RELATIONSHIPS BETWEEN DEEP AND SURFACE PROCESSES IN SUBDUCTION ZONES: INSIGHTS FROM NUMERICAL MODELS

Abstract

It is widely accepted that the gravitational instabilities of cold, dense, subducting lithospheres constitute the main driving force for plate tectonics. Unravelling the complex processes of subduction zones which originate from the down-going plate and its interplay with the surrounding mantle remains challenging. Due to the lack of direct observations and measurements of the internal Earth, it is difficult to understand how slabs behave at depth. However, seismological and gravimetric measurements provide indirect evidences about the mantle structure and its movement. On the other hand, petrology and geochemistry provide insights on the composition of the Earth interior. Although, these data may be considered as a snapshot of the ongoing processes. The use of numerical modelling is thus a useful ingredient to provide a dynamic picture of the interior of the Earth, integrating the available indirect observables. This thesis is focused on: (1) analysing how the surface fingerprint of the upper plate may be affected by the deep slab dynamics, especially once slabs starts to interact with the mantle transition zone. Thanks to 2D numerical models, it is systematically investigated the role of the mantle rheology (Clapeyron slopes and lower mantle viscosity) as well as plate strength (initial thermal age of plates) on the slab dynamics. In the chapter 3, four categories of slab dynamics are found resulting from the combination of the geodynamics variables mentioned above. These slab categories depict that transient slab dynamics (i.e., slab folding, slab avalanche) have a profound impact on the large-scale dynamic topography of the upper plate in time by pulling up and down the upper plate. Additionally, the dynamic topography shows a tilting of the upper plate towards and fromward the trench following the slab folding at depth. (2) 2D numerical outcomes which are exported to the analysis of the Neotethys subduction system, in particular of the central Iran from ~100 to 30 Ma. Results show that the slab folding behaviour into the transition zone is a reasonable scenario for explaining the particular episodic back-arc closure and opening of the central Iran as well as large-marine flooding events in the Eocene and early Miocene, before the onset of the collision. (3) 3D numerical models of subduction with the aim to understand if and how the slab may evolve lateral with simple model of flattening slab into the mantle transition zone. These models demonstrate that the width of the subduction system have an impact on the way that slab propagate down into the mantle in the centre of wide model domain and stagnate on the leading plate edge. Moreover, I highlighted that the lateral 3 variability of slab morphology from penetrating to stagnation into the mantle transition zone control the dynamic topography of the upper plate. Finally (4) this thesis ends with outlooks about more complex 3D models set-up which may help us to better constrains the lateral variability of deep slab morphology and thus better understand specific surface fingerprints characterizing the natural prototype.