mega-thRust Earthquake SystEm Theory

Plate tectonics and mantle convection drive motions and slip buildup along plate boundaries. The locations where slip will be released during earthquakes are inferred from satellite observations. However, inversions ignore plate tectonic and mantle mechanics and much of the available data. Recent vertical velocities suggest megathrust earthquakes in subduction zones could rupture twice closer or farther from the coast. That could make the difference between shock and world-class disaster for one third of the world’s population. To set this straight and unlock earthquake sciences, I propose a paradigm shift by imaging, quantifying and understanding processes as part of an earthquake system. This multi-scale plate tectonic concept couples novel and known interactions between subducting and overriding plates, mantle and the megathrust from milli-seconds to millions of years. Imaging this multi-scale, visco-elasto-plastic system poses a major computational challenge that I will tackle by adapting the latest computer science advances in automatic differentiation and high-performance computing on graphical processing units to earthquake sequences. Adjoint-based slip and medium inversions and an Ensemble Kalman Smoother will assimilate all observations into physical system models. Application and validation in five natural laboratories will quantify system processes, which will be integrated into a unified theory quantifying surface velocities with time.This integrative approach allows for the first probabilistic seismic hazard assessment based on first physical principles. RESET’s Earthquake System concept, methods and theory will equip solid Earth scientists with what is needed to explain recent observations and enable efficient, system-based quantification of future seismicity.