Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
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
Magnitude
6 5 4
0.1
Distance (km)
10
100
1000
Developer Scope
Applicable distance range
Select preferred distance measure 0 200 km
Site classification
Select preferred site classification scheme Classification scheme need not include very soft soil Provide translation to NEHRP site categories
Recordings
Negligible embedment effects Negligible soil-structure interaction effects Generally reliable path and site metadata
Moment Magnitude
1.0
10.0
100.0
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
Campbell-Bozorgnia Findings
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
Campbell-Bozorgnia Findings
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
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)
= = = = = = = =
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)
f1 (M)
M RRUP
f1 (M)
Style-of-Faulting Term
= reverse-faulting indicator variable = normal-faulting indicator variable = depth to top of coseismic rupture (km)
f1 (M)
Hanging-Wall Term
= = = = = =
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)
= = = =
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)
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)
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
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
l a u d i s e R
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
l a u d i s e R
200
A G -1 P -2 -3 0 250 500 750 1000 1250 30-m Shear-Wave Velocity (m/sec) 1500
1500
l a u d i s e R
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
0.8
-3 0
RRUP=0.1 RRUP=0.1
10
-1
10
10
-1
10-2
50
10-2
10
Acceleration (g)
200
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
10-1
M = 8.0
10-1
M = 8.0
Acceleration (g)
10-2 1
10
100
10-2 1
10
100
RRUP=0.1 RRUP=0.1
10-1
10
10-1
10 50
10-2
200
10-2
50 200
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA M ay 06)
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
M = 8.0
10
-1
M = 8.0
10
-1
Acceleration (g)
7.0
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA May 06)
5.0
6.0
10-2 1
10
100
10-2 1
10
100
10 -1
M = 8.0
10-1
M = 8.0
Acceleration (g)
10 -2 1
5.0
7.0 6.0
10
100
10-2 1
10
100
10
10 -1
M = 8.0
10 -1
M = 8.0
Acceleration (g)
7.0
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA Jan 06)
7.0
10 -2 1
5.0
6.0
10
100
10 -2 1
5.0
6.0
10
100
10-1
M = 8.0
10-1
M = 8.0
Acceleration (g)
7.0
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA May 06)
5.0
6.0
10-2 1
10
100
10-2 1
10
100
10-1
M = 8.0
10-1
M = 8.0
Acceleration (g)
7.0
Campbell & Bozorgnia (2003) Campbell & Bozorgnia (NGA May 06)
5.0
6.0
10-2 1
10
100
10-2 1
10
100
Site Factor
0.01
0.1
1.5
560
1.0
1130
Campbell & Bozorgnia (NGA, 2006) Choi & Stewart (2005) NEHRP (2003)
Campbell & Bozorgnia (NGA, 2006) Choi & Stewart (2005) NEHRP (2003)
0.5 0.01
0.1
0.0 0.01
0.1
f1 (M)
Sediment-Depth Factor
10
PGA, VS30=360
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)
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
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=1500
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)
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
SA(0.2s)
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.)
0.1
0.1
SA(1.0s)
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.01
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(3.0s)
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.)
0.1
0.1
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
0.01
0.1
0.1
Vs30 = 180 Vs30 = 360 Vs30 = 760 Vs30 =1500
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
Comparison with Ambraseys et al. (2005) Empirical Ground Motion Model for Europe and Middle East
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 -1
M = 8.0
10 -1
M = 8.0
Acceleration (g)
10 1
-2
10
100
10 1
-2
5.0
6.0
7.0
10
100
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
M = 8.0
10 -1
M = 8.0
10 -1
Acceleration (g)
10 -2 1
10
100
10 -2 1
10
100
RRUP=0.1
10 -1
10
10 -1
10
10 -2
50
10 -2
50 200
Ambraseys et al. (2005) Campbell & Bozorgnia (2006)
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
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
1.4 1.2
560 1130
1.5
Site Factor
Site Factor
1130
0.1
0.1
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)
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
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
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-Faulting Coefficient
Unsmoothed Regression Coefficients vs. Period
Regression Coefficient C8
0 -0.1
Coefficient
Period (sec)
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
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
Normal-Faulting Questions
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
Normal-Faulting Questions
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
Normal-Faulting Questions
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
Normal-faulting effects
11% lower than strike slip for T < 0.09s Effect phases out at T > 0.2s
SA(0.2s)
1.5 1.0 0.5 0.0 -0.5
Other California Taiwan Western U.S. Alaska
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
Other California Taiwan Western U.S. Alaska
SA(3.0s)
1.5 1.0 0.5 0.0 -0.5
Other California Taiwan Western U.S. Alaska
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(0.2s)
1.5 1.0 0.5 0.0 -0.5
Non-extensional Extensional Extensional?
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
Non-extensional Extensional Extensional?
SA(3.0s)
1.5 1.0 0.5 0.0 -0.5
Non-extensional Extensional Extensional?
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(0.2s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
Intra-event Residual
SA(1.0s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
Intra-event Residual
SA(0.2s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
Intra-event Residual
SA(1.0s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
Intra-event Residual
SA(0.2s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
Intra-event Residual
-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
Other California Taiwan Western U.S.
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
Other California Taiwan Western U.S.
Intra-event Residual
-2.0 -3.0 0 10 20 30 40 50
-2.0 -3.0 0 10 20 30 40 50
SA(0.2s)
3.0 2.0 1.0 0.0 -1.0
Off Rupture to North Off Rupture to South Footwall Hanging Wall
Intra-event Residual
SA(1.0s)
3.0 2.0 1.0 0.0 -1.0
Off Rupture to North Off Rupture to South Footwall Hanging Wall
SA(3.0s)
3.0 2.0 1.0 0.0 -1.0
Off Rupture to North Off Rupture to South Footwall Hanging Wall
Intra-event Residual
SA(0.2s)
3.0
Hanging Wall Footwall
Intra-event Residual
SA(1.0s)
3.0
Hanging Wall Footwall
SA(3.0s)
3.0
Hanging Wall Footwall
Intra-event Residual
SA(0.2s)
3.0
Off Rupture to North Off Rupture to South
Intra-event Residual
SA(1.0s)
3.0
Off Rupture to North Off Rupture to South
SA(3.0s)
3.0
Off Rupture to North Off Rupture to South
Intra-event Residual
SA(0.2s)
1.5 1.0 0.5 0.0 -0.5
Strike Slip Reverse Normal
Inter-event Residual
SA(1.0s)
1.5 1.0 0.5 0.0 -0.5
Strike Slip Reverse Normal
SA(3.0s)
1.5 1.0 0.5 0.0 -0.5
Strike Slip Reverse Normal
Inter-event Residual
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!