Characteristics of Earthquake Input

• Earthquake occurrence and return period

• Attenuation relationships

• Earthquake response spectra

• Design spectra: Force reduction factors, earthquake records, …etc

• Other Intensity Measures: Duration and number of cycles of

earthquakes ground motions, ..etc
 Earthquake Occurrence and Return Period
§ Occurrence of earthquakes are usually modelled in probabilistic seismic hazard
assessment as a Poisson process.

§ Each earthquake is assumed to occur independently of history or any other seismic

event. This is time-independent hazard. This model cannot include fore- and after-
shocks.

(N ×t r ) n e - N × t where P is the probability of having n

P [m > M, t r ] =
r
earthquakes with magnitudes m greater
n! than M over a reference time period tr in
a given area. The parameter N is the
expected number of occurrences (mean
rate) per unit time for that area.

ln N = a - b × M Frequency-magnitude relationship, in which a

and b are model constants.

P [m > M, t r ] = 1 - e - e (a - b × M) × t r
By combining these two above

1 tr
TR = = - The return period TR
N ln (1 - P )
 Problem

A long-span suspension bridge is going to be built in an active seismic region.

The structural earthquake engineer can choose the design ground-motion
parameter with respect to three return periods: 475, 950 and 2,500 years.
Which is the most suitable return period to select and why? What is the
associated probability of the peak ground acceleration for the return period
being exceeded?

1 tr
TR = =-
N ln (1 - P )
 Attenuation Relationships (Strong-motion Models)
§ Attenuation relationships, recently referred to as ground motion or strong-motion
models, are analytical expressions specific to a site describing ground motion
variation with at least magnitude and source distance.

log(Y ) = log (b1 ) + log[ f1 (M )] + log[ f 2 (R )] + log[ f 3 (M, R )] + log[ f 4 (E i )] + log (ε )

where Y is the ground motion parameter to be computed and b1 is a scaling factor.
The second-to-fourth terms on the right hand-side are function fi of the magnitude
M, source-to-site distance R and possible source, site and geologic structure
effects Ei. Uncertainty and errors are quantified through the parameter ε.
1.60
Ms=5.5_Rock Ms=5.5_Stiff Soil
1.40
Peak Horizontal Acceleration (g)

Ms=6.5_Rock Ms=6.5_Stiff Soil

1.20 Ms=7.5_Rock Ms=7.5_Stiff Soil

1.00
Focal depth = 3.5 km
0.80

0.60

0.40

0.20

0.00
0.1 1 10 100
Distance (km)

Attenuation of peak ground horizontal acceleration: effect of magnitude (left) and focal depth (right)
 Attenuation Relationships

► Attenuation relationships for Europe

log (PGA )h = 2.522 - 0.142 × M w + (0.314 × M w - 3.184) × log 57.76 + d 2 + 0.137 × SS + 0.050 × SA - 0.084 × FN + 0.062 × FT - 0.044 × FO
§ in which the epicentral distance d is in kms.
(PGA in m/sec2)
§ the coefficients SA and Ss are given as a function of the soil type.
Fault mechanism FN FT FO
Soil type Shear wave velocity, vS (in m/s) SA SS
Normal 1 0 0
Rock vs > 750 0 0
Odd 0 1 0
Stiff 360 < vs ≤ 750 1 0
Soft 180 < vs ≤ 360 0 1 Strike-slip 0 0 0
Thrust 0 0 1
► World-wide attenuation relationship
log (PGA ) = -0.659 + 0.202 × M S - 0.0238 × d + 0.020 × S A + 0.029 × SS
(PGA in m/sec2)
in which the epicentral distance d is in kms.

Coefficients SA and SS in the above equations account for the effects of soil condition.
Soil type Shear wave velocity, vS (in m/s) SA SS
Rock vs > 750 0 0
Stiff 360 < vs ≤ 750 1 0
Soft 180 < vs ≤ 360 0 1
Very Soft vs ≤ 180 0 1
 Response Spectra (See Lecture Notes)
§ Besides ground motion parameters (peak ground acceleration PGA, peak ground velocity PGV,
peak ground displacement PGD), response spectra is commonly used to define ground motion
intensity.

§ A response spectrum is a plot of the maxima of the acceleration, velocity and displacement
response for single degree of freedom (SDOF) systems with various natural periods when
subjected to an earthquake ground motion.

§ The equation of dynamic equilibrium for linear elastic systems with mass m, stiffness k and
damping c is as follows:
m !u! + c u! + k u = -m !u! g

§ The maximum value of the displacement Sd, defined as ‘spectral displacement’, is equal to:
ét ù
êò u g × e
- ξ ω (t - τ)
!! sin[ω d (t - τ)] × dτ ú (1.)
ë0 û max
S d = [u(t)]max =
in which t is a time variable, ω and ω are the natural w frequencies of the undamped and damped
d
system, respectively. ξ is the viscous damping ratio.

§ The maximum velocity Sv (or pseudo-velocity) can be approximated by :

S v = ω × Sd (2.)
§ The ‘spectral pseudo-acceleration’ Sa is as follows:

Sa = ω × S v = ω 2 × Sd (3.)
 Elastic and Inelastic Spectra (continued)
i. Select the earthquake record from data-banks;
ii. Select a T-ξ pair, i.e. the fundamental period of vibration and the damping ratio for
the SDOF;
iii. Select a numerical method (e.g., Newmark iteration approach) to integrate the
equation of motion;
iv. Compute the response time history for the given earthquake record and dynamic
properties of the SDOF; ét
ò
-ξ ω
v. Compute the spectral displacement Sd, pseudo-velocity Sv and pseudo- ê !u! g × e
ë0
acceleration Sa; S = [u(t)] = d max
vi. Select a new T-ξ pair and repeat steps (i) to (v);
S v = ω × Sd
vii. Plot the maxima response versus the fundamental period.
1.00
Sa = ω × S v = ω 2 × Sd
0.75

0.50 amax = Sa(T1=0.5sec) = 0.83g 1.00

2.0
Acceleration (g)

Critical damping x=5% 0.25 0.75

1.8
Spectral Acceleration (g)

0.00 0.50
1.5

Acceleration (g)
0.0 5.0 10.0 15.0 20.0 0.25 amax = Sa(T2=1.5sec) = 0.44g
-0.25
1.3
amax = Sa(T1=0.5sec) = 0.83g 0.00
-0.50
1.0 M 0.0 5.0 10.0 15.0 20.0
-0.25
amax = Sa(T2=1.5sec) = 0.44g
-0.75 k1 T 1= 0.5 sec
0.8 ` = 5% -0.50 M
-1.00 k2
0.5 -0.75 T2 = 1.5 sec
PGA
Time (seconds)
° = 5%
0.3 -1.00
Time (seconds)
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Natural Period of Vibration (seconds)
 Factors Influencing Response Spectra
§ Magnitude, distance and site conditions are the most influential factors influencing
the shape of earthquake spectra. Others include Rupture directivity, Wave travel
path, Source mechanism and characteristics, far- and near-field …etc
3.00 3.00
M=5.5 M=6.0 M=7.0 d = 5 km d = 20 km d = 50 km

2.50 2.50
Ambraseys et al . (1996)
Ambraseys et al . (1996)
2.00 2.00

Elastic
Sa / PGA

Sa / PGA
1.50 1.50

1.00
Spectra
1.00

SCALED SCALED
0.50 0.50

0.00 0.00
0.00 0.50 1.00 1.50 2.00 0.00 0.50 1.00 1.50 2.00
Period (seconds) Period (seconds)

0.40
Rock Stiff Soft
• The effect of distance here can be compensated for by
0.35
scaling. However, this might not be the case of inelastic and
degrading systems, where the duration will have an effect.
Absolute Acceleration (g)

0.30

0.25
• The acceleration amplifications for soft soils extend over a
0.20 larger period range than the amplifications for the other two
0.15 soil categories.
0.10
• The longer the predominant period of vibration of the site the
0.05 longer is the period at which the response spectrum peak
0.00 occurs.
0.00 0.50 1.00 1.50 2.00
Period (seconds)
 Features of Response Spectra
2.0 • The value of Sa is equal to the peak ground
1% 5% 10% 15% 20%
acceleration PGA at T=0 seconds and for long
1.5
Elastic Response periods the response tends to zero asymptotically.
Spectral Acceleration (g)

1.0
Spectra for 1940
Imperial Valley EQ • By increasing the damping smoother curves are
0.5 (FAR FAULT) derived.

0.0
0.0 1.0 2.0 3.0 4.0
• Spectral ordinates for all damping levels increase
Period (s) with the period from zero to some maximum value
150.0
1% 5% 10% 15% 20%
and then descend to converge at the values of
PGD and PGV, respectively, at long periods.
Spectral Velocity (cm/s)

100.0

• Differences in shape between far- and near-field

50.0
response spectra are related to the frequency
content of the input motion. The former are
0.0
0.0 1.0 2.0 3.0 4.0
generally broad-band signals while the latter are
Period (s)
narrow-band, pulse-like records.
50.0 2.0
1% 5% 10% 15% 20% 1% 5% 10% 15% 20%
40.0
Spectral Displacement (cm)

1.5
Elastic Response
Spectral Acceleration (g)

30.0
1.0
Spectra for 1994
20.0 Northridge EQ
10.0
0.5 (NEAR FAULT)
0.0 0.0
0.0 1.0 2.0 3.0 4.0 0.0 1.0 2.0 3.0 4.0
Period (s)
Period (s)
 Elastic Spectra versus Inelastic Spectra
Lower accelerations and hence force levels are generated in the inelastic systems. The latter
is due to the energy absorption by hysteresis.

The scaling factor between elastic and inelastic spectra changes with period, ductility factor
and earthquake.
3,0 3,0

Elastic Ductility 2 Ductility 4 Elastic Ductility 2 Ductility 4

2,5 2,5
Spectral Acceleration (g)

Spectral Acceleration (g)

2,0 2,0

1,5 1,5

1,0 1,0

0,5 0,5

0,0 0,0
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0
Period (s) Period (s)

Elastic and constant ductility spectra for the 1994 Northridge (left) and 1995 Kobe (right) earthquakes

Question: Which curve do you prefer for your building design ?

 Design Spectra in Earthquake Codes
§ Design spectra include features that are decided upon by code committees or other
interested parties.

§ For design applications, response spectra obtained from records with similar
characteristics are averaged and smoothed. Smoothing is necessary because of the
difficulties encountered in determining the exact periods of the structure.

§ For a given period, the elastic spectral ordinate should be divided by the inelastic
counterpart for a value of ductility expected for the structural system under
consideration.
S a elastic (T) the ratio of the elastic strength demand to the inelastic
Force Reduction Factor = strength demand for a specified constant ductility μ.
S a inelastic (T)
Newmark and Hall (1982): The force reduction factor Rμ is defined as the ratio of the maximum elastic force
to the yield force required for limiting the maximum inelastic response to a displacement ductility μ.

R μ =1 when T < 0.05 s

a linear interpolation is suggested
R μ = 2μ - 1 when 0.12 s < T < 0.5 s
for intermediate periods.
Rμ = μ when T > 1.0 s

R=fcn (T, μ, soil conditions, strain hardening, hysteretic parameters, damping types,…)
 Comparison between Different Force Reduction Factors in the Literature
4.0
• As ductility increases, R increases.
3.0
• As T increases, R increases.
Force Reduction Factor

2.0 • Variations are significant mainly for low

Newmark & Hall
Krawinkler & Nassar period structural systems.
Miranda & Bertero
1.0 Vidic et al. • For long-period systems, all the
Ductility µ=2
Borzi & Elnashai (accurate) formulations show an asymptote value
0.0
Borzi & Elnashai (approximate)
of the target ductility.
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Period (seconds)

8.0

6.0
Force Reduction Factor

Newmark & Hall

4.0 Krawinkler & Nassar

Miranda & Bertero
Vidic et al.
2.0 Ductility µ=6 Borzi & Elnashai (accurate)
Borzi & Elnashai (approximate)

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Period (seconds)
 Design Spectra in Earthquake Codes (continued)
§ The general procedure for generating statistically derived design spectra is as
follows:
i. Select a set of ground motions on the basis of their magnitude, distance and site conditions;
ii. Generate response spectra in terms of acceleration, velocity and displacement;
iii. Average the response spectra derived in Step (ii);
iv. Evaluate the design response spectrum with desired probability of exceedence on the basis of the
relationships derived in the previous steps.

§ Standard design response spectra

• Elastic design spectra are usually available and ready to use in the codes (results of the
statistical procedure above). For example, AFAD website provides Turkish elastic design
spectra plots for 4 different exceedance levels (scenarios) for a specific location.

• Inelastic design spectra can be obtained either directly or by scaling elastic spectra
through force reduction factors R presented earlier. These R values are usually available in
the code.

• The basic curves employ 5% damping; however, simplified expressions exist to obtain
spectra for different damping values.

• The standard design response spectra are based on fixed spectral shapes which vary as a
function of the soil site conditions, e.g. rock, stiff and soft.

§ Site-specific design spectra: These may be generated by employing response spectral

attenuation relationships or by advanced numerical modelling of the energy release and travel
path.
 Earthquake Records http://peer.berkeley.edu/smcat/;
§ Instead of design spectra, codes can also recommend http://db.cosmos-eq.org;
use of a set of site specific ground motion recordings
for special cases.
§ Three approaches in obtaining of EQ records:

1. From strong ground motion databanks, natural recordings

can be obtained for the region of interest. In the absence
of a strong-motion databank for a specific region, other
strong-motion data with similar regional characteristics
(such magnitude-distance-soil conditions, PGA/PGV
values, intra versus inter-plate mechnaism…) can be
checked for the selection of natural records.
• High PGA/PGV (larger than 1.2) : near-source shallow
earthquakes or records measured on rock
0.30
• Low PGA/PGV (less tan 0.8) : Deep or distant
earthquakes, or records measured on soft ground 0.20
2. Another approach is to generate artificial recordings by
random vibr. theory that fit, with a certain degree of

Acceleration (g)
0.10

approximation, a target spectrum.

0.00
0.00 5.00 10.00 15.00
3. Use of mathematical source models to generate strong- -0.10
motion-like time series is increasing in popularity since
-0.20
the ensuing records resemble natural records more than
signals generated to fit a target spectrum. -0.30
Time (s)
§ It is sometimes necessary to scale earthquake
artificial records generally exhibit a larger number of
records in a manner that takes into account the cycles than natural records and hence they may lead to
periods of vibrations of the structure under unrealistic (excessive) seismic demand for inelastic
consideration. structural systems
2023 Kahramanmaras-Pazarcık Earthquake (Mw 7.8)
6 3,5
EQ1-EW EQ1-Vertical
5 EQ1-NS 3 TEC Design Vertical DD1
TEC Design Horizontal DD1 TEC Design Vertical DD2
2,5
4 TEC Design Horizontal DD2

2
Sa (g)

Sa (g)
3
1,5
2
1
1
0,5
0
0 1 2 3 0
T(s) 0 1 T(s) 2 3

(a)vertical
Comparison of horizontal and Stationspectral
No:3126acceleration
Horizontal and vertical
values components
from the recording of the first
earthquake at Station 3126 in Antakya City Center with the design values from TEC2019

2023 KAHRMAANMARAS EARTHQUAKES-1st PAZARCIK EQ

Rep. Station Dist.to the Epicenter PGA-NS PGA-EW
PGA-Vert. (g)
Number (km) (g) (g)
3126 143.5 1.234 1.050 1.092
 Vertical Component of Ground Motion
§ The vertical component of ground motion can be linearly amplified from bedrock to
the surface to very high levels, leading to the widely observed high v/h ratios
greater than 1.0 near the EQ source, sometimes reaching the value of 2.0.
§ Sample of earthquakes with high vertical component
Earthquake Country Date Peak Horizontal Peak Vertical Vertical-to-Horizontal
Acceleration (g) Acceleration (g) Ratio
Gazli ex-USSR 17.5.1976 0.622 1.353 2.17
Coyote Lake USA 6.8.1979 0.256 0.420 1.64
Loma Prieta USA 17.10.1989 0.424 0.514 1.21

§ Northridge records possessing a strong vertical component

Station d (km) HPGA (g) VPGA (g) VPGA / HPGA
Tarzana, Cedar Hill Nursery 5 1.82 1.18 0.65
Arleta, Nordhoff Avenue Fire Station 10 0.35 0.59 1.69
Sylmar, County Hospital 16 0.91 0.60 0.66
Newhall, LA County Fire Station 20 0.63 0.62 0.98

§ Kobe records possessing a strong vertical component

• During the 1995 Hyogo-ken Nanbu (Kobe, Japan) earthquake, liquefaction at the vertical
array at Port Island caused an abrupt reduction in the horizontal ground shaking, but the
vertical motion continued to be amplified through the liquefied layer.
Station d (km) HPGA (g) VPGA (g) VPGA / HPGA
JMA Station 18 0.84 0.34 0.41
Port Island Array 20 0.35 0.57 1.63
Kobe University 25 0.31 0.43 1.39
Key: HPGA = Horizontal Peak Ground Acceleration; VPGA = Vertical Peak Ground Acceleration
 Vertical Component of Ground Motion (continued)
Attenuation of V / H ratio for 84.1% confidence limit
All 104 records Thrust 53 records Strike-Slip 43 records Europe 23 records
s =0.48 s =0.46 s =0.53 s =0.36

MS d=0.0 d=15.0 d=0.0 d=15.0 d=0.0 d=15.0 d=0.0 d=15.0

6.0 1.28 1.15 1.43 0.96 1.32 1.32 0.77 0.75
6.5 1.37 1.24 1.54 1.07 1.40 1.40 0.98 0.97
7.0 1.46 1.32 1.64 1.17 1.48 1.48 1.20 1.18
7.5 1.54 1.41 1.75 1.28 1.56 1.56 1.41 1.39

Key : d = Source distance (in kms); σ = Standard deviation from regression analysis

§ The computed values clearly show that the assumption that the vertical peak is
2/3 of the horizontal component suggested by can be a serious underestimate.

2023 KAHRMAANMARAS EARTHQUAKES-1st PAZARCIK EQ

Rep. Station Dist.to the Epicenter PGA-NS PGA-EW
PGA-Vert. (g)
Number (km) (g) (g)
3126 143.5 1.234 1.050 1.092
 Intensity Measures for Earthquake Inputs

A large number of definitions of strong Arias intensity

ground motion durations have been
proposed, such as: p tr
AI =
2g ò0
a 2 ( t ) dt

• Bracketed duration;
• Uniform duration; Matsumura spectrum intensities
• Significant duration.
2E i
Ve =
Kawashima and Aizawa Intensity m
Housner spectrum intensity
1
h(T, x, n ) =
1 + e(T, x ) × (n - 1) 2.5

n: number of
SI H = ò
0.1
S v (T, ξ) × dT
80 x response cycles
e(T, x ) = 0.0815 × T 0.349 Intensity scales of Nau and Hall
60 x + 1
0.185
SI a = ò S v (T, ξ) × dT for 0.118 s < T £ 0.5 s
0.028
The best reliable IM is the one that is well
correlated to accumulating damage under 2.000
SI v = ò S v (T, ξ) × dT for 0.5 s < T £ 5.0 s
cyclic seismic loadings ! 0.285

12.500
This IM can be then used to choose a suite of SI d = ò S v (T, ξ) × dT for 5.0 s < T £ 14.085 s
4.167
earthquake records for design applications
Inelastic Spectral Values