Thermal structure and metamorphic evolution of subducting slabs

Abstract

Subducting lithospheric plates are the cool, downwelling limbs of mantle convection and the negative buoyancy of subducting slabs (slab pull) drives plate tectonics [Forsyth and Uyeda, 1975]. Subduction zones are regions of intense earthquake activity, explosive volcanism, and complex mass transfer between the crust, mantle, hydrosphere, and atmosphere. In this contribution, I present subduction-zone thermal models that provide a framework for discussing the petrological and seismological processes that occur in subducting slabs (defined herein as the subducting sediments, oceanic crust, and oceanic mantle). Specific issues to be discussed include uncertainties regarding mantle-wedge convection, metamorphic reactions in the subducting plate, the origin of arc magmas, and subduction-zone earthquakes. GENERAL OBSERVATIONS REGARDING SUBDUCTION ZONES Subducting slabs are cool because oceanic lithosphere, the cold upper boundary layer of Earth's internal convection, descends into the mantle more rapidly than heat conduction warms the slab. The chilling effect of subduction is recorded by surface heat flux measurements < 0.03 W m-2 in subduction-zone forearcs (one-half of the average global surface heat flux). In well-studied subduction zones like Cascadia, forearc heat flux systematically decreases from the trench to the volcanic front [Hyndman and Wang, 1995].

Thermal structure due to solid-state flow in the mantle wedge beneath arcs

“…Thermal models with symbols: Open squares and circles, NE Japan and Cascadia, nonlinear rheology [van Keken et al, 2002]; open diamonds and closed circles, 40 km and 70 km decoupling models, NE Japan [Furukawa, 1993a]; closed squares, fast subduction of thin plate, assuming mantle potential temperature of 1400°C [Kincaid and Sacks, 1997], heavy solid line with no symbols, isoviscous corner flow, NE Japan [van Keken et al, 2002]. Other thermal models: NE and SW Japan [Peacock and Wang, 1999]; Izu-Bonin ; Aleutians [Peacock and Hyndman, 1999]; 100 km decoupling model, NE Japan [Furukawa, 1993a]; slow subduction of thin plate, old plate and young plate, plus fast subduction of old plate and young plate, all assuming potential temperature of 1400°C [Kincaid and Sacks, 1997]; Alaska Range hot and cold models [Ponko and Peacock, 1995]; fast and slow subduction, with and without shear heating [Peacock, 1996]; general models [Peacock, 1990a]. Triangular grey field encloses geotherms inferred from heat flow data in the Oregon Cascades arc [Blackwell et al, 1982]; although interpretation of heat flow data is often controversial, the inferred geotherm is broadly consistent with metamorphic PT estimates for arc crust.…”

“…These calculations are essentially of three types: (1) analytical approximations including various assumptions about coupling between the subducted crust and the overlying mantle and about convection in the mantle wedge [e.g., Davies, 1999;Molnar and England, 1995;Molnar and England, 1990], (2) purely plate-driven models with uniform viscosity, in which the thermal regime is calculated numerically using analytical expressions for corner flow in the mantle wedge, with model results depending on various input parameters including the thickness of the arc "lithosphere" and the depth of coupling between subducting crust and overlying mantle [e.g., Peacock, 2002;Peacock and Hyndman, 1999;Peacock and Wang, 1999;Iwamori, 1997;Peacock, 1996;Ponko and Peacock, 1995;Peacock et al, 1994;Pearce et al, 1992;Peacock, 1991;Peacock, 1990a;Peacock, 1990b], and (3) dynamic models in which the mantle flow field as well as the thermal regime are calculated numerically, with model results depending on parameters such as thermal buoyancy, chemical buoyancy and mantle viscosity [van Keken et al, 2002;Furukawa and Tatsumi, 1999;Kincaid and Sacks, 1997;Furukawa, 1993a;Furukawa, 1993b;Davies and Stevenson, 1992]. These models differ in many respects, but most agree that subduction of oceanic crust that is more than 20 million years old at down-dip rates greater than 20 km/Myr will not produce temperatures at the top of the subducting plate that are high enough to allow fluid-saturated melting of sediment or basalt.…”

Seismic anisotropy in the Izu‐Bonin subduction system

“…We cannot resolve whether this pattern is due to a change in the olivine slip system at depths of 300 km or greater [i.e., Mainprice et al, 2005]. The rapid changes in splitting are likely not due to a shift from ''dry'' type A to ''wet'' type B olivine fabric [Jung and Karato, 2001], as the Izu-Bonin mantle wedge does not possess the high stresses and water content required for type B fabric development [e.g., Peacock, 2003;Hyndman and Peacock, 2003]. Developing a firstorder change in olivine fabric is difficult over such short (<50 km) spatial scales of mantle flow [e.g., Kaminski, 2002] and would likely not be resolvable with shear wave splitting observations [Lassak et al, 2004].…”

Arc Crust-Magma Interaction in the Andean Southern Volcanic Zone from Thermobarometry, Mineral Composition, Radiogenic Isotope and Rare Earth Element Systematics of the Azufre-Planchon-Peteroa Volcanic Complex, Chile

“…The subducting ocean lithospheric Nazca plate in the CSVZ has an age range of 18 to 25 Ma and the Nazca plate in the TSVZ ranges from 35 to 45 Ma while the rate of subduction in both regions remains the same of ~ 7-9 cm/yr (Jarrard 1986;Dewey & Lamb 1992;DeMets et al 1990). Given those parameters the thermal structure of the subducting slab in the CSVZ would be predicted to be the same or hotter than the thermal structure of the subducting plate beneath the TSVZ (Peacock, 2003). As a result the magmas generated in the CSVZ should have equal if not a greater AOC and sediment signatures than the magmas from the TSVZ, which is not observed.…”

“…Although it is clear that Nazca Plate sediments are subducted and contribute to basaltic magmas generated, their limited along-trench variability and low isotopic contrast to erupted basalts make it an unlikely source that controls the along-arc isotopic regional trends. Quaternary however, in the SVZ there is no strong evidence for systematic and consistent along-arc changes to this slab component (Hickey et al 1984(Hickey et al , 1986Hickey-Vargas 1989, 2003Sigmarsson et al 1990;Leeman et al, 1995). The subducting ocean lithospheric Nazca plate in the CSVZ has an age range of 18 to 25 Ma and the Nazca plate in the TSVZ ranges from 35 to 45 Ma while the rate of subduction in both regions remains the same of ~ 7-9 cm/yr (Jarrard 1986;Dewey & Lamb 1992;DeMets et al 1990).…”