THE INSTITUTION OF ENGINEERS, MALAYSIA

2-DAY SYMPOSIUM AND WORKSHOP ON

EARTHQUAKE ENGINEERING IN MALAYSIA AND ASIA PACIFIC REGION
6 7 December 2011

Local Site Effects on

Earthquake Loading Model in

Regions of Low-to-Moderate-Seismicity
Hing-Ho Tsang & Nelson Lam
2-DAY SYMPOSIUM / WORKSHOP ON
EARTHQUAKE ENGINEERING IN MALAYSIA AND ASIA PACIFIC REGION
April 10 & 11, 2013

Content
n

1. Basic Phenomena of Local Site Effects

2. Effects on Building and Foundation Structures

3. Key Concepts for Site Effects Evaluation

4. Design Spectrum Models in Codes of Practice

5. Deficiencies with Existing Codified Models

Amplification Effects by Near-Surface Sediments

Seismic
wave
Bedrock

Poorly
Well
Bay mud
consolidated consolidated (water
sediments saturated)
sediments

Depending on the dynamic properties of soil sediments,

the ground motion can be significantly modified
(amplitude, frequency content, duration).

Map of the eastern San Francisco Bay Area showing surface geology,
and accelerograms revealing the relative amplitude of waves passing
through different materials. (Source: USGS)

Hazard at a Site
Large scale hazard maps refer to average / reference subsoil conditions.
Local subsoil conditions may lead to differences in the level of ground
shaking over small areas, which cannot be shown on large scale map.

Average change in Intensity for various subsoil conditions

Subsoil

Average change in Intensity

Rock

-1

Firm sediments

Loose sediments
(sand, alluvial deposit)

+1

Saturated sediments or artificial

filled ground (reclaimed land)

+2

Seismic Intensity Shaking Map for

San Francisco, CA (1989 Loma Prieta earthquake)

Soil Response at Treasure Island, 1989 Loma Prieta Earthquake

The case of distant, long-duration earthquakes

The classical case: Mexico 1985

In former times, Mexico was considered unique, but this type of

earthquake may affect many cities all over the world ...

Ahmedabad, India
(2001)

The Gujarat Earthquake

in India, Jan 26, 2001:
Another example of
long-distance long-period
earthquake effects

Long-distance Long-period Effects

in 2004 Aceh Earthquake
(Lam, Balendra, Wilson, Venkatesan, 2009)
10

Response Spectral Velocity (mm/s)

X10

Soft Soil Site

in Singapore

Borelog at the recording station

3.5 m

Silty Sand

0.1
18 m

Very soft clay

4m

0.01

Rock spectrum (recorded)

15 m

Silty Sand
Very stiff silty sand

Soil spectrum (recorded)

Granite

0.001
0.01

0.10

1.00

10.00

100.00

Natural Period (seconds)

(a) Recorded from Aceh earthquake

l Velocity (mm/s)

100

10

X10
Borelog at the recording station

3.5 m

18

Silty Sand

Very soft clay

Content
n

1. Basic Phenomena of Local Site Effects

2. Effects on Building and Foundation Structures

3. Key Concepts for Site Effects Evaluation

4. Design Spectrum Models in Codes of Practice

5. Deficiencies with Existing Codified Models

Firm Soil
Strata
Hard Soil or
Bedrock

Shallow
Foundation

Soft Soil
Strata
Bedrock

End-Bearing
Pile Foundation

Dynamic Soil-Foundation-Structure Interaction Analysis

2-D Finite Element Modelling
(Tsang et al., 2012)

Bedrock motions
are required.
t2

Rubber-soil Mixture

t1

Original Soil

t1
t2

both horizontal
Tranmitting
Base

&
vertical shakings

Dynamic Soil-Foundation-Structure Interaction Analysis

3-D Finite Element Modelling
(Xu, Tsang, Lo, 2013)

Bedrock motions
are required.

Dynamic SFSI Analysis

Analytical Modelling
-

Lumped-mass model

-
-

Inertial effects
Kinematic effects

Dynamic SFSI Analysis

Analytical Modelling
-
-

Inertial effects
On both superstructure
and foundation

Surface motions
are required
Site response
analysis
Site-specific design
spectrum model

-
-

Dynamic SFSI Analysis

Analytical Modelling
-
-

Kinematic effects
On foundation

Loading induced by
soil movement

Mass of superstructure
is ignored

Kinematic Effects
-

Piles deflect less

than the soil mass

Free-field
soil deformation
profile is required
Site response
analysis has to
be conducted

Castelli & Maugeri (2009)

Kinematic Effects
-
-

Castelli & Maugeri (2009)

Pseudo-static approach
Soil deformation modelled by p-y spring

Dynamic SoilFoundation-Structure
Interaction Analysis
Illustration of
Combined Effects on
Bending Moment at
Pile Head

Castelli & Maugeri (2009)

10

Content
n

1. Basic Phenomena of Local Site Effects

2. Effects on Building and Foundation Structures

3. Key Concepts for Site Effects Evaluation

4. Design Spectrum Models in Codes of Practice

5. Deficiencies with Existing Codified Models

Local Site (Soil) Effects

A
HS

SOIL

soil
Vsoil

BEDROCK

Seismic
Impedance

rock
Vrock

=.V

ROCK
OUTCROP
(Reference Site)

Seismic Waves
Uniform soil layer on elastic rock subjected to
vertically propagating seismic waves

11

July 22, 2004 19:46 WSPC/124-JEE

00157

665

1.2
1 .0
0.8

Peninsular Malaysia
Local Scenario: R = 30 km

Elgamal & He (2004)

R = 5 km
R = 10 km

2 /3

R = 40 km

0.4

0.6

R = 20 km

Insignificant for most structures

0 .0

0.2

V ertical / H o rizo n tal S p ectral R atio

1.4

1.6

Vertical Earthquake Ground Motion Records

Vertical Ground
Motions

0 .0 1

0 .1 0

1 .0 0

1 0 .0 0

P erio d T (s)

Fig. 2. V/H spectral ratio for the 1994 Northridge earthquake motions at different epicentre
Horizontal shaking is the primary cause of damage,
distances of R [Bozorgnia et al., 1995].

as structures are more vulnerable to horizontal motion.

Thus the factor of 2/3 underestimates the effects of vertical motion at short
periods and overestimates the effects at long periods. Accordingly, Papazoglou and
Elnashai [1996] recommended that vertical motion be considered separately in designing structures. In order to define such design spectra, Elnashai and Papazoglou
[1997] studied the near-field vertical response spectra at 0%, 2% and 5% damping
using 35 earthquake records. These records were selected from 15 earthquakes based
on peak vertical acceleration (PVA) 0.3 g, focal depth h 25 km, and surface
wave magnitude Ms 5.0.
In view of the findings reported by these pioneering studies, this paper attempts
to shed light on some related aspects. In this regard, response spectra are estimated
based on a larger data set including near-field and far-field ground motions from
California earthquakes (Table 1). As can be seen from Table 1, 111 strong free
field motion records (PVA 0.1 g) were employed from six different California
earthquakes. Some of this data was used to examine the trend of PVA with
distance. Response spectra and their upper bound were evaluated at different levels
of damping based on the employed data set.
Refraction of seismic waves
In addition, available downhole array vertical
records were
studied
to investigate
through
a series
of successively
the variations with depth of vertical ground motion characteristics. Currently, this
softer (lower shear wave
data remains scarce only allowing preliminary observations. As lateral site response
velocity,
VS) surface
layers.
analysis is usually based on a 1D wave propagation model, an attempt
was made
to
employ a similar technique for vertical response, calibrated by the available vertical
Note that orientation of ray
array motions. Limitations of this 1D approach will be briefly outlined, denoting
path becomes closer to vertical.
the need for further research [Beresnev et al., 2002].

Local Site Response

Site Response Analysis can be simplified by numerically modelling

the 1-D vertical propagation of seismic waves through a simplified,
horizontally-layered model of the weathered rock and soil at the site.

12

Earthquake Ground Motion Modelling

Stochastic Simulation of Seismological Model

A ( f ) = C M0 S ( f ) G An( f ) P( f ) V( f ) F ( f )
Source
Source
Properties

Local

Path

Crustal Quality
Q0 factor

Site

Crustal SWV & Soil SWV &

Attenuation
Plasticity
Site
Local

Path
Source

Site-Specific Site Response Analysis using

Computer Programs, e.g. SHAKE and SIREN
0.02
0.01
0
-0.01 0

Time (s)
10

20

30

40

50

60

70

80

-0.02
-0.03

Acceleration (m/s/s)

Acceleration (m/s/s)

0.03

0.06
0.04
0.02

Time (s)

0
-0.02 0
-0.04
-0.06

Bedrock Input Motions

from GENQKE
4

10

12

14

16

18

13

20

Subsurface profile at Richmond field station, California.

Soil Resonance Effects

Site Natural Period : TS = 4 HS / VS

t1

Energy loss
(Soil damping)

t2 t3

t4

Soil

HS

VS
Seismic
Impedance
= V .

Energy loss
(Radiation damping)
Incident
Wave

Rock (VR)

14

Nonlinear Site Effects

5:1

3:1

2:1

Approximate relationship between peak ground acceleration

(PGA) on rock and soft soil sites (after Idriss, 1990).

Dynamic Properties
of Soil Sediments
Energy Dissipated

Max. Strain Energy

Equivalent (or Secant)

Shear Modulus &
Damping Ratio

15

Higher plasticity (cohesive) soils

tend to behave linearly to higher
shear strain, compared with soils
with lower plasticity index (PI).

G/Gmax

Effects of
Plasticity of Soil

0.5

Increasing
Plasticity
0
0.001

Damping Ratio (%)

0.01

0.1

Shear strain (%)

25

Increasing
Plasticity

20

Vucetic, M. and Dobry, R.,

Effect of Soil Plasticity on
Cyclic Response,
J. Geotech. Eng., ASCE,
Vol. 117, No. 1, p. 89-107, 1991.

15
10
5
0
0.001

0.01

0.1

Shear strain (%)

Content
n

1. Basic Phenomena of Local Site Effects

2. Effects on Building and Foundation Structures

3. Key Concepts for Site Effects Evaluation

4. Design Spectrum Models in Codes of Practice

5. Deficiencies with Existing Codified Models

16

Design Spectrum (DS) Model in

Eurocode 8 (2004) & Australian Standard (2007)
RSA

TC = T1 = 0.3 s (AS)

RSAmax

= 0.25 s (EC Type 2)

= 0.4 s (EC Type 1)

for rock sites

TD = T2 = 1.5 s (AS)

1
T

PGA

= 1.2 s (EC Type 2)

= 2.0 s (EC Type 1)

1
T2

EC8 (2004)
Type 1
RSA
RSAmax =
2.5*S*PGA

S*PGA

1
T2

1
T

T1

T2 = 2.0 s

17

IBC (2012) Design Spectrum (since 1997 UBC)

Fa and Fv are site dependent coefficients
Ss and S1 are mapped spectral accelerations

RSA

at short (0.2 s) and long (1.0 s) periods

FaSs

Fa S s
2.5

0.2 T1

ASCE/SEI Standard 7-10:

Fv S1
T

T1 =

Fv S1T2
T2

Long-period
transition
period 4.0 s

Fv S1
Fa S s
T2

Site Classification in IBC (2012) (since 1997 UBC)

SITE
CLASS

SOIL PROFILE
NAME

Vs , 30 (m/s) *

SPT
N-Value *

Undrained Shear
Strength * (kPa)

HARD ROCK

> 1500

Not
Applicable

Not
Applicable

ROCK

760 1500

Not
Applicable

Not
Applicable

VERY DENSE
SOIL AND SOFT
ROCK

360 760

> 50

> 100

STIFF SOIL

180 360

15 50

50 - 100

SOFT SOIL

< 180

< 15

< 50

SPECIAL SOILS REQUIRING SITE-SPECIFIC EVALUATION

* in top 30 m of site profile

18

Site Classification in IBC, NBCC and EC

The following method is used to determine the average shear wave
velocity (SWV) of the top 30m Vs , 30 of the site profile:
n

Vs ,30 =

Notes: The same equation applies to

the computation of the values of SPTN and undrained shear strength.

i =1
n

di

i =1 Vs ,i

WHERE:
di = thickness of Layer i between 0 and 30 m.
V s ,i = shear wave velocity in Layer i in m/s.

Site Coefficients in IBC

(a) Short-period site coefficient Fa
Mapped Spectral Acceleration at Short Period

Site Class /
Soil Profile Type

Ss 25% g

Ss = 50% g

Ss = 75% g

Ss = 100% g

Ss 125% g

0.8

0.8

0.8

0.8

0.8

1.0

1.0

1.0

1.0

1.0

1.2

1.2

1.1

1.0

1.0

1.6

1.4

1.2

1.1

1.0

2.5

1.7

1.2

0.9

0.9

(b) Long-period site coefficient Fv

Mapped Spectral Acceleration at 1.0 s

Site Class /
Soil Profile Type

S1 10% g

S1 = 20% g

S1 = 30% g

S1 = 40% g

S1 50% g

0.8

0.8

0.8

0.8

0.8

1.0

1.0

1.0

1.0

1.0

1.7

1.6

1.5

1.4

1.3

2.4

2.0

1.8

1.6

1.5

3.5

3.2

2.8

2.4

2.4

(*) Site-specific geotechnical investigation and dynamic site response analyses shall be performed.

19

IBC-2012 (comparing site classes)

Acceleration Response Spectra
Hard (rock)

0.213 sec. 1.065 sec.

(0.9)
(0.825)
(0.75)
(0.6)

0.107 sec.

PGA

Site Class A

Site Class A - E
PGA = 0.3 g
Class B
Elastic ; = 5%

0.533 sec.

Site Class E
Soft (soil)
E
A

D
BC

AS (2007) and NZS (2004)

n

Sites are classified into FIVE site classes.

For rock sites, average SWV over the top 30 m is used.

For soil sites, site natural period TS , depths of soils HS ,

undrained shear strength and SPTN values are used.

The basic parameter for site classification in the standard is

site natural period.

The approach adopted in IBC of placing deep stiff soil sites in

Site Class D stiff soil, with relatively short-period spectral
characteristics may be non-conservative in the medium to long
period spectral range.

20

Comparison of DS for IBC Class E Soil Site

2

Comparison
DS Shape
T1 ~ 120%ofdifference
IBC

Normalised RSA

1.5

S ~ 2.45 5.84
NZS

S =1

GB

IBC / NBCC

EC

AS

NZS

UHS for HK

TCL06

(Rock Site)

S ~ 3.37 8.30

GB

0.5
HK-UHS
Rock

0
0

2
T (s)

Content
n

1. Basic Phenomena of Local Site Effects

2. Effects on Building and Foundation Structures

3. Key Concepts for Site Effects Evaluation

4. Design Spectrum Models in Codes of Practice

5. Deficiencies with Existing Codified Models

21

M6.3 Christchurch EQ

NZS 1170.5:2004

February 21, 2011

10 km SE of Christchurch
Depth: 5 km

Design Level (RP=500yrs)

PGA = 0.22 g
Collapse Prevention
(RP=2500yrs)
PGA = 0.4 g

No. of Deaths = 185

Collapse of Modern RC Building

Hazard Map

Recorded Ground Motions

NZS 1170.5:2004
PGA for Collapse Prevention (RP=2500yrs) = 40% g
The Largest Short-Period Site Factor = 1.28
The Largest RSAmax for Collapse Prevention = 120% g

Station

Distance (km)

PGA (%g)

RSA (0.3s) (%g)

PRPC

8.3

66

184

CCCC

8.4

48

185

CHHC

9.6

36

263

SHLC

11

34

251

RHSC

13.7
24.9

29

230
159

KPOC

21

22

Recorded Response Spectra

Implied Site Factor for Soft Soil Sites
in NZ Standard at T = 1.01.5 s ~ 2.0
Recorded Site Factor
at T = 1.01.5 s ~ 4.0
Exceeded by
~ 100% !!

Site Coefficients in IBC

(a) Short-period site coefficient Fa
Mapped Spectral Acceleration at Short Period

Site Class /
Soil Profile Type

Ss 25% g

Ss = 50% g

Ss = 75% g

Ss = 100% g

Ss 125% g

0.8

0.8

0.8

0.8

0.8

1.0

1.0

1.0

1.0

1.0

1.2

1.2

1.1

1.0

1.0

1.6

1.4

1.2

1.1

1.0

2.5

1.7

1.2

0.9

0.9

(b) Long-period site coefficient Fv

Mapped Spectral Acceleration at 1.0 s

Site Class /
Soil Profile Type

S1 10% g

S1 = 20% g

S1 = 30% g

S1 = 40% g

S1 50% g

0.8

0.8

0.8

0.8

0.8

1.0

1.0

1.0

1.0

1.0

1.7

1.6

1.5

1.4

1.3

2.4

2.0

1.8

1.6

1.5

3.5

3.2

2.8

2.4

2.4

(*) Site-specific geotechnical investigation and dynamic site response analyses shall be performed.

23

International Building Code (IBC)

Spectral Amplification Factor
at periods of 0.2 and 1.0 s
n

Based on empirical data

(at rock PGA = 0.1 g) and
supplemented by computer
simulations (using SHAKE)
rock PGA up to 0.5 g

Borcherdt (1994)
1989 Loma Prieta EQ

Take into account shaking level,

shear stiffness (VS) of soil

For example:

Vs , 30 = 250 m/s (Class D)

Fv = 2.4
Data scatter from 1.0 to 6.0

Borcherdt (2002)
1994 Northridge EQ

H.B. Seed et al. (1976, BSSA)

Site-Dependent Spectra for Earthquake-Resistant Design
Recorded Normalised Spectral Acceleration for Soft to Medium Stiff Clay
Maximum
Normalised RSAmax
at Individual Site ~ 4

Average
Normalised
RSAmax ~ 2

Prof. H.B. Seed: indicating the need for caution

in the use of statistical mean relationships

24

Why Soil Thickness HS or Site Period TS is not

parameterised in typical Codes of Practice?

Between 1976 and 1994, both stiffness (G or Vs) and soil depth Hs
were taken into account (based on Seed et al.).

Dobry et al. (2000):

Previous code, relying on qualitative descriptions of the
soil and thus being more ambiguous, did require total soil
thickness down to greater depths (200 ft), certainly play a
role in local site response
V s , 30 is clearly measurable in the field, thus removing the
ambiguity makes it much more feasible to obtain necessary
information for the site
Nowadays, soil properties down to bedrock are readily
available from site investigation.

The spikes in a spectrum due to soil resonance can be suppressed

by energy dissipation (high damping) in ductile structures (as in
regions of high seismicity, e.g. California & Japan).
Acceleration Response Spectrum Sa
1995 M6.8 Kobe, Japan Earthquake

Displacement Response Spectrum Sd

1995 M6.8 Kobe, Japan Earthquake

25

Conclusions
n

Distant Earthquake Motions in Deep Soft Soil Sites

Inertial and Kinematic Effects in SFS Interaction

Both Soil Stiffness and Thickness are Essential

Large Inter-Model Discrepancies amongst Codes

Soil Thickness not yet Parameterised in Codes

End of Presentation on
Local Site Effects on
Earthquake Loading Model in
Regions of Low-to-Moderate-Seismicity
THE INSTITUTION OF ENGINEERS, MALAYSIA

2-DAY SYMPOSIUM AND WORKSHOP ON

RTHQUAKE ENGINEERING IN MALAYSIA AND ASIA PACIFIC REGION
6 7 December 2011

26