HR: 08:05h
AN: S41B-01 INVITED     [PDF]
TI: Energy Partition and Variability of Earthquakes
AU: * Kanamori, H
EM: hiroo@gps.caltech.edu
AF: California Institute of Technology, Seismological Laboratory MS252-21, Pasadena, CA 91125 United States
AB: During an earthquake the potential energy (strain energy + gravitational energy + rotational energy) is released, and the released potential energy ($\Delta W$) is partitioned into radiated energy ($E_R$), fracture energy ($E_G$), and thermal energy ($E _H$). How $\Delta W$ is partitioned into these energies controls the behavior of an earthquake. The merit of the slip-weakening concept is that only $E_R$ and $E_G$ control the dynamics, and $E_H$ can be treated separately to discuss the thermal characteristics of an earthquake. In general, if $E_G/E_R$ is small, the event is ``brittle", if $E_G /E_R$ is large, the event is ``quasi static" or, in more common terms, ``slow earthquakes" or ``creep". If $E_H$ is very large, the event may well be called a thermal runaway rather than an earthquake. The difference in energy partition has important implications for the rupture initiation, evolution and excitation of long-period ground motions from very large earthquakes. We review the current state of knowledge on this problem in light of seismological observations and the basic physics of fracture. With seismological methods, we can measure only $E_R$ and the lower-bound of $\Delta W$, $\Delta W_0$, and estimation of other energies involves many assumptions. $E_R$: Although $E_R$ can be directly measured from the radiated waves, its determination is difficult because a large fraction of energy radiated at the source is attenuated during propagation. With the commonly used teleseismic and regional methods, only for events with $M_W>7$ and $M_W>4$, respectively, we can directly measure more than 10% of the total radiated energy. The rest must be estimated after correction for attenuation. Thus, large uncertainties are involved, especially for small earthquakes. $\Delta W_0$: To estimate $\Delta W_0$, estimation of the source dimension is required. Again, only for large earthquakes, the source dimension can be estimated reliably. With the source dimension, the static stress drop, $\Delta \sigma_S$, and $\Delta W_0$, can be estimated. $E_G$: Seismologically, $E_G$ is the energy mechanically dissipated during faulting. In the context of the slip-weakening model, $E_G$ can be estimated from $\Delta W_0$ and $E_R$. Alternatively, $E_G$ can be estimated from the laboratory data on the surface energy, the grain size and the total volume of newly formed fault gouge. This method suggests that, for crustal earthquakes, $E_G/E_R$ is very small, less than 0.2 even for extreme cases, for earthquakes with $M_W>7$. This is consistent with the $E_G$ estimated with seismological methods, and the fast rupture speeds during most large earthquakes. For shallow subduction-zone earthquakes, $E_G/E_R$ varies substantially depending on the tectonic environments. $E_H$: Direct estimation of $E_H$ is difficult. However, even with modest friction, $E_H$ can be very large, enough to melt or even dissociate a significant amount of material near the slip zone for large events with large slip, and the associated thermal effects may have significant effects on fault dynamics. The energy partition varies significantly for different types of earthquakes, e.g. large earthquakes on mature faults, large earthquakes on faults with low slip rates, subduction-zone earthquakes, deep focus earthquakes etc; this variability manifests itself in the difference in the evolution of seismic slip pattern. The different behaviors will be illustrated using the examples for large earthquakes, including, the 2001 Kunlun, the 1998 Balleny Is., the 1994 Bolivia, the 2001 India earthquake, the 1999 Chi-Chi, and the 2002 Denali earthquakes.
DE: 7200 SEISMOLOGY
DE: 7209 Earthquake dynamics and mechanics
DE: 7215 Earthquake parameters
DE: 7260 Theory and modeling
DE: 8164 Stresses--crust and lithosphere
SC: Seismology [S]
MN: 2003 Fall Meeting