The Southern California Earthquake Center Information Technology Research Initiative

Transcription

1 The Southern California Earthquake Center Information Technology Research Initiative Toward a Collaboratory for System-Level Earthquake Science Tom Jordan USC Kim Olsen - UCSB 4th Meeting of the US-Japan Natural Resources Panel on Earthquake Research Morioka, November 6-9, 2002

2 Collaboratory Concept The fusion of computers and electronic communications has the potential to dramatically enhance the output and productivity of U.S. researchers. A major step toward realizing that potential can come from combining the interests of the scientific community at large with those of the computer science and engineering community to create integrated, tool-oriented computing and communication systems to support scientific collaboration. Such systems can be called collaboratories. From National Collaboratories: Applying Information Technology for Scientific Research, Computer Science and Telecommunications Board, National Research Council,

3 SCEC Collaboratory An information infrastructure organized and maintained to support the distributed scientific activities and product development essential to seismic hazard analysis and emergency response to earthquake disasters. 3

4 Southern California: a Natural Laboratory for Understanding Seismic Hazard and Managing Risk Tectonic diversity Complex fault network High seismic activity Excellent geologic exposure Rich data sources Large urban population with densely built environment high risk Extensive research program coordinated by Southern California Earthquake Center (SCEC) under NSF and USGS sponsorship 4

5 Southern California Earthquake Center Core Institutions California Institute of Technology Columbia University Harvard University Massachusetts Institute of Technology San Diego State University Stanford University U.S. Geological Survey (3 offices) University of California, Los Angeles University of California, San Diego University of California, Santa Barbara University of Nevada, Reno University of Southern California (lead) Consortium of 14 core institutions and 26 other participating organizations, founded as an NSF STC in 1991 Co-funded by NSF and USGS under the National Earthquake Hazards Reduction Program (NEHRP) Mission: Gather all kinds of data on earthquakes in Southern California Integrate information into a comprehensive, physics-based understanding of earthquake phenomena Communicate understanding to end-users and the general public to increase earthquake awareness, reduce economic losses, and save lives 5

6 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 SDSC SCEC/ITR Project USGS Information Science ISI SCEC Institutions IRIS Earth Science 6

7 Problem Focus of SCEC ITR Use physics-based earthquake forecasting and wavefield simulation to improve seismic hazard analysis for performance-based design. 7

8 Seismicity (ANSS) Components of Seismic Hazard Analysis Paleoseismology Local site effects Geologic structure (USArray( USArray) Faults (USArray( USArray) Seismic Hazard Model Stress Rupture transfer dynamics (InSAR,, PBO, & SAFOD) Crustal (SAFOD, ANSS, Crustal Seismic velocity motion (PBO) & USArray) deformation (InSAR( InSAR) structure (USArray( USArray) 8

9 Computational Pathways Pathway 1: Standard Seismic Hazard Analysis Earthquake Forecast Model Attenuation Relationship 1 Intensity Measures 9

10 Computational Pathways Pathway 1: Standard Seismic Hazard Analysis Earthquake Forecast Model Attenuation Relationship 1 Intensity Measures 9

11 Computational Pathways Pathway 2: 1: Ground Standard motion Seismic Hazard simulation Analysis Unified Structural Representation Faults Motions Stresses Anelastic model AWM SRM Ground Motions Earthquake Forecast Model Attenuation Relationship 2 1 Intensity Measures AWM = Anelastic Wave Model SRM = Site Response Model 9

12 Computational Pathways Pathway 2: 1: Ground Standard motion Seismic Hazard simulation Analysis Unified Structural Representation Faults Motions Stresses Anelastic model AWM SRM Ground Motions Earthquake Forecast Model Attenuation Relationship 2 1 Intensity Measures AWM = Anelastic Wave Model SRM = Site Response Model 9

13 Computational Pathways Pathway 2: 1: 3: Physics-based Ground Standard motion Seismic Hazard simulation earthquake Analysis forecasting Unified Structural Representation Faults Motions Stresses Anelastic model FSM 3 RDM AWM SRM 2 Ground Motions Earthquake Forecast Model Attenuation Relationship Intensity Measures 1 FSM = Fault System Model RDM = Rupture Dynamics Model AWM = Anelastic Wave Model SRM = Site Response Model 9

14 Computational Pathways Pathway 2: 1: 3: Physics-based Ground Standard motion Seismic Hazard simulation earthquake Analysis forecasting Unified Structural Representation Faults Motions Stresses Anelastic model FSM 3 RDM AWM SRM 2 Ground Motions Earthquake Forecast Model Attenuation Relationship Intensity Measures 1 FSM = Fault System Model RDM = Rupture Dynamics Model AWM = Anelastic Wave Model SRM = Site Response Model 9

15 Computational Pathways Pathway 2: 4: 1: 3: Physics-based Ground Standard motion Seismic Hazard simulation earthquake inverse problem Analysis forecasting Other Data Geology Geodesy Unified Structural Representation Faults Motions Stresses Anelastic model Invert 4 FSM 3 RDM AWM SRM 2 Ground Motions Earthquake Forecast Model Attenuation Relationship Intensity Measures 1 FSM = Fault System Model RDM = Rupture Dynamics Model AWM = Anelastic Wave Model SRM = Site Response Model 9

16 Computational Pathways Pathway 2: 4: 1: 3: Physics-based Ground Standard motion Seismic Hazard simulation earthquake inverse problem Analysis forecasting Other Data Geology Geodesy Unified Structural Representation Faults Motions Stresses Anelastic model Invert 4 FSM 3 RDM AWM SRM 2 Ground Motions Earthquake Forecast Model Attenuation Relationship Intensity Measures 1 FSM = Fault System Model RDM = Rupture Dynamics Model AWM = Anelastic Wave Model SRM = Site Response Model 9

17 Science Goals 10

18 Science Goals Construct an open-source, object-oriented, and web-enabled framework for SHA computations that can incorporate a variety of earthquake forecast models, intensity-measure relationships, and site-response models 10

19 Science Goals Construct an open-source, object-oriented, and web-enabled framework for SHA computations that can incorporate a variety of earthquake forecast models, intensity-measure relationships, and site-response models Utilize the predictive power of dynamic-rupture and wavefield simulations in modeling time-dependent ground motion for scenario earthquakes and constructing intensity-measure relationships 10

20 Science Goals Construct an open-source, object-oriented, and web-enabled framework for SHA computations that can incorporate a variety of earthquake forecast models, intensity-measure relationships, and site-response models Utilize the predictive power of dynamic-rupture and wavefield simulations in modeling time-dependent ground motion for scenario earthquakes and constructing intensity-measure relationships Incorporate fault-system models into time-dependent earthquake forecasts 10

21 ITR Goals To develop an information infrastructure for system-level earthquake science to create a SCEC collaboratory that can: 11

22 ITR Goals To develop an information infrastructure for system-level earthquake science to create a SCEC collaboratory that can: Capture and manipulate the knowledge that will permit a variety of users with different levels of sophistication to configure complex computational pathways. 11

23 ITR Goals To develop an information infrastructure for system-level earthquake science to create a SCEC collaboratory that can: Capture and manipulate the knowledge that will permit a variety of users with different levels of sophistication to configure complex computational pathways. Enable execution of physics-based simulations and data inversions that incorporate advances in fault-system dynamics, rupture dynamics, wave propagation, and non-linear site response. 11

24 ITR Goals To develop an information infrastructure for system-level earthquake science to create a SCEC collaboratory that can: Capture and manipulate the knowledge that will permit a variety of users with different levels of sophistication to configure complex computational pathways. Enable execution of physics-based simulations and data inversions that incorporate advances in fault-system dynamics, rupture dynamics, wave propagation, and non-linear site response. Manage large, distributed collections of simulation results, as well as the large sets of geologic, geodetic and seismologic data required to validate the simulations and constrain parameter values. 11

25 ITR Goals To develop an information infrastructure for system-level earthquake science to create a SCEC collaboratory that can: Capture and manipulate the knowledge that will permit a variety of users with different levels of sophistication to configure complex computational pathways. Enable execution of physics-based simulations and data inversions that incorporate advances in fault-system dynamics, rupture dynamics, wave propagation, and non-linear site response. Manage large, distributed collections of simulation results, as well as the large sets of geologic, geodetic and seismologic data required to validate the simulations and constrain parameter values. Provide access to SHA products and methodologies to end-users outside of the SCEC community, including practicing engineers, emergency managers, decision-makers, and the general public. 11

26 Educational Goals 12

27 Educational Goals Cross-train earth scientists and computer scientists Terminology and problem orientation Methodology Current capabilities and research goals 12

28 Educational Goals Cross-train earth scientists and computer scientists Terminology and problem orientation Methodology Current capabilities and research goals Provide IT tools for the SCEC communication, education, and outreach mission Better public access to earthquake information Knowledge transfer to end-users in engineering, emergency response, and public policy 12

29 SCEC Collaboratory An information infrastructure for system-level earthquake science KNOWLEDGE REPRESENTATION & REASONING Knowledge Server Knowledge base access, Inference Translation Services Syntactic & semantic translation Knowledge Base DIGITAL LIBRARIES Navigation & Queries Versioning, Topic maps Mediated Collections Federated access Data Collections Ontologies Curated taxonomies, Relations & constraints Pathway Models Pathway templates, Models of simulation codes FSM RDM Data & Simulation Products Code Repositories AWM SRM KNOWLEDGE ACQUISITION Acquisition Interfaces Dialog planning, Pathway construction strategies Pathway Assembly Template instantiation, Resource selection, Constraint checking Users GRID Pathway Execution Policy, Data ingest, Repository access Grid Services Compute & storage management, Security Pathway Instantiations Computing Storage 13

30 Short-Term Objectives 14

31 Short-Term Objectives Development and verification of the computational modules 14

32 Short-Term Objectives Development and verification of the computational modules Standardization of data structures and interfaces needed for model interoperability 14

33 Short-Term Objectives Development and verification of the computational modules Standardization of data structures and interfaces needed for model interoperability Development of object classes, control vocabularies, and ontologies for knowledge management and model interoperability 14

34 Short-Term Objectives Development and verification of the computational modules Standardization of data structures and interfaces needed for model interoperability Development of object classes, control vocabularies, and ontologies for knowledge management and model interoperability Construction of SCEC computational and data grid testbeds 14

35 Short-Term Objectives Development and verification of the computational modules Standardization of data structures and interfaces needed for model interoperability Development of object classes, control vocabularies, and ontologies for knowledge management and model interoperability Construction of SCEC computational and data grid testbeds Development of user interfaces for knowledge ingest and acquisition, code execution, and visualization 14

36 Application Targets for KR&R Ontology construction and management Extension of IRIS s FISSURES seismological data model Development of a comprehensive earthquake ontology Management of complex collections Pathway 1 model components Pathway 2 simulations Ingest of geologic data into fault activity data base SHA Input validation and error advice Evaluation of alternative models Incorporation of Pathway 2 15

37 Distributed Operations of Code with Knowledgebased descriptions for Earthquake Research (DOCKER) Ties model descriptions to overarching SCEC ontology Enforces proper use of code through knowledge-based constraint reasoning (Powerloom) Guides users to make appropriate use of models Suggests alternative models more appropriate for user s analysis Supports distributed access to models and code through a layered view of service-based interaction (eventually) through the Open Grid Services Architecture (OSGA) Facilitates code publication by generating the code wrappers that enable the code to function at appropriate service layers 16

38 Application of KR&R to SHA Input validation and error advice Model verified for magnitudes

39 Application of KR&R to SHA Input validation and error advice Model verified for magnitudes 7.0 User attempt to enter a magnitude of

40 Application of KR&R to SHA Input validation and error advice Model verified for magnitudes 7.0 User attempt to enter a magnitude of Warning: The magnitude of 8.2 exceeds the limits of this model s magnitude parameter (7.0). For best results, choose a magnitude less than or equal to 7.0 Standard Warning 17

41 Application of KR&R to SHA Input validation and error advice Model verified for magnitudes 7.0 User attempt to enter a magnitude of Warning: The magnitude of 8.2 exceeds the limits of this model s magnitude Warning: parameter The magnitude (7.0). of 8.2 Options: exceeds the limits of this model s (1) Accept possibly magnitude inaccurate parameter results(7.0). (2) Choose a magnitude For best results, less than choose or equal a to 7.0 (3) Use a different magnitude model less than or equal to 7.0 A&S 97 with magnitude 8.2 and soil type = rock Standard Warning Steidl 2000 with magnitude 8.2, site type = Q Warning Using KR&R 17

42 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Yes 18 No Miller time

43 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Yes 18 No Miller time

44 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Yes 18 No Miller time

45 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Breakdown of contributions Example: Fault x critical (accounts for 50% of hazard) 18 No Yes Miller time

46 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Breakdown of contributions Example: Fault x critical (accounts for 50% of hazard) Would you like to perform physicsbased simulation? 18 No Yes Miller time

47 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Breakdown of contributions Example: Fault x critical (accounts for 50% of hazard) Would you like to perform physicsbased simulation? 18 No Yes Miller time

48 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Breakdown of contributions Example: Fault x critical (accounts for 50% of hazard) Would you like to perform physicsbased simulation? 18 No Yes Miller time

49 Fault x scenario PGA s Pathway 1 - SHA Pathway 2 - AWM Attenuation relations Hazard map Breakdown of contributions Example: Fault x critical (accounts for 50% of hazard) Would you like to perform physicsbased simulation? 18 No Yes Miller time

50 SCEC Collaboratory An information infrastructure for system-level earthquake science KNOWLEDGE REPRESENTATION & REASONING Knowledge Server Knowledge base access, Inference Translation Services Syntactic & semantic translation Knowledge Base DIGITAL LIBRARIES Navigation & Queries Versioning, Topic maps Mediated Collections Federated access Data Collections Ontologies Curated taxonomies, Relations & constraints Pathway Models Pathway templates, Models of simulation codes FSM RDM Data & Simulation Products Code Repositories AWM SRM KNOWLEDGE ACQUISITION Acquisition Interfaces Dialog planning, Pathway construction strategies Pathway Assembly Template instantiation, Resource selection, Constraint checking Users GRID Pathway Execution Policy, Data ingest, Repository access Grid Services Compute & storage management, Security Pathway Instantiations Computing Storage 19

51 (1) USER SCEC Computational Grid Testbed (1) Scientist issues a request (compute or data retrieval) to "Job Manager" (2) (3) (2) Job Manager talks to a Testbed computer via GRID service communication protocals. (3) Testbed computer performs the requested actions. 20

52 (1) USER SCEC Computational Grid Testbed SCEC Pathway (1) Scientist issues a request (compute or data retrieval) to "Job Manager" (2) Future complex pathways require a more versatile Job Manager. (3) (2) Job Manager talks to a Testbed computer via GRID service communication protocals. (3) Testbed computer performs the requested actions. 20

53 Parallelized and Scalable AWM 21

54 Parallelized and Scalable AWM 21

55 Parallelized and Scalable AWM s FD3D performance 1000 O Linux cluster (USC) o TCS1 (PSC) o LINUX (USC) O Compaq cluster (PSC) seconds PE s 22

56 Visualization 23

57

58 25

59 26

60 27

61 28

62 29

63 30

64 31

65 32

66 33

67 34

68 35

69 36

70 37

71 38

72 39

73 40

74 41

75 42

76 43

77 44

78 45

79 46

80 47

81 48

82 49

83 50

84 51

85 52

86 53

87 54

88 55

89 56

90 57

91 58

92 59

93 60

94 61

95 62

96 63

97 64

98 65

99 66

100 67

101 68

102 69

103 70

104 71

105 72

106 73

107 Lessons Learned So Far 74

108 Lessons Learned So Far Well-defined system-level problems such as SHA provide the focus needed for collaboratory development 74

109 Lessons Learned So Far Well-defined system-level problems such as SHA provide the focus needed for collaboratory development Interoperability is the key problem for information flow in the system-level approach to SHA 74

110 Lessons Learned So Far Well-defined system-level problems such as SHA provide the focus needed for collaboratory development Interoperability is the key problem for information flow in the system-level approach to SHA Development of domain ontologies should lead efforts to construct computational pathways in system-level science 74

111 Lessons Learned So Far Well-defined system-level problems such as SHA provide the focus needed for collaboratory development Interoperability is the key problem for information flow in the system-level approach to SHA Development of domain ontologies should lead efforts to construct computational pathways in system-level science KR&R tools will be required for curation of complex collections managed by SCEC Collaboratory 74

112 Lessons Learned So Far Well-defined system-level problems such as SHA provide the focus needed for collaboratory development Interoperability is the key problem for information flow in the system-level approach to SHA Development of domain ontologies should lead efforts to construct computational pathways in system-level science KR&R tools will be required for curation of complex collections managed by SCEC Collaboratory Computational and data grids offer great advantages for distributed scientific communities such as SCEC 74

113 Typical Questions 75

114 Typical Questions What the hell is an ontology? 75

115 Typical Questions What the hell is an ontology? What can it do for me? 75

116 A Simple Ontology Velocity Model 76

117 A Simple Ontology Velocity Model is_a Isotropic Model disjunct is_a Anisotropic Model Construction, Part 1: A A seismic velocity model is either isotropic or anisotropic. 77

118 A Simple Ontology is_a Velocity Model has is_a has is_a Elasticity Properties is_a Isotropic Model disjunct Anisotropic Model Isotropic Symmetry disjunct Anisotropic Symmetry Construction, Part 2: Elastic properties are either isotropic or anisotropic. An isotropic model has isotropic elastic properties. An anisotropic model has anisotropic elastic properties. 78

119 A Simple Ontology is_a Velocity Model has is_a has is_a Elasticity Properties is_a Isotropic Model disjunct Anisotropic Model Isotropic Symmetry disjunct Anisotropic Symmetry is_a Hexagonal Symmetry is_a Construction, Part 3: Hexagonal symmetry is a special case of anisotropic symmetry. Transversely isotropic symmetry is a special case of hexagonal symmetry. Transversely Isotropic Symmetry 79

120 A Simple Ontology is_a Velocity Model has is_a has is_a Elasticity Properties is_a Isotropic Model disjunct Anisotropic Model Isotropic Symmetry disjunct Anisotropic Symmetry is_a is_a PREM has Hexagonal Symmetry is_a Consider a particular model: PREM is an anisotropic model with transversely isotropic symmetry. y. Transversely Isotropic Symmetry 80

121 A Simple Ontology is_a Velocity Model has is_a has is_a Elasticity Properties is_a Isotropic Model disjunct Anisotropic Model Isotropic Symmetry disjunct Anisotropic Symmetry is_a has_not is_a PREM has has Hexagonal Symmetry is_a KR&R classifiers and inference engines (e.g., PowerLoom) ) can automatically infer new relationships: PREM does not have isotropic symmetry. PREM has hexagonal symmetry. Transversely Isotropic Symmetry 81

122 A Simple Ontology is_a Velocity Model has is_a has Elasticity Properties is_a Isotropic Model disjunct Anisotropic Model Isotropic Symmetry disjunct Anisotropic Symmetry is_a Azimuthally Anisotropic Model PREM is_a disjunct is_a Radially Anisotropic Model has has is_a Hexagonal Symmetry is_a Transversely Isotropic Symmetry Consider the addition of new terms: An anisotropic model is either azimuthally anisotropic or radially anisotropic. A radially anisotropic model has transversely isotropic symmetry. An azimuthally anisotropic model does not have transversely isotropic symmetry. 82

123 A Simple Ontology is_a Velocity Model has is_a has is_a Elasticity Properties is_a Isotropic Model disjunct Anisotropic Model Isotropic Symmetry disjunct Anisotropic Symmetry is_a is_a is_a Azimuthally Anisotropic Model disjunct PREM is_a Radially Anisotropic Model has Hexagonal Symmetry is_a Transversely Isotropic Symmetry KR&R classifier can automatically position new concepts in taxonomy and infer new relationship: PREM is a radially anisotropic model. 83