### Research

The main subject is "dynamic rupture process of earthquake source." An earthquake source, which is something between frictional slip and fracture of intact rock, is quite complex and diverse. There are many basic questions unsolved. One important question is what makes the difference between small and large events. Why one earthquake can be so large while many remain small. The simple answer is that is the nature of complex process, but some properties of earthquake also suggest that this is not the complete answer.

We analyze the detail of real earthquake sources using observed seismic data, recent dense and high-quality digital records, to see what actually happened during the events. From a gigantic event of magnitude 8 to a rock burst in mine of magnitude 0, we develop tools for waveform inversion and apply for each event. Some parameters that represent average property of an event are also important. Among them, we are now interested in seismic energy, which is energy radiated and transported to far-field.

We guess the background physics from detailed source process. Friction law governing the source process is the important target of our study. For example, calculating energy balance during rupture, we can estimate fracture surface energy on the fault plane. Numerical simulations based on assumed fault geometry and frictional properties provide clues for understanding dynamic rupture process.

It is also important to utilize our knowledge for disaster prevention. We have tried a near-realtime waveform analysis and studies of dynamic source parameters for more reliable strong motion simulation.

Page top## Earthquake Dynamics

An earthquake ruptures following natural laws of fracture and friction, which are poorly understood. We analyze seismic waveform to reveal detailed source process. These source process implies the governing law of earthquake ruptures. One realization is the relation between fault slip and stress, which is a kind of friction law or constitutive relation. The important parameters, such as critical slip distance and surface fracture energy, can be estimated using source models. These parameters are also important to assess the energy balance during an earthquake rupture process.

The left figure shows distribution of surface fracture energy estimated for the 1995 Kobe, Japan, earthquake (Ide, 2003). Surface fracture energy determines local rupture speed and whether rupture extend or not locally. One of large area of fracture energy corresponds to Akashi straits, a local geographic irregularity. The order of fracture energy is MJ, which is larger than lab. experiments by 2-4 orders. This is an example suggesting that the fracture energy scales with earthquake size.

Page top## Seismic Energy

Usually, earthquake size is measured by seismic moment, which is a static measure of faulting process. It is also well known even the earthquakes of same seismic moment look quite different. The left figure is an example of seismic moment time function for a pair of events having the same seismic moment (Mw 7.6); the 1992 Nicaragua earthquake (Ide et al., 1993) and the 1993 Kushiro-Oki earthquake (Takeo et al., 1993). The difference of characteristic time of two events is about 10 times. The former is a rare 'tsunami' earthquake and the later is a kind of high stress drop event at intermediate depth. This illustrates the difficulties of measuring earthquakes just by one parameter.

While seismic moment is a static parameter, seismic energy, energy radiated by an earthquake and carried by far-field body waves, is another parameter which describes the power of seismic waves. In the above example, the seismic energy of the Kushiro-Oki earthquake is about 100 times larger than that of the Nicaragua earthquake. However, it is not easy to estimate seismic energy with high reliability and this is a problem we are studying now. A present question about the seismic energy is whether the ratio of seismic energy to seismic moment is constant or not. This looks quite simple and easy to solve, but there are many factors to solve before rigid conclusion. We compiled several studies of seismic energy and made the following figure. This is the ratio of seismic energy to seismic moment, Es/Mo, as a function of seismic moment. Our conclusion so far is there is not a well resolved breaks for the constant Es/Mo ratio scaling (Ide and Beroza, 2001). Nevertheless, it is unlikely that such simple scaling hold down to very small earthquakes and we are doing similar research to find a clear evidence of break of this scaling.

Page top## Waveform Inversion

The left movie is made from the result of inversion analysis of strong motion waveform for the 1995 Kobe, Japan, earthquake (Ide and Takeo, 1997). This earthquake ruptured an area of about 50 km x 20 km within 10+ seconds. The color scales with local slip rate. The maximum of slip is about 2 m and the rupture propagation velocity is about 3 km/s. We can see a bilateral rupture from the Akashi straits toward Awaji Island and the city of Kobe. Similar analysis is possible for many earthquakes and such complexity of seismic source seems to exist in smaller events. Increasing quality and quantity of seismic data eneble us to reveal finer complexities in future.