Detailed Session Information

Session 4.1:

Inversion and assimilation of geodetic data

Convenors: T. Sagiya (Japan), B. Minster (USA)

Session Overview: Jean-Bernard Minster, UCSD (15 minutes)

In Session 4.1, we will discuss how to incorporate various observations 
into numerical simulations. Following are the main themes to be discussed:
  1. Data assimilation technique used in earthquake simulations
  2. Combination of geodetic observations
  3. Application of inversion techniques in earthquake simulations
  4. Database system for earthquake simulations

1. Data assimilation technique used in earthquake simulations:

In order to conduct realistic simulations, we definitely need to 
incorporate observations into numerical models. Data assimilation is 
therefore an indispensable part connecting observation and numerical 
models. The procedure includes updating observables in numerical models, 
providing constraints on the models, and revising the models themselves. 
Various kinds of analysis techniques, such as geophysical inversion, 
simulated annealing, and genetic algorithms, may be used in data 
assimilation. We will discuss these methods and try to formulate a 
general framework for the overall problem to be solved. 
In addition, to improve the ability of simulation models to forecast 
future crustal activities, data assimilation must be conducted in real 
time. We will discuss what we can learn from meteorology and 
oceanography in terms of real time data assimilation.

2. Combination of geodetic observations:

Nowadays various types of geodetic observation are available such as GPS,
InSAR, VLBI, SLR, borehole strainmeter, creep meter, etc. Sometimes we 
also  need to incorporate traditional geodetic data that are sparsely 
sampled in time and perhaps of lesser quality, because these data span 
a long time period and include precious records of rare seismic events. 
Each type of observation offers its own advantages so that efficient and 
optimal combinations of these data are desirable. We will discuss what 
types of data are necessary in earthquake simulation and how we can 
combine them.

3. Application of inversion techniques in earthquake simulations:

Inversion technique is an important tool to incorporate various 
observations into numerical models. Among various inversion techniques, 
geodetic inversion has a special significance because we have a explicit 
representation theorem between deformation of the earth and earthquake 
sources based on dislocation theory. We will discuss usage as well as 
limitations of inversion techniques in relation to earthquake 
simulations. We also discuss technical and theoretical development 
required for the  earthquake simulations.

4. Database system for earthquake simulations:

In order to conduct data assimilation in an efficient way, it is 
essential to have a well-designed database system. Database systems are 
expected to contain all the information about earth structure, various 
types of observations, parameters controlling fault friction and failure 
laws, and  so on. In order to be effective, databases should be designed 
so as to guarantee quick and efficient response to various queries. We 
will also discuss how to deal with incoming observation data in real 

Session Plenary: (1 Hour)

David Jackson, UCLA (30 minutes)
Geophysical inverse problem (title is tentative)

Takeshi Sagiya, GSI (30 minutes)
Continuous GPS Array of Japan and its Application 
to Crustal Activity Modelling

Detailed Session: (2.5 Hours)

1. Data assimilation in earthquake simulations

Jean-Bernard Minster and Michael Watkins (10 minutes)
Data Assimilation and Options in Earthquake Modeling

2. Combination of different geodetic observations

Andrea Donellan, Gregory Lyzenga, Don Argus, Gilles Peltzer, 
Jay Parker, Fran Webb, Mike Heflin, and James Zumberge (15 minutes)
Use of GPS and InSAR technology and its further development in 
earthquake modelling

3. Application of inversion techniques in earthquake simulations

Mitsuhiro Matsu'ura, Akira Nishitani, and Yukitoshi Fukahata (15 minutes)
Slip history during one earthquake cycle at the Nankai subduction zone 
inferred from the inversion analysis of levelling data with a 
viscoelastic response function

Peng Cheng Liu and Ralph Archuleta (15 minutes)
Nonlinear Inversion for Earthquake Rupture Parameters

4. Database system for earthquake simulations

Akitake Makinouchi (10 minutes)
Database system for the Earth Simulator

Discussion and comments (1 hour 25 minutes)