103/25/04 NGA Workshop Parameterization of Basin Response Based on 3D Simulations by PEER/SCEC 3D Ground Motion Project Team PI: Steven M. Day San Diego.

1 03/25/04 NGA Workshop

Parameterization of Basin Response Based on 3D Simulations

byPEER/SCEC 3D Ground Motion Project Team

PI: Steven M. DaySan Diego State University

March 25, 2004

2 03/25/04 NGA Workshop

Simulations Completed

– 85 different earthquake simulations

• 10 faults from SCEC Community Fault Model

• 6 rupture scenarios for each (hypocenter and slip model variations)

• 10 cross-check simulations (1 per fault)

• 15 1D reference simulations

3 03/25/04 NGA Workshop

Faults Modeled

• 1. Sierra Madre (7.0)• 2. Santa Monica SW (6.3)• 3. Hollywood (6.4)• 4. Raymond (6.6)• 5. Puente Hills I (6.8)• 6. Puente Hills II (6.7)• 7. Puente Hills (all) (7.1)• 8. Compton (6.9)• 9. Newport-Inglewood (6.9)• 10. Whittier (6.7)

4 03/25/04 NGA Workshop

Six Rupture Scenarios Per Fault

• 2 hypocenters

– 1/4 fault-length from each end

– 7/10 fault-width down dip

• 3 slip models

– Constructed following Somerville (1999)

– Constant rupture velocity (2.8 km/s)

– Rise time scaled to empirical formula:

Log(Tr)=0.5(Mw+10.7) + log(2.9x10-9)

5 03/25/04 NGA Workshop

Coordination SchemeUCB -LLNL UCSB CMU URS (RG) URS (AP)

S. Madre F,R,S C

S. Mon. F,R C

HollyW F,R C

Raym F,R C

P.Hills6.8 F,R C

P.Hills6.7 F,R C

P.Hills7.1 F,R,R,S C

Comp F,R,S C

N-I N. R,S F C

Whit N. R F C

F = 6 3D scenariosC = single cross-check

R = 1D rock reference simulation

S = 1D basin-profile simulation

6 03/25/04 NGA Workshop

Output

• Full time histories

• 3 velocity components

• 1600 surface points per simulation

• Basin and rock sites sampled

• ~300,000 synthetic time histories and associated metadata in digital library

7 03/25/04 NGA Workshop

Cross-check (cont’d)

• Sierra Madre Scenario

• Compares FD and FE codes at 16 sites (N-S component):

– FD (UCB/LLNL) red– FE (CMU) green

8 03/25/04 NGA Workshop

Response Spectra Data Set

• synthetic Sa ordinates (2-10 second period range)

• source distances

• local basin depth measures (depths to 1.0, 1.5, 2.5, and 3.5 km/s isosurfaces)

• Sa files available on web

9 03/25/04 NGA Workshop

3D÷1D_rock Sa Ratios

Frequency (Hz)

ln(S

a R

atio

)F

requ

ency

(H

z)

10 03/25/04 NGA Workshop

Vertically Incident SH Response

ln(S

a R

atio

)F

requ

ency

(H

z)

Frequency (Hz)

11 03/25/04 NGA Workshop

Basin Depth Effect Curves

12 03/25/04 NGA Workshop

Curve Fits to Basin Depth Effect(Separately Optimized at Each Period)

13 03/25/04 NGA Workshop

Curve Fits to Basin Depth Effect(6-parameter model for Depth and Period Dependence)

14 03/25/04 NGA Workshop

Basin Effect Relative to 1-D Soil (2000 m Depth to Isosurface)

15 03/25/04 NGA Workshop

1D Rock Simulations vs A-S Regression Model

Period (sec)

Und

erpr

edic

tion

Fac

tor

16 03/25/04 NGA Workshop

Transfer Function from “Very Hard Rock” to Boore/Joyner “Generic Rock”

17 03/25/04 NGA Workshop

18 03/25/04 NGA Workshop

1.5 Km/s Isosurface Map

19 03/25/04 NGA Workshop

8 Sec Sa Mean Residual Map

20 03/25/04 NGA Workshop

5 Sec Sa Mean Residual Map

21 03/25/04 NGA Workshop

3 Sec Sa Mean Residual Map

22 03/25/04 NGA Workshop

RMS Residuals

23 03/25/04 NGA Workshop

8 sec Sa RMS Residual Map

24 03/25/04 NGA Workshop

5 sec Sa RMS Residual Map

25 03/25/04 NGA Workshop

3 sec Sa RMS Residual Map

26 03/25/04 NGA Workshop

Summary• Source-averaged 3D effect is largely captured by basin depth

term (depth to 1.5 km/s isosurface)

• Mean and variance are period-dependent

• Results almost certainly double-count effects partially represented in “rock” regression equations

• With ~500 m depth sites (instead of 0 depth) taken as “rock” reference:– absolute amplitudes at long period (5 sec) come into agreement with

A-S rock regression (i.e., under-prediction eliminated)– Addition under-prediction at shorter periods probably partly a source

effect (which would be removed by our analysis of ratios)– Maximum basin effect reduced to ~2 (@ 2 sec) to ~3 (@ 10 sec)

27 03/25/04 NGA Workshop

Summary (cont’d)

• Little or no systematic basin-edge effect in source-averaged residuals

• Likewise, no clear basin-edge effect in source-averaged standard deviations

28 03/25/04 NGA Workshop

Directions for Additional Work

• Analysis of current synthetic data set for

– Basin-specific (e.g., L.A., San Fernando, San Gabriel) variations– Event-specific basin effects

• Simulations for additional regions (e.g., Santa Clara Valley? Imperial Valley? others) to examine transportability of results

• Push simulations to ~1 Hz