PEER PSHA Code Verification Project Christie Hale Norm Abrahamson Yousef Bozorgnia
Agenda Introduction and Approach Details on Three Tests Which codes completed which tests?
Probabilistic Seismic Hazard Analysis Loop over SOURCES Loop over MAGNITUDE Loop over RUPTURE AREA Loop over RUPTURE LOCATION Loop over GROUND MOTION MODELS end loop end loop end loop end loop end loop
Codes and Participants Code Participant Affiliation CRISIS Mario Ordaz Universidad Nacional Autónoma de México EZ-FRISK Jason Altekruse Fugro FRISK88 Jason Altekruse Fugro HAZ38-URS Patricia Thomas URS HAZ45 Norm Abrahamson Pacific Gas and Electric Co. HAZ45b Nick Gregor Bechtel HazMapEQ Trey Apel Risk Management Solutions OpenQuake Marco Pagani Global Earthquake Model OpenSHA Peter Powers U.S. Geological Survey PROBHAZ Roland LaForge Fugro / Nuclear Regulatory Commission PROSIT Daniel Duggan AIR Worldwide RIZZO-HAZARD Jose Blanco RIZZO Associates SISMIC Manuela Villani Arup THAZ Arash Zandieh Lettis Consultants International, Inc. XCD55, HAZ51, TREE51 Valentina Montaldo Amec Foster Wheeler
Approach Instructions developed / distributed Participants run their own codes Results compiled, blind feedback loop Meetings to discuss results and differences in approaches Tests re-run, sometimes several iterations Acceptance criteria
Approach Instructions developed / distributed Participants run their own codes Results compiled, blind feedback loop Meetings to discuss results and differences in approaches Tests re-run, sometimes several iterations Acceptance criteria
Verification Tests and Descriptions Set 1 Test Description 1.1 Rate calculation 1.2 Rupture location variability 1.3 Rupture area variability 1.4 Dipping fault 1.5 Truncated exponential magnitude pdf 1.6 Truncated normal magnitude pdf 1.7 Youngs and Coppersmith magnitude pdf 1.8a-c Ground motion variability 1.10 Areal source, single depth 1.11 Areal source, depth range Set 2 Set 3 2.1 Multiple sources and deaggregation 2.2a-d NGA-West2 ground motion models, SS 2.3a-d NGA-West2 ground motion models, RV 2.4a-b Hypocenter distribution 2.5a-b Upper tails, mixture model 3.1a-b Bending fault 3.2 Logic tree, percentiles 3.3 Intraslab zone 3.4 Areal source, virtual faults
Verification Tests and Descriptions Set 1 Test Description 1.1 Rate calculation 1.2 Rupture location variability 1.3 Rupture area variability 1.4 Dipping fault 1.5 Truncated exponential magnitude pdf 1.6 Truncated normal magnitude pdf 1.7 Youngs and Coppersmith magnitude pdf 1.8a-c Ground motion variability 1.10 Areal source, single depth 1.11 Areal source, depth range Set 2 Set 3 2.1 Multiple sources and deaggregation 2.2a-d NGA-West2 ground motion models, SS 2.3a-d NGA-West2 ground motion models, RV 2.4a-b Hypocenter distribution 2.5a-b Upper tails, mixture model 3.1a-b Bending fault 3.2 Logic tree, percentiles 3.3 Intraslab zone 3.4 Areal source, virtual faults
Test 2.2a-d NGA-West2 ground motion models, SS Instructions Magnitude: truncated exponential M min = 5.0, M max = 7.0 b-value = 0.9 Source: Fault, L = 85 km fault plane depths = 0-12 km strike-slip, dip 90 slip rate = 2 mm/yr N (above) (looking North) 5 4 3 2 6 1 Ground motion models: a. ASK 14 b. BSSA 14 c. CB 14 d. CY 14 σ untruncated Damping ratio = 5% V S30 = 760 m/s V S30 is measured Z 1.0 = 0.048 km Z 2.5 = 0.607 km Region = California 5 4 3 2 1 5 km 10 km 10 km 42.5 km 15 km 25 km 6 1 km 5 km 85 km 12 km
Test 2.2a-d NGA-West2 ground motion models, SS First Submissions Range: 145%
Test 2.2a-d NGA-West2 ground motion models, SS Fixes to bugs and programming errors ASK14: Z 1.0 should be in kilometers, typo in ASK14 list of parameters ASK14: Error in PEER Spreadsheet related to R y0 distance (HW only) ASK14, BSSA14, and CY14: For ASK14 and BSSA14 Z 1.0 in kilometers, CY14 Z 1.0 in meters BSSA14: Coefficient values out to four decimal points, truncated to three decimal places in error BSSA14: Other models, shortest period is T = 0.01 s, should be used for PGA. BSSA14 coefficients for both T = 0.01 s and PGA CB14: Rock PGA, A 1100, should calculate Z 2.5 from equation 33 or 34 in CB14 CY14: ΔZ TOR calculated from Z TOR and E[Z TOR ], not the same as Z TOR
Test 2.2a-d NGA-West2 ground motion models, SS Final Results Range: 1%
Test 2.2a-d NGA-West2 ground motion models, SS Final Results Range: 1%
Test 3.3 Intraslab zone What are we testing?
Test 3.3 Intraslab zone Instructions (above) (looking North) N Magnitude: truncated exponential M min = 5.0, M max = 7.0 b-value = 0.8 25 km 1 2 Source: Intraslab zone N(M min >5) = 0.013 eq/yr slab thickness = 12.5 km 100 km 1 25 km 2 30 103.46 km 45 Ground motion model: Zhao et al 2006 Site Class I, Rock σ untruncated 125 km (projection of top of slab) 12.5 km
Test 3.3 Intraslab Zone First Submissions (sort of) Range: 325%
Test 3.3 Intraslab zone a) Areal zones filled with point sources b) Top or middle of slab as listric fault c) Fill slab with pseudo faults d) Areal zones filled with virtual faults
Test 3.3 Intraslab zone h a) Areal zones filled with point sources 1 Slab boundary M 6.5 point source Hypocenter Rupture is a point source larger R RUP distances lower median ground motions lower hazard smaller h distances (hypocenter depth) lower median ground motions lower hazard c) Fill slab with pseudo faults 1 h Slab boundary M 6.5 rupture Hypocenter Rupture has finite dimensions Rupture dip = 35 relative to slab (65 abs) No lateral leakage of rupture
Test 3.3 Intraslab Zone Results from different modeling approaches Range: 204%
Test 3.3 Intraslab Zone Final Results
Test 3.4 Areal source zone with virtual faults Instructions Use virtual faults to account for the rupture dimensions (above) (looking North) Magnitude: truncated exponential M min = 5.0, M max = 6.5 b-value = 0.9 N Source: Area source zone N(M min >5) = 0.0395 eq/yr seismogenic zone: 5-25 km Area Ground motion model: Chiou and Youngs 2014 σ untruncated Damping ratio = 5% V S30 = 760 m/s V S30 is measured Z 1.0 = 0.048 km Z 2.5 = 0.607 km Region = California Fault styles: Strike slip 60%, Normal 20%, Reverse 20% 50 km 25 km 1 2 3 4 r = 100 km 5 km Note: figures not to scale seismogenic zone 25 km Dip angles: Strike slip 90, Normal 60, Reverse 30
Test 3.4 Areal source zone with virtual faults First Submissions (sort of) Range: 23%
Test 3.4 Areal source zone with virtual faults a) Boundaries define hypocenters b) Boundaries define hypocenters, but upper limit is strict 0 km 0 km 5 km 5 km M 6 rupture seismogenic zone seismogenic zone 25 km 25 km c) Boundaries define rupture limits 0 km 5 km d) Boundaries define top of ruptures 0 km 5 km seismogenic zone seismogenic zone 25 km 25 km
Test 3.4 Areal source zone with virtual faults Z TOR a) Boundaries define hypocenters R RUP 0 km 5 km seismogenic zone 25 km smaller Z TOR distances lower median ground motions lower hazard smaller R RUP distances higher median ground motions higher hazard d) Boundaries define top of ruptures Z TOR R RUP 0 km 5 km seismogenic zone 25 km
Test 3.4 Areal source zone with virtual faults Results from different modeling approaches all sources contributing to hazard (expect to see Z TOR differences here) close-in sources dominate hazard (expect to see R RUP differences here)
Test 3.4 Areal source zone with virtual faults Results from different modeling approaches Range: 18%
Test 3.4 Areal source zone with virtual faults Final Results
Verification Tests - concluding remarks Benchmark answers for all tests in Set 1 and Set 2 Set 3 differences are well-understood need input from source characterization experts need more detailed specifications in hazard input document Improved PSHA codes, many changes to codes: 1) added features 2) enhancements 3) fixes to bugs and programming errors Documentation PEER report and electronic supplements Future Verification Tests (NGA-East ground motion models, Directivity)
Which codes completed which tests? 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.1 0 1.11 2.1 2.2 2.3 2.4a 2.4 b 2.5 3.1 3.2 3.3 3.4 CRISIS EZ-FRISK FRISK88 HAZ38-URS HAZ45 HAZ45b HazMapEQ OpenQuake OpenSHA PROBHAZ PROSIT RIZZO-HAZARD SISMIC THAZ XCD55, HAZ51, TREE51 same as HAZ45
Thank you