DP RC - Geological Survey of Japan / AISTDP RC Disaster Prevention Research Center, National Cheng...

DP RCDisaster Prevention Research Center,National Cheng Kung University, Taiwan

How to choose the sensitive site to earthquakes?How to choose the sensitive site to earthquakes?Studies of spatial sensitivity of the Studies of spatial sensitivity of the

hydrological response to earthquakeshydrological response to earthquakes

* * WenWen--Chi LaiChi Lai1,21,2, , ChjengChjeng--LunLun ShiehShieh1,31,3, , KuoKuo--Chin HsuChin Hsu22,,MingMing--JameJame HorngHorng44, Kai, Kai--WenWen KuoKuo55

1. Disaster Prevention Research Center, NCKU 2. Department of Resources Engineering, NCKU3. Department of Hydraulic and Ocean Engineering, NCKU4. Water Resource Agency, Ministry of Economic Affair, Taiwan5. Central Weather Bureau, Ministry of Transportation and Communications

50km

########

########

#### ####

#######

##

####

#

###

## #####

####

###

######## ##

#####

####

#####

###### ###

##

####

#####

##

##

###

###########

#### ## ############

#####

######### ##

#######

###### #####

######## ####

##

######

####

#### ###

#### ###

## ####

####

############# ######

#######

######

##

###

## ###

##

##

###

####

#######

####

######## ##

##

#

#

###

############

##

############

###

#

###

###

####

##

#####

#

#

###

#####

###

####

#

##

##

#####

######

####

###

######

#########

##

######

#

##

######

###

###

####

#######

########

#####

##

########

####

######

##

##

########

## #######

#### #

#

###

#

##

#

######

###

####

#

##

##########

1. Motive1. Motive

▼▼1.0

▼2.0

▼

▼

▼

▼

▼

▼

▼8.0

▲1.0

◎JUS◎TWN◎WUJ

◎CHK

◎HRK◎TLO

◎HUL▼

◎CHS

◎HUT

◎DHR

◎SBT◎TNS◎LUJ◎NBA

◎JDR

TUY ◎

2006/04/01 Taitung 2006/04/01 Taitung Earthquake, Taiwan Earthquake, Taiwan

MMLL 6.46.4

Volumetric Strain(micro str)

＋＋: Dilation : Dilation －－: Contraction : Contraction

MICAP-G(Naito & Yoshikawa, 1999)

Choosing the sensitive Choosing the sensitive wells for wells for best coveragebest coverage of of the seismic activity area. the seismic activity area.

Some wells seems more Some wells seems more sensitive to some special sensitive to some special areaarea, and not so sensitive , and not so sensitive to others.to others.

Most observed changes Most observed changes fitted to fitted to strain fieldstrain field, but , but some wells always not.some wells always not.

33

DP RC2.Objective2.Objective

Use the recorded coseismic groundwater level Use the recorded coseismic groundwater level changes to study the three problems:changes to study the three problems:

包包11①①Why the cosesimic groundwater level changes Why the cosesimic groundwater level changes happen in some earthquakes and didnhappen in some earthquakes and didn’’t in other t in other earthquakes? earthquakes? ~ Criteria of the detectability.~ Criteria of the detectability.

QQ2 2 ①①Where are the high / low sensitivity areas? Where are the high / low sensitivity areas? ~ Spatial distribution of the detectability. ~ Spatial distribution of the detectability.

QQ3 3 ①① What the reasons made the different responses to What the reasons made the different responses to the earthquake? the earthquake? ~ Structural anisotropy / mechanical heterogeneity~ Structural anisotropy / mechanical heterogeneity

44

DP RCCriteria by the Moments & DistancesCriteria by the Moments & Distances

1 10 100 1000Distance (km)

2

3

4

5

6

7

8

Mag

nitu

de(M

L)

Undetectable

Detactable

M=2.2*Log ( D )+1.5

◆◆ HULHUL◆◆ CHSCHS◆◆ TWNTWN◆◆ LUJLUJ

Obs. well

Data Period:Data Period:

2003~20052003~2005

Detected records:Detected records:

103103

Earthquake events:Earthquake events:

125Should consider the spatial (directional) Should consider the spatial (directional) relationship, not only distance (1D)relationship, not only distance (1D) 125

55

DP RC

Comparison of the theoretic and Comparison of the theoretic and observed responses observed responses

(c) LUJ (d) NAB

-50 -40 -30 -20 -10 0 10

-10

-8

-6

-4

-2

0

2

-10 0 10 20

-2

-1

0

1

2

n = 1

n = 30

0 20 40 60 80 100 120

-30

-25

-20

-15

-10

-5

0

-5 -4 -3 -2 -1 0

-2

-1

0

1

2

n = 1

n = 1

n = 1

n = 60

(a) HUL (b) TWN0 20 40 60 80

-3

-2

-1

0

1

2

n = 1

n = 60

0 30 60 90 120 150

-1.5

-1

-0.5

0

0.5

1

n = 1

n = 5

-6 -4 -2 0 2 4

-1

-0.5

0

0.5

1

n = 1

n = 300

-4 -3 -2 -1 0 1

-1

-0.5

0

0.5

1

n = 1

(g) TLO (h) SIP

(e) HRD (f) DHR

Am

p. O

f Gro

undw

ater

Lev

el C

hg. (

cm)

Am

p. O

f Gro

undw

ater

Lev

el C

hg. (

cm)

Static Volumetric Strain (10Static Volumetric Strain (10--88))

66

DP RCPhilippine Sea Plate

Philippine Sea Plate

Eurasian Plate

Eurasian Plate

Central Mountain

Central Mountain

Range(Accretional

Range(Accretional

Wedge)Wedge)

Luzon Arc

Luzon Arc

Ryuku Arc

Ryuku Arc

TaipeiTaipei

Kaoshiung

Kaoshiung

TainanTainan

TaichungTaichung

TaitungTaitung

HulienHulien

IlanIlan

JDR

CHS

DHR

TLO

SBTTUS

LUJ

NBATYE

HRKTUK

HTN

JUS

TWNWUJ

HUL

Homogeneous

HeterogenousEQEQ11

DetectDetect

EQEQ22

NoNo--DetectDetect

77

DP RC2.Objective2.Objective

Use the recorded coseismic groundwater level Use the recorded coseismic groundwater level changes to study the three problems:changes to study the three problems:

包包11①①Why the cosesimic groundwater level changes Why the cosesimic groundwater level changes happen in some earthquakes and didnhappen in some earthquakes and didn’’t in other t in other earthquakes? ~ Criteria of the detectability.earthquakes? ~ Criteria of the detectability.

QQ2 2 ①①Where are the high / low sensitivity areas? Where are the high / low sensitivity areas? ~ Spatial distribution of the detectability.~ Spatial distribution of the detectability.

QQ3 3 ①① What the reasons made the different responses to What the reasons made the different responses to the earthquake? the earthquake? ~ Structural anisotropy / mechanical heterogeneity~ Structural anisotropy / mechanical heterogeneity

SStrategtrategy of sensitive sites choosingy of sensitive sites choosing

88

DP RC

2001

2002

2003

2004

CHK

HULHUL

JDRNBA

LUJLUJTUS

DHR

TLO

2005

CHSCHS

JUS

SBT

HRT

TUY

HRKTWNTWN WUJ

3.Observation3.ObservationObservation WellObservation Well

44 observation wellobservation well上上1515detectable recordsdetectable records

Observation PeriodObservation Period

20032003--20062006 MMLL上上55EarthquakeEarthquakeㄨㄨTotal Total 125125Events.Events.

Total Total 8383 coseismic changes, coseismic changes, step changes (S) step changes (S) 4545 records, records, oscillation (O) oscillation (O) 3838 recordsrecords

Observation Network (01’~05’)

50km

99

DP RC3.Observation3.Observation

6/10/03M6.5

12/10/03M6.6

5/19/04M6.5

11/8/04M6.7

Events of the earthquake ML >3 in Taiwan 03’~04’

CMT solution and distribution of events3

4

5

6

7

03/1

03/2

03/3

03/4

03/5

03/6

03/7

03/8

03/9

03/1

003

/11

03/1

204

/104

/204

/304

/404

/504

/604

/704

/804

/904

/10

04/1

1

Mag

nitu

deM

agni

tude

MMLL 3~3.93~3.9 4~4.94~4.9 55~5.9~5.9 上上6.06.0

20032003 118118 181181 4343 33

20042004 8686 125125 2525 55

20052005 277277 140140 2424 33

1010

DP RC4.Methdology (1/3)4.Methdology (1/3)

Variogram Variogram ( ~( ~ covariance in space)covariance in space)The variogram is a measure of disThe variogram is a measure of dis--similarity similarity between two points in space separated by a between two points in space separated by a distance distance hh..

……..(1)(1)

22末末((hh): Variogram value): Variogram valueZ(Z(uu): value of the specified variate): value of the specified variate

Z(Z(u+hu+h): value with spacing ): value with spacing hhVar [ ]: Var [ ]: variance operatorvariance operator

( ) ( )[ ]uZhuZVarh −+=)(2γ

Z(u)

Z(u+h)

Z(u+2h)

1111

DP RCDistribution models of the Distribution models of the variogramvariogram

(D G Rossiter, 2006)

1. Curve Fitting1. Curve Fitting2. Choose suitable model2. Choose suitable model3. Finding spatial parameters3. Finding spatial parameters

1212


KrigingKriging ((MatheronMatheron, G., 1962), G., 1962)( ~( ~ Interpolation data by distribution model)Interpolation data by distribution model)The Kriging methods are for data interpolation inThe Kriging methods are for data interpolation in Best, Best, Linear, Unbiased, EstimateLinear, Unbiased, Estimate ( ( BLUE BLUE )) assumption.assumption.

…………. (2). (2)

犯犯00: estimate value: estimate valueN(N(犯犯00): numbers of neighborhood): numbers of neighborhood甲甲ii: weighting: weighting犯犯ii : sample value: sample value

∑=

=)N(

1iii0

0

ων

νν

ν1

ν2

ν3

ν4

ν5

ν7

ν8νν00

1313


Indicator Indicator KrigingKriging ((JournelJournel, 1983), 1983)

The main difficult of the spatial analysis is theThe main difficult of the spatial analysis is the samples samples number is limited (number is limited (又又20~3020~30)). We solve the problem by . We solve the problem by consider the consider the undetectable eventsundetectable events, then the samples increase , then the samples increase to all earthquakes to all earthquakes ((125125))..

Indicator KrigingIndicator Kriging’’s made index transfer ( s made index transfer ( Yes: 1 Yes: 1 ㄩㄩNo: 0No: 0 ) ) result shows result shows ProbabilitiesProbabilities (the probability that the grade is (the probability that the grade is above the above the detect criteriadetect criteria: exp. : exp. 11) or) or ProportionsProportions (the (the proportion of the block above the proportion of the block above the detect criteriadetect criteria : exp. : exp. 11 on on data support).data support).

(a) Original data (b) Variogram and modeling(a) Original data (b) Variogram and modeling5.Result: HUL well (1/2)5.Result: HUL well (1/2)

Step (15)Step (15)Oscillation (22)Oscillation (22)Undetectable(88)Undetectable(88)

N = 125 N = 125

Exponential Model

50km

5.Result: HUL well (2/2)5.Result: HUL well (2/2)03/1~06/11 ML≧5.0Tectonic control caseTectonic control case

•• Mostly Mostly ＜＜ 30% 30% for M5for M5

•• Sensitive to eastern Sensitive to eastern offshore( offshore( 50% 50% ), low to ), low to western Taiwan. western Taiwan.

•• Low sensitive to Low sensitive to northern Taiwannorthern Taiwan

•• Difference responses Difference responses in in Philippine Sea PlatePhilippine Sea Plateand and Eurasian PlateEurasian Plate

Philippine Sea Philippine Sea PlatePlate

Eurasian PlateEurasian Plate

Step (15)Step (15)Oscillation (22)Oscillation (22)Undetectable(88)Undetectable(88)

N = 12550km

1616

DP RC

5.Result: TWN well (1/2)5.Result: TWN well (1/2)

N = 125 N = 125

Step (11)Step (11)Oscillation (4)Oscillation (4)Undetectable (110)Undetectable (110)

(a)Original data (b)Variogram and modeling(a)Original data (b)Variogram and modeling

Exponential Model

50km

5.Result: TWN well (2/2)5.Result: TWN well (2/2)

N = 125

Homogeneous caseHomogeneous case•• Sensitive to Sensitive to eastern eastern offshore area (40%)offshore area (40%)..

•• Responses to southResponses to south--western earthquakes.western earthquakes.

•• Homogeneous responseHomogeneous responseto crustal strain.to crustal strain.

•• Sensitive well, low noise Sensitive well, low noise and few oscillation and few oscillation record (ground motion)record (ground motion)


03/1~06/11 ML≧5.0

50km

(a) Original data (b) Variogram and modeling(a) Original data (b) Variogram and modeling5.Result: LUJ well (1/2)5.Result: LUJ well (1/2)

N = 125 N = 125


Gaussian Model

50km

5.Result: LUJ well (2/2)5.Result: LUJ well (2/2)

N = 125

Structural control caseStructural control case•• Mostly Mostly 20~40%20~40% for M5for M5•• Sensitive to southSensitive to south--western Taiwanwestern Taiwan(>60%)(>60%) , , low in eastlow in east--west trend. west trend.

•• Sensitive well, low noise Sensitive well, low noise and few oscillation and few oscillation record (ground motion)record (ground motion)

•• FaultFault--Barrier effect Barrier effect •• FaultFault--Conduit effectConduit effect


03/1~06/11 ML≧5.0

50km

5.Result: CHS well (1/2)5.Result: CHS well (1/2)

N = 125 N = 125


(a) Original data (b) Variogram and modeling(a) Original data (b) Variogram and modeling

Exponential Model

50km

5.Result: CHS well (2/2)5.Result: CHS well (2/2)

N = 125

Structural control caseStructural control case•• Mostly Mostly ＜＜30% 30% for M5for M5•• Sensitive to southern Sensitive to southern Taiwan Taiwan (50~70%)(50~70%). .

•• Sensitive well, larger Sensitive well, larger noise and lots oscillation noise and lots oscillation record (ground motion)record (ground motion)

•• FaultFault--Barrier effectBarrier effect, , made low transmit of the made low transmit of the strain.strain.


03/1~06/11 ML≧5.0

50km

2222

DP RC6.Conclusion6.Conclusion

From the spatial analysis of the detectability, they shows From the spatial analysis of the detectability, they shows the the highly anisotropy and heterogeneityhighly anisotropy and heterogeneity in three wells. in three wells. They could partly explain the different responses of They could partly explain the different responses of earthquake induced groundwater changes. earthquake induced groundwater changes.

Tectonic and structural geologyTectonic and structural geology setting could be the setting could be the main reason control the spatial difference of earthquake main reason control the spatial difference of earthquake induced groundwater changes (induced groundwater changes (FaultFault--Barrier EffectBarrier Effect). ).

The strain modelThe strain model usually could explain the type of the usually could explain the type of the coseismic change, but the amplitudes usually not fit to coseismic change, but the amplitudes usually not fit to the homogeneous assumption. The the homogeneous assumption. The Structural Structural anisotropy and mechanical heterogeneityanisotropy and mechanical heterogeneity should be should be consider to improve the volumetric strain estimation. consider to improve the volumetric strain estimation.

DP RC - Geological Survey of Japan / AISTDP RC Disaster Prevention Research Center, National Cheng...

Documents

Transcript of DP RC - Geological Survey of Japan / AISTDP RC Disaster Prevention Research Center, National Cheng...