Origin of paired high pressure–ultrahigh-temperature orogens: a ridge subduction and slab window model


  • M. Santosh,

    1. Department of Natural Environmental Science, Faculty of Science, Kochi University, Kochi 780-8520, Japan
    Search for more papers by this author
  • T. Kusky

    1. Three Gorges Geohazard Research Center and State Key Lab for Geological Processes and Mineral Resources, China University of Geosciences, Wuhan 430074, Hubei Province, China
    Search for more papers by this author

Prof. Madhava Warrier Santosh, Department of Earth Environmental Science, Kochi University, Faculty of Science, Akebono-cho 2-5-1, Kochi 780-8520, Japan. E-mail: santosh@cc.kochi-u.ac.jp


Terra Nova, 22, 35–42, 2010


Ultrahigh-temperature (UHT) and high pressure (HP) metamorphic rocks generated at different times in Earth history form paired suites within the same tectonic belts in several regions. We evaluate the thermal regimes and fluid circulation patters in different plate tectonic settings and propose a new model involving ridge subduction where the slab window places hot asthenosphere against the base of the overlying plate and provides an ideal setting for the supply of heat and CO2-rich fluids at relatively shallow domains, which can explain the formation and preservation of UHT metamorphic rocks in the roots of the arc and forearc. The plate geometry below the slab window would inhibit the normal cooling induced by the slab, and a temporary deprivation of slab-derived sediments and hydrous fluids. Our analysis provides a unified model for paired UHT and HP metamorphic orogens and presents a key to the associated thermal and fluid regimes.


Metamorphism at elevated pressure–temperature conditions and distinct fluid regimes generates ultrahigh-pressure (UHP), high pressure (HP) and ultrahigh-temperature (UHT) rocks (Harley, 2004; Brown, 2006, 2007, 2009; Kelsey, 2008; Santosh and Omori, 2008a,b; Santos et al., 2009; Santosh et al., 2009a; Zhang et al., 2009). Although previous studies have correlated these different orogens to distinct tectonic settings, recent investigations record high-pressure and UHT conditions from the same tectonic environment, within the same orogenic belt and, in some cases, even within the same samples. Most of these examples come from major collisional zones along which continental blocks were sutured at different times in Earth history, including from the Neoarchean HP–UHT rocks of the Limpopo Belt in South Africa (Tsunogae and van Reenen, 2006), the Palaeoproterozoic HP (Guo et al., 2004; O’Brien et al., 2005) and UHT (Santosh et al., 2007a,b, 2008, 2009b) belts in the North China Craton, and the HP–UHT belts in the Neoproterozoic Eastern Ghats Belt as well as the Cambrian Gondwana suture in southern India (Shimpo et al., 2006; Tsunogae et al., 2008a; Santosh et al., 2009c). Notably, the thermal regimes for these paired HP and UHT rocks are similar (∼1000 °C), and they also share a common fluid history characterized by CO2-rich composition (Tsunogae et al., 2008b; Santosh et al., 2009a), albeit with a contrasting pressure record varying from 15 to 20 kbar for the HP rocks down to less than 8 kbar for the UHT rocks (Fig. 1). The formation of such paired HP–UHT metamorphic belts has important implications in terms of understanding the tectonic settings, particularly, the plate tectonic architecture in which these orogens develop.

Figure 1.

 Pressure–temperature (PT) diagram showing the field of paired HP–UHT orogens (thick box). The ranges of isochores for CO2-rich fluid inclusions in HP–UHT rocks (references in text) are also shown, together with the highest and the lowest density values. The thick broken line represents a decompression P–T path corresponding to the exhumation of the orogens. Fields of extreme crustal metamorphism (shaded regions) include ultrahigh-pressure, high pressure and UHT domains. The facies boundaries and assemblages are after Liou et al. (1998), Harley (1998, 2004), Okamoto and Maruyama (1999), and Kelsey (2008). Figure constructed after Brown (2006), and Sajeev and Santosh (2006). The CO2 isochores are from Santosh et al. (2009e). UHP, ultrahigh-pressure; HP, high pressure; UHT, ultrahigh-temperature; AM, amphibolite; Amp-EC, amphibolite–eclogite; BS, blueschist; EA, epidote–amphibolite; EC, eclogite; Ep-EC, epidote–eclogite; GR, granulite; GS, greenschist; HGR, high-grade granulite; Lw-EC, lawsonite–eclogite; Coe, coesite; Qtz, quartz; Jd, jadeite; Ab, albite; And, andalusite; Ky, kyanite; Sil, sillimanite; Opx, orthopyroxene; Spr, sapphirine; Spl, spinel.

In this study, we evaluated the various plate tectonic settings that could account for the thermal, baric and fluid regimes in paired HP–UHT associations and propose a new model involving ridge subduction and slab window effects to explain the enigma.

Tectonic settings of HP–UHT metamorphism

It has been suggested that the generation of regional UHT metamorphic belts requires that the crust has achieved an advective geothermal gradient steeper than approximately 20 °C km−1 (Brown, 2007). One of the commonly invoked tectonic settings for the UHT thermal conditions to be achieved are the back-arcs of active accretionary–extensional margins (Collins, 2002), following the observation that many modern back-arc basins are regions of thin, weak crust with relatively high heat flow. Brown (2006) correlated UHT metamorphic rocks of Pan-African mobile belts to inverted and thickened back-arc basins. Many workers therefore consider that modern-day back-arc basins are candidates for UHT metamorphism and future sites of mobile belts (Brown, 2006, 2007, 2009; Kelsey, 2008).

We briefly evaluate the principal plate tectonic settings in relation to the formation of HP–UHT orogens. Figure 2(a) shows a schematic illustration where an oceanic plate is born at a ridge and becomes hydrated, followed by subduction where it releases water through dehydration reactions. The residual mass after the release of arc magmas is dragged down, finally becoming horizontal and stagnant at the mantle transition zone at 410–660 km depth (e.g. Maruyama et al., 2009). Under the back-arc region, rising hydrous plumes generate a major zone of hydration, such as in the case of present day Japan (cf. Maruyama et al., 2007, 2009; Santosh et al., 2009a). Although these are regions of thin and weakened crust with relatively high heat flow, the fluids in this zone are dominantly hydrous, rather than CO2-dominated, as also attested by geophysical studies in modern-day subduction systems (e.g. Zhao et al., 1994, 2007; Hasegawa et al., 2009; Maruyama et al., 2009). However, HP–UHT rocks are characterized by the preservation of dry mineral assemblages and the common association of CO2-rich fluid inclusions (e.g. Ohyama et al., 2008; Santosh and Omori, 2008a; Santosh et al., 2008; Tsunogae et al., 2008a,b). Moreover, UHT rocks are often associated with accretionary complexes together with continental margin sequences (e.g. Santosh et al., 2009b,c).

Figure 2.

 Fluid regimes in convergent and divergent plate tectonic settings as discussed in the text. (a) Subduction zone fluid factories. Recent models suggest the deeply subducted TTG (tonalite–trondhjemite–granodiorite) materials (trench sediments) accumulate in the mantle transition zone, remaining stagnant, behaving in a manner similar to a second continental crust on the mantle transition zone, and functioning as a critical heat source to generate hot upwellings (plumes) (Kawai et al., 2009; Senshu et al., 2009). Note that the domain below back-arc basin is a dominant zone of hydration. (b) Enlarged segment of the subduction zone showing the MMF (metasomatic–metamorphic factory) and SZMF (subduction zone magma factory). The location of generation of high PT metamorphic rocks is marked by a star.

Whereas the role of fluid and/or melt (particularly CO2) plays an important role in the generation of HP–UHT rocks, the more critical aspect is the heat source. Recent works (e.g. Whittington et al., 2009) suggest that the lower crust has a lower thermal diffusivity than has been previously employed in geodynamic models. The new finding implies that the lower crust is able to retain heat for longer time and that the underlying mantle has a higher mean temperature. In terms of UHT metamorphism, the recent work of McKenzie and Priestley (2008) indicates that if the temperature dependence of thermal conductivity is taken into account in geodynamic models, UHT conditions can be replicated. Therefore, the heat required for the generation of some of the T > 900 °C crustal rocks could be a function of the inherent properties and characteristics of the continental crust. These new studies offer various alternatives to generate UHT conditions.

In a normal subduction regime, an oceanic slab slides beneath the convergent margin and most of the sediments overlying the slab are typically accreted to the overlying arc, but, in some cases, there is considerable flux of hydrous material through subduction erosion where the sediments are subducted. Examples are known from the Central American trench (Ranero and von Huene, 2000), some of the missing strata from the southern Alaska forearc may have been subducted (Kusky et al., 1997a) and some of the arcs in the SW Pacific are experiencing considerable amounts of ongoing subduction erosion (von Huene and Scholl, 1993; von Huene et al., 2004; Santosh et al., 2009d; Yamamoto et al., 2009). In this tectonic setting, the subducted rocks are subjected to HT–HP regimes in different zones and the cooling effect of the slab keeps temperatures low in the forearc region. In the adjacent arc, however, arc magmas intruded near the base of the crust result in HT metamorphism, forming a paired metamorphic belt (sensuMiyashiro, 1961; see also Brown, 2009).

Figure 2(b) shows an enlarged section of a subduction zone where the triangular corner of the wedge defines a metasomatic–metamorphic factory (MMF) and the middle domain represents a subduction zone magma factory (SZMF). From deep to shallow levels, the width of the MMF, which is a function of the subduction angle of the oceanic plate, exerts a significant control on the style of tectonic exhumation and the channel flow or extrusion of the regional metamorphic belts against the downgoing subducting slab (Cloos, 1984Gerya et al., 2008; Warren et al., 2008; Maruyama et al., 2009). The convection in the magma factory is driven by a rising plume generated at about 200 km depth. The ubiquitous occurrence of hydrous fluids underneath the SZMF lowers the mantle viscosity in the magma factory. The counter flow convection against the subducting oceanic slab moves in a clockwise sense delivering subduction zone magma under the volcanic front and subsequently moves down in the deep mantle attached to the top of the subducting slab. This convection pattern imparts higher temperatures towards the deeper domains of the subduction zone, aiding in the generation of relatively dry, high PT metamorphic mineral assemblages. On the contrary, the convection in the MMF is geometrically opposite, with an anticlockwise sense above the neutral boundary (Maruyama et al., 2009). Water derived by the dehydration of the oceanic slab and underlying hydrated mantle moves upward and enters the MMF domain. Recent two-dimensional numerical modelling (Gerya et al., 2008) of early continental collision associated with subduction of the lithospheric mantle also suggests that extensive recrystallization and metamorphism of subducted trench sediments lead to the anticlockwise convective flow, promoting the subducted and metamorphosed rocks to be returned to the surface.

Temperatures are generally low in convergent settings characterized by a subduction zone geotherm. To generate UHT conditions of up to or even above 1000 °C, pressures much higher than those normally recorded from UHT granulites are required. Therefore, Santosh and Omori (2008b) postulated a post-collisional extensional setting as a potential location for the formation of UHT granulites, which has implications for the generation of UHT rocks in the subduction channel. The several UHT localities reported from the Cambrian collisional suture in southern India occur in the vicinity of the palaeo-subduction zone (e.g. Santosh et al., 2009c). In the case of extensional orogeny at divergent plate boundaries such as the continental rift zone in the present day African rift valley, the heat and volatiles supplied by rising plumes contribute to the generation of dry UHT assemblages in the lower crust. However, the extensional orogeny is incapable of exposing the lowermost crust and therefore it is difficult to observe the UHT rocks and they remain in a (temporarily) concealed state.

It has been proposed that a carbonated tectosphere (sub-continental lithospheric mantle or continental keel; Jordan, 1988) can function as an important reservoir of anhydrous fluids, particularly CO2 (Santosh et al., 2009a,d). Figure 3(a) shows a schematic illustration of the thermal and material erosion of tectosphere by a rising plume. The high temperature plume thermally erodes the tectosphere and causes metasomatic replacement. This model thus accounts for both the thermal regime and CO2-dominated nature of the fluids in the tectosphere. The North China Craton provides a typical example for extensive erosion of the sub-continental lithospheric mantle (e.g. Kusky et al., 2007 and references therein). An interpretative sketch of P-wave velocity tomographic image along North 40° latitude (after Tian et al., 2009) below the Palaeoproterozoic Ordos Block in the North China Craton is shown in Fig. 3b. The P-wave velocity perturbations clearly define a rising hot asthenosphere that erodes the overlying tectosphere.

Figure 3.

 (a) Cartoon illustrating the concept of carbonated tectosphere being eroded by rising plumes which supply heat and CO2 to generate UHT metamorphic orogens within extensional settings. (b) Sketch of P-wave velocity tomographic image along North 40° latitude (after Tian et al., 2009) below the Palaeoproterozoic Ordos Block in the North China Craton. A prominent asthenospheric upwelling is seen, which erodes the overlying tectosphere. HP–UHT orogens occur immediately above this zone and suggest that a similar asthenospheric upwelling and resultant heat and CO2 input in the past might have been instrumental in stabilizing the anhydrous assemblages.

Ridge subduction and slab window model

Ridge subduction is a term used to describe the interaction of an oceanic spreading centre with a subduction zone. Once a ridge enters the subduction realm, the spreading centre ceases to exist, and in its place, an ever widening ‘slab window’ opens downdip of the trench (Bradley et al., 2003; Goldbarb et al., 2004). The slab window functions as the interface between the base of the overriding plate and the hot asthenosphere that wells up from beneath the two subducted, but still diverging plates. The upwelling mantle material that fills the slab window would normally trigger partial melting, but once the slab window opens beneath the convergent margin, a switch over occurs with increasing depth. In shallow levels of the accretionary prism, melt still rises, mixes with, and partially melts the accretionary wedge material, forming hybrid magmas (e.g. Lytwyn et al., 1997, 2000; Kusky et al., 2003). The high topographic features of the subducting ridge erode the accretionary wedge and cause slivers of the wedge to be subducted, metamorphosed and, in some cases, extruded back towards the surface, such as in the metamorphic cover over the Resurrection ophiolite and perhaps in the Chugach metamorphic complex in Alaska (Kusky and Young, 1999; Sisson et al., 2003). Migration of the near trench magmatic/metamorphic event along the trench is a fingerprint of ridge subduction, because the kinematics of the ridge subduction is controlled by the relative motions of the three plates involved (e.g. Bradley et al., 2003).

In our present model (Fig. 4), at deeper levels of ridge subduction, near the base of the lithosphere and in the lower domains, the slab window places hot asthenosphere against the base of the overlying plate and sublithospheric mantle, in a region that would normally be cooled by the slab. Without the hydrous fluids being released from the slabs, the magmatic activity in the arc shuts off while the slab window is passing beneath any part of the arc (e.g. Bradley et al., 2003). We propose that this domain provides an ideal setting for the preservation of UHT metamorphic rocks in the deep roots of the arc and forearc. This zone is essentially dry because there is no slab and therefore there are no hydrous fluids expelled from the slab. The dominant fluids in this zone are those left from previous hydration by the older slabs or those coming from deeper domains in the asthenosphere or lower mantle, which presumably would be CO2-rich. In the case of Alaska, ridge subduction and slab window phenomena resulted in the rocks in the accretionary complex being heated to temperatures above 650 °C at pressures as low as 2.5 kbar (e.g. Sisson et al., 2003). Since the rising asthenosphere in the slab window has temperatures over 1200 °C, UHT metamorphism (>1000 °C) at relatively low pressures (down to 8 kbar) is a likely scenario at deeper domains. Below this part of the slab window, the geometry is such that hot asthenosphere from the slab window is juxtaposed with hot asthenosphere of the overriding plate. This would inhibit the normal cooling induced by the slab and a temporary deprivation of slab-derived sediments and fluids, further aiding in the generation and preservation of high- and UHT dry metamorphic assemblages. Slab windows typically migrate along the entire length of the convergent margin and extend to beneath the arc and beyond as far as the slabs penetrate into the mantle. Some of the present day examples include the South Chile rise and Woodlark Basin (e.g. Johnston and Thorkelson, 1997; Behrmann et al., 2004). The recent numerical modelling presented by Omori et al. (2009) for regional metamorphism beneath the Japanese islands also provides an ongoing case characterized by a duality of high temperature and HP metamorphism.

Figure 4.

 Ridge subduction and slab window model proposed in this study linking HP and UHT metamorphism. On either side of the migrating triple junction, the overriding plate experiences normal HP metamorphism, with HT and UHT metamorphism beneath the arc. During ridge subduction, the near trench area is briefly heated to HT conditions and the deeper parts of the overriding plate are affected by dry HT–UHT metamorphism. See text for discussion. The numbers refer to the locations of HT–UHT/HP zones.

Paired metamorphic belts comprising a low thermal gradient metamorphic belt outboard and a high thermal gradient metamorphic belt inboard may record orogen-parallel terrane migration and juxtaposition by accretion of contemporary belts of contrasting type. Brown (2009) proposed a wider definition of paired metamorphism to incorporate all types of dual metamorphic belts with an additional feature of ridge subduction, which may be reflected in the pattern of high dT/dP metamorphism and associated magmatism. Paired HP–HT/UHT belts occur along the trace of some of the major suture zones formed by convergent tectonics to generate supercontinent amalgams such as the Palaeoproterozoic Columbia suture in North China Craton, the Neoproterozoic Rodinia suture in the Eastern Ghats Belt and the Cambrian Gondwana suture in southern India (e.g. Kusky and Santosh, 2009; Santosh et al., 2009c). Available geochronological data suggest a time interval between the various HT/UHT and HP events within the individual segments (e.g. Guo et al., 2004; Santosh et al., 2009b,c), in accordance with the diachronous nature predicted in the case of ridge suduction because of migration along the margin, as illustrated from the modern examples of Alaska and elsewhere (e.g. Bradley et al., 2003; Kusky et al., 2003 or Bradley et al., 2003).

We have shown in Fig. 5 the case from southern India where paired UHT–HP rocks occur within the Palghat–Cauvery Suture Zone, a trace of the Cambrian Gondwana collisional suture (Collins et al., 2007). The development of a wide accretionary belt with typical features of ocean plate stratigraphy associated with the closure of the Mozambique Ocean in the late Neoproterozoic has been recently recorded from this zone (Santosh et al., 2009c). The nature and distribution of the rock types in this zone with charnockites and granites at the higher crustal level followed by mafic/ultramafic rocks and HP–UHT paired sequences towards the deeper level broadly correspond with the southward polarity of ridge subduction and possible slab window opening. Thus, one of the possible scenarios for the HP–UHT conditions in a CO2-dominated fluid regime recorded from these rocks would be a model where the slab window was placed against a hot asthenosphere, although alternate explanations are also possible.

Figure 5.

 Ridge subduction model to explain the paired HP–UHT metamorphic rocks within the Palghat–Cauvery Suture Zone in the Cambrian Gondwana collisional suture of southern India. (a) Cartoon illustrating the pattern of ridge subduction and slab window opening (after Bradley et al., 2003). (b) Part of the central Gondwana assembly showing the trace of Late Neoproterozoic–Cambrian suture formed by the closure of Mozambique Ocean. The location of the Palghat–Cauvery Suture Zone (PCSZ) is also shown. (c) Cartoon illustrating the subduction-collision-accretion tectonics along the PCSZ (after Santosh et al., 2009c) with a southward polarity of the subducting oceanic plate. The major location of accretionary belts, dismembered ophiolite and exhumed HP–UHP orogens are also shown. (d) Geological map of the PCSZ area showing the locations of HP–UHT granulites. The proposed slab window is shown by thick dashed lines.

Considerable disruption of the accretionary wedge during ridge subduction (Kusky et al., 1997a,b) leads to much higher sedimentation rates in the trench and changes in the style of subduction that lead to both greater growth of the wedge and subduction of sediments to greater depths. The deeply subducted material undergoes high and UHP metamorphism. The plate tectonic scenario proposed in this study provides a unified model that links extreme metamorphism with subduction processes and offers a viable explanation for the occurrence of exhumed HP–UHT orogens in convergent settings.


We thank Terra Nova reviewers Dave Kelsey, Gary Ernst, Taras Gerya and two anonymous referees for their comments. We also thank Editor Prof. Alfred Kroner for helpful handing of the manuscript.