Dynamics of small-scale subduction systems: a numerical and
analogue approach
The
theory of plate tectonics was well established around 1960. This theory
describes the outer shell of the Earth as a number of thin, rigid plates
(lithosphere) that are continuously in relative motion above the Earth’s mantle
that behaves like a fluid. The relative plate velocities are around 5 cm/year. This
implies that most of the tectonic processes and its geological consequences cannot
be studied in real time since they occur on the order of millions years. With
the aim of studying the dynamics of these long-term processes, analogue
experiments (scaled models in a laboratory) and 2D/3D numerical simulations are
commonly used. Combining computational and laboratory experiments to study a
specific geodynamic process may provide a complete evolutionary study,
complementing each method’s weaknesses and strengths. While laboratory experiments
provide the physical realism and a high resolution, numerical models allow
quantifying the physical parameters characterizing the evolution of the system
that cannot be obtained from laboratory experiments.
 |
| Figure 1. Scheme of a typical subduction zone. |
One
of these large-scale tectonic processes is subduction. Subduction occurs along convergent
plate boundaries where one plate (subducting plate) descends beneath another
(overriding plate; Figure 1). Understanding subduction processes is of great
interest among scientists since a large fraction of earthquakes, volcanic
eruptions and ore deposits occur along these convergent boundaries. Some of
these subduction zones are more complex than others involving two subducting
plates in opposite direction. This type of double polarity subduction system is
observed and proposed to occur in several regions of the Earth as in northern
Italy or the Western Mediterranean, but the dynamics and physical parameters
characterizing the evolution of such systems are poorly studied.
This
work is based on 3D numerical and analogue experiments of small spatial scale
subduction systems. Several single (one plate) and double (two plates) subduction
models have been performed and analyzed. The objective is to compare and
complement both methods to provide new insights into the analogue modelling of
subduction systems and to better understand the main factors characterizing the
evolution of double polarity subduction systems and related mantle flow.
 |
| Figure 2. Scheme of double polarity subduction experiment performed in the laboratory. Plate 1 and Plate 2 subduct in opposite directions. |
At
large temporal scale, the
dynamic evolution of the Earth’s interior can be modeled as a viscous flow
problem. Accordingly, laboratory experiments consist on both linear viscous
syrup and silicone putty representing the mantle and the subducting plates,
respectively. For simplicity, the overriding plate is ignored. Single and
double subduction models are performed in a Plexiglas tank of 150 x 150 x 50 cm3
(Figure 2). Plates are fixed at their trailing edge to enforce rollback
behavior (retreating plates) and the 660 km lower mantle discontinuity is
simulated by placing a fixed based at 11 cm depth. The width of the plates
varies from 10 cm to 30 cm (600km to 1800 km in nature). Subduction is started
by pushing down manually the leading edge of the plates into the syrup and the
process continues due to the higher density of the plates with respect to the
syrup.
Numerical
model setup is similar to the above described laboratory experiments. Several 3D
numerical simulations of single and double subduction systems with varying size
of the box domain, boundary conditions, viscosity and plate thickness have been
performed.
Laboratory
experiments of double-polarity subduction show that trench (the contact between
the subducting plate and the surface) velocities increase while trenches are
approaching (phase 2) and decrease when trenches diverge (phase 3). This effect,
produced by the asymmetrical pattern of the induced mantle flow, does not occur
in single subduction models. Moreover, both single and double subduction models
show that trench curvature increases linearly with time showing an unusual
strong curvature for the wide plate models (30 cm; ̴1800 km in nature) comparing with previous
laboratory experiments of single subduction.
 |
Figure 3. Temporal
evolution (top view) of double polarity subduction model with 10 cm
wide plates
carried out in the laboratory and by numerical modelling. Color arrows indicate
the mantle velocity field at 5 cm depth.
|
On the
other hand, numerical results show that variable box sizes do not produce major
differences in the evolution of a double polarity subduction system. A box domain of 80 x 80 cm
is enough to simulate accurately the laboratory experiment showing similar
mantle flow pattern than in the large box model (Figure 3). Moreover, the
interaction between the return mantle flow in a double subduction systems is
studied by quantifying the stress and velocity field in the mantle (Figures 3&4).
Our results show that two flow cells in opposite direction occur in the
inter-plate region, decreasing in size during phase 2 (approaching trenches)
and increasing during phase 3 (diverging trenches). Finally, numerical models
of single subduction indicate that a thinner plate fits better the observations
made from laboratory experiments arising the question whether the thickness of
viscous plates may be modified in the laboratory during experiment preparation
(Figure 5).
 |
Figure 4. Double subduction
model with 10 cm wide plates carried
out by numerical modelling. Color arrows
show the mantle velocity field at
3 cm depth during plates intersection.
|
 |
| Figure 5. Analogue and numerical single subduction models of 30 cm wide plates at late stage of the evolution. |
No comments:
Post a Comment