Valid HTML 4.01!

8.4 Analogic physical modelling

Based on the constatation that exhumation does occur during continuing plate convergence, that the exhumation of the subducted crust is frequently associated with major normal faults (STDS) which accomodate a displacement of up to several tens of kilometres, that these faults are subparallel and dip in the same direction as major thrust faults (MCT) and that both type of faults operate simultaneously, Chemenda et al. (1995) propose an elegant experimental model (Fig. 8.4) which may partly explain the mechanism leading to exhumation in a compressive system.

In this model, plate collision and underthrusting is simulated in laboratory by taking different hydrocarbon compositional systems that reproduce the physical plastic properties of a lithosphere composed of a mantle layer and two crustal layers «floating» on a liquid asthenosphere (water). The collision is driven by a piston and erosion is simulated by scraping off the excess relief (corresponding to values above 6-8 km) formed by this experimental collision.

This experiment shows that at an initial stage of the collision, all layers of the continental lithosphere subduct into the mantle, but with the crust subducting slightly more slowly. At a certain point, the subduction reaches a critical stage, where the crustal sheet stops completely and fails in front of the subducted lithoshere.This sheet then begins to move back up to the surface, in an opposite direction to the rest of the crustal layer and mantle which continue to subduct. The rising crustal sheet overthrusts the subducting plate and a normal shear motion develops along the upper surface of the sheet.

The physical laws that rule this model are quite simple, subduction of the crust is driven both by the force exerted by the piston and by the drag force from the underlying mantle (gravity). These forces are counteracted by the buoyancy force which grows with increasing volume of subducted crust (Archimede's principle) until the upper crust layer fails and a slab of upper crust is detached from the still subducting lithosphere. Push and drag forces do not affect this slab of detached crust any more, whilst buoyancy force remains the same. It is this force which causes the subsequent exhumation of this detached crust slab. Exhumation continues until the now diminishing buoyancy force is counteracted by frictional forces on the sides of the rising slab and by its own weight. As subduction of the lithosphere goes on, a new slab of upper crust is detached and undergoes the same exhumation processes as the former slab, and so on.

The autors of this model have applied these experimental results to the specific case of the Himalaya (Fig. 8.4) and have obtained comparable values in the exhumation rate, timing of events and amount of displacement.

This model is limited, however in its analogy with the reality as it considers the HHCS as a rigid slab and does not incorporate the effects of pure shear, the changes in viscosity due to increasing or diminishing temperature and the probable lubrification of the STDS through partial melting. Moreover, the STDS and the MCT are considered as parallel although they have been proven to converge.

In several ways, the Himalayan orogen can be compared to the Alps. Since the early investigation in the mid-XIX century, the geological knowledge of the Alps has been steadily refined by a large number of geologists resulting in a fairly good comprehension of the mechanism and processes that lead to the building of this mountain belt. Recent deep seismic survey of the Ecors-Crop and NFP-20 programs (Frei et al., 1990; Marchant et al., 1993; Epard and Escher, 1996; Escher and Beaumont, 1997; Steck et al., 1997) allowed to control, to some extent, the geometric extrapolation and reconstruction of the deep Alpine structures. These interpretations of what happened
deep below the Alps provide a good illustration of collisional tectonics and can be extrapolated to the Himalaya, if one makes abstraction of the fact that, unlike to the Alps, Himalayan tectonics are not characterized by a basement-cover relation.

On the basis of the precursory work of Ramsay (1980), Epard and Escher (1996) propose a generic two-dimensional model to explain the possible relationship between superficial thrust sheet and deeper fold nappes and apply this model to the specific case of the Western Alps to demonstrate that the simultaneous formation of these two types of structures at different levels of the crust is possible. Contrary to the gravity collapse and ductile extrusion flow models, which isolate the upper crustal structures and associate extension with gravific instability or a lateral pressure gradient due to the building of a marked topographic relief, the model of Epard and Escher considers the topographic relief to be of minor geologic importance compared to the 40-60 km deep root zone (70 - 80 km in the Himalaya). As with the model of Chemenda et al (1995), lithospheric subduction is here considered to play a major role both for underthrusting and exhumation in the formation of nappes.

This model aims essentially at explaining the initial formation of ductile basement nappes, but the authors suggest, in agreement with Dietrich and Casey (1989), that the simple shear deformation in the lower structural units is often superposed by pure shear. They propose that the effect of an heterogeneous wedge-shaped pure shear would cause an upward extrusion of lower-level rocks and that this mechanism may be responsible for bringing near the surface high-pressure rocks already at an early stage of the continental subduction.

These final considerations of Epard and Escher's can be applied to the Himalaya, because the HHCS was affected by pervasive simple shear, and the effect of pure shear on the wedge-shaped HHCS is a very plausible mechanism for the lateral extrusion of this high-grade domain and could thus explain a relative extentional sense of shear along the STDS.

A model (Fig. 8.5) that associates the effects of both simple shear and pure shear during simultaneous underplating of the lithosphere and exhumation of the upper crust is proposed for the Alps by Escher and Beaumont (1997). These authors take up and elaborate the model of Epard and Escher (1996) to give a geometrical and mechanical explanation for the initiation and subsequent exhumation and stacking of basement nappes. Their model is based on several geological factors and previously published ideas and should also partly be relevant for the Himalaya.

An essential consideration is that originally deep basement rocks display a penetrative early foliation and stretching lineation, implying strong deformation throughout most of the volume rock. Consequently, it is unrealistic to represent large deep-seated rock bodies as having been displaced over significant distances without strong internal deformation.

T he model of Escher and Beaumont for the Alps (Fig. 8.5) ressembles that of Chemenda et al (1995) inasmuch as they also suggest that a slice of upper continental crust is detached at the front of the subducted lithospere and that this slab moves back to the surface while the rest of the lithosphere continues to be subducted. The mechanism of ascent is, however, different as it is proposed that buoyancy forces and erosion only assist the process but are not the dominant factor. Instead it is proposed, similarly to the channel flow model, that exhumation of the detached crustal slab is the result of combined simple shear and pure shear. In other words, once a crustal slab is detached from the subducting lithosphere, further descent of this slab is hindered at a certain point and these now high-grade rocks are then flattened through pure shear between the two converging plates. The flattening of the slab implies a decrease in thickness and an increase in length. The additional lenght cannot be accomodated downwards and the slab is thus laterally expulsed or «squeezed out» to the surface. (see also Steck et al., 1998, where geometrical proofs are given to support this model for the
Tso-Morari region, south of the Indus Suture Zone)



Ductile extrusion-channel flow model Discussion forward

©Pierre Dèzes