Exploring the effects of sub-surface fault geometry on rupture propagation and long-term fault behavior

Geometric complexity plays an important role for a fault’s seismotectonic behavior as it affects the initiation, propagation and termination of an earthquake and influences stress-slip relationships, fault-segment size, and the probability of multi-segment rupture. Consequently, geometric fault complexity is studied intensively and increasingly incorporated into computational earthquake rupture simulations. These efforts reveal a problem: While a natural fault’s geometry may be well quantifiable at the surface (i.e., the fault trace), its down-dip buried portion cannot be well constrained. At this point, it is not clear how this epistemic uncertainty affects the propagation of individual ruptures and a fault’s seismotectonic behavior (e.g., large-earthquake recurrence).

We address this issue computationally with a physics-based multi-cycle earthquake rupture simulator (MCQsim), enabling us to investigate various aspects of rupture propagation and earthquake cycle in a controlled environment (e.g., with well constrained fault geometry). We approximate fault geometric complexity as a 2-D random field using the “random midpoint displacement” method which allows us to represent fault roughness (i.e., incorporate its epistemic uncertainty) while keeping the fault surface trace unchanged. 

Using MCQsim, we create 20kyr-long synthetic earthquake catalogs for strike-slip faults that share the same complex fault surface trace but have different sub-surface fault geometries. We analyze the resulting variations in single-event rupture propagation (i.e., the kinematic source model) and long-term seismotectonic behavior. We find that kinematic source models of individual events differ substantially between different realizations of sub-surface geometry. However, the long-term seismotectonic behavior (e.g., large-earthquake recurrence) does not differ as much and is less sensitive to the epistemic uncertainties of sub-surface fault geometry.