|
Book -- July 1 1999 deadline
|
|
Computing in Science and Engineering Journal Issue
|
|
KDI -- May 17 deadline
|
|
$100K -- Initial Project
|
|
International Links -- Japan, Australia
|
|
Multiple scales are critical and computation good approach
|
|
Reductionist (normal science) v. constructionist (computing -- emergent phenomena)
|
|
Role of Observation and Model
|
|
How do you evaluate probabilistic predictions
|
Components of GEMCI -- must be modular
-
data assimilation, analysis, web-based solvers, visualization, data, modules for friction etc.
|
|
|
Working groups at Queensland Meeting Jan-Feb 99
-
data assimilation
-
algorithms
-
applications
-
microscopic physics
-
macroscopic physics
|
|
Why GEM -- Analogy to Weather community
|
GeoVision 2000: Future of field -- computation
-
Out of ~10 initiatives, only one solid earth
|
|
Santa Fe October 97: first SCEC contacts
|
|
San Francisco December98: GEM is a geophysical enterprise and not a specialized simulation activity
|
March 10 99 USC SCEC Steering committee
-
GEM part of earthquake physics but perhaps too big?
|
Future of SCEC: 3 more years to go but NSF will not fund existing centers
-
Goal is that NSF is catalyst for new ideas
-
Working with NSF on new funding models
-
Significant personnel changes at NSF
|
|
climate, oil reservoir are related more mature fields
|
traditional approach based on "fundamental laws" limited
-
Need close coupling with observation
|
|
Seismic wave propagation through complex media successful
|
|
fault rupture immature - why?
|
|
Has 6 major questions to be answered
|
|
Need to standardize problems
|
Encourage collaboration in numerical simulations of earthquake system
-
5 proposed areas: Seismic Wave propagation, Stress transfer, Rupture dynamics on single faults(quite sophisticated 3D calculations now), Earthquake sequences in fault systems, Damage mechanics
|
|
|
Klein: Role of Simple Models? Do these first?
-
Use this to test importance of friction laws etc.
-
look at hierarchy of overlapping models: slider, CA, full featured continuum simulation
-
robust features in simple models can either be wrong or irrelevant!
-
To get believable results, one must have multiple efforts
|
|
Whitcomb: Climate community has a multiscale activity where simulations at several scales are linked via boundary conditions
|
|
Before large scale simulations, need to clarify disagreements in basic simulations ?
|
Sammis: science is critical
-
science is simple and elegant
-
makes predictions
-
can be falsified
|
Large catalog of observations in need of modeling
-
data patterns (in seismicity) suggest there is something to predict
|
Lisa Grant, Kerry Sieh on integrating data with GEM
-
larger earthquakes are most important societally
-
data + models,GEMCI gives validated predictions
|
|
Many gaps in data space needed for GEM -- 10 year proposal from Kerry
|
|
They need good database technology (GEMXML)
|
|
|
|
NPACI set up to help with computer time
|
|
Has presentation from ACES on role of HPCC in earthquake modeling
|
|
Relation to national approaches
|
Identify Disciplines, Questions, "Collaborations" (a.k.a. multidisciplinary efforts to do particular application areas)
-
Seismic wave propagation -- SCEC "shoot outs"
-
Stress Transfer -- "SCEC workshop"
-
Rupture Dynamics, Earthquake Sequences, Damage mechanics less advanced
|
|
Catalog science and resources, via question and collaboration 2D matrix
|
|
Correlation (the result) versus interactions (direct link)
|
|
Japan will build 1010 element FEM models of mantle
|
|
Interactive models useful for Stress Transfer
|
|
Would like shoot out on same problem of different single fault rupture dynamics
|
|
After shocks important -- not present in simple models
|
|
Damage mechanics is study of properties of material as a result of earthquake related phenomena
|
|
Map collaborations to multiscale diagram
|
|
I suggest integration as additional needed activity -- not well received
|
|
Activity relating applications (called collaborations) to questions
|
|
Skepticism by Leon Knopoff on utility of large scale computation before we understand basic principles
|
|
Prediction (probabilityᢈ%?) versus Forecasting!
|
3 Basic Science
-
Seismic wave
-
Single Fault
-
Regional Fault
|
|
Interactions
|
3 Stages of computing
-
Computational Collaboratory
-
Data Inversion
-
Data Assimilation
|
|
Make case based on this vision but don't prejudge implementation
|
|
National focus requires care to be attentive
|
|
April 21-26 meeting
|
|
Technical meeting later
|
|
NPACI Earth Science group would like to continue
|
Two issues addressed this meeting
-
KDI Scope
-
SCEC Involvement
|
KDI wants infrastructure and a pilot project to test it
-
Stress/Seismicity, Rupture Mechanics, Waves, Data mining
-
Emphasize that many different approaches to simulation
|
Outside Index
Summary of Material
Denote Foils where Image Critical
Denote Foils where HTML is sufficient
