|
Thrust Area 5: Earthquake Physics
The primary objective of the Earthquake Physics Group is to provide theoretical foundations and physical models to support advances in earthquake hazard assessment. In pursuit of this goal, the Working Group conducts research aimed at achieving the following: (i) Improved models of time-dependent rupture and slip in individual earthquakes. These models aid estimates of earthquake potential, constrain ground motion simulation methodologies, and reveal fundamental source processes by connecting kinematical observations (from seismology, geodesy, and geology) with fault-zone physics. (ii) Improved models for the space-time evolution of earthquake sequences, fault system structure, and regional stress fields. These models are aimed at improving our understanding of the nature of short- and long-range interactions and correlations among seismic events, and have the long-term goal of improving time-dependent probabilistic estimates of earthquake potential.
Overview
Progress on this program during 1999 includes theoretical advances, advances in numerical modeling methodology, and progress in the observational testing and validation of rupture models. On the theoretical side, work on the emergence of continuum complexity in earthquake sequences has achieved an important consensus (Harvard, Lamont). Additional theoretical work has addressed the origin of short slip rise times (Heaton pulses) and clarified the parameter range over which short slip pulses can arise due to friction with strong velocity dependence (UCSB). Theoretical work has also substantiated our earlier results demonstrating that short slip pulses arise naturally in response to heterogeneity of fault zone stress and strength, even with conventional friction laws (UCSB, SDSU). Progress has also been made in using a continuum approach to model the development and evolution of fault systems, leading to new insights into the onset of disorder, accelerated moment release, and long-range stress correlations (USC). An important achievement this year in numerical modeling methodology was the development of an integrated methodology for dynamic simulation of the earthquake sequences in a faulted continuum, including interseismic tectonic loading, nucleation and dynamic rupture, and post-seismic deformation, creep, and reloading (Harvard). Also significant was development of a regularized numerical method for modeling rupture along a material interface (Harvard). In the area of observational assessment of models, major efforts are in progress to test fully dynamic models for the 1992 Landers, 1994 Northridge, 1999 Izmit, and 1999 Chi-chi (Taiwan) earthquakes (UCSB, SDSU). A few selected highlights are described further below.
Criticality of Rupture Dynamics
Dynamic simulations by the UCSB group (Olsen and colleagues) have led to the proposal that earthquakes occur when stress conditions are near a critical condition that can be described by a single non-dimensional parameter. This idea has been explored for both the Landers and Northridge earthquakes. For the Northridge earthquake, they find that a critical balance between initial conditions and friction parameters has to be met in order to obtain a moment as well as a final slip distribution in agreement with kinematic slip inversion results. The model rupture process is strongly controlled by the average stress and connectivity of high-stress patches on the fault. In particular, a strong connectivity of the high-stress patches is required in order to promote the rupture propagation from the initial nucleation to the remaining part of the fault. Similar dynamic modeling of the Landers rupture, constrained by strong motion observations, demonstrated that rupture evolution is highly sensitive to prestress distribution, and also requires a critical level of prestress and connectivity.
Slip Pulses
Ongoing theoretical and observational studies by the Working Group have lead to an improved understanding of the conditions required for rupture to occur in a pulse mode (Heaton pulse). Earlier studies (Rice and colleagues) had shown that strongly velocity-weakening friction, plus so-called understressing conditions, promote the pulse mode. 3D studies this year by Nielsen and colleagues at UCSB, using a different parameterization of the fault, lead to similar conclusions: the pulse mode emerges in a velocity-weakening model in a restricted parameter range, with lower stress favoring the pulse mode. Simulations of the Northridge earthquake by Nielsen and Olsen, however, require that only a small amount of velocity weakening occur. In their models of the Northridge and Landers earthquakes, pulse-like behavior reflects instead the length scales of stress and strength distributions. This supports earlier SCEC studies (Day and colleagues) who concluded that stress/strength heterogeneity could account fully for slip pulse observations.
Continuum Complexity
Efforts to understand the origin of complexity in earthquake sequences have been a major focus of the Working Group almost since the inception of SCEC. A key theoretical goal has been to determine under what conditions complexity can emerge in the absence of geometrical disorder, i.e., solely from the nonlinear dynamics of frictional contact within an elastodynamic continuum. Shaw and Rice have completed an important synthesis which gives the most comprehensive answer to date to this question. Complexity they designate type I, a broad distribution of large event sizes with nonperiodic recurrence, occurs when the modeled region is very long, along strike, compared to the seismogenic layer thickness. Complexity of type II, numerous small events showing a power law distribution, occurs only in a restricted range of parameter space. Complex constitutive laws, with small strength drop at small slips (or rates) and large strength drop at larger slips (or rates) are required. Nucleation from slip weakening and time dependent weakening lead to similar large scale behavior.
In a related development, a numerical method developed by Lapusta and colleagues provides a new tool for rigorous elastodynamic simulation of earthquake sequences under conditions of rate/state friction. The method treats accurately, within a single computational procedure, loading intervals of thousands of years. It can model, for each earthquake episode, initially aseismic accelerating slip prior to dynamic rupture, the rupture propagation itself, rapid postseismic deformation which follows, and also ongoing creep slippage throughout the loading period in velocity strengthening fault regions.
Rupture of Segmented Faults
Since 1991, SCEC modelers have been examining the conditions under which earthquakes are likely to rupture through segment boundaries such as stepovers. This question is important to seismic hazard assessment, since it determines the probable sizes of future earthquakes on a fault system. That modeling has shown that stepovers of a km or two are relatively weak barriers to rupture, likely to break in large earthquakes (as subsequently occurred in the 1992 Landers earthquake, for example). During the past year, the modeling has been extended to 3D strikeslip systems (Harris and Day), and also extended to include segmented thrust fault systems (Magistrale and Day). The Izmit, Turkey earthquake of 1999 provides a opportunity to further test the 3D models, and that work is in progress.
|
system
Southern California Earthquake Center
@ 