How sensitive is earthquake ground motion to source ... - EFISPEC

Anastasia Kiratzi. 4 ... 10. 100. 315−360 deg. Frequency (Hz). TST0/TST5. 0.5. 1.0. 4.0. 1. 10. 100 ..... to the South cause a localized band of large values close to ...Missing:
17MB taille 2 téléchargements 281 vues
How sensitive is earthquake ground motion to source parameters? Insights from a numerical study in the Mygdonian basin 1,3 1 1 2 Emmanuel Chaljub , Emeline Maufroy , Florent de Martin , Fabrice Hollender , Cédric Guyonnet-Benaize3, Maria Manakou4, Alexandros Savvaidis5, Anastasia Kiratzi4, Zaferia Roumelioti4, and Nikos Theodoulidis5

Amplification measured by SSR

Reciprocity-based calculations 0.005

b

c

H1k D10k Az 320

4.0

1.0 225−270 deg.

4.0

45−90 deg. 100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

4.0

1.0

4.0

4.0

100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

1 0.5

270−315 deg.

100

CMP T, DIST. = 5000 m, DEPTH = 1000 m

1.0 225−270 deg.

45−90 deg.

10

1.0

4.0

1.0 225−270 deg.

4.0

45−90 deg. 100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

10

10

1 0.5

270−315 deg.

100

CMP T, DIST. = 10000 m, DEPTH = 1000 m

1 0.5

4.0

1 1.0 4.0 0.5 Frequency (Hz)

1 0.5

4.0

0−45 deg. 100

1 1.0 4.0 0.5 Frequency (Hz)

315−360 deg.

1.0

4.0

4.0

100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

100

10

10

1 0.5

270−315 deg.

100

CMP T, DIST. = 15000 m, DEPTH = 1000 m

1.0 225−270 deg.

45−90 deg.

4.0

1 0.5

0−45 deg.

100

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

10

CMP T, DIST. = 20000 m, DEPTH = 1000 m

1.0 225−270 deg.

4.0

45−90 deg.

100

1 0.5

ALL AZIMUTHS

100

100

100

100

100

100

100

100

100

100

100

100

100

10

10

10

10

10

10

10

10

10

10

10

10

10

10

10

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

4.0

1 0.5

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

1 0.5

4.0

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

1 0.5

4.0

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

4.0

1 0.5

1.0

4.0

180−225 deg.

1 0.5

100

100

100

100

100

100

100

100

100

100

10

10

10

10

10

10

10

10

10

10

1 4.0 0.5

1.0

1.0

1 0.5

4.0

1.0

1 4.0 0.5

1.0

1 0.5

4.0

1.0

1 4.0 0.5

1.0

1 0.5

4.0

1 4.0 0.5

1.0

1.0

1 0.5

4.0

24

10

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

1.0 225−270 deg.

4.0

100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

4.0

Sigma-duration

Simulation of real events duration

1.0

1 4.0 0.5

1.0 135−180 deg.

1.0

1 4.0 0.5

1.0 90−135 deg.

1.0

4.0

4.0

4.0

S2

1.0 225−270 deg.

4.0

45−90 deg. 100

10

10

1 0.5

1.0 90−135 deg.

1 0.5

270−315 deg.

100

ALL AZIMUTHS

1.0

4.0

1.0 225−270 deg.

4.0

45−90 deg. 100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

10

1 0.5

1 0.5

4.0

10

1 1.0 4.0 0.5 Frequency (Hz)

315−360 deg. TST0/TST5

10

270−315 deg.

100

CMP T, DIST. = 10000 m, DEPTH = 5000 m

1 0.5

4.0

1 1.0 4.0 0.5 Frequency (Hz)

TST0/TST5

10

0−45 deg. 100

1.0

4.0

4.0

100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

100

10

10

1 0.5

270−315 deg.

100

CMP T, DIST. = 15000 m, DEPTH = 5000 m

1.0 225−270 deg.

45−90 deg.

4.0

1 0.5

0−45 deg.

100

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

10

CMP T, DIST. = 20000 m, DEPTH = 5000 m

1.0 225−270 deg.

4.0

45−90 deg.

100

1 0.5

ALL AZIMUTHS

100

100

100

100

100

100

100

100

100

100

100

100

100

100

10

10

10

10

10

10

10

10

10

10

10

10

10

10

10

1 0.5

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1 0.5

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

1 0.5

4.0

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

1 0.5

4.0

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

4.0

1 0.5

1.0

4.0

180−225 deg.

1 0.5

100

100

100

100

100

100

100

100

10

10

10

10

10

10

10

10

10

10

1 4.0 0.5

1.0

100

10

10

1 0.5

100 10

1.0

1 0.5

4.0

0−45 deg.

100

270−315 deg.

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

1.0

1 4.0 0.5

315−360 deg.

4.0

45−90 deg. 100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

100

10

10

1 0.5

4.0

1 0.5

1.0

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

1.0

1 4.0 0.5

315−360 deg.

4.0

45−90 deg. 100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

100

10

10

1 0.5

1.0 225−270 deg.

1 0.5

4.0

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

1 4.0 0.5

1.0

315−360 deg.

4.0

45−90 deg. 100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

100

10

10

1 0.5

1 0.5

4.0

1.0

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

1.0

1 4.0 0.5

315−360 deg.

4.0

45−90 deg. 100

10

10

1 0.5

ALL AZIMUTHS

1.0 90−135 deg.

100

10

10

1 0.5

4.0

1 0.5

1.0 135−180 deg.

1.0

1 4.0 0.5

1 1.0 4.0 0.5 Frequency (Hz)

1.0

4.0

4.0

45−90 deg.

10 1 0.5

ALL AZIMUTHS

100

100

100

100

100

100

100

100

100

100

100

100

100

100

10

10

10

10

10

10

10

10

10

10

10

10

10

10

10

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

4.0

1 0.5

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

1 0.5

4.0

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

1 0.5

4.0

1.0

4.0

180−225 deg.

1 0.5

1.0 135−180 deg.

1 4.0 0.5

1.0

4.0

1 0.5

1.0

4.0

180−225 deg.

1 0.5

100

100

100

100

100

100

100

100

100

100

10

10

10

10

10

10

10

10

10

10

1.0

1 4.0 0.5

1.0

1 0.5

4.0

1.0

1 4.0 0.5

1.0

1 0.5

4.0

D=5 km

1.0

1 4.0 0.5

1.0

1 0.5

4.0

D=10 km Epicentral Distance D

1 4.0 0.5

1.0

1.0

1 0.5

4.0

D=15 km

1.0

4.0

100

100

1 0.5

1.0

4.0

4.0

CMP T, DIST. = 20000 m, DEPTH = 15000 m

1.0 225−270 deg.

1.0 90−135 deg.

0−45 deg.

100

270−315 deg.

100

CMP T, DIST. = 15000 m, DEPTH = 15000 m

1.0 225−270 deg.

1 0.5

4.0

0−45 deg.

100

270−315 deg.

100

CMP T, DIST. = 10000 m, DEPTH = 15000 m

1 0.5

4.0

1.0

0−45 deg.

100

270−315 deg.

100

CMP T, DIST. = 5000 m, DEPTH = 15000 m

1.0 225−270 deg.

1 0.5

4.0

0−45 deg.

100

270−315 deg.

100

CMP T, DIST. = 2500 m, DEPTH = 15000 m

1.0 225−270 deg.

4.0

100

315−360 deg.

1 0.5

1.0

100

1 0.5

1 4.0 0.5

1.0 135−180 deg.

1.0

1 4.0 0.5

1.0 90−135 deg.

1.0

4.0

4.0

4.0

D=20 km

The top Figs. show the amplification at TST (measured by the SSR between the surface and downhole [-197 m] TST receivers) for various source depths and epicentral distances. Each subfigure displays the SSR for azimuths gathered in 45 deg. bins. The SSR are quite stable, although some azimuthal variability is observed in the frequency range dominated by surface waves (around 1 Hz, between the fundamental and first harmonic resonance frequencies). Note that this variability does not decrease with distance.

Comparison to observations

1

10

0

10 0.5

1

10

0

0.6

0.7

0.8

0.9 1.0

Frequency (Hz)

2.0

3.0

4.0

10 0.5

0.6

2

c

T component

b

1D median REAL DATA standard deviation

Amplification ratio TST0/TST5

Amplification ratio TST0/TST5 Earthquake #5 − component Y

Earthquake #4 − component X

Earthquake #3 − component Y

4.0

10

315−360 deg. 100

S5

S4

1

10

0.7

0.8

0.9 1.0

Frequency (Hz)

2.0

3.0

4.0

−1

1−D R0 R1 L0 L1

1

10

0

10 0.5

0.6

0.7

0.8 0.9 1.0

Frequency (Hz)

2.0

3.0

4.0

10 −1

10

0

10

10

Amplitude of response spectra

−2

10

Amplitude of response spectra

10

Amplitude of response spectra

Amplitude of response spectra

−2

0

−2

10

−1

10

−3

10

−3

10

−1

10 10

Frequency (Hz)

0

1

10

0

10

0

10

Frequency (Hz)

0

1

10

0

10

0

10

Frequency (Hz)

10

Frequency (Hz)

Earthquake #3 − component Y

TST0/TST5 spectral ratio

Earthquake #2 − component Y

TST0/TST5 spectral ratio

TST0/TST5 spectral ratio

1

10

10

0

10

Frequency (Hz)

0

1

10

0

10

0

10

Frequency (Hz)

10

Frequency (Hz)

Earthquake #4 − component X

TST0/TST5 spectral ratio

10

0

Frequency (Hz)

Earthquake #5 − component Y

TST0/TST5 spectral ratio

0

Earthquake #1 − component Y

SSR TST0/TST5

S3

Earthquake #2 − component Y

Earthquake #1 − component Y

Amplitude of response spectra

Acc response spectrum

Varying the locations of real events

1.0

0−45 deg. 100

100

a

S1

1 0.5

1 1.0 4.0 0.5 Frequency (Hz)

CMP T, DIST. = 5000 m, DEPTH = 5000 m

D=2.5 km

Note that the deviation of PGA clearly increases with distance for the shallow sources (depth < 5 km). Note also that the duration of ground motion seems to be controlled by the local basin response rather than by the distance to the source.

1 0.5

270−315 deg.

100

CMP T, DIST. = 2500 m, DEPTH = 5000 m

1 0.5

45−90 deg.

10

TST0/TST5

TST0/TST5

TST0/TST5 100

Sigma-PGA The left Figs. show the evolution with epicentral distance of the median (and of the deviation from it) of PGA and duration computed for the 980 sources. The curves are gathered by source depth values.

S3

1 0.5

270−315 deg.

10

315−360 deg. 100

TST0/TST5

20

10

0−45 deg. 100

TST0/TST5

16

10

315−360 deg. 100

TST0/TST5

12

Time (s)

0−45 deg. 100

TST0/TST5

8

315−360 deg. 100

1 0.5

We designed a spectral element mesh of the Mygdonian basin which is refined both horizontally and vertically (see Figs. a and b). The mesh includes surface topography and the size of the elements is adapted to the S wavelengths in the basin (see Fig. c). The size of the computational domain is 47 km x 64 km in EW and NS directions, and 30 km in the vertical direction. The mesh is made of 4.7 millions elements and contains 314 millions gridpoints. The velocity model in the basin is defined by two linear gradients, one between the surface and the premygdonian-mygdonian limit and the second one down to the sediment-bedrock interface. The surface S wave velocity is 137 m/s and the grid allows accurate calculations for frequencies up to 4 Hz. .

S1

1 0.5

1 1.0 4.0 0.5 Frequency (Hz)

10

315−360 deg. 100

Amplification ratio TST0/TST5 or group velocity (dam/s)

PGA

S2

1 0.5

0−45 deg. 100

100

TST0/TST5

-0.005

We computed the response at the TST station (red triangle) due to 980 sources located at epicentral distances D=2.5, 5, 10, 15, 20, 30 and 40 km and depths (relative to sea level) 1, 2.5, 5, 10 and 15 km. The source is a vertical strike-slip which is rotated such that the TST station is in the maximum of the S-radiation pattern. We computed the median PGA and duration for each circle of sources. Figs a and c show the deviation from the median of PGA and duration for shallow sources (1 km). Fig. b shows the time series of ground acceleration at TST for 2 sources with different backazimuths. Note the slight anti-correlation of PGA and duration caused by the different arrival times of surface waves diffracted off the Northern basin edge.

We computed the response of the basin to five real events, which occured in the last years. The source parameters are given in the table above and are represented as beachballs on the situation map shown on the left. These events were well recorded by the Euroseistest accelerometric array which is centered on the TST station (red triangle). The circular crosses indicate the positions of the sources that were considered in reciprocity-based calculations. The maps of peak ground velocity are shown for the five events (S1 to S5). Note the differences in the distribution of PGV following the position of the sources. In particular, the events located to the South cause a localized band of large values close to the basin edge. Note also the distortion of the distribution of peak values outside the basin, due to surface topography.

10

315−360 deg. 100

100

1 0.5

h=5 km

0.000

4

S4

10

270−315 deg.

100

CMP T, DIST. = 2500 m, DEPTH = 1000 m

1 0.5

h=15 km

H1k D10k Az 140

Source depth h

0.010

Jan 01 (001), 1970 00:00:00.012

S5

10

1.0

0−45 deg. 100

-0.010

-0.010

c

100

1 1.0 4.0 0.5 Frequency (Hz)

315−360 deg. 100

-0.005

We present results showing (1) the sensitivity of ground motion parameters to the location of the five real seismic sources; and (2) the variability of the amplification caused by site effects, as measured by standard spectral ratios, to the source characteristics

b

1 0.5

1 0.5

0.005

a

10

0.000

Two kinds of simulations are performed: (1) direct simulations of the surface ground motion for real regional events having various back azimuth with respect to the center of the basin; (2) reciprocity-based calculations where the ground motion due to 980 different seismic sources is computed at a few stations in the basin. In the reciprocity-based calculations, we consider epicentral distances varying from 2.5 km to 40 km, source depths from 1 km to 15 km and we span the range of possible back-azimuths with a 10 degree bin.

Spectral Element calculations

10

270−315 deg.

h=1 km

a

0.010

0−45 deg. 100

TST0/TST5

TST0/TST5

315−360 deg. 100

TST0/TST5

Understanding the origin of the variability of earthquake ground motion is critical for seismic hazard assessment. Here we present the results of a numerical analysis of the sensitivity of earthquake ground motion to seismic source parameters, focusing on the Mygdonian basin near Thessaloniki (Greece). We use an extended model of the basin (65 km [EW] x 50 km [NS]) which has been elaborated during the Euroseistest Verification and Validation Project. The numerical simulations are performed with two independent codes, both implementing the Spectral Element Method. They rely on a robust, semi-automated, mesh design strategy together with a simple homogenization procedure to define a smooth velocity model of the basin. Our simulations are accurate up to 4 Hz, and include the effects of surface topography and of intrinsic attenuation.

TST0/TST5

Abstract

TST0/TST5

1: ISTerre, Grenoble University, France, 2: BRGM, Orléans, France, 3: CEA, Cadarache, France, 4: Aristotle University,Thessaloniki, Greece, 5: ITSAK, Thessaloniki, Greece

1

10

Acknowledgments 0

10

0

10

Frequency (Hz)

The central Fig. b shows the amplification curves for the 980 sources. The solid black line is the median and the dotted lines are the 16th and 84th percentiles (corresponding to median+/- 1 sigma for a normal or lognormal distribution). The green line is the theoretical 1D amplification for a vertically incident SH wave. Note the larger variability in the frequency range between the resonance frequencies, which is caused by local surface waves. Fig. a shows the median (and median +/- 1 sigma) observed SSR ratios measured at TST for a set of more than 20 real events with even azimuthal coverage. The larger variability around 1 Hz is clearly present. Fig. c displays the group velocity of the fundamental and first higher mode of Rayleigh and Love modes at TST (based upon the model used in the calculations). The larger variability in the synthetic SSR seems to coincide with a Airy phase of the first harmonic of Rayleigh waves.

0

10

Frequency (Hz)

For each of the 5 real events, we computed the basin response at TST (GL and GL-197m) for 125 sources which hypocenter was shifted by +/- 1 km or +/- 2 km in the X,Y,Z directions. The variability of the acceleration spectra (top) and of the amplification measured by Standard Spectral Ratio (between TST[GL] and TST[GL-197] is shown for the most energetic horizontal component. Colors indicate source depth from blue (deep) to red (shallow). Note the large variability for the shallow distant source (S1) and the deep close source (S5).

A. Kiratzi, N. Theodoulidis, Z. Roumelioti and A. Savvaidis acknowledge partial financial support from the Research Funding Program THALES co-financed by the European Union (ESF) and Greek national funds. The direct calculations were done with the efispec software on the Froggy platform of the Grenoble HPC center CIMENT. The reciprocity-based calculations were done with the specfem software on the Curie machine (TGCC,CEA) under GENCI project t2014046060.