SCEC Award Number 16274 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Long-term behavior of faults with heterogeneous strength
Investigator(s)
Name Organization
Nadia Lapusta California Institute of Technology
Other Participants PhD student Stephen Perry
SCEC Priorities 3c, 3e, 6b SCEC Groups Simulators, GMP, FARM
Report Due Date 03/15/2017 Date Report Submitted 05/10/2017
Project Abstract
 Dynamic rupture simulations suggest that fault heterogeneity can strongly influence dynamic rupture and earth-quake patterns. Its effects are typically studied in simulations of isolated dynamic events. To study the long-term effects of heterogeneity, we simulate earthquake sequences and slow slip in fault models with laboratory-derived friction laws, including enhanced co-seismic weakening due to shear heating. We find that large earthquake events can penetrate into deeper creeping regions, if enhanced co-seismic weakening is activated. Such deeper penetration results in lack of seismicity concentration at depth in the interseismic period; the seismicity would be expected otherwise to concentrate at the bottom of the seismogenic zone. This simulated behavior is consistent with observations for several major SAF fault segments with large historical events, suggesting that those segments can host large events with deeper slip. Such models feature a broad locked-creeping transition at the deeper fault extensions, with the upper limit of the transition getting shallower with time and promoting microseismicity, and the bottom of the transition getting deeper. A single locking depth estimated using a commonly used dislocation model deepens with interseismic time and approximately corresponds to the depth with creeping rates equal to the half of the long-term fault slip rates. It also approximately corresponds to the depth extent of large earthquakes. Despite the increasing dynamic weakening in larger events, the events of different sizes have comparable static stress drops, due to systematic variations in average initial stresses.
Intellectual Merit Dynamic rupture simulations suggest that fault heterogeneity can strongly influence dynamic rupture and earth-quake patterns. Its effects are typically studied in simulations of isolated dynamic events. To study the long-term effects of heterogeneity, we simulate earthquake sequences, slow slip, and their interaction in fault models with laboratory-derived friction laws, including enhanced co-seismic weakening due to shear heating. Our simulations resolve all the stages of every earthquake in detail, including nucleation, dynamic rupture propagation and arrest, as well as reproduce post-seismic slippage and interseismic creep. We find that large earthquake events can pene-trate into deeper creeping regions, if enhanced co-seismic weakening is activated. Such deeper penetration results in lack of seismicity concentration at depth in the interseismic period; the seismicity would be expected otherwise to concentrate at the bottom of the seismogenic zone. This simulated behavior is consistent with observations for several major SAF fault segments with large historical events, suggesting that those segments can host large events with deeper slip. Such models feature a broad locked-creeping transition at the deeper fault extensions, with the upper limit of the transition getting shallower with time and promoting microseismicity, and the bottom of the tran-sition getting deeper. A single locking depth estimated using a commonly used dislocation model deepens with interseismic time and approximately corresponds to the depth with creeping rates equal to the half of the long-term fault slip rates. Despite the increasing dynamic weakening in larger events, the events of different sizes have com-parable static stress drops, due to systematic variations in average initial stresses.
Broader Impacts Large-scale dynamic rupture simulations carried out by SCEC teams have the potential to provide novel and critical information for the assessment of seismic hazard in Southern California. The results of this project, when further developed, would (a) provide better understanding of the long-term behavior of faults, including nucleation conditions and seismicity at rheological boundaries; (b) provide better as-sessment of seismic hazard and evaluation of possible extreme events, based on physical models and integrations of laboratory, field and seismological studies; and (c) contribute to the development of real-istic scaling laws for large events. Graduate students have gained valuable research experience by par-ticipating in the project and interacting with the SCEC community.
Exemplary Figure Figure 1. Connecting the depth limits of interseismic locking, microseismicity, and large earthquakes in models of long-term fault slip (modified from Jiang and Lapusta, 2016; 2017). (Top panels) Examples of fault models which incorporate depth-variable friction properties. In both cases, the low-velocity rate-and-state properties are the same, with velocity-weakening (VW) region embedded into velocity-strengthening areas. However, the extent of dynamic weakening (DW) that occurs co-seismically differs, being limited to the VW region in one model (M2, left) and extending deeper into the VS area in the other (M4, right). (Middle panels) The associated coseismic slip distribution during a typical large event (red color scale) and microseismicity before the event (blue circles) and after the event (red circles). Clearly, the microseismicity is suppressed for the model with deeper slip penetration. (Bottom panels) Evolution of the locked-creeping transition in the two models (M2 and M4), as well as in the third model (M1) which has no DW. The depths D0.1, D0.5, and D0.9, at which slip rates reach 0.1 Vpl, 0.5 Vpl, and 0.9 Vpl, respectively, where Vpl is the loading rate, are shown in red, orange, and yellow. The corresponding geodetic locking depth Dglock, estimated from interseismic surface velocity using a 2-D elastic dislocation model and typical data uncertanties, is shown in blue. The black dashed line corresponds to the VW/VS transition. D0.1 corresponds to the microseismicity-promoting front; note that it is mostly below the VW region in M4, suppressing microseismicity. D0.1 and D0.9 diverge toward the late interseismic period, due to the shrinkage of the effectively locked zone and the expansion of the slip deficit zone. Dglock approximately corresponds to D0.5, deepens with time, and bounds the depth limit and potency release of large earthquakes in our models.