We present the first attempt to explain slow earthquakes as cascading thermal-mechanical instabilities. To attain this goal we investigate brittle-ductile coupled thermal-mechanical simulation on vastly different time scales. The largest scale model consists of a cross section of a randomly perturbed elasto-visco-plastic continental lithosphere on the order of 100 × 100 km scale with no other initial structures. The smallest scale model investigates a km-scale subsection of the large model and has a local resolution of 40 × 40 m. The model is subject to a constant extension velocity applied on either side. We assume a free top surface and with a zero tangential stress along the other boundaries. Extension is driven by velocity boundary conditions of 1 cm/a applied on either side of the model. This is the simplest boundary condition, and makes it an ideal starting point for understanding the behavior of a natural system with multiscale brittle-ductile coupling. Localization feedback is observed as faulting in the brittle upper crust and ductile shearing in an elasto-viscoplastic lower crust. In this process brittle faulting may rupture at seismogenic rates, e.g., at 102-103 ms-1, whereas viscous shear zones propagate at much slower rates, up to 3 × 10-9 ms-1. This sharp contrast in the strain rates leads to complex short-time-scale interactions at the brittle-ductile transition. We exploit the multiscale capabilities from our new simulations for understanding the underlying thermo-mechanics, spanning vastly different, time- and length-scales.
Bibliographical noteFunding Information:
We are grateful for discussions with Bruce Hobbs and Charley Kameyama. This research has been supported by NSF’s ITR and Math-Geo programs and the Western Australian Premiers Fellowship program, the University of Western Australia and the CSIRO Exploration and Mining Division.
- Brittle-ductile transition
- Fast instabilities
- Numerical modelling
- Slow earthquakes