SCEC Community Modeling Environment SCECCME Status Update for June 2004

1
SCEC Community Modeling Environment (SCEC/CME )
Status Update for June 2004

• Philip Maechling
• Southern California Earthquake Center
• University of Southern California

2
Outline

• SCEC/CME Project Goals
• Recent Activities
• Project Results
• Project Discussion

3
Geoscience Components of Seismic Hazard Analysis
Seismicity (ANSS)
Paleoseismology
Geologic structure (USArray)
Local site effects
Faults (USArray)
Seismic Hazard Model
Stress transfer (InSAR, PBO, SAFOD)
Rupture dynamics (SAFOD, ANSS, USArray)
Crustal motion (PBO)
Crustal deformation (InSAR)
Seismic velocity structure (USArray)
4
SCEC/ITR Project
Goal To develop a cyberinfrastructure that can
support system-level earthquake science the
SCEC Collaboratory
Funding 10M grant over 5 yrs from NSF/ITR
program (CISE and Geoscience Directorates) Start
date Oct 1, 2001
NSF
SCEC/ITR Project
USGS
ISI
Information Science
Earth Science
SDSC
IRIS
SCEC Institutions
5
SCEC Computational Pathways
1. Standardized Seismic Hazard Analysis 2. Ground
motion simulation 3. Physics-based earthquake
forecasting 4. Ground motion inverse problem
RDM
FSM
3
AWP Anelastic Wave Propagation SRM Site
Response Model
FSM Fault System Model RDM Rupture Dynamics
Model
6
SCEC/CME Project Schedule
June - 2004 32 months out of 60 months completed
Oct-2002
Oct-2003
Oct-2004
Oct-2001
Oct-2005
Oct-2006
7
Outline

• SCEC/CME Project Goals
• Recent Project Activity
• Project Results
• Project Discussion

8
SCEC/CME NSF Mid-Term Review

• NSF Mid-Term Project Review held at SCEC on March
  3-4, 2004.
• Project Review report was largely positive and
  anticipates the successful completion of the
  Project based on current status.

9
SCEC/CME NSF Mid-Term Review Integration
Demonstrations

• System Integration Demonstration shown at
  Mid-Term
• Pathway 1 demonstration showing creation of
  computation pathway using Ontology-based tool
  (CAT) and execution using grid.

10
SCEC/CME NSF Mid-Term Review Integration
Demonstrations

• System Integration Demonstration shown at
  Mid-Term
• Pathway 2 demonstration showing scenario event,
  simulation using 2 AWM codes, results stored in
  Digital Library, and ability to show results in
  multiple formats including 4D viz., Hazard Maps,
  and seismograms.

11
SCEC/CME Build Schedule
SCEC/CME Build Numbers
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15
Oct-2002
Oct-2003
Oct-2004
Oct-2001
Oct-2005
Oct-2006
March 30 - 2004 SCEC/CME Release 5 Completed
12
Outline

• SCEC/CME Project Goals
• Recent Activities
• Project Results
• Project Discussion

13
NSF Research Goals - People

• Developing a diverse, competitive, and
  globally-engaged U.S. workforce of scientists,
  engineers, technologists and well-prepared
  citizens.

14
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of this approach.
Our undergraduate research group has developed a
very impressive 3-D, GeoWall-based, visualization
tool called LA3D. This software has been
distributed to SCEC researcher for use in
visualization 3D fault catalogs in Southern
California.
The SCEC/CME Project provides research
opportunities to undergraduate students through
the SCEC UseIT (Undergraduate Summer in
Earthquake Information Technology) Program.
Our SCEC UseIT students bring diverse educational
backgrounds including geoscience, computer
science, pre-law, cinema, and history to our
summer-time research program. The program has
generated significant interest in geoscience
research for students in unrelated disciplines.
15
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of this approach.
Graduate and undergraduate students on the
SCEC/CME project have published research papers
and abstracts in several areas including
Geoscience and Computer Science as well as
participating in scientific meeting and Project
meetings.
The SCEC/CME Project has assembled a strong and
diverse collaboration of experienced
Geoscientists, Computer Scientists, graduate
students, and undergraduates that encourages
student participation in Seismic Hazard Analysis
related research and also encourages interaction
between scientific disciplines.
The SCEC/CME Project provides socially relevant
scientific problems as research subjects for
undergraduate and graduate research efforts. The
Project provides students with access to leading
scientific experts and to leading edge technology.
16
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of this approach.
Images and Movies produced by our SCEC UseIT
Intern program have been used in widely
disseminated Earthquake Preparedness publications
such Written in Stone a publicly televised
Seismic Hazard education program.
By generating interesting, colorful, informative,
and scientifically accurate images, maps, and
movies, our SCEC UseIT Intern program
communicates scientific information about Seismic
Hazards to the general public.
Communication of scientifically accurate Seismic
Hazard Information to the general public can help
the large population in Southern California, and
elsewhere, prepare their families and homes for
the next large earthquake in the region.
17
NSF Research Goals - Ideas

• Enabling discovery across the frontier of
  science and engineering, connected to learning,
  innovation, and service to society.

18
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
The series of of increasingly complex
computational pathways we are using as our
research roadmap has established a conceptual
framework for improving Seismic Hazard Analysis.
SCEC/CME researchers have initiated work in each
pathway work area.
SCEC/CME researchers have developed a conceptual
framework for improving Seismic Hazard Analysis
through a series of four, increasingly complex,
computational pathways. Each pathway integrates
an additional geophysical model into our SHA
computations. In addition, we are constructing
the computing and data management capabilities in
a grid-based, computing environment that we call
a Community Modeling Environment.
Our computational pathways converge into a common
endpoint which is socially useful information on
ground motion (Intensity Measures). The SCEC/CME
Computational Pathway model provides a roadmap
for integrating results from previously
independent work areas within geoscience.
19
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
The OpenSHA code developed on the SCEC/CME
Project is currently available for use by
scientists. It is a well-validated, open-source
suite of software that supports the most widely
used Earthquake Rupture Forecasts and Attenuation
Relationships.
SCEC/CME researchers have developed a general
framework (called OpenSHA) for Probabilistic
Seismic Hazard Analysis calculations that can
support the wide variety of existing Attenuation
Relationships, and Earthquake Rupture forecasts
and will support application of new approaches
such as time-dependent Earthquake Rupture
forecasts and Intensity Measure Relationships
built on Wave Propagation codes.
Probabilistic Seismic Hazard Analysis (PSHA)
calculations are needed in such socially valuable
areas as National Seismic Hazard Maps. The
SCEC/CME OpenSHA software automates how PSHA
calculations are done and can significantly
change how such hazard maps are created.
20
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
The SCEC/CME Project has demonstrated a
Compositional Analysis Tool that interacts with a
Knowledge base containing information about the
Projects computational modules and data types.
Through a web-based user interface, a
non-sophisticated user can construct and execute
a grid-based Seismic Hazard Map calculation.
SCEC/CME researchers have applied AI planning
technology to the creation of scientific
computational workflows. A scientist can identify
a starting data type and a desired data type, and
our ontology-based Compositional Analysis Tool
(CAT) will provide suggestions, error, and
warning messages that help the researcher
assemble the appropriate modules and data types
into a valid computational workflow.
Applying AI planning technology to scientific
workflow construction can be applied in any
scientific discipline in which a series of
calculations, data conversions, and data
reductions are performed.
21
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
SCEC/CME researchers have run simulations for
actual Los Angeles area earthquakes and have
validated their results against observed
seismograms for those events. This provides
validation for the ground motions predicted by
these codes for scenario earthquakes, that is,
earthquakes that have been simulated, but that
have not actually occurred.
SCEC/CME researchers are running a large number
of Earthquake Wave Propagation simulation for
both real, and scenario, earthquakes in the Los
Angeles Basin. These simulations produce
synthetic ground motion records comparable to
observed seismograms.
Because the wave propagation simulation codes
used to run these simulations are well-tested and
validated, the ground motions predicted by these
simulations are valuable information to
Earthquake Engineers working on buildings in the
Los Angeles area.
22
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
SCEC/CME Pathway 4 researchers have used data
inversion techniques to calculate Fréchet
sensitivity kernels. These kernels provide
scientists with insights into how to improve the
velocity models of the Los Angeles that are used
in many geophysical calculations.
SCEC/CME researchers are using data inversion
techniques to improve geophysical models, such as
a crustal velocity models, commonly used in
seismological research. This data inversion
technique establishes a feedback loop between
observed data and simulated data in which
observed data helps to improve geophysical models.
Complex and computationally expensive, data
inversion techniques are technology drivers for
other geophysical research activities. As
SCEC/CME researchers address the computational,
and data management, requirements in this area,
they are finding solutions other groups can also
apply.
23
NSF Research Goals - Tools

• Providing broadly accessible, state-of-the-art
  SE facilities, tools, and other infrastructure
  that enable discovery, learning and innovation.

24
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
The SCEC/CME OpenSHA Probabilistic Seismic Hazard
Analysis (PSHA) software supports the Earthquake
Rupture Forecasts (ERF) used in National Seismic
Hazard Maps including the most recent
Frankel-2002 Earthquake Rupture Forecast.
SCEC/CME researchers have developed OpenSHA
software that supports Probabilistic Seismic
Hazard Analysis calculations. This software
provides access, in a single software system, to
a wide variety of Earthquake and Attenuation
models that have been developed and used in PSHA
studies over the last 20 years.
By hosting a wide variety of Earthquake Rupture
Forecasts and Attenuation Relationships in a
single software system, Probabilistic Seismic
Hazard researchers can combine Earthquake Rupture
Forecasts and Attenuation Relations in ways never
before possible.
25
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
The SCEC/CME software called OpenSHA can produce
ground motion predictions for scenario
earthquakes in Los Angeles, and other, areas. The
flexibility of the code allows scientists to use
the attenuation relationships and geophysical
models of their choice in the calculations.
Predictive ground motions for Scenario
earthquakes (earthquakes that have not actually
occurred) are used in a variety of Emergency
Management applications. Our OpenSHA software
provides an easy-to-use, highly flexible, tool
for these calculations.
The SCEC/CME OpenSHA software integrates a
variety of attenuation relationships and
geophysical models into a single software system
that can be used to calculate predictive ground
motions for scenario earthquakes.
26
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
SCEC/CME researcher have exported our Scenario
Shakemap ground motion predictions in GIS
Shape-file format. Through this format, we can
export our results to FEMAs Hazus system. Our
underlying SHA models provide significantly more
capabilities than the SHA capabilities currently
available in these Hazard Estimation Tools.
SCEC/CME researchers have worked with Emergency
Management personnel so that predictive ground
motion data sets generated by our OpenSHA tools
can be imported into Emergency Management tools
such as Hazus.
Integrating OpenSHA ground motion results with
Emergency Management tools such as Hazus provides
Emergency Management Agency with the best Ground
Motion Modeling together with the large
infrastructure and fragility information
available in their existing Loss Estimate Tools.
27
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
SCEC/CME researchers have developed an
E-Tree-based mesh generation approach that can
support a dense mesh at the surface and a coarse
mesh at depth. Earthquake wave propagation
simulations using this approach have been run for
very large meshes with good results.
SCEC/CME researchers have developed a highly
efficient, octree-based, mesh generation software
toolkit called the E-Tree library. This software
is used to create very large, sparse, velocity
meshes for use by Finite Element-based Anelastic
Wave Propagation simulation software.
The E-Tree software is particularly useful in
generation of meshes in which large areas have
very similar values, and other areas have great
variation in values. This is the case in velocity
models of Southern California where surface
velocity varies greatly, but velocity at depth
does not.
28
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
The SCEC/CME Project recently released an initial
version of the UseIT LA3D software to SCEC
researchers for use as a desktop application that
allows researchers to visualize 3D Fault
geometries. The interactive nature of the
software allows users to zoom, rotate, and
capture images while viewing data.
Scientists want to visualize many geophysical
data sets such as earthquakes, faults, velocity
structures, and topography in 3D. SCEC/CME UseIT
Interns have developed an inexpensive,
GeoWall-based, visualization software tool
capable of displaying all these dataset, and
more, in 3D.
Our software development approach, utilizing
cross-platform Java language, and inexpensive
GeoWall display hardware provides a flexible,
extensible, software platform for a large number
of 3D scientific visualization projects.
29
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
SCEC/CME IT researchers have developed a Grid
Visualization Toolkit (GVT) that supports
interactive visualization of very large datasets.
The software separates the computationally
expensive feature extraction processing from
the display processing. By connecting display
computers and processing nodes via a grid, users
can interact with very large datasets.
SCEC/CME researchers have developed a software
toolkit called the Grid Visualization Toolkit
(GVT) for visualizing very large, time-varying,
volumetric data sets. By integrating interactive
visualization software with Grid-based
communication codes, a highly distributed
visualization system have been developed.
Large, time-varying, volumetric data sets are
common in seismology, climatology, oceanography,
and other disciplines. Tools for interactive
visualization of large datasets can provide
researchers with new insights into these very
large datasets.
30
NSF Grant Number EAR-0122464PIs Thomas H.
Jordan, J. Bernard Minster, Carl Kesselman,
Reagan Moore Institution Southern California
Earthquake CenterTitle The SCEC Community
Modeling Environment An Information
Infrastructure for System-Level Earthquake
Research
The SCEC/CME objectives are (a) to develop a
physics-based, system-level, approach to Seismic
Hazard Analysis (SHA), and (b) to develop the
Cyberinfrastructure required to support the
computational, data management, and usability
requirements of our approach.
SCEC/CME researchers ran a very large rupture
dynamic simulation on at San Diego Supercomputer
Center that resulted in a 10TB data set. We used
the TeraGrid visualization tools to provide
multiple views of the data set to geoscientists.
SCEC/CME researchers use the computational, data
management, and visualization capabilities of the
TeraGrid to run large Earthquake Rupture
Simulations. The computational capabilities of
the TeraGrid allow us to work at a larger scale
than previously possible. The TeraGrid
visualization tools provide alternative views of
the important data products.
As SCEC/CME researchers use the TeraGrid system,
we are beginning to develop computational
expertise that is valuable to other parts of the
geoscience community. The SCEC/CME Project
represents a resource to other geoscience
communities interested in high performance
computing on the TeraGrid.
31
Outline

• SCEC/CME Project Goals
• Recent Activities
• Project Results
• Project Discussion

32
Leadership Discussion

• Desire by some Project participants for more
  detailed system architecture to which all
  development is done.

33
SCEC/CME Layered Software Architecture
Intelligent Assistant
UI
UI
Web Portal Applications
UI
Workflow Management
UI
Applications
Web Services
Grid Services (Resource, Knowledge Mgmt)
Digital Library Services
Computing Resources
KB
Metadata
SRB
Data
34
Leadership Discussion

• Wide Range of efforts underway (including some
  that involve administration and maintenance).

35
Scenario-Oriented Interface3D Ground Motion
Project for LA Basin
36
Automated Pathway 2 Simulations
37
IRIS Synthetic and Observed Seismogram Access
(SOSA) Tool
38
GIS Interface to Vertical Motion, FAD, and CFM
Data
39
Coupling Simulations on Different Grids and
Spatial or Temporal Scales
40
TeraShake Simulation Planning
41
Leadership Discussion

• Primary Project Management Issue is keeping two
  Customers happy
• NSF wants Research Results.
• SCEC Researchers want Operational
  Hardware/Software Infrastructure.