286 Design of 2D numerical geodynamic models
and Schubert, 2002): 0
◦
C at the surface linearly increasing to 1380
◦
Cat
100 km depth. The temperature structure of the oceanic plate corresponds to its age
(Eq. (17.1)). A gradual linear transition from the oceanic to the continental geotherm
is prescribed within a 50 km wide area at the two ocean/continent boundaries. The
initial temperature gradient in the asthenospheric mantle is 0.5
◦
C/km. Due to thick-
ening in response to shortening of the model with time, the temperature imposed for
the lower boundary condition should increase with time according to the adiabatic
gradient.
In this numerical model, the mechanisms driving subduction are a combined
‘plate push’ (prescribed constant convergence velocity at the right boundary) and
‘slab pull’ (temperature induced density contrast between the subducted lithosphere
and surrounding mantle, see above retreating subduction model example). This
type of boundary condition is widely applied in numerical models of subduction
and collision (e.g., Burov et al., 2001; Burg and Gerya, 2005; Faccenda et al.,
2008b; Warren et al., 2008), assuming that in the globally confined system of
plates, the ‘external push’ imposed on a plate (coming from a different slab) can
be significant. A spontaneously increasing slab pull mainly regulates slab bending
dynamics and delamination of the slab from the overriding plate (e.g., Gerya et al.,
2008b). Indeed, similar (and in relation) to the subduction initiation problem, the
issue of choosing proper convergence conditions is not fully resolved yet and the
external push at the initial stages of convergence may be exaggerated compared to
nature.
Figure 17.5(b) shows the final stage of an experiment for continental collision
with a 200 km wide intermediate oceanic plate (30 Myr cooling age) moving
leftward at a constant velocity of 3 cm/yr due to the model shortening imposed
from the right boundary. Results are computed with the code Collision.m.The
development of the continental collision zone is associated with deep (>100 km)
subduction of the continental crust underneath the orogen (Fig. 17.5(b)). This
feature appears in many numerical models of continental collision (e.g. Burov
et al., 2001; Gerya et al., 2008b; Warren et al., 2008) including those driven by slab
pull rather than by a prescribed convergence velocity (e.g. Faccenda et al., 2008b;
Baumann et al., 2009). This fits findings of ultrahigh-pressure (UHP) rocks (e.g.
Chopin, 2003; Liou et al., 2004) that contain metamorphic diamonds (e.g., Rosen
et al., 1972; Sobolev and Shatsky, 1990; Dobrzhinetskaya et al., 1995; Massonne,
1999) and coesite (e.g., Chopin, 1984; Smith, 1984) in Phanerozoic collision belts.
The topography development predicted with our simplified model (Fig. 17.6)also
appears realistic and demonstrates the growth of a positive topography (up to
2500 m above the water level, Eq. (17.9)) after the beginning of collision at around
5–7 Myr. Growth of the elevated region is associated with erosion and deposition
of sediments on both sides of the ‘orogen’ (see black material in Fig. 17.5(b)).