and negative slope are very analogous to those for the dehydration reaction serpentine to forsterite + enstatite + water. This strongly suggests that these earthquakes are the result of the dehydration of serpentinized mantle within the downgoing oceanic plate. This explanation assumes that the oceanic mantle is serpentinized to a depth of several tens of kilometers, whereas hydrothermal circulation and alteration at mid-ocean ridges is thought to be restricted to the crust. However, the normal faulting associated with the outer rise and bending of the oceanic lithosphere oceanward of the trench may well permit ingress of seawater and hydration of the lithosphere to depths of tens of kilometers (Peacock, 2001).

Below 300 km (zone "d" in Fig. 9.8) the earthquake mechanism is believed to be a result of the sudden phase change from olivine to spinel structure, producing transformational or anticrack faulting. This takes place by rapid shearing of the crystal lattice along planes on which minute spinel crystals have grown (Green, 1994). At normal mantle temperatures this phase change occurs at a depth of approximately 400 km (Sections 2.8.5, 9.5). However, the anomalously low temperatures in the core of a downgoing slab enable olivine to exist metastably to greater depths, potentially until it reaches a temperature of about 700°C (Wiens et al., 1993). In old, rapidly subducting slabs this may, exceptionally, be at a depth of approximately 670 km, explaining the termination of subduction zone seismicity at this depth. It is also probable that a similar transformation from enstatite to ilmenite contributes to subduction zone seismicity in this depth range (Hogrefe et al., 1994). The phase changes that occur in the slab at a depth of approximately 700 km (Sections 2.8.5, 9.5) are thought to produce fine-grained materials that behave in a superplastic manner and thus cannot generate earthquakes (Ito & Sato, 1991).

The deep events of regions "c" and "d" (Fig. 9.8) are characterized by principal stress directions that are either parallel or orthogonal to the dip of the descending plate (Isacks et al., 1969) (Fig. 9.13). Consequently, the nodal planes determined by focal mechanism solutions do not correspond to the dip of the Benioff zone or a plane perpendicular to it. The principal stress directions show that the descending plate is thrown a b c d a b c d

High strength

Figure 9.14 A model of stress distributions in the descending lithosphere. Solid circles, extensional stress down dip; open circles, compressional stress down dip (redrawn from Isacks & Molnar, 1969, with permission from Nature 223,1121-4. Copyright © 1969 Macmillan Publishers Ltd).

High strength

Figure 9.14 A model of stress distributions in the descending lithosphere. Solid circles, extensional stress down dip; open circles, compressional stress down dip (redrawn from Isacks & Molnar, 1969, with permission from Nature 223,1121-4. Copyright © 1969 Macmillan Publishers Ltd).

into either down-dip compression or extension. Isacks & Molnar (1969) have suggested that the distribution of stress type in the seismic zone may result from the degree of resistance experienced by the plate during its descent, and Spence (1987) has described this resistance in terms of the net effect of ridge push and slab pull forces (Section 12.6). In Fig. 9.14a the plate is sinking through the asthenosphere because of its negative buoyancy and is thrown into down-dip tension as its descent is unimpeded. In Fig. 9.14b the bottom of the plate approaches the mesosphere, which resists descent and throws the leading tip into compression. As the plate sinks further (Fig. 9.14c), the mesosphere prevents further descent and supports the lower margin of the plate so that the majority of the seismic zone experiences compression. In Fig. 9.14d a section of the downgoing slab has decoupled so that the upper portion of the plate is thrown into tension and the lower portion into compression. A global summary of the stress directions determined from focal mechanism solutions (Isacks & Molnar, 1971) is shown in Fig. 9.15.

The stress distributions shown in Fig. 9.14b,d provide a possible explanation for the seismic gaps observed along the middle parts of the Benioff zone at certain trenches, such as the Peru-Chile Trench (Figs 9.15), where it is known that the slab is continuous (James & Snoke, 1990). A further type of seismic gap appears to be present in some island arcs at shallow depths. Figure 9.16 shows sections through the Benioff zone at the Aleutian-Alaska arc (Jacob et al., 1977). There is a prominent gap in seismicity between the trench and a point about halfway towards the volcanic arc that becomes progressively greater from west to east. The angle of underthrusting is very shallow in this region. The probable cause of this seismic gap and shallow underthrusting is the presence of copious quantities of terrigenous sediments within the trench that become increasingly abundant towards that section of the trench adjacent to Alaska. The unconsolidated nature of these sediments probably prevents any build-up of the strain energy necessary to initiate earthquakes, and their high positive buoyancy may force the subducting plate to descend at an anomalously shallow angle.

In reviewing the data for numerous subduction zones, Fukao et al. (2001) noted that subducted slabs are either deflected horizontally within or just beneath the transition zone, or penetrate the 660 km discontinuity and descend into the lower mantle (Plate 9.2 between pp. 244 and 245). Beneath Chile, the Aleutians, southern Kurile, and Izu-Bonin the slabs appear to flatten out within the transition zone, whereas beneath the Aegean, central Japan, Indonesia, and Central America they penetrate deep into the lower mantle. The slab beneath Tonga both flattens out within the transition zone and extends into the lower mantle (van der Hilst, 1995) (Plate 9.2e between pp. 244 and 245). There is no relationship between the age of a subducting slab and penetration into the lower mantle. Some researchers maintain that in places there is evidence for the slabs descending throughout the lower mantle to the core-mantle boundary (Section 12.8.2); others consider that there is little evidence for slab penetration beneath 1700 km depth (Karason & van der Hilst, 2000). The possible implications of these tomographic results for convection in the mantle are considered in Section 12.9.

Was this article helpful?

+1 0
How To Have A Perfect Boating Experience

How To Have A Perfect Boating Experience

Lets start by identifying what exactly certain boats are. Sometimes the terminology can get lost on beginners, so well look at some of the most common boats and what theyre called. These boats are exactly what the name implies. They are meant to be used for fishing. Most fishing boats are powered by outboard motors, and many also have a trolling motor mounted on the bow. Bass boats can be made of aluminium or fibreglass.

Get My Free Ebook


Responses

  • Rosaria
    What is oceanic subduction zone gravity anomaly?
    3 years ago

Post a comment