Thermomechanical modelling of slab detachment

https://doi.org/10.1016/j.epsl.2004.07.022Get rights and content

Abstract

Slab detachment or breakoff is appreciated as an important geological process, as shown by recent tomographic imaging. Using a 2-D upper-mantle model 660-km deep and 2000-km wide, we have investigated with a 2-D finite-difference and marker-in-cell numerical technique the multi-resolutional character of the thermomechanical phenomena related to this complex geological process. Our experiments show that this process can be initiated in form of slab necking by a prolonged (8–30 My) period of slab weakening due to thermal diffusion (<20 °C/My) after cessation of active subduction. The rapid detachment process takes place over a few million years and is accelerated by non-Newtonian strain-rate softening and focused thermal erosion (>60 °C/My) due to strong positive thermal feedback from shear heating. Detached slab fragments sink rapidly with a tendency for coherent rotation. The influence of temperature- and pressure-dependent thermal conductivity on the process of thermal weakening of the slab is quite significant. This supports the idea that the breakoff process is triggered by thermal diffusion on a time scale linearly dependent on heat conductivity. Rapid topographic changes and increasing volcanic activities due to the melting of subducted oceanic crust are possible scenarios of this vigorously driven geodynamic process.

Introduction

Slab detachment or slab breakoff is a realistic geological process originally inferred seismically from gaps in the hypocentral distributions associated with subducted slabs (e.g., [1]) and supported by both theoretical considerations [2], [3], [4] and seismic tomography studies [5], [6], [7], [8]. This process can involve mantle lithosphere of continental (e.g., [9]), oceanic (e.g., [7]) or continental margin (e.g., [4], [10], [11], [12]) origin. Slab detachment is commonly suggested to be associated with the early stages of continental collision (e.g., [4], [13]), due to a decrease in the subduction rate damped by the positive buoyancy of continental lithosphere introduced into the subduction region. Slab detachment and slab fragmentation may also be induced by a decrease in subduction rate in other geodynamic settings. This process can thus play an influential role along slab edges (e.g., [7]).

In spite of past significant progress in our understanding of slab detachment from both analytical and numerical studies (e.g., [4], [9], [13], [14], [15], [16], [17], [18]), there have been no studies on the dynamics of this process with a realistic rheological–petrological model to account for the effects of slab melting, as well as adiabatic and shear heating (e.g., [17], [19]). Most previous studies have used a static situation with simplified rheological models (e.g., [4], [13], [14], [18]). The dynamic modeling presented by Pysklywec et al. [9] and Houseman and Gubbins [16], which considered different modes of deformation of subducting lithosphere, is also based on a simplified rheology, and does not account for thermal effects such as thermal diffusion and shear heating, which, as we will demonstrate, turn out to be extremely important.

In this paper, we will investigate, by means of a systematic, high resolution 2-D study, the dynamics of detachment of oceanic slabs due to thermal diffusion after cessation of active subduction. Our primary goal is to study the dynamics, geometry and modes of slab detachment, and to take into account the effects of pressure-, temperature-, and strain-rate-dependent rheology of the mantle and oceanic crust, of pressure- and temperature-dependent thermal conductivity (Table 1), and of shear and adiabatic heating as well as partial melting of the oceanic crust.

Section snippets

Initial and boundary conditions of the 2-D model

Fig. 1 displays our 2-D numerical model, which is specifically designed for studying dynamic processes in a subducted slab after cessation of active subduction. A nonuniform rectangular grid with variable 2–10-km resolution is designed in such a way as to provide the highest resolution of 2×2 km in the upper central, 400-km-wide and 200-km-deep area of the model (Fig. 1a), where the detachment process is expected to be localized. The initial thermal structure of subducted slab is defined by the

Model for partial melting

Melting of subducted oceanic crust to produce characteristic melts is an important process at slab edges formed during slab detachment (e.g., [7], [29]). We adopt a model [28] that allows for melting of subducted oceanic crust in the PT region between the wet solidus and dry liquidus of basalt (Table 1). As a first approximation, the degree of melting is assumed to increase linearly with the temperature according to the relationsM=0atT<Tsolidus,M=(TTsolidus)/(TliquidusTsolidus)atTsolidus<T<T

Rheological model

The viscosity for viscous flow in the upper mantle depends primarily on the strain-rate and the temperature and is given in terms of the second invariant of the deviatoric strain-rate tensors [31] byηcreep=(εII)(1n)/2nF(AD)1/nexp[(E+PΔV)/nRT],where ɛII=(1/2)ɛijɛij is the second invariant of the strain rate tensor, with dimension s−2; E is the activation energy in kJ mol−1; ΔV is the activation volume, J MPa−1 mol−1; AD is the material constant in Pan s−1, n is the stress exponent; and R

Conservation equations and numerical implementation

We have considered two-dimensional creeping flow wherein both thermal and chemical buoyant forces are included. The conservation of mass is approximated by the incompressible continuity equationvx/x+vz/z=0.The 2-D Stokes equations for creeping flow take the form:σxx/x+σxz/z=P/xσzz/z+σxz/x=P/zgρ(T,C,M).The density ρ(T,C,M) depends explicitly on the temperature, the composition, and the degree of melting (see Section 3).

We also employ viscous rheological constitutive

Results from numerical experiments

We have garnered results from well over 50 different numerical models (see description of selected representative runs in Table 2). These simulations have been calculated over a finite-difference grid with 441×187 irregularly spaced Eulerian points (Fig. 1a). In contrast to our recent high-resolution numerical study of multiscale dynamics of hydrous cold plumes at subduction zones [36], simplified lithological structure of our model (Fig. 1a) allowed us to use a moderate number of markers (∼5

Discussion, conclusions and perspectives

In this work, we have laid out the numerical setup for a realistic thermomechanical study of the slab detachment process. By dint of a wide parametric study, we have demonstrated the dynamic feasibility of the detachment process as caused by thermal diffusion of subducted slabs after cessation of active subduction. Our experiments show that this process can be initiated in form of slab necking, associated with partial melting of the subducted oceanic crust. The detachment process is accelerated

Acknowledgements

This work was supported by ETH Research Grant TH-12/04-1, by RFBR grants #03-05-64633 and #1645-2003-5, by an Alexander von Humboldt Foundation Research Fellowship to TVG, by the geophysics program of the National Science Foundation and by the German Science Foundation within SFB 526. Arne P. Willner, Klaus Regenauer-Lieb and Anne M. Hofmeister are thanked for discussions and comments. Constructive reviews by G.A. Houseman and an anonymous reviewer are appreciated.

References (45)

  • B. Isacks et al.

    Mantle earthquake mechanisms and the sinking of the lithosphere

    Nature

    (1969)
  • P.E. Sacks et al.

    Delamination in collisional orogens

    Geology

    (1990)
  • F. Von Blanckenburg et al.

    Slab breakoff: a model for syncollisional magmatism and tectonics in the Alps

    Tectonics

    (1995)
  • J.H. Davies et al.

    Slab breakoff: a model of lithospheric detachment and its test in the magmatism and deformation of collisional orogens

    Earth Planet. Sci. Lett.

    (1995)
  • M.J.R. Wortel et al.

    Subduction and slab detachment in the Mediterranean–Carpathian region

    Science

    (2000)
  • P.F. Xu et al.

    Seismic tomography showing, subduction and slab breakoff of the Yangtze block beneath the Dabie–Sulu orogenic belt

    Chin. Sci. Bull.

    (2000)
  • V. Levin et al.

    Seismic evidence for catastrophic slab loss beneath Kamchatka

    Nature

    (2002)
  • R.D. Rogers et al.

    Epeirogenic uplift above a detached slab in northern Central America

    Geology

    (2002)
  • R.N. Pysklywec et al.

    Modeling the behavior of the continental mantle lithosphere during plate convergence

    Geology

    (2000)
  • A.T. Ismail-Zadeh et al.

    Stress in the descending relic slab beneath the Vrancea region, Romania

    Pure Appl. Geophys.

    (2000)
  • F. Wenzel et al.

    Seismotectonics of the Romanian Vrancea area

  • S.Y.M. Wong A Ton et al.

    Slab detachment in continental collision zones: an analysis of controlling parameters

    Geophys. Res. Lett.

    (1997)
  • Cited by (221)

    View all citing articles on Scopus
    View full text