3D Numerical Modelling of the Wilson Cycle: A Study of Structural Inheritance in the Lithosphere

Abstract

The Wilson Cycle entails the cyclic closing and opening of oceans and is a fundamental component of plate tectonics. It implicitly emphasis the importance of tectonic inheritance in geodynamics, which has been confirmed by recent studies. For example, reactivation of extensional inheritance increases the width of orogens and, in turn, inherited heterogeneity of the lithosphere strongly controls the dynamics of rifting and continental breakup. However, the effects of inheritance on subduction, in particular on subduction initiation, remains mostly unkown. In this thesis a high-resolution 3D thermomechanical model - I3ELVIS - is used to study at the scale of the lithosphere the different stages of the Wilson Cycle in a self-consistent way. The model goes through rifting, continental breakup, seafloor spreading, inversion from divergence to convergence, subduction and/or obduction. Focus lies on the understanding of subduction initiation. This work builds on the recent advances in numerical modelling (e.g. grain size rheology) which allow realistic representations of the long-term evolution of the lithosphere. Results show that in the framework of the Wilson Cycle, off-ridge intra-oceanic subduction initiation occurs over a wide range of rheological parameters and kinematics conditions. They show that the initiation of subduction is an intrinsically three-dimensional process controlled by three main factors: (1) the inherited compositional and thermal heterogeneity of the ridge and oceanic lithosphere resulting from rifting and seafloor spreading; (2) the obliquity of the ridge with respect to the convergence direction; (3) the duration of transition from plate divergence to convergence. The modelled mechanisms are consistent with geological records, for example in the Oman subduction-obduction system. Exploring near-ridge subduction initiation enables studying the formation of contra-dipping slabs and their interplay during continuing subduction. Alternating subduction polarity along suture zones has been suggested in several orogenic systems. Yet, the mechanisms prompting such a geometric inversion and the subsequent interplay between the dipping slabs have been sparsely studied. The width of the slab segments is delimited by transform faults inherited from the rifting and ocean floor spreading stages. The models show that the number of contra-dipping slab segments depends mainly on the size of the oceanic basin, the asymmetry of the ridge and variations in kinematic inversion from divergence to convergence. Convergence velocity has been identified as a second order parameter. The geometry of the linking zone between contra-dipping slab segments varies between two end-members governed by the lateral coupling between the adjacent slab segments: (1) Coupled slabs generate wide, arcuate linking zones holding two-sided subduction; (2) Decoupled slabs generate narrow transform fault zones against which one-sided, contra-dipping slabs abuts. Modelled obduction occurs across a wide range of geodynamic settings and convergence velocities when combined with preceding near-ridge subduction initiation. These findings constitute an important advance compared to previous models which either relied on very specific geodynamic settings to produce obduction or were unable to reproduce large (>100km) ophiolites. Furthermore, models of this study elucidated two obduction-controlling factors: (1) Reactivation of geometrical and thermal structures inherited from divergent stages in the continental margin: A "cold" Moho favours long ophiolites (>100km). A "hot" Moho favours small ophiolites. (2) Convergence velocity, which also affects the size of the obducted ophiolite. Long and thin ophiolites are associated with low-strength oceanic lithosphere and slow convergence. Conversely, thick and long ophiolites are associated with fast convergence. These findings hold over a large range of geodynamic settings and produce ophiolite morphologies consistent with those observed in Oman and Southern Quebec. Specifically, both the morphology (i.e. length of >100 km and thickness of ~15 km) as well as the strong inverted thermal gradients match natural examples. Subduction initiation at passive margin is explored by imposing a quiescence period after seafloor spreading. The weak ridge is healed and convergence strain can localize at the continental margin. Results emphasize the importance of tectonic inheritance on subduction initiation at passive margins. Two types of structural inheritance are investigated: (1) Hydrated weak zones and (2) grain-size reduction, the latter by means of a grain-damage model. Subduction initiation at passive margins by reactivation of hydrated weak zones depends on two pre-requisite initial conditions: that hydrated zones are located directly below the Moho and that the lower crust of the continental margin is initially hydrated. Stress concentration in the lithospheric mantle during rifting can induce localized grain-size reduction, which prompts the formation of long-lived damaged zones above the lithosphere-asthenosphere boundary. Because this boundary is deflected by the asthenosphere upwelling during rifting, the damage-zones systematically dip away from the ocean. The reactivation of the damaged-zones as reverse zones during convergence allow for subduction initiation at the passive margin. Results show that this subduction mechanism is controlled by the continental Moho temperature. With a high Moho temperature (700° C) the weak zones partially heal through grain coarsening during seafloor-spreading. The resulting coarse grain, hence strong mantle prevents subduction initiation.