Qua uantification n of of Top opography Ef Effect on on Se - - PowerPoint PPT Presentation

qua uantification n of of top opography ef effect on on
SMART_READER_LITE
LIVE PREVIEW

Qua uantification n of of Top opography Ef Effect on on Se - - PowerPoint PPT Presentation

Qua uantification n of of Top opography Ef Effect on on Se Seismi mic Grou ound nd M Mot otion on: A Case se Study i y in Norther ern T Taiwan Yi-Ching Lo 1 , Li Zhao 2 , and Shu-Huei Hung 1 1. Department of Geosciences, National


slide-1
SLIDE 1

Yi-Ching Lo1, Li Zhao2, and Shu-Huei Hung1

  • 1. Department of Geosciences, National Taiwan University, Taiwan
  • 2. Institute of Earth Sciences, Academia Sinica, Taiwan

August 15, 2016

Qua uantification n of

  • f Top
  • pography Ef

Effect on

  • n

Se Seismi mic Grou

  • und

nd M Mot

  • tion
  • n:

A Case se Study i y in Norther ern T Taiwan

slide-2
SLIDE 2
  • The Earth is not flat. Some regions have drastic variation in surface

relief, e.g. ±4 km in Taiwan region over ~100-km distance.

  • Topography affects wave propagation and therefore arrival times

and amplitudes of seismic waves

  • Neglecting topography:
  • Topography-induced travel time anomalies introduce biases in

seismic tomography models

  • Topography-induced amplitude anomalies lead to unrealistic

ground motion predictions

Motiva vations s for S Studying T Topography E y Effect ct

slide-3
SLIDE 3
  • Surface topography

influences the intensity of ground motion by focusing, defocusing and scattering of seismic waves.

  • Topography has been

ignored in most ground motion studies, leading to biases in PGV and PGA predictions.

Stu Studying T g Topo pogr graph aphy Effec ects ts o

  • n Ground M

d Moti tion by Nu Numerical S l Sim imula latio ions

Numerical simulations show that the effect of topography on PGV predictions can be up to ±50% for ground motion of ~0.5 Hz (Ma et al., 2007; Lee et al., 2009).

Lee et al. (2009) Ma et al. (2007)

Southern California Northern Taiwan

SAF SAF SAF SGM SGM SGM

slide-4
SLIDE 4
  • FDM code by Zhang et al. (2012)
  • Grid spacing: 300 m horizontally and

variable vertically (171.4 m near surface and increasing to 782.6 m at ~ 60-km).

  • Accurate waveforms up to 0.8 Hz
  • FD simulations on the IES HPC cluster
  • Newton’s second law:
  • Hooke’s Law:
  • Surface topography : curvilinear grids

transformed to Cartesian grids

Fin init ite-difference M ce Method ( (FDM)

, t ρ ∂ = ∇⋅ + ∂ v σ f 1 : [ ( ) ] , 2

T

t ∂ ′ = ∇ + ∇ − ∂ σ C v v m

, t ξ η ζ ∂ ∂ ∂ ∂ = + + + ∂ ∂ ∂ ∂ U U U U A B C F   

Zhang et al. (2012)

slide-5
SLIDE 5
  • Regional model from travel time tomography

(Kuo-Chen et al., 2012).

  • ETOPO1 topography: 1 arc-minute (~1.85 km).

Up to ~6 km topography contrast over 100-km distance in northern Taiwan.

Ta Taiwan: 3 3D Mode del a and To d Topo pography

  • 2
  • 1.5
  • 1
  • 0.5

0.5 1 1.5 2 2.5 3

ETOPO1 VP

2 3 4 5 6 7 8 20 10 30 40 Depth (km) (km/s)

VS

2 3 4 5

1.5 2.5 3.5 4.5

20 10 30 40 Depth (km) (km/s)

slide-6
SLIDE 6
  • Use 54 x 3 sources (at depths 7, 15 and 23

km) to examine P and S waves coming from a variety of directions.

  • Compute waveforms from the 162 sources to

all 38 stations with and without topography and measure the differences in their P- and S- wave travel times and amplitudes.

Sour Sources a and nd Statio tions

38 Stations (BATS and TAIGER)

slide-7
SLIDE 7

Compar ariso son o

  • f W

Wave veforms ( s (Z componen ent)

Black: “record” (with topography) Red: “Synthetics” (flat surface)

10 20 30 40 50 60 70 80 9

) c e s ( e m i T

B P N A

1 8 7 / 8 9

1 N S N

6 1 2 / 4 5

2 N S N

6 3 2 / 3 5

3 N S N

9 8 / 1 5

4 N S N

7 3 / 1 5

5 N S N

4 3 7 / 3 5

6 N S N

4 3 1 / 5 5

7 N S N

1 1 1 / 5 5

8 N S N

7 2 1 / 6 5

9 N S N

5 7 2 / 1 6

1 N S N

6 4 / 9 6

1 1 N S N

9 7 6 / 4 7

2 1 N S N

7 1 8 / 6 7

3 1 N S N

7 2 9 / 9 7

4 1 N S N

4 3 2 / 9

5 1 N S N

8 9 6 1 / 6 9

B C B S

1 7 / 6 2 1

O T A T

3 2 1 / 5 8

1 N G T

8 / 8 2 1

2 N G T

7 4 / 1 2 1

3 N G T

8 4 2 / 2 1 1

4 N G T

8 1 2 / 4 1

5 N G T

1 2 9 / 6 9

6 N G T

9 7 7 1 / 8

8 N G T

9 1 8 1 / 6

9 N G T

3 8 3 / 8 4

1 1 N G T

6 1 7 / 6 6

2 1 N G T

4 9 2 / 4 6

3 1 N G T

1 7 / 3 5

4 1 N G T

3 2 2 / 4

5 1 N G T

2 4 8 / 1 1 1

6 1 N G T

9 9 7 / 7 1

7 1 N G T 8 1 N G T

7 2 3 1 / 4 8

9 1 N G T 2 2 N G T

6 4 1 / 2 9

B S F W

5 3 4 / 2 7

B N H Y

5 3 9 / 5 8 8 9 6 1 / 6 9 7 2 9 / 9 7

distance (km)/station elevation (m)

slide-8
SLIDE 8

P S

Me Measuri ring T Tim ime and A Amplit itude A Anomali lies

  • “Record” and “synthetics” with and

without topography

  • Compute synthetic autocorrelation CA and

record-synthetic crosscorrelation CC

  • Bandpass filter the correlations around

3 frequencies: 0.1Hz, 0.2Hz and 0.5Hz

  • Time anomaly: lag time of

Amplitude anomaly:

P wave S wave

Broadband CC Broadband CC 0.1 Hz CC 0.1 Hz CC 0.2 Hz CC 0.2 Hz CC 0.5 Hz CC 0.5 Hz CC

slide-9
SLIDE 9

Z-comp. P (broadband) Z-comp. P (0.1Hz/10sec)

Freq equency-dependence ce o

  • f Topography E

y Effec ect

P-wave arrival time is more delayed at higher elevation and longer period, up to 0.4 sec.

Z-comp. P (0.2Hz/5sec) Z-comp. P (0.5Hz/2sec) N-comp. S (broadband) N-comp. S (0.1Hz/10sec) N-comp. S (0.2Hz/5sec) N-comp. S (0.5Hz/2sec)

δt δt

δlnA S-wave arrival time is more delayed at higher elevation and longer period, up to 0.7 s. Topography effect on S-wave amplitude has a complex pattern.

N-comp. S (broadband) N-comp. S (0.1Hz/10sec) N-comp. S (0.2Hz/5sec) N-comp. S (0.5Hz/2sec)

slide-10
SLIDE 10
  • P-wave delay times due to topography can be

up to ~0.4 s.

  • Delay times from topography have no clear

relationship with epicentral distance; increase with elevation.

  • Delay times have larger variation at longer

periods (red and green symbols) due to finite- frequency effect.

P-wav wave T Trav avel Time A e Anomalies d due to to T Topo pogr graph phy

Epicentral distance (km) δt (sec) Station elevation (m) δt (sec) broadband 0.1 Hz 0.2 Hz 0.5 Hz

δt vs. epicentral distance δt vs. station elevation

broadband 0.1 Hz 0.2 Hz 0.5 Hz

slide-11
SLIDE 11
  • S-wave delay times due to topography can be

up to ~0.7 s.

  • Delay times from topography have no clear

relationship with epicentral distance; increase with elevation.

  • Delay times have larger variation at longer

periods (red and green symbols) due to finite- frequency effect.

S-wav wave T Trav avel Time A e Anomalies d due to to T Topo pogr graph phy

δt vs. epicentral distance

Epicentral distance (km) δt (sec) broadband 0.1 Hz 0.2 Hz 0.5 Hz Station elevation (m)

δt vs. station elevation

δt (sec) broadband 0.1 Hz 0.2 Hz 0.5 Hz

slide-12
SLIDE 12
  • S-wave amplitude anomalies due to

topography have slightly negative mean, i.e. amplitude reduction.

  • Topography effect on amplitudes has no

clear relation with either distance or station elevation.

S-wav wave A Amplitude de Anomalies d due to to T Topo pogr graph phy

δlnA vs. epicentral distance

δlnA Epicentral distance (km) broadband 0.1 Hz 0.2 Hz 0.5 Hz

δlnA vs. station elevation

δlnA Station elevation (m) broadband 0.1 Hz 0.2 Hz 0.5 Hz

slide-13
SLIDE 13
  • In northern Taiwan, topography induces

up to 0.4 s and 0.7 s in P- and S-wave delay times, respectively, and up to 80% in S-wave amplitude anomaly.

  • P- and S-wave delay times have positive

means and no dependence on distance

  • Amplitude anomalies for S wave have

slightly negative mean and no dependence

  • n distance.

Topography E y Effects v s vs. Distan ance ce

P wave travel time anomaly S wave travel time anomaly S wave amplitude anomaly

Source Depths: 7 km 15 km 23 km

slide-14
SLIDE 14

Topography E Effects ts v vs. . Statio tion E Ele levatio tion

P wave travel time anomaly S wave travel time anomaly S wave amplitude anomaly

  • Delay times for P and S waves increase

with station elevation.

  • Increase of S-wave delay times with

station elevation is nearly twice as fast as P wave delay times.

  • Amplitude anomalies for S wave have no

dependence on station elevation. Source Depths: 7 km 15 km 23 km

slide-15
SLIDE 15

Topography-induced delay times δt :

  • No clear variation in the topography-induced P-wave delay times with back azimuth.
  • Clearly positive with larger values at high-elevation YHNB and smaller values at low-

elevation SBCB.

SBCB : Western Foothills with a low elevation of ~70.6 m TATO : Taipei Basin with an elevation of ~123.0 m YHNB : Central Range with a relatively high elevation of 934.9 m.

Varia riatio ion wi with th B Bac ack A Azimuth: P-wav wave Trav avel T Times

Source Depths 7 km 15 km 23 km

slide-16
SLIDE 16

Topography-induced amplitude anomalies δlnA:

  • Amplitude anomalies show complex variation with azimuth.
  • A majority of amplitude anomalies are negative (reduction).

Varia riatio ion wi with th B Bac ack A Azimuth: S-wav wave A Amplitude des

Source Depths 7 km 15 km 23 km

slide-17
SLIDE 17

 In northern Taiwan, topography induces up to 0.4 s and 0.7 s in P-

and S-wave delay times, respectively, and up to 80% in S-wave amplitude anomaly.

 Topography effect leads to positive anomalies in P- and S-wave

travel times (delays), and slightly negative anomalies in S-wave amplitudes (reduction).

 Delay times increase with station elevation but have no dependence

  • n epicentral distance. Amplitude anomalies do not depend on either

station elevation or epicentral distance.

 Topography-induced travel time delays do not vary with azimuth;

whereas topography effect on amplitude varies with azimuth in a complex fashion.

Summa mmary