Attenuation Relationships Campbell Bozorgnia

Campbell-Bozorgnia NGA Empirical Ground Motion Model for the Average Horizontal Component of PGA, PGV and SA at Selected
Spectral Periods Ranging from 0.0110 Seconds

Kenneth W. Campbell(1) and Yousef Bozorgnia(2)
Inc., Beaverton, Oregon (2) PEER, University of California, Berkeley, California
(1) EQECAT
Workshop on Implementation of the Next Generation Attenuation Relationships (NGA) in the 2007 Revision of the National Seismic Hazard Maps PEER Center, Richmond, California September 2526 , 2006
Next Generation Attenuation Project

Partnered research program
PEER Lifelines Program Pacific Earthquake Engineering Research Center (PEER) California Energy Commission (CEC) California Department of Transportation (Caltrans) Pacific Gas and Electric Company (PG&E) U.S. Geological Survey (USGS) Southern California Research Center (SCEC)
Multi-disciplinary, 3-year project bringing together

Seismologists Geophysicists Geologists Geotechnical engineers Structural engineers
NGA Project Details

NGA empirical ground motion model developers
Abrahamson & Silva (updating their 1997 model) Boore & Atkinson (updating Boore et al., 1997 model) Campbell & Bozorgnia (updating their 2003 model) Chiou & Youngs (updating Sadigh et al., 1997 model) Idriss (updating his 1993 & 1996 models) Worldwide earthquakes Shallow crustal earthquakes in active tectonic regions Metadata (e.g., magnitude, distance, etc.) Uniformly processed strong-motion recordings
All developers started with a common database
NGA Project Details

NGA developers applied their own selection criteria to the common database, with the proviso that
Criteria are explicitly defined and documented Criteria are shared with other developers Reasons for excluding data are justifiable Other developers are notified if metadata is modified
NGA supporting studies

1-D ground-motion simulations of rock ground motion 3-D ground-motion simulations of basin response 1-D ground-motion simulations of shallow site response
NGA Project Database

NGA strong-motion database:
172 worldwide earthquakes 1,400 recording stations 3,500 multicomponent strongmotion recordings Over 100 parameters describing source, path, and site conditions
8
Magnitude
6 5 4
0.1
Distance (km)
10
100
1000
Previous Data New Data
NGA Project Requirements

Ground motion model
Provide a model for median ground motion Provide a model for aleatory standard deviation
Ground motion parameters

Horizontal components (Geometric Mean, FN, FP) PGA, PGV, PGD (optional) Spectral acceleration (minimum of 20 periods from 0 10 s)
Applicable magnitude range

Use moment magnitude 5.0 8.5 (strike-slip faulting) 5.0 8.0 (reverse faulting)
Developer Scope
Applicable distance range
Select preferred distance measure 0 200 km
Style of faulting (fault mechanism)

Strike slip Reverse Normal
Site classification
Select preferred site classification scheme Classification scheme need not include very soft soil Provide translation to NEHRP site categories
Some Critical Issues Addressed

Over-saturation at short distances
Data generally required it regardless of functional form Issue for short periods, short distances, large magnitudes Developers were divided on whether to allow it Some concern about stability in future events We allow saturation but not over-saturation We initially proposed it to model breakdown in self-similarity We abandoned its use because of data inconsistencies Its effect trades-off with magnitude scaling in the regression We will reconsider its use once inconsistencies are resolved
Rupture aspect ratio (L/W)

Depth to coseismic rupture
Higher ground motions from buried & blind reverse faults are empirically supported and included in our model Also supported by laboratory and numerical ground-motion simulations when weak upper layers are included Not empirically supported for many strike-slip faults We include buried effects for reverse faults, implying similar ground motions for large surface-rupturing reverse and strike-slip events
Extrapolation to 10-second period

Small-to-moderate events are not reliable to long periods Large events provide empirical constraints to 10 seconds There is no developer consensus on how to extrapolate We extrapolate to small magnitudes using seismological constraints and assuming constant displacement beyond a corner period

Functional form of aleatory standard deviation
Should it be a function of magnitude or amplitude? Results support independence on M, dependence on PGA Large near-fault scatter in 2004 Parkfield EQ is consistent with predicted uncertainty We include decreasing aleatory uncertainty with lower VS30 and higher rock PGA
Treatment of epistemic uncertainty

Should it be a function of magnitude and distance? Developers agree the use of multiple models is insufficient Developers agree it should be increased in parameter ranges where data are limited and models are extrapolated We will develop an epistemic uncertainty model using bootstrap and jackknife statistical methods

Normal-faulting earthquakes
Should normal-faulting events have lower ground motions and hanging-wall effects Spudich et al. (1999) found strike-slip and normal-faulting ground motions were lower for extensional regions as compared to strikeslip events in both extensional and non-extensional regions Ambraseys et al. (2005) found normal-faulting events to have lower ground motions than strike-slip events at short periods Differences between extensional and non-extensional regimes are not empirically supported after including the effects of normal faulting Empirical results support statistically insignificant reduction at short periods and increase on hanging wall; laboratory and numerical modeling support hanging wall effects but are inconsistent regarding lower ground motions We include hanging-wall effects and slight reduction at short periods

Use of large worldwide earthquakes
Does the use of large worldwide earthquakes bias the predictions in California and the WUS? Data are not sufficient to resolve this issue statistically Studies have shown that ground motions from moderate sized events in similar tectonic regimes are similar in U.S., Japan, Europe and Taiwan, inferring the same might be true for large events Many studies outside of the U.S. use near-source data from events in the U.S. Topic is controversial, but some seismologists suggest there is no theoretical basis for assuming ground motions are not similar in similar tectonic regimes worldwide As in our previous models (19812003), we include large worldwide events to constrain magnitude scaling at M > 7.0, mitigated by disallowing over-saturation when predicted by the regression
General Data Selection Criteria

Earthquakes
Located in shallow continental lithosphere Located in tectonically active regions Generally reliable earthquake metadata
Recordings
Negligible embedment effects Negligible soil-structure interaction effects Generally reliable path and site metadata
CB-NGA Specific Exclusion Criteria

Only one horizontal and/or vertical component No measured or estimated value of VS30 No rake angle, focal mechanism, or P-T plunge Focus/rupture in deep crust, oceanic plate, or SCR Aftershocks (but not triggered events) Poorly recorded events (depending on magnitude)
N < 5 for M < 5.0 N < 3 for 5.0 M < 6.0 N < 2 for 6.0 M < 7.0 and all RRUP > 60 km
CB-NGA Specific Exclusion Criteria

Non-free-field recordings
Building basement or superstructure Embedded or downhole Toe, base, or crest of dam (but not abutments)
Demonstrated strong topographic effects

Tarzana Cedar Hill Nursery Pacoima Dam upper-left abutment
Recordings considered unreliable

LDGO recordings from 1999 Duzce, Turkey earthquake Quality D recordings from 1999 Chi-Chi earthquake
Distribution of Selected Recordings

64 Earthquakes, 1561 Recordings
DATABASE
8.0 7.5 7.0 6.5 6.0 5.5
Moment Magnitude
5.0 4.5 4.0 0.1
Other California Taiwan Western U.S. Alaska
1.0
10.0
100.0
Closest Distance to Rupture (km)
Campbell-Bozorgnia Findings
GM scaling with magnitude
Scaling decreases at large M, small RRUP, and short periods Short periods saturate at small RRUP for M > 6.5
GM scaling with distance

Rate of attenuation decreases with increasing M All periods saturate for small RRUP
GM scaling with fault parameters

Higher GM for buried and blind reverse faults GM same for strike slip and surface-rupturing reverse faults Higher GM on hanging-wall of reverse-faults (PGA 1g) All these effects become negligible at long periods
GM scaling with shallow soil conditions
GM decreases with decreasing VS30 (linear part) GM decreases with increasing rock PGA (nonlinear part) Nonlinear scaling constrained by 1-D site response simulations
GM scaling with sediment depth

GM increases for depths > 3 km (basin effects) GM decreases for depths < 1 km (shallow sediment effects) Basin effects constrained by 3-D ground motion simulations
Aleatory standard deviation

Based on better recorded events Only weakly dependent on magnitude and period Larger at larger magnitudes compared to C-B (2003) Smaller at smaller magnitudes compared to C-B (2003)
Epistemic uncertainty
Underestimated by use of multiple ground motion models Need to specify a minimum value of standard deviation Need to develop a separate model of standard deviation
Analysis Methodology
Functional form development
Exploratory data analysis (analysis of residuals) Past experience (personal and literature review) Developer interaction meetings Theoretical studies (site response, basin effects) Two-step regression analysis (Boore et al., 1997) Weighted nonlinear least squares Intra-event terms fit in first step Inter-event terms fit in second step
Regression analysis (exploratory phase)
Analysis Methodology
Regression analysis (final phase)
Random effects regression (Abrahamson & Youngs, 1992) Maximum likelihood method with random effects Iteratively smoothed regression coefficients Start with least correlated coefficients Smooth observed coefficient trend with period Constrain coefficients as necessary Compensate for over-saturation Control behavior at long periods
f1 (M)
Generalized Functional Form
ln Y fmag fdis fflt fhng fsite fdep
= = = = = = = =
natural log of ground motion earthquake magnitude term source-site distance term style-of-faulting term hanging-wall term shallow site conditions term sediment depth term random error term: Normal (0, T)
f1 (M)
Earthquake Magnitude Term
M = moment magnitude c0c12 = regression coefficients
f1 (M)
Source-Site Distance Term
M RRUP
= moment magnitude = closest distance to fault rupture (km)
f1 (M)
Style-of-Faulting Term
FRV FNM ZTOR
= reverse-faulting indicator variable = normal-faulting indicator variable = depth to top of coseismic rupture (km)
f1 (M)
Hanging-Wall Term
M RRUP RJB FRV ZTOR
= = = = = =
moment magnitude closest distance to fault rupture (km) closest surface distance to fault rupture (km) reverse faulting indicator variable depth to top of coseismic rupture (km) dip of rupture plane (degrees)
f1 (M)
Shallow Site Conditions Term
VS30 A1100 k1, k2 c, n
= = = =
average 30-m shear-wave velocity (m/s) PGA for VS30 = 1100 m/s (rock PGA, g) theoretical period-dependent coefficients theoretical period-independent coefficients (theoretically constrained nonlinear soil model from Walling & Abrahamson, 2006)
f1 (M)
Sediment Depth Term
Z2.5 k3
= depth to 2.5 km/s S-wave velocity (km) = theoretical period-dependent coefficient (theoretically constrained 3-D basin model from Day et al., 2005)
f1 (M)
Random Error Term
T , t, s r
= = = = = =
total aleatory standard deviation (S.D.) inter-event and intra-event S.D. of geom. mean inter-event and intra-event S.E. of regression rate of change of fsite w.r.t. ln A1100 (rock PGA) S.D. of the random (arbitrary) horizontal comp. 0 for geometric mean; 1 for random component
General Limits of Applicability

Tectonic environment
Shallow continental lithosphere Active tectonic regions
Magnitude
M = 4.0 8.5 (strike-slip faulting) M = 4.0 8.0 (reverse faulting) M = 4.0 7.5 (normal faulting)
Distance
RRUP = 0 200 km
Shallow site conditions

VS30 = 180 1500 m/s (NEHRP B D)
General Limits of Applicability

Sediment Depth
Z2.5 = 0 6 km
Depth to top of rupture

ZTOR = 0 20 km
Dip of rupture plane

= 15 90
l a u d i s e R
Residuals (Intra/Inter) vs. Magnitude

PGA, All Faults SA(1.0s), All Faults
3 2 1 0 3 2 ) s 1 0 . 0 1 ( -1 l A a S -2 u
l a u d i s e R
A G -1 P -2 l a u d i s e R -3 4.0 4.5 5.0 PGA, All Faults 7.0 7.5 8.0 5.5 6.0 6.5 Moment Magnitude 3 2 1 0
A G -1 P -2 -3 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Moment Magnitude
d -3 i 4.0 4.5 5.0 5.5 6.0 Faults7.0 7.5 8.0 SA(1.0s), All 6.5 s Moment Magnitude e 3 R 2 ) s 1 0 . 0 1 ( -1 A S -2 -3 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Moment Magnitude
l a u d i s e R
Total Residuals vs. Distance and VS30

PGA, All Faults SA(1.0s), All Faults
3 2 1 0 3 2 ) s 1 0 . 0 1 ( -1 l A a S -2 u d -3 i 0 s e 3 R 2 ) s 1 0 . 0 1 ( -1 A S -2 -3 0
l a u d i s e R
A G -1 P -2 l a u d i s e R -3 0 3 2 1 0 50 PGA, All Faults 150 100 Rupture Distance (km ) 200
50 100 SA(1.0s), All Faults150 Rupture Distance (km )
200
A G -1 P -2 -3 0 250 500 750 1000 1250 30-m Shear-Wave Velocity (m/sec) 1500
250 500 750 1000 1250 30-m Shear-Wave Velocity (m/sec)
1500
l a u d i s e R
Total Residuals vs. Rock PGA and Z2.5

PGA, All Faults, Vs30<360 SA(1.0s), All Faults, Vs30<360
3 2 1 0 3 2 ) s 1 0 . 0 1 ( -1 l A a S -2 u d -3 i 0.0 s e 3 R 2 ) s 1 0 . 0 1 ( -1 A S -2
l a u d i s e R
A G -1 P -2 l a u d i s e R -3 0.0 3 2 1 0 0.2 PGA, All Faults 0.6 0.4 PGA on Vs30=1100 m/sec (g) 0.8
0.2 0.4 SA(1.0s), All Faults0.6 PGA on Vs30=1100 m/sec (g)
0.8
A G -1 P -2 -3 0 1 2 3 4 Sedim Depth (km ent ) 5 6
-3 0
2 3 4 Sedim Depth (km ent )
Strike Slip/Surface Reverse NEHRP BC

PGA, Strike Slip, VS30=760
10
0
SA(1.0s), Strike Slip, VS30=760

10
0
RRUP=0.1 RRUP=0.1
10
-1
10
10
-1
10-2
50
10-2
10
Acceleration (g) Acceleration (g)
Acceleration (g)
200
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA M ay 06)
50
200
10-3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
10-3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
Moment Magnitude
Moment Magnitude

10
0
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA May 06)

10
0
10-1
M = 8.0
10-1
M = 8.0
Acceleration (g)
7.0 5.0 6.0
7.0 5.0 6.0
10-2 1
10
100
10-2 1
10
100
Strike Slip/Surface Reverse NEHRP D

10
0

10
0
RRUP=0.1 RRUP=0.1
10-1
10
10-1
10 50
10-2
200
10-2
50 200
Acceleration (g)
10-3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
10-3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
Moment Magnitude
Moment Magnitude

10
0

10
0
M = 8.0
10
-1
M = 8.0
10
-1
Acceleration (g)
7.0
7.0 5.0 6.0
5.0
6.0
10-2 1
10
100
10-2 1
10
100
Surface Reverse FW vs. HW

10
0
PGA, Reverse, FW, ZTOR=0, VS30=760

10
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA Jan 06)
SA(1.0s), Reverse, FW, ZTOR=0, VS30=760

0
10 -1
M = 8.0
10-1
M = 8.0
Acceleration (g)
10 -2 1
5.0
7.0 6.0
7.0 5.0 6.0
10
100
10-2 1
10
100
10
PGA, Reverse, HW, ZTOR=0, VS30=760

10
SA(1.0s), Reverse, HW, ZTOR=0, VS30=760

10 -1
M = 8.0
10 -1
M = 8.0
Acceleration (g)
7.0
7.0
10 -2 1
5.0
6.0
10
100
10 -2 1
5.0
6.0
10
100
Buried Reverse FW vs. HW

PGA, Reverse, FW, ZTOR=5, VS30=760
10
0
SA(1.0s), Reverse, FW, ZTOR=5, VS30=760

10
0
10-1
M = 8.0
10-1
M = 8.0
Acceleration (g)
7.0
7.0 5.0 6.0
5.0
6.0
10-2 1
10
100
10-2 1
10
100
PGA, Reverse, HW, ZTOR=5, VS30=760

10
0
SA(1.0s), Reverse, HW, ZTOR=5, VS30=760

10
0
10-1
M = 8.0
10-1
M = 8.0
Acceleration (g)
7.0
7.0 5.0 6.0
5.0
6.0
10-2 1
10
100
10-2 1
10
100
Site Effects Shallow Site Conditions

PGA, Nonlinear Site Effects
2.0 1.8 Vs30 = 180 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.001 0.01 0.1 1
270 360 560 760 1130 1500
SA(1.0s), Nonlinear Site Effects

4.0 3.5 3.0 2.5
360 270 Vs30 = 180
2.0 1.5 1.0 0.5 0.001

560 760 1130 1500
Site Factor
Site Factor Site Factor Relative
0.01
0.1
PGA on Rock (g)
PGA on Rock (g)

2.5 3.5 3.0 2.0
Vs30 = 270

Vs30 = 270
2.5 2.0 1.5 1.0 0.5

1130 560
1.5
560
1.0
1130
Campbell & Bozorgnia (NGA, 2006) Choi & Stewart (2005) NEHRP (2003)
Campbell & Bozorgnia (NGA, 2006) Choi & Stewart (2005) NEHRP (2003)
Site Factor Relative
0.5 0.01
0.1
0.0 0.01
0.1
PGA on Rock (g)
PGA on Rock (g)
f1 (M)
Site Effects Sediment Depth

Shallow Sediment and Basin Effects
2.5 2.0 1.5 1.0 0.5 0.0
Sediment-Depth Factor
PGA SA(1.0s) SA(3.0s)
10
Depth to 2.5 km/s S-Wave Velocity (km)
Standard Deviation vs. Magnitude

PGA, VS30=180
0.8 0.7 0.6 0.5 0.4
C&B C&B C&B C&B 2006 2006 2006 2006 (Rrup (Rrup (Rrup (Rrup = = = = 0) 10) 50) 200) C&B C&B C&B C&B C&B 2003 2003 2003 2003 2003 (Rseis = 3) (Rseis = 10.4) (Rseis = 50.1) (Rseis = 200) (Mw-based)
PGA, VS30=360
0.8 0.7 0.6 0.5 0.4
Standard Deviation (Ln) Standard Deviation (Ln)
Standard Deviation (Ln)
0.3 0.2 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
0.3 0.2 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
Moment Magnitude
Moment Magnitude
PGA, VS30=760
0.8 0.7 0.6 0.5 0.4
PGA, VS30=1500
0.8 0.7 0.6 0.5 0.4
0.3 0.2 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
0.3 0.2 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
Moment Magnitude
Moment Magnitude
Standard Deviation vs. Rock PGA

PGA
0.8 0.7 0.6 0.5 0.4 0.3
Vs30 Vs30 Vs30 Vs30 Vs30 Vs30 Vs30 Vs30 = = = = = = = = 180 (Geo.) 360 (Geo.) 760 (Geo.) 1500 (Geo.) 180 (Ran.) 360 (Ran.) 760 (Ran.) 1500 (Ran.)
SA(0.2s)
0.8 0.7 0.6 0.5 0.4 0.3
Standard Deviation (Ln) Standard Deviation (Ln)
0.2 0.1 0.01
0.2 0.1 0.01
0.1
0.1
PGA on Rock with Vs30 = 1100 m/s (g)
SA(1.0s)
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.01
SA(3.0s)
0.8 0.7 0.6 0.5 0.4 0.3
0.2 0.1 0.01
0.1
0.1
2004 Parkfield Earthquake (M = 6.0)

PGA, Parkfield, Vs30=360
10 0 10 0
SA(1.0s), Parkfield, Vs30=360
10 -1
10 -1
10 -2
10 -2
Acceleration (g)
Acceleration (g)
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA Oct 05)
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA Oct 05)
10 -3 1
10
100
10 -3 1
10
100
Closest Distance to Fault (km)
Closest Distance to Fault (km)
Predicted Acceleration Spectra

Strike Slip, RRUP = 10 km, VS30 = 760 m/s
1
Spectral Acceleration (g)
0.1 M=5.0 M=6.0 M=7.0 M=8.0 0.01 0.1 Period (sec) 1 10
0.01

Strike Slip, M = 7.0, VS30 = 760 m/s
1
RRUP = 10
RRUP = 40 RRUP = 200
0.1
0.01 0.01 0.1 Period (sec) 1 10

Strike Slip, M = 7.0, RRUP = 10 km
1
0.1
Vs30 = 180 Vs30 = 360 Vs30 = 760 Vs30 =1500
0.01 0.01 0.1 Period (sec) 1 10
Predicted Spectral Displacement

Strike Slip, RRUP = 10 km, VS30 = 760 m/s
100
Spectral Displacement (cm)
10 1 0.1 0.01 0.001 M 5.0 M 6.0 M 7.0 M 8.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Period (sec) 7.0 8.0 9.0 10.0
0.0001
Work in Progress
Finish smoothing regression coefficients
Smooth predicted response spectra Extrapolate to 10-second period
Develop relationship for other components

Fault-normal (FN) Fault-parallel (FP) Vertical
Add source directivity term

Geometric method (e.g., Abrahamson and Somerville) Isochrone method (e.g., Spudich)
Extend to other regions and tectonic environments

Subduction zones
Comparison with Ambraseys et al. (2005) Empirical Ground Motion Model for Europe and Middle East
Strike Slip NEHRP B (Rock)

10
0

10
RRUP=0.1

0
10 -1
10
10 -1
RRUP=0.1
10 -2
50
Ambraseys et al. (2005) Campbell & Bozorgnia (2006)
10 -2
10
Acceleration (g)
200
10 -3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
10 -3 4.0
50
200
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
Moment Magnitude
Moment Magnitude
10

10

Ambraseys (2005) Campbell & Bozorgnia (2006)
10 -1
M = 8.0
10 -1
M = 8.0
Acceleration (g)
7.0 5.0 6.0
10 1
-2
10
100
10 1
-2
5.0
6.0
7.0
10
100
Strike Slip NEHRP C (Stiff Soil)

10
0

10
RRUP=0.1

0
10 -1
10
10 -1
RRUP=0.1
10 -2
50
10 -2
10
Acceleration (g)
200
10 -3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
10 -3 4.0
50
200
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
Moment Magnitude
Moment Magnitude
10

10

Ambraseys (2005) Campbell & Bozorgnia (2006)
M = 8.0
10 -1
M = 8.0
10 -1
Acceleration (g)
7.0 5.0 6.0
7.0 5.0 6.0
10 -2 1
10
100
10 -2 1
10
100
Strike Slip NEHRP D (Soft Soil)

10
0

10
RRUP=0.1
RRUP=0.1
10 -1
10
10 -1
10
10 -2
50
10 -2
50 200
Acceleration (g)
200
10 -3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
10 -3 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
Moment Magnitude
Moment Magnitude
10

10

M = 8.0 M = 8.0
10 -1
10 -1
Acceleration (g)
7.0
7.0
Ambraeys et al. (2005) Campbell & Bozorgnia (2006)
10 -2 1
5.0
6.0
10
100
10 -2 1
5.0
6.0
10
100
Site Effects Shallow Site Conditions

1.8 1.6
Vs30 = 270

3.0 2.5 Vs30 = 270 2.0
560
1.4 1.2
560 1130
1.5
Site Factor
Site Factor
1.0 0.8 0.6 0.01
1.0 0.5 0.0 0.01
1130
Campbell & Bozorgnia (2006) Ambraseys et al. (2005) NEHRP (2003)
Campbell & Bozorgnia (2006) Ambraseys et al. (2005) NEHRP (2003)
0.1
0.1
PGA on Rock (g)
PGA on Rock (g)
Standard Deviation vs. Magnitude

PGA
0.8 0.7 0.6 0.5 0.4 0.3 0.2 4.5 5.0
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (2006) Ambraseys et al. (2005)
SA(1.0s)
0.8 0.7 0.6 0.5 0.4 0.3 0.2 4.5
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (2006) Ambraseys et al. (2005)
Standard Deviation (nat. log)
Standard Deviation (nat. log)
5.5
6.0
6.5
7.0
7.5
8.0
8.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
Moment Magnitude
Moment Magnitude
Conclusions Strike Slip

Comparison with Ambraseys et al. (2005) is best for:
NEHRP B site conditions (Rock, VS30 = 1130 m/s) M = 5.0 6.5 RRUP = 10 100 km
Comparison with Ambraseys et al. (2005) deteriorates for softer site conditions
NGA model supports nonlinear soil behavior Ambraseys model assumes linear soil behavior
Comparison with Ambraseys et al. (2005) deteriorates for large magnitudes and close distances
NGA model supports nonlinear magnitude scaling Ambraseys model assumes linear magnitude scaling
Conclusions Strike Slip

Ambraseys et al. (2005) has larger standard deviations
NGA model applied stricter data selection criteria based on a database that was rich in metadata Ambraseys model used all relevant data based on a database with limited metadata
Preliminary conclusions
Need to compare NGA model directly with Ambraseys data Campbell-Bozorgnia model appears to be appropriate for use for shallow crustal earthquakes in active tectonic regions of Europe By analogy other NGA models are also likely to be appropriate in this region
Normal Fault Issues
Normal and Normal-Oblique Events

Date 1979 1980 1980 1981 1983 1984 1987 1990 1992 1995 1995 Earthquake Norcia, Italy Mammoth Lakes, California Irpinia, Italy Corinth, Greece Bora Peak, Idaho Lazio-Abruzzo, Italy Edgecumbe, New Zealand Griva, Greece Little Skull Mtn., Nevada Kozani, Greece Dinar, Turkey M 5.9 6.1 6.9 6.6 6.9 5.8 6.6 6.1 5.7 6.4 6.4 RRUP (km) 19 36 5 15 8 60 10 83 85 19 51 16 69 29 16 100 20 79 3 44 No. 3 3 13 1 2 5 2 1 8 3 2
Normal-Faulting Coefficient
Unsmoothed Regression Coefficients vs. Period
Regression Coefficient C8
0 -0.1
Coefficient
-0.2 -0.3 -0.4 -0.5 -0.6 0.01 0.1 1 10
Period (sec)
HW Factor for Normal Faulting

PGA HW Coefficient Similar to Reverse Faulting
Overlay Plot
1.5
PGA Residual
0.5
-0.5
-1
-1.5
-2
.2
.4
.6
.8
Hanging-Wall Factor
Reverse-Faulting HW Coefficient
Unsmoothed Regression Coefficients vs. Period
Regression Coefficient C9
0.8 0.6
Coefficient
0.4 0.2 0 -0.2 -0.4 0.01 0.1 1 10
Period (sec)
Normal-Faulting Questions
Should hanging-wall effects be included for normalfaulting events?
Supported by preliminary results for PGA Based on only a limited number of recordings
Should hanging-wall effects for normal-faulting be similar to reverse-faulting?

Supported by preliminary results for PGA PGA HW term is similar to that for reverse-faulting events Regression coefficient is not statistically significant
Should normal-faulting events have lower short-period ground motion than strike-slip events?
Ambraseys et al. (2005) Europe and Middle East Both normal and strike slip from extensional regions 8% to 24% lower than strike slip for T < 0.13s Effect decreases to 0 between T = 0.130.32s Short-period effect is statistically significant Campbell and Bozorgnia (NGA) No distinction between extensional and non-extensional 9% to 12% lower than strike slip for T < 0.08s Effect decreases to 0 between T = 0.080.2s Effect increases to 30% from T = 0.22.0s Short-period effect is marginally statistically significant
Should normal-faulting events have lower short-period ground motion than strike-slip events?
Spudich et al. (1999) Both normal and strike slip from extensional regions Many events from Californias Imperial Valley 15% to 25% lower than strike slip for T < 0.15s Effect decreases to 0 between T = 0.150.35s Effect increases to 35% between T = 1.22.0s Long-period effect is not significant (IV events?) Spudich et al. (1999) Compared to Boore et al. (1997) soil Slightly lower than BJF97 strike slip for T < 0.15s Effect increases to 33% from T = 0.152.0s Long-period effect is statistically significant
Should normal-faulting events have lower long-period ground motion than strike-slip events?
Ambraseys et al. (2005) predicts similar amplitudes Recordings come from same regions Suggests similar static stress drops Campbell and Bozorgnia (NGA) predicts different amplitudes Recordings come from different regions Could be systematic differences in sediment depths Spudich et al. (1999) predict different amplitudes than BJF97 Recordings come from different regions Effects are nearly identical to NGA results Could be systematic differences in sediment depths
Tentative Conclusions: Normal Faulting

Hanging-wall effects
Assumed to be the same as reverse faults 63% higher than footwall for T < 1.7s Effect phases out at T > 3.8s
Normal-faulting effects
11% lower than strike slip for T < 0.09s Effect phases out at T > 0.2s
Regional Bias Issues
Inter-event Residuals by Region

PGA
1.5 1.0 0.5 0.0 -0.5
SA(0.2s)
1.5 1.0 0.5 0.0 -0.5
Inter-event Residual Inter-event Residual
Inter-event Residual
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Moment Magnitude
Moment Magnitude
SA(1.0s)
1.5 1.0 0.5 0.0 -0.5
SA(3.0s)
1.5 1.0 0.5 0.0 -0.5
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Moment Magnitude
Moment Magnitude
Inter-event Residuals by Stress Regime

PGA
1.5 1.0 0.5 0.0 -0.5
Non-extensional Extensional Extensional?
SA(0.2s)
1.5 1.0 0.5 0.0 -0.5
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Moment Magnitude
Moment Magnitude
SA(1.0s)
1.5 1.0 0.5 0.0 -0.5
SA(3.0s)
1.5 1.0 0.5 0.0 -0.5
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
-1.0 -1.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Moment Magnitude
Moment Magnitude
Intra-event Residuals by Region

PGA
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
SA(0.2s)
3.0 2.0 1.0 0.0 -1.0
Intra-event Residual Intra-event Residual
Intra-event Residual
-2.0 -3.0 0.1 1 10 100
-2.0 -3.0 0.1 1 10 100
SA(1.0s)
3.0 2.0 1.0 0.0 -1.0
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
-2.0 -3.0 0.1 1 10 100
-2.0 -3.0 0.1 1 10 100
Intra-event Residuals by VS30

PGA
3.0 2.0 1.0 0.0 -1.0
SA(0.2s)
3.0 2.0 1.0 0.0 -1.0
-2.0 E -3.0 100 D C B 1000 A
-2.0 E -3.0 100 D C B 1000 A
Shear-Wave Velocity in Top 30m (m/sec)
SA(1.0s)
3.0 2.0 1.0 0.0 -1.0
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
-2.0 E -3.0 100 D C B 1000 A
-2.0 E -3.0 100 D C B 1000 A
Near-Source Bias Issues
Intra-event Residuals Within 50 km

PGA
3.0 2.0 1.0 0.0 -1.0
SA(0.2s)
3.0 2.0 1.0 0.0 -1.0
-2.0 -3.0 0 10 20 30 40 50
-2.0 -3.0 0 10 20 30 40 50
SA(1.0s)
3.0 2.0 1.0 0.0 -1.0
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
-2.0 -3.0 0 10 20 30 40 50
-2.0 -3.0 0 10 20 30 40 50
Chi-Chi Earthquake Issues
Intra-event Residuals by Distance

PGA
3.0 2.0 1.0 0.0 -1.0
Off Rupture to North Off Rupture to South Footwall Hanging Wall
SA(0.2s)
3.0 2.0 1.0 0.0 -1.0
-2.0 -3.0 0 25 50 75 100 125 150 175 200
-2.0 -3.0 0 25 50 75 100 125 150 175 200
SA(1.0s)
3.0 2.0 1.0 0.0 -1.0
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
-2.0 -3.0 0 25 50 75 100 125 150 175 200
-2.0 -3.0 0 25 50 75 100 125 150 175 200
Intra-event Residuals by HW/FW

PGA
3.0
Hanging Wall Footwall
SA(0.2s)
3.0
2.0 1.0 0.0 -1.0
2.0 1.0 0.0 -1.0
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
SA(1.0s)
3.0
SA(3.0s)
3.0
2.0 1.0 0.0 -1.0
2.0 1.0 0.0 -1.0
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
Intra-event Residuals by Direction

PGA
3.0
Off Rupture to North Off Rupture to South
SA(0.2s)
3.0
2.0 1.0 0.0 -1.0
2.0 1.0 0.0 -1.0
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
SA(1.0s)
3.0
SA(3.0s)
3.0
2.0 1.0 0.0 -1.0
2.0 1.0 0.0 -1.0
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
-2.0 -3.0 -40 -30 -20 -10 0 10 20 30 40
Buried vs. Surface Faulting Issues
Inter-event Residuals by Rupture Depth

PGA
1.5 1.0 0.5 0.0 -0.5
Strike Slip Reverse Normal
SA(0.2s)
1.5 1.0 0.5 0.0 -0.5
-1.0 -1.5 0.0 5.0 10.0 15.0 20.0
-1.0 -1.5 0.0 5.0 10.0 15.0 20.0
Depth to Top of Rupture (km)
SA(1.0s)
1.5 1.0 0.5 0.0 -0.5
SA(3.0s)
1.5 1.0 0.5 0.0 -0.5
-1.0 -1.5 0.0 5.0 10.0 15.0 20.0
-1.0 -1.5 0.0 5.0 10.0 15.0 20.0
Conclusions
NGA project represents a significant improvement in:
Quantity and quality of ground-motion recordings and associated metadata Number of near-source recordings from large-magnitude earthquakes Number of independent variables from which to choose Degree of developer and principle investigator interaction Availability of supporting theoretical and numerical studies Peer review through workshops, conference presentations and formal review Scientific basis of the Campbell-Bozorgnia empirical ground motion model
Recommendation
We believe that our 2006 NGA empirical ground motion model is scientifically superior to our 1997 and 2003 empirical ground motion models and should be considered to supersede them
Thank You!

Attenuation Relationships Campbell Bozorgnia

Caricato da

Informazioni sul documento

Descrizione originale:

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Attenuation Relationships Campbell Bozorgnia

Caricato da

Copyright:

Formati disponibili

Campbell-Bozorgnia NGA Empirical Ground Motion Model for the Average Horizontal Component of PGA, PGV and SA at Selected

Spectral Periods Ranging from 0.0110 Seconds

Next Generation Attenuation Project

Multi-disciplinary, 3-year project bringing together

NGA Project Details

All developers started with a common database

NGA Project Details

NGA supporting studies

NGA Project Database

Previous Data New Data

NGA Project Requirements

Ground motion parameters

Applicable magnitude range

Style of faulting (fault mechanism)

Some Critical Issues Addressed

Rupture aspect ratio (L/W)

Some Critical Issues Addressed

Extrapolation to 10-second period

Some Critical Issues Addressed

Treatment of epistemic uncertainty

Some Critical Issues Addressed

Some Critical Issues Addressed

General Data Selection Criteria

CB-NGA Specific Exclusion Criteria

CB-NGA Specific Exclusion Criteria

Demonstrated strong topographic effects

Recordings considered unreliable

Distribution of Selected Recordings

5.0 4.5 4.0 0.1

Other California Taiwan Western U.S. Alaska

Closest Distance to Rupture (km)

GM scaling with distance

GM scaling with fault parameters

GM scaling with sediment depth

Aleatory standard deviation

Regression analysis (exploratory phase)

Generalized Functional Form

ln Y fmag fdis fflt fhng fsite fdep

Earthquake Magnitude Term

M = moment magnitude c0c12 = regression coefficients

Source-Site Distance Term

= moment magnitude = closest distance to fault rupture (km)

FRV FNM ZTOR

M RRUP RJB FRV ZTOR

Shallow Site Conditions Term

VS30 A1100 k1, k2 c, n

Sediment Depth Term

Random Error Term

General Limits of Applicability

Shallow site conditions

General Limits of Applicability

Depth to top of rupture

Dip of rupture plane

Residuals (Intra/Inter) vs. Magnitude

Total Residuals vs. Distance and VS30

A G -1 P -2 l a u d i s e R -3 0 3 2 1 0 50 PGA, All Faults 150 100 Rupture Distance (km ) 200

50 100 SA(1.0s), All Faults150 Rupture Distance (km )

250 500 750 1000 1250 30-m Shear-Wave Velocity (m/sec)

Total Residuals vs. Rock PGA and Z2.5

0.2 0.4 SA(1.0s), All Faults0.6 PGA on Vs30=1100 m/sec (g)

A G -1 P -2 -3 0 1 2 3 4 Sedim Depth (km ent ) 5 6

2 3 4 Sedim Depth (km ent )

Strike Slip/Surface Reverse NEHRP BC