RESEARCH ACTIVITIES IN 2001

Summary of US-Japan Cooperative Research for Urban Earthquake Disaster Mitigation 2nd Workshop on Prediction of Strong Ground Motions in Urban Regions, June 9th, 2001

1. what we have done

1-1. total framework of strong ground motion prediction

* proposal of recipe of strong ground motion prediction

how to set source model
how to put path and site effect
treatment of radiation pattern

1-1-1. how to set source model for strong ground motion prediction

* usage of geological information

segmentation of active fault traces
accumulated surface displacement pattern

1-1-2. how to put path and site effect in strong ground motion prediction

1-1-3. treatment of radiation pattern in strong ground motion prediction

* introduction of frequency-dependent radiation pattern

investigation of frequency-dependence of radiation pattern using high-density strong ground motion network data

2. issues and important points we have realized through the project

* number of asperities

3 asperities is enough or should we put more asperities for M 8 earthquakes ?
use of self-similar distribution of slip ?

* Is M 7.5 earthquake really the most dangerous as shown by Dr. Somerville ?

This is related to the problem of the number of the asperities.
Should we put smaller asperities (which generate short-period velocity pulses)for the M 8 earthquake or not ?

* fault segmentation

In case the fault plane has plural segmentations, how should we put asperities according to scaling relation?
(1) regard each segment as a separated fault plane and apply scaling relation to each segment (Dr. Irikura's recipe)
(2) regard the group of segments as a single fault plane and apply scaling relation.
Scaling relation (Somerville et al., 1999) is constructed without a concept of segmentation. At present, we have two actual source inversion results for the earthquakes with segmentations: the 1992 Landers and the 1995 Hyogo-ken Nanbu earthquakes. We should check the scaling relation for these two earthquakes in case of taking segments into consideration.

* difference between surface and buried faulting

importance of dynamic simulation of rupture process to give constrains on how to set physically acceptable source model
shape of slip velocity function
variable rupture velocity
variable rise time
relation between surface slip amount and slip amount of deeper area
depth-dependent stress drop
difference between surface and buried faulting

* target frequency in the numerical simulation of wave propagation in 3D basin is becoming higher

We already have satisfactory codes for simulation.
Ability of computers is becoming satisfactory (??).
Detailed information of underground structure and source is necessary.

* importance of accurate underground structure information

seismic velocity structure survey
reflection and refraction survey
array observation of microtremors
use of downhole array data (e.g. KiK-net, Sendai, ...)
calibration of velocity structure by simulating waveforms of small-size events
--> use if high-density strong ground motion observation network shallow Q-structure
use of downhole array data (e.g. KiK-net, Sendai, ...)
(but we have to be careful to contribution of scattering effect when assuming 1-D structure)

* usefulness of high-density strong ground motion network (c.g. K-net and KiK-net) data

investigation of frequency-dependence of radiation pattern
calibration of underground structure
accurate source model used for the calibration of underground structure

3. what we have to do from now

* continuous investigation of difference between surface and buried faulting

investigation from observation: waveform records and source inversion results
investigation from theory: dynamic simulation of rupture process

* use of high-density strong ground motion network (c.g. K-net and KiK-net) data

investigation of frequency-dependence of radiation pattern
get accurate source model of small-size earthquake whose records are used for the calibration of underground structure

* continuous development of dynamic rupture process simulation to get information

which can constrain source model from the physical viewpoint

* continuous effort to bring the deterministic simulation to higher frequency

necessity of underground structure survey to get detailed information
calibration by simulating waveforms from small-size event whose source model is determined accurately.

* continuous investigation of frequency-dependence of radiation pattern

use of high-density strong ground motion network (c.g. K-NET and KiK-net) data
difference of frequency-dependence of radiation pattern among various regions

* continuous investigation of slab earthquakes

Is scaling relation different from inland earthquake necessary ?
high stress drop ?
large high-frequency radiation ?
deep Q-structure must