Seismic Hazard Examples

1
Seismic Hazard Examples

• Two Sites
• Site in Pacific Northwest
• Site in San Francisco Bay Area
• Topics
• Source characterization
• Ground Motion models
• Hazard calculation
• Scenario Design spectra
• Time Histories

2
Source Characterization

• Earthquake Rates
• Historical seismicity
• Earthquake catalogs
• Slip-rates
• Long term slip rates from geology
• Recurrence intervals of large earthquakes
• From paleoseismic data

3
Historical Seismicity

• Catalog completeness
• Minimum magnitude detection threshold has been
  reduced over time
• Mgt6 complete for 100s of years
• Mgt3 complete for 10s of years depending on the
  seismic instrument coverage
• Aftershocks
• Standard practice is to remove aftershocks from
  catalog because they are dependent
• Aftershocks happen, shouldnt they be included?

4
Magnitude Recurrence Models

• Gutenberg-Richter (truncated exponential)
• Applicable to regions
• Pure characteristic (maximum magnitude)
• Applicable to individual faults that generate
  only large earthquakes (above the background)
• Composite exp-char (characteristic)
• Applicable to individual faults

5
Smoothing of Seismicity

• Two Approaches Commonly used
• Source zones with uniform seismicity
• Smoothes the earthquakes over the zone
• Spatial smoothing without fixed zones
• Key Issue for site region
• How much smoothing should be done?
• (methods for testing the amount of smoothing are
  being developed)

6
Ground Motion Regions

• Three main tectonic categories
• Shallow crustal earthquakes in active tectonic
  regions
• Shallow crustal earthquakes in stable continental
  regions
• Subduction earthquakes
• Interface
• Intraslab

7
PNW example
8
Source Parameter Summary
9
Ground Motion Models

• Both subduction and crustal sources
• Subduction
• Crustal

10
Subduction Models

• Youngs et al (0.5)
• Atkinson Boore (0.5)
• Gregor et al 2002 (0.0)
• This model is more work

11
NGA Models for Shallow Crustal Earthquakes in
Active Regions

• Five new models
• Abrahamson Silva
• Boore Atkinson
• Campbell Bozorgnia
• Chiou Youngs
• Idriss
• Parameters
• Average horizontal spectral acceleration
• Periods from 0 to 10 sec
• Applicable to all fault types (SS,RV, NML)

12
Site Classification in NGA Models

• VS30
• Abrahamson Silva
• Boore Atkinson
• Campbell Bozorgnia
• Chiou Youngs
• Rock class
• Idriss

13
Parameters in NGA Models
14
Hazard Curves
PGA
T2 sec
15
PGA Hazard by GM Model
Subduction models
Crustal Models
16
T2 sec Hazard by GM Model
Subduction models
Crustal Models
17
Hazard Uncertainty
PGA
T2 sec
18
UHS MCE (Cascadia)
MCE M9.0 R55 km
19
UHS MCE (intra-slab)
MCE M7.5 R40 km
20
Deaggregation (2000 yrs)
T2 sec Mode M8.5-9.0 R50-60 km
PGA Mode M6.5-7.0 R50-80 km
21
DeaggregartionMean M, R,Epsilon
2000 Years
22
Method for Computing Scenario Expected Spectra

• (1) Compute epsilon value needed to scale the
  median Sa(To) to the UHS(To)

23
Method for Computing Scenario Expected Spectra

• (2) Compute expected epsilon value at other
  periods

24
Scenario Eqk Expected Spectra
(3) Compute Spectrum
25
Scenario Design Spectra
Broaden to reduce number of scenarios to consider
26
Time Histories Using Spectral Matching

• Extend Spectrum to 10 seconds
• Select reference time histories
• Match using time domain method
• Plots to check the matching
• Response spectrum
• Acc, Vel, Disp time history of reference and
  matched
• Husid plot for reference and matched
• Fourier amplitude spectra of reference and
  matched

27
Extend to 10 seconds
28
Select Reference Time Histories

• Scenario 1 (T0.2)
• M7,2 R50 km, Depth50km, intra-slab

29
Initial Scaling to PGA
30
Spectral MatchingNisqually
31
Husid Plot
32
Compare Acc Time Histories
33
Compare Vel Time Histories
34
Compare Dis Time Histories
35
Fourier Amplitude Spectra
Reference
Modified
36
SF Bay Area Example
37
Fault Segmentation

• Geologists often break a long fault into
  sub-segments
• Based on segmentation points that may stop a
  rupture
• Segmentation
• Reduces the magnitudes of characteristic
  earthquakes
• Increases the rate of characteristic earthquakes
• Not always conservative to assume no segmentation
• Issue
• How often does the rupture break through a
  segment boundary?
• Multi-segment ruptures

38
Hayward/RC Example(includes fault memory)
39
Hazard Curves
PGA
T1 sec
40
Hazard by GM Model
PGA
T1 sec
41
Hazard Uncertainty
PGA
T1 sec
42
UHS MCE (Hayward/RC)
MCE M7.25 R3.5 km SS
43
Deaggregation (2000 yrs)
T2 sec Mode M6.5-7.5 R0-5 km
PGA Mode M6.5-7.0 R0-5 km
44
DeaggregartionMean M, R,Epsilon
2000 Years
45
Scenario Eqk Expected Spectra
46
Scenario Design Spectra
47
Time Histories Using Scaling

• Select reference time histories
• Scale by a constant
• Plots to check the matching
• Response spectrum
• Acc, Vel, Disp time histories of scaled ground
  motion
• Husid plots
• Fourier amplitude spectra

48
Selection and Scaling

• PEER/Lifelines and COSMOS have been working on
  this over last three years for buildings
• Project will be started this summer for earth
  dams
• Key Issue
• What is the objective of the time history
  analysis
• Average response
• Variability of response
• Key Results
• Including the gross non-linear behavior of the
  system, we can be much smarter in the selection
  of the recording
• Better to scale a good record by a large
  factor, than to use a bad record with a scale
  factor near 1.
• Dont fear large scale factors

49
Select Reference Time Histories

• Scenario 1 (T0.2)
• M6.75 R4 km, crustal, SS, rock
• Factors considered
• Shallow crustal
• Mag 6.5 - 7.0
• Dist 0-15 km
• Site VS30 gt 350 m/s
• Site class is not critical for spectral matching
• Lowest useable freq lt 0.25 Hz
• Other factors less important
• Style-of-faulting

50
Select Using Spectral Shape
51
Select Reference Time Histories

• Scenario 1 (T0.2)
• M6.75 R4 km, crustal, SS

52
Scaling

• How should the scaling be done?
• Scale to a single period
• Scale to a range of periods
• In this example, use a period range of 0.4-2 sec

53
Scaled Ground Motions (1)
54
Scaled Acc
Set 1
Set 2
Set 3
55
Scaled Vel
Set 1
Set 2
Set 3
56
Husid Plots
57
Fourier Spectra of Scaled GM
58
How to Use PSHA for Dams?

• Acceptable return period currently not set by
  regulators
• Group of utility owners is starting a project to
  develop a return period for active regions to
  propose to regulators
• Starting with comprehensive PHSA studies for each
  utility
• Will conduct risk analyses to evaluate the chance
  of failure (rapid uncontrolled release of water)
• Relate design ground motion return period to
  failure probability