Kompetenznetzwerk für wissenschaftliches Höchstleistungsrechnen in Bayern


NewWave: New Roads for Computational Wave Propagation


Prof. Dr. Heiner Igel
LMU München


During the first year of GeoPF- NewWave we developed realistic physics-based dynamic rupture earthquake scenarios shedding light on devastating, and puzzling, recent earthquakes. We focus on two events in very different tectonic settings: i) the 2004 Sumatra-Andaman Mw 9.1-9.3 megathrust earthquake and ii) the 2016 Mw 7.8 Kaikoura earthquake. We solve the coupled dynamic rupture and wave propagation problem using the freely available software SeisSol (Dumbser and Käser (2006), Pelties et al. (2014), https://github.com/SeisSol/SeisSol). SeisSol employs fully adaptive, unstructured tetrahedral meshes to combine geometrically complex 3D geological structures, nonlinear rheologies and high-order accurate propagation of seismic waves. Our multi-physics earthquake scenarios illustrate how physics-based, dynamic source modeling advance our understanding of the fundamental physics governing complex earthquake behaviour in a subduction setting or for complex cascading rupture sequences. The developed 3D dynamic rupture models feature a new degree of realism and are constrained by independent observational data sets. Physics-based modeling, reveals unexpected features and constrains competing hypotheses. In both cases, we show that detailed fault geometries, combined with smoothly initialized regional stress states allow explaining rupture characteristics to first-order. Our models also demonstrates the effects of laboratory constrained physics of frictional failure of rock, such as off-fault plasticity or rapidly weakening friction, on rupture dynamics. During this project period, we verified the multi-physics capabilities of SeisSol (Harris et al., 2018). We also developed efficient algorithms for constraining the initial conditions of large-scale dynamic rupture models in terms of fault stress and strength presented in Ulrich et al. (2018). In Aochi et al. (2017), we discuss the stress accumulation level, one of the key input parameter in our simulations, and propose a procedure to constrain it. Being able to identify the conditions that lead to large previous earthquake is an important step towards generating realistic physics-based scenarios of future earthquakes in order to complement empirical seismic hazard assessment methods. We furthermore developed off-line coupling methodologies to combine the earthquake simulation tool SeisSol to global seismic wave propagation methods and tsunami simulation codes. Developments made during the project should facilitate setting-up realistic physic-based earthquake models shortly after an earthquake occurs, making physics-based interpretations an important part of the rapid earthquake response toolset.