Andréa Tommasi (Géosciences Montpellier): "Strain localization in nature, experiments, and models"

Strain localization is the rule rather than the exception in the lithosphere. Yet, predicting strain localization at different spatial scales in polycrystalline materials (rocks, but also ice and metals), which deform by ductile processes (without loss of continuity), as 70-90% of the lithosphere, remains a real scientific challenge. More than 50 years after the scientific revolution that established Plate Tectonics as the main paradigm in Earth Sciences, we are still not able to correctly model the deformation of tectonic plates, which is extremely heterogeneous at all scales. This hinders our predictive power on:

1.how and why Plate Tectonics started,

2.the dynamics of subduction zones, mid-ocean ridges, and mountain belts,

3.rifting and the architecture of passive margins,

4. faults/earthquake dynamics, which is modified by coupling with ductile deformation at depth.

In this presentation, I will review observations of ductile strain localization at various spatial scales in nature and experiments and how strain localization is currently obtained in geodynamical models. Then I will discuss why self-consistent prediction of strain localization in geodynamical models remains a challenge and present the solutions we are developing in the frame of the ERC RhEoVOLUTION to model how heterogeneity and anisotropy in the mechanical behavior of rocks control strain localization from the cm to the tens of km scale in the Earth.