TaiwanSSHACLevel3StudyCurrentGMCLogicTreeandHazardFeedback

TaiwanSSHACLevel3PSHAStudyWorkshop#3,June19-23,2017

Taipei,Taiwan

PTINorman Abrahamson

Outline

n Basecasen Hazardcurvesanddeaggregation(PGA,2s)n CurrentGMC LogicTreen GMCSensitivity

– Tornadoplot– ForAdjustedModels- median

2

GMCCrustalbasecaseSpectralFrequency

MedianGMPE

SingleStationSigmaModel

AdditionalEpistemicUncertainty

PGA

SA[5.0Hz]

SA[0.5Hz]

ASK14(0.2)

(0.2)

(0.2)

(0.2)

(0.2)

BSSA14

CB14

CY14

I14

(1.0)SWUS,CEN

(1.0)0.0

(1.0)

Formofln(SA)

Distribution

3

Traditionalnormal

GMCSubduction basecaseSpectralFrequency

AdditionalEpistemicUncertainty

PGA

SA[5.0Hz]

SA[0.5Hz]

(0.5)

(0.5)

BCH

LL08

(1.0)0.0

MedianGMPE

4

(1.0)BCH, CEN

(1.0)

Traditionalnormal

SingleStationSigmaModel

Formofln(SA)

Distribution

HAZARDDEAGGREGATION

Becausethesourcesisunderworking,theSSCteamsuggestedthatweusedthesourcessameasWS#2.

5

NPP1- PGA- 5Hz– 0.5Hz

6

PGA 2.0s

Period Source M R

PGA

Shanchiao (42%) 6.6-7.4 0-10

Intraslab (31%) 7.4-8.0 70-100

Interface(15%) 8.0-8.2 60-80

2s

Intraslab (52%) 7.5-8.2 150-230

Interface(30%) 8.4-9.0 60-80

Shanchiao (17%) 7.0-7.4 0-10

NPP2- PGA- 5Hz– 0.5Hz

7

PGA 2.0s

Period Source M R

PGAShanchiao (62%) 6.4-7.4 0-10

STII(20%) 6.4-7.4 0-10

2s

Shanchiao (38%) 7.0-7.4 0-10

Intraslab (24%) 7.5-8.2 150-230

Interface(22%) 8.4-9.0 50-80

STII(16%) 7.0-7.4 0-10

NPP3- PGA- 5Hz– 0.5Hz

8

PGA 2.0s

Period Source M R

PGAInterface(53%) 8-8.8 10-20

Hengchun (43%) 6.4-7.2 0-10

2sInterface(46%) 8-8.8 10-20

Hengchun (48%) 6.6-7.6 0-10

NPP4- PGA- 5Hz– 0.5Hz

9

PGA 2.0s

Period Source M R

PGA

Intraslab (37%) 7.5-8.0 70-100

Interface(32%) 8.2-9.0 40-70

Shallowarea(26%) 5-7 0-20

2sInterface(61%) 8-9 40-60

Intraslab (33%) 7.5-8.2 150-230

CURRENTGMC LOGICTREE

10

GMCSensitivityresults– Crustal– PGA(WM#3)

11

Node GMPEsusedinSensitivityGMRatiogreater than10%

NPP1 NPP2 NPP3 NPP4

1 GMPEforMedian ✔ ✔ ✔ ✔

2 Additional EpistemicUncertaintyforMedian ✔ ✔ ✔ ✔

3 SigmaSS ✔ ✔ ✔ ✔

4 FormofDistribution ofln(SA)

5 Splayfault(only forNPP3)

6 Deepevents

7 Dipimplementationforlistric fault(only forNPP2)

8 Directivitymodel (only forperiodlargerthan0.5sec)

9 StyleofFaulting ✔ ✔ ✔ ✔

10 Hanging-wallEffect ✔ ✔

✔

✔

✔

✔

✔

GMCSensitivityresults– Crustal– 2sec(WM#3)

12

Node GMPEsusedinSensitivityGMRatiogreater than10%

NPP1 NPP2 NPP3 NPP4

1 GMPEforMedian ✔ ✔ ✔ ✔

2 Additional EpistemicUncertaintyforMedian ✔ ✔ ✔ ✔

3 SigmaSS ✔ ✔ ✔ ✔

4 FormofDistribution ofln(SA)

5 Splayfault(only forNPP3)

6 Deepevents

7 Dipimplementationforlistric fault(only forNPP2)

8 Directivitymodel

9 StyleofFaulting ✔ ✔ ✔ ✔

10 Hanging-wallEffect ✔

✔

✔

✔

✔

✔

LogicTreeoftheMedianforCrustalSource

13

Model 1

...

Model 2

...

Model ...

...

Model 18

...

Median forVertical DipSS Faults

Mag. Dep.

...

Mag. Indep.

...

SOFFactors - I

Lower

0.2

Center

0.6

Upper

0.2

SOFFactors - II

Lower

0.2

Center

0.6

Upper

0.2

SOFFactors - II

Model 1

0.2

Model 2

0.2

Model 3

0.2

Model 4

0.2

Model 5

0.2

HWFactor

Model 1

1.0

Listric FaultFactor

Logic Tree of the Median for Crustal Source

17

14

LogicTreeoftheSigmaSS forCrustalSource

Global...

Taiwan...

TauModel-I

Mag. Dep.1.0

TauModel-II

Mag. Dep....

Mag. Indep....

TauModel-II

Lower0.2Center0.6Upper0.2

TauModel-III

Lower0.2Center0.6Upper0.2

TauModel-III

Lower0.2Center0.6Upper0.2

TauModel-III

Taiwan1.0

PhiSSModel-I

Mag. Dep....

Mag. Indep....

PhiSSModel-II

Lower0.2Center0.6Upper0.2

PhiSSModel-III

Lower0.2Center0.6Upper0.2

PhiSSModel-III

Model 11.0

ln(Sa)Distribution

Logic Tree for the SigmaSS for Crustal Sources

Wechoosetotestsensitivity

GMCSensitivityresults– Subduction – PGA

15

Node GMPEsusedinSensitivityGMRatiogreater than10%

NPP1 NPP2 NPP3 NPP4

1 GMPEforMedian ✔ ✔ ✔

2 Additional EpistemicUncertaintyforMedian ✔ ✔ ✔ ✔

3 SigmaSS

4 FormofDistribution ofln(SA)

5 Depthscalingforintraslab

6 LargeMagscalingforintraslab ✔

7 EdgeEffectforinterface

✔

✔

✔

GMCSensitivityresults– Subduction – 2sec

16

Node GMPEsusedinSensitivityGMRatiogreater than10%

NPP1 NPP2 NPP3 NPP4

1 GMPEforMedian ✔ ✔ ✔ ✔

2 Additional EpistemicUncertaintyforMedian ✔ ✔ ✔ ✔

3 SigmaSS

4 FormofDistribution ofln(SA)

5 Depthscalingforintraslab

6 LargeMagscalingforintraslab

7 EdgeEffectforinterface

✔

✔

✔

17

LogicTreeoftheMedianforSubductionSource

Model 1

...

Model 2

...

Model ...

...

Model 18

...

Median

Model 1

1.0

Edge Effectfor Interface

Logic Tree of the Median for Subduction Source

17

18

LogicTreeoftheSigmaSS forSubductionSource

Global...

Taiwan...

TauModel-I

Mag. Indep.1.0

TauModel-II

Mag. Dep....

Mag. Indep....

TauModel-II

Lower0.2Center0.6Upper0.2

TauModel-III

Lower0.2Center0.6Upper0.2

TauModel-III

Lower0.2Center0.6Upper0.2

TauModel-III

Taiwan1.0

PhiSSModel-I

Mag. Dep....

Mag. Indep....

PhiSSModel-II

Lower0.2Center0.6Upper0.2

PhiSSModel-III

Lower0.2Center0.6Upper0.2

PhiSSModel-III

Model 11.0

ln(Sa)Distribution

Logic Tree for the SigmaSS for Subduction Sources

Wechoosetotestsensitivity

CRUSTALGMCSENSITIVITY

19

GMC—Crustal Sensitivity case(WS#3)MedianGMPE

EpistemicNode SOFNode

(1.0)

(1.0)

SWUSHW1

HWEffectNode

SpectralFrequency

PGA

SA[5.0Hz]

SA[0.5Hz] SWUSHW5

NoSOFeffect

(1.0)

NoHWeffect

Basecase5 models

ASB14

NewTaiwanmodelChao17

(1.0)0.0

Formofln(SA)

Distribution

(1.0)

(1.0)

SWUS,CEN(1.0)

PhiSSlow+Taulow(1.0)

PhiSSlow+Taumed(1.0)

PhiSSlow+TauHigh(1.0)

PhiSSmed+Taulow(1.0)

PhiSSmed+Taumed(1.0)

PhiSSmed+TauHigh(1.0)

PhiSSHigh+Taulow(1.0)

PhiSSHigh+Taumed(1.0)

PhiSSHigh+TauHigh(1.0)

SingleStationSigmaModel

normal

mixture

Published

MagDep.Low

MagDep.Cen

MagDep.High

MagInDep.Low

MagInDep.Cen

MagInDep.High

7Adjustedmodels

BI14

Node1:CrustalGMPEforMedianRegion Branch GMPEsusedinSensitivity

Global model(Basecase)

12345

ASK14(Abrahamson, etal.,2014)BSSA14(Boore etal.,2014)CB14(CampbellandBozorgnia,2014)CY14(Chiou andYoungs,2014)I14(Idriss,2014)

Europe(SIGMAprojectmodels)

67

ASB14(Akkaretal.,2014)BI14(Bindietal.,2014)

Adjustedmodel 891011121314

ASK14_C01BSSA14_C01CB14_C01CY14_C01I14_C01ASB14_C01BI14_C01

Taiwannewmodel 15 Chao17

21

Node2:StyleofFaulting

22

Branch GMPEs Value

1 Basecase5models Published

2 Basecase5models NoSoF

3 Basecase5models Magnitudedependent Center

4 Basecase5models Magnitudedependent High

5 Basecase5models Magnitudedependent Low

6 Basecase5models Magnitudeindependent Center

7 Basecase5models Magnitudeindependent High

8 Basecase5models Magnitudeindependent Low

Node3:Hanging-walleffect

Branch GMPEs Value

1 Basecase5models Published

1 Basecase5models NoHanging-wall effect

2 Basecase5models SWUSHWModel1

3 Basecase5models SWUSHWModel5

23

Node4:CrustalSigmaSS

n Usethe3TaiwanMagnitudeIndependent Phi_SSmodelswiththe3GlobalMagnitudedependenttaumodels(9total).

24

Branch Value

123

TaiwanphiSS low&globaltaulowTaiwanphiSS low&globaltaumedTaiwanphiSS low&globaltauhigh

456

TaiwanphiSS med&global taulowTaiwanphiSS med&global taumedTaiwanphiSS med&global tauhigh

789

TaiwanphiSS high&global taulowTaiwanphiSS high&global taumedTaiwanphiSS high&global tauhigh

Node5:CrustalFormofDistributionofln(SA)

25

Branch Value

1 NormalDistribution

2 SWUSMixtureModel

26

27

SUBDUCTIONGMCSENSITIVITY

28

GMC—Subduction Sensitivity case(WS#3)

29

SpectralFrequency

PGA

SA[5.0Hz]

SA[0.5Hz]

MedianGMPE

EpistemicNode

(1.0)0.0

Basecase2 models

NewTaiwanmodelChao17

AdjustedmodelsAGA16_C04

ZH06

Formofln(SA)

Distribution

(1.0)

(1.0)

SWUS,CEN(1.0)

PhiSSlow+Taulow(1.0)

PhiSSlow+Taumed(1.0)

PhiSSlow+TauHigh(1.0)

PhiSSmed+Taulow(1.0)

PhiSSmed+Taumed(1.0)

PhiSSmed+TauHigh(1.0)

PhiSSHigh+Taulow(1.0)

PhiSSHigh+Taumed(1.0)

PhiSSHigh+TauHigh(1.0)

SingleStationSigmaModel

normal

mixture

Node1:SubductionGMPEforMedianRegion Branch GMPEsusedinSensitivity

Global 1 BCH(BCHydro,2016)

Taiwan 2 LL08(LinandLee,2008)

Japan 3 ZH06(Zhaoetal.,2006)

Taiwannewmodel 4 Chao17

Adjusted model 5 AGA16_C04

30

Node2:Subduction SigmaSS

n Usethe3TaiwanMagnitudeIndependent Phi_SSmodelswiththe3GlobalMagnitudeindependenttaumodels(9total).

31

Branch Value

123

TaiwanphiSS low&globaltaulowTaiwanphiSS low&globaltaumedTaiwanphiSS low&globaltauhigh

456

TaiwanphiSS med&global taulowTaiwanphiSS med&global taumedTaiwanphiSS med&global tauhigh

789

TaiwanphiSS high&global taulowTaiwanphiSS high&global taumedTaiwanphiSS high&global tauhigh

Node3:Subduction FormofDistributionofln(SA)

32

Branch Value

1 NormalDistribution

2 SWUSMixtureModel

33

34

GMCSENSITIVITYRESULTSForAdjustedModels- median

35

0.5

LL08

BCH

ZH06

Chao17

AGA16C01

NPP1

36

0.5

ASK14CY14CB14BSSA14I14ASB14BI14ASB14.C01BI14.C01ASK14.C01CY14.C01BSSA14.C01I14.C01CB14.C01Chao17

PGA

2sec

PGA

2sec

Crustal

Subduction2

2

1

1

Upper:Originalmodels

Lower:Adjustedmodels

new

new

0.5

LL08

BCH

ZH06

Chao17

AGA16C01

NPP2

37

0.5

ASK14CY14CB14BSSA14I14ASB14BI14ASB14.C01BI14.C01ASK14.C01CY14.C01BSSA14.C01I14.C01CB14.C01Chao17

PGA

2sec

PGA

2sec

Crustal

Subduction2

2

1

1

0.5

LL08

BCH

ZH06

Chao17

AGA16C01

NPP3

38

0.5

ASK14CY14CB14BSSA14I14ASB14BI14ASB14.C01BI14.C01ASK14.C01CY14.C01BSSA14.C01I14.C01CB14.C01Chao17

PGA

2sec

PGA

2sec

Crustal

Subduction2

2

1

1

0.5

LL08

BCH

ZH06

Chao17

AGA16C01

NPP4

39

0.5

ASK14CY14CB14BSSA14I14ASB14BI14ASB14.C01BI14.C01ASK14.C01CY14.C01BSSA14.C01I14.C01CB14.C01Chao17

PGA

2sec

PGA

2sec

Crustal

Subduction2

2

1

1

40

Thankyou!

BACKUP

41

Interface,plotfordistancescalingZtor =depth=0km,Mag=8.4

42

Interface,plotfordistancescalingZtor =depth=20km,Mag=8.4

43

Interface,plotfordistancescalingZtor =depth=50km,Mag=8.4

44

Interface,plotfordistancescalingZtor =2,depth=20km,Mag=8.4

45

NPP1- PGA- 5Hz– 0.5Hz

46

PGA 0.2s 2.0s

NPP2- PGA- 5Hz– 0.5Hz

47

PGA 0.2s 2.0s

NPP3- PGA- 5Hz– 0.5Hz

48

PGA 0.2s 2.0s

NPP4- PGA- 5Hz– 0.5Hz

49

PGA 0.2s 2.0s

GMCSensitivitycases– Crustal

50

Node SensitivityCase

1 GMPEforMedian

DonebeforeWM#3

2 Additional EpistemicUncertaintyforMedian

3 SigmaSS

4 FormofDistribution ofln(SA)

5 Splayfault

6 Deepevents

7 Dipimplementation forlistric fault

8 Directivitymodel

9 StyleofFaulting

10 Hanging-wallEffect

GMCSensitivitycases– Subduction

51

Node SensitivityCase

1 GMPEforMedian

DonebeforeWM#3

2 Additional EpistemicUncertaintyforMedian

3 SigmaSS

4 FormofDistribution ofln(SA)

5 Depthscalingforintraslab

6 LargeMagscalingforintraslab

7 EdgeEffectforinterface

52

53

54

55