Modeling of the subsurface structure from the seismic bedrock to the ground surface

for a broadband strong motion evaluation in the Kanto area, Japan

Shigeki Senna, Atsushi Wakai, Kaoru Jin, Takhiro Maeda, Hiroyuki Fujiwara

National Research Institute For Earth Science and Disaster Resilience

5th IASPEI/IAEE International Symposium: Effects of Surface Geology on Seismic Motion

Taipei Taiwan August 15-17, 2016

Introduction

In Japan, In order to estimate damages caused by strong ground motions from a mega-thrust earthquake, it’s important to evaluate broadband strong motion in wide area. To realize it, it’s necessary to sophisticate subsurface structure models, on which shallow and deep subsurface structures are integrated .

We think that a period characteristic and an amplification characteristic are the most important for modeling of subsurface structure. We make high-accuracy structure models, and we're available for the earthquake damage estimate of the building.

2

Contents

1) Subsurface structure model construction procedure 2) 3D Shallow subsurface structure models(Initial) 3) 3D Deep subsurface structure models (Initial) 4) Microtremor observation methods 5) Joint inversion and verification methods(Final) 6) Other topics(microtremor observation systems)

3

About

Borehole Data

Microtremor Data

Collected by Geological and Physical Data

Microtremor and Seismic Data ( Single point observation)

Initial Structure Model (Geological Models)

Microtremor(Array)

Phase Velocity Structure Model

Final Structure Model

H/V, R/V Spectrum

(About 250m grid models)

(About 250m grid model)

1) Subsurface structure model construction procedure

4

2) Outlines of 3D Shallow subsurface structure Models

・ Various methods for making Models. Conventional method using geomorphologic classifications :J-SHIS（250m grid） Detailed methods using a lot of boring data :recent NIED projects (250m grid） → following description

PGV site amplification factor (ARV) converted from Vs30

5

S-wave velocity layered models

Laterally continuous layers

with each N-Value

3-D geological structure

Relatinships of

N-Value and Vs

Boring columns

(stratigraphic divisions,N-Value)

Flow chart of modeling the Shallow subsurface structure models

Various Civil engineering projects ⇒ Boring data ⇒ Database(Geo-station)

6

Number of data are over 500,000 data

http://www.geo-stn.bosai.go.jp

Development of Integrated Geophysical and Geological Information Database

Geo-Station(NIED)

被害予測イメージ（モデル建物Aの場合）

崩壊大破中破小破軽微

NIED-DB

AIST-DB

PWRI-DB

MLIT

KuniJiban

GJI-DB

geological information

soil dynamics

ERI of

Tokyo University

Tokyo Institute of

Technology

Municipalities

DB

underground

structure

7

Initial shallow models (Geological models)

Urayasu City Hall

Collected borehole data in the Kanto district(N=200,000)

Chiyoda

Geological models

Soil models(250m grid)

Stratigraphic classification

Engineering Bedrock(Vs350)

Number of data(/250m grid)

Alt

itu

de

(m)

8

3) Outline of 3D Deep subsurface structure model

・Based on geological data and geophysical exploration data.

・Composed of different seismic velocity layers.

・Vertically divided into cubes with about 1km mesh surface.

1km

1km

9

Flow chart of modeling the Deep subsurface structure models

Subsurface structure data ・Deep borehole, Logging

・Seismic reflection , refraction survey

・Micro tremor survey

・Gravity survey

Surface geological data ・Topographical Map ・Geological Map

Distribution of physical properties ・Seismic wave velocity ・Density

Structure of Sedimentary layers ・depth ・thickness ･faluts ・folds

Layered structure model of physical property (0th-order geological model）

Simulation of seismic waveform using the Layered Model

Model modification compared to observed waveform data

Deep subsurface structure model for Strong-motion Evaluation (1st-order model and so on） = JIVSM(2012) 10

138.5 139 139.5 140 140.5 141 141.5

J-SHIS

34.5

35

35.5

36

36.5

37

138.5 139 139.5 140 140.5 141 141.5

Joint Inv.

34.5

35

35.5

36

36.5

37

0

10

20

40

60

100

200

400

800

1500

2000

2500

3000

3500

Depth(m)

Layer Vs(km/s) Vp(km/s) ρ(g/cm3)

1 0.35 1.6 1.85

2 0.40 1.6 1.85

3 0.45 1.7 1.90

4 0.50 1.8 1.90

5 0.55 1.8 1.90

6 0.60 2.0 1.90

7 0.65 2.0 1.95

8 0.70 2.1 2.00

9 0.75 2.1 2.00

10 0.80 2.2 2.00

11 0.85 2.3 2.05

12 0.90 2.4 2.05

13 0.95 2.4 2.10

14 1.0 2.5 2.10

15 1.1 2.5 2.15

16 1.2 2.6 2.15

17 1.3 2.7 2.20

18 1.4 3.0 2.25

19 1.5 3.2 2.25

20 1.6 3.4 2.30

21 1.7 3.5 2.30

22 1.8 3.6 2.35

23 1.9 3.7 2.35

24 2.0 3.8 2.40

25 2.1 4.0 2.40

26 2.1 4.0 2.40

27 2.7 5.0 2.50

28 2.9 4.6 2.55

29 2.7 5.0 2.50

30 3.1 5.5 2.60

31 3.2 5.5 2.65

<Deep model>

<Shallow model>

Yellow hatch are physical-properties value of the Kanto district.

Seismic bedrock Vs=3200(m/s)

upper depth Level(GL-m)

138.5 139 139.5 140 140.5 141 141.5

J-SHIS

34.5

35

35.5

36

36.5

37

138.5 139 139.5 140 140.5 141 141.5

Joint Inv.

34.5

35

35.5

36

36.5

37

0

10

20

40

60

100

200

400

800

1500

2000

2500

3000

3500

Depth(m)Initial value of physical properties of deep and shallow models

11 11

Integration model (Shallow and Deep)

This model is ‘Initial model’ 12

Geological strata line landfill , loam Alluvium, Diluvium etc.

Miniture and Irregular Array Method

(ex.Cho et al.(2013))

Normal Size Array Method

(R=800,400,200,100,50,25m) ＜Observation Spec＞ ・Seismometers = 6units ・Sampling=200(Hz)

・Observation Time 15(min.)

＜Observation Spec＞ ・Seismometers = 7～10units

・Sampling=100Hz ・Observation Time 30～80(min.)

R=800m

R=400m

R=200m

13

4) Microtremor observation methods

Dispersion analysis SPAC and (nc-)CCA( < 200m) methods

Miniature array Illegular array

Microremor observation for a subsurface structure model construction （ Two types of observation method ）

Miniature and Irregular array ( ～ Vs700(m/s)) Normal size array ( Vs300 ～ Vs3000(m/s))

Miniature and Irregular Array about every 1km Nomal Size Array about every 5km

14

2014（black） 992 points 2015(red） 7532 points 2016（blue） 5500 points 2017（green） 3500 points

2014（black） 76points 2015（red） 244points 2016（blue） 103points Data base （green） 460points

15

5) Correction of subsurface structure models using joint inversion method

15

JIVSM(2009-2012)

Final structure models(Deep Model Part)

138.5 139 139.5 140 140.5 141 141.5

J-SHIS

34.5

35

35.5

36

36.5

37

138.5 139 139.5 140 140.5 141 141.5

Joint Inv.

34.5

35

35.5

36

36.5

37

0

10

20

40

60

100

200

400

800

1500

2000

2500

3000

3500

Depth(m)

This Study Model(v7.4) J-SHIS Model(Initial)

Predominant period( >2s )

(sec)

138.5 139 139.5 140 140.5 141 141.5

J-SHIS

34.5

35

35.5

36

36.5

37

138.5 139 139.5 140 140.5 141 141.5

Joint Inv.

34.5

35

35.5

36

36.5

37

0

10

20

40

60

100

200

400

800

1500

2000

2500

3000

3500

Depth(m)

Top surface depth of the Vs=900(m/s) layer

Top surface depth of the Vs=3,200(m/s) (Seismic bedrock)

※ Vs≦3,200(m/s)

NIED

Mt. Tskuba

NARITA Airport

HANEDA Airport

16

The final structure models (Shallow model Part)

This Study Structure Models

AVS30(m/s) Predominant Period (sec)

AVS30(m/s)

Geomorphological Classification Models

(reference information)

※ Vs≦350(m/s)

NIED

Mt. Tskuba

17

Arakawa river

Edogawa river

Tone river

Tokyo bay’s Reclaimed land

Arakawa river

Edogawa river

Tone river

Tokyo bay’s Reclaimed land

Arakawa river

Edogawa river

Tone river

Tokyo bay’s Reclaimed land

Matsuoka and Wakamatsu(2013)

Verification analysis of models

Short Period Side （～1.0(s)）

Long Period Side （1.0(s)～10(s)）

Finite Difference Method (FDM)

One dimensional multi-reflection

Theory

18

Mountain

Verification of period and amplifying characteristics by FDM ①

Structure model for FDM

Grid size(m) number of grids Interval(s)

shallow part Shallow part Deep part

dx1 dy1 dz1 nx1 ny1 nz1 nx2 ny2 nz2

70 70 35 3789 4146 231 1263 1382 400 0.003125

SCEC(Southern California Earthquake Center) GOF（＝goodness-of-fit：𝐆𝐎𝐅 = 𝐥𝐧(𝐝𝐚𝐭𝐚/𝐦𝐨𝐝𝐞𝐥)）

※ Two horizontal components synthesis of fourier spectrum is used

※hypocenter is JMA , Hypocenter mechanism and seismic moment depend on F-net.

ＦＤＭ Aoi and Fujiwara(1999)

19

𝐂𝐆𝐎𝐅 =𝟏

𝟐𝐥𝐧(𝐝𝐚𝐭 𝐚 𝐦 𝐨𝐝𝐞𝐥) +

𝟏

𝟐𝐥𝐧(𝐝𝐚𝐭 𝐚 𝐦 𝐨𝐝𝐞𝐥)

（Dreger et al., 2015）

mean S.D.

Verification of period and amplifying characteristics by FDM ②

Result of joint inversion

KiK-net IBRH20 (HASAKI-2) KiK-net CHBH13(NARITA)

V7_4 : This Study Model JIVSM ： Japan Integrated Velocity Structure Model Ver.1(2009) CDMC ： Cabinet Office, Government of Japan Model(2012) JSHIS ： J-SHIS(ver2.0)(2009) 20

Result of joint inversion

Initial model

This study model

Obs. Dispersion,H/V

Initial model

This study model

Obs. Dispersion,H/V

Verification of period and amplifying characteristics by FDM ③ （Comparison with other past models）

21

V7_4 This Study Model CDMC ： Cabinet Office, Government of Japan Model(2012) JSHIS J-SHIS(ver2.0)(2009)

Verification of period and amplifying characteristics by FDM ③

（Comparison with other past models）

Average of all K-NET and KiK-net points(197points) in Kanto area. The distribution of the mean and the S.D.

22

V7.4 : This Study Model JIVSM ： Japan Integrated Velocity Structure Model Ver.1(2009-2012) CDMC ： Cabinet Office Government(Disaster Management) Model(2012-2016) JSHIS ： J-SHIS(ver2.0) Model (2009)

mean S.D.

Chiba Prefecture(1997) Dis

trib

uti

on

of

Bo

ugu

er

gr

avit

y an

om

aly(

mgl

)

Model(v7_4)

P-wave velocity (m/s)

Comparison of cross-sections by reflection method seismic exploration and velocity structure models by microtremor

Reflection method seismic exploration result

Bouguer gravity anomaly 23

Conclusions

24

・ We constructed the detailed subsurface structure models from seismic bedrock to ground surface by a 250m mesh in the whole of Kanto region. ・ About accuracy of broadband seismic ground motions computed with Finite Difference Method(FDM), these models are improved in comparison with the existing models. Above all, period and amplification characteristics are improved drastically, from 2 to 10 seconds, in deeper layers than engineering bedrock surface.

Reconsideration of subsurface structure models around engineering bedrock (from Vs350m/s to Vs500m/s) including shallower ground. (Verification of models using records of miniature and irregular array microtremor observation conducted at an interval of 1km)

●Future tasks

6) Other topics(microtremor system)

Please visit poster session ‘P201D’ 25 25

Thank you for your attention.