LIST OF Figures
Figure 1 Earthquake
epicenters and location of accelerograph stations.
Figure 2 Comparison
of PGA values measured at different distances with known attenuation
relationships.
Figure 3 Estimation
of the directivity effect for the Nuweiba main shock 22.11.95 (from Gitterman
et al., 1996a).
Figure 4 Accelerograms
at the EIL station of the Nuweiba mainshock and two aftershocks to the same
scale (a) and to different scales (b). Similar waveforms were
observed for the mainshock and nearby aftershock 26.12.95, as against the
aftershock of 23.11.95 located on the eastern bank of the Gulf.
Figure 5 Pseudo-acceleration
response spectra of the mainshock EW component for different damping
values (a) and the average estimation over both horizontal components of
the mainshock and the aftershock-1 (b). Variation ±s is shown by dotted lines. Building code 413 curve for S2 soil
conditions is also presented.
Figure 6 Spectra
(a), spectral H/V ratios (b) and the average ratio (c) at the EIL station for
the mainshock 22.11.95 (window t = 20 sec is shown by arrows).
Figure 7 Spectra
(a) and spectral H/V ratios (b) at the EIL station for the aftershock 23.11.95
(t = 20 sec).
Figure 8 Spectra
(a) and spectral H/V ratios (b) at the EIL station for the aftershock 26.12.95
(t = 20 sec).
Figure 9 Spectral
H/V ratios at the EIL station for the Nuweiba mainshock and two aftershocks (a)
and an average ratio over the mainshock and the aftershock2 (b).
Figure 10 Computer
simulation (by SHAKE, Su=10) of ground motions at the outcropped granite
basement underlying soft sediment layers (b), when EW recording of the
mainshock (on the surface) was used as input (a).
Figure 11 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the SVT (Shivta) station for
the mainshock 22.11.95 (t=30 sec). If exchange NS and V components, the
ratios become more reasonable (d).
Figure 12 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the GMK station for the
mainshock 22.11.95 (t=40 sec). (The raw data record was presented by two
files which were concatenated).
Figure 13 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the GMK station for the
aftershock-1 23.11.95 (t=40 sec).
Figure 14 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the MIZ (Mizpe Shalem) station
for the mainshock 22.11.95 (t=13 sec).
Figure 15 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the ASQ (Ashquelon) station for
the mainshock 22.11.95 (t=20 sec).
Figure 16 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the ALM (Almog) station for the
mainshock 22.11.95 (t=30 sec).
Figure 17 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the HAD (Hadera) station for
the mainshock 22.11.95 (t=30 sec).
Figure 18 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the ALN (Alonim) station for
the mainshock 22.11.95 (t=30 sec).
Figure 19 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the HAC (Haifa) station for the
mainshock 22.11.95 (t=30 sec).
Figure 20 Accelerograms
at the GOS (Hagoshrim) station for the mainshock 22.11.95.
Figure 21 Site
response estimation at the Eilat station by microtremor measurements (from
Gitterman et al., 1996): a - map of the area, b - average
H/V spectral ratio.
Figure 22 Time
windows on accelerograms of the mainshock and aftershock-2 selected for
spectral analysis.
Figure 23 H/V
spectral ratios for different windows on accelerograms of the Nuweiba mainshock
(a) and aftershock-2 (b).
Figure 24 Seismograms
of two weak earthquakes recorded at the Eilat accelerograph station.
Figure 25 Spectra
(a) and spectral H/V ratios (b) for the weak earthquake2 recorded at the Eilat
accelerograph station.
Figure 26 Analytical estimation
of site response at the Eilat station by the Rayleigh spectral ratio
(ellipticity) based on the subsurface model of the site.
Figure 27 Shift
of the spectra dominant frequency for several successive recording windows,
demonstrating an anomaly change for the Nuweiba mainshock (a) and a normal
change for aftershock-2 (b) and weak earthquake2 (c).
Figure 28 Non-linear
transfer functions (b) calculated by the Joyner program using accelerograms of
the mainshock and the aftershock as input for the program (a).
Figure 29 Accelerograms
at stations TMR, RDG and LAV for the Cyprus earthquake 9.10.96 (spectral
analysis windows are shown by arrows).
Figure 30 Spectra
(a) and spectral H/V ratios (b) at the RDG station for the Cyprus earthquake.
Figure 31 Pseudo-acceleration
response spectra of the Cyprus earthquake, station RDG, EW component, for
different damping values. Building code 413 curve for S2 soil conditions is
also shown.
Figure 32 Spectra
(a) and spectral H/V ratios (b) at the LAV station for the Cyprus earthquake.
Figure 33 Spectra
(a) and spectral H/V ratios (b) at the TMR station for the Cyprus earthquake.
Figure 34. Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the GMK station for the Cyprus
earthquake.
Figure 35 Spectral
H/V ratios at the GMK station for the Nuweiba mainshock and aftershock-1, and
the Cyprus earthquake (a) and an average ratio over all three events and both
horizontal components.
Figure 36 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the IZR station for the
24.08.84 earthquake (Galilee) (t = 7 sec).
Figure 37 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the HAT station for the
24.08.84 earthquake (t = 4 sec).
Figure 38 Accelerograms
at the DA2 station for the 27.04.87 earthquake. Low signal amplitudes and poor
waveforms make unreliable any spectral ratio estimates.
Figure 39 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the TUG station for the
23.10.87 earthquake (t = 8 sec).
Figure 40 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the DA1 station for the
23.10.87 earthquake (t = 7 sec).
Figure 41 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the DA2 station for the
23.10.87 earthquake (t = 10 sec). It should be noted that weaker spectra
smoothing increases
amplification factors (d).
Figure 42 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the DA3 station for the
23.10.87 earthquake (t = 23 sec).
Figure 43 Accelerograms
at the BET station for 3.01.89 and 6.01.89 earthquakes.
Figure 44 Spectra
(a), spectral H/V ratios (b) and an average ratio (c) at the BET station for
3.01.89 and 6.01.89 earthquakes (t = 7 sec).
Figure 45 Accelerograms
at the NET and MIF stations (a), and spectral H/V ratios at the NET (t = 7 sec)
and MIF (t = 10 sec) stations (b) for the 28.09.91 earthquake.
Figure 46 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the ALM station for the 2.08.93
earthquake (t = 8 sec).
Figure 47 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the JER station for the 2.08.93
earthquake (t = 7 sec).
Figure 48 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the KIT station for the 26.03.97 earthquake (t = 10 sec).
Figure 49 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the GOS station for the 26.03.97 earthquake (t = 10 sec).
Figure 50 Accelerograms
(a), spectra (b) and spectral H/V ratios (c) at the ZEF station for the 26.03.97 (Beirut) earthquake (t = 8 sec).
Figure 51 Horizontal
pseudo-acceleration response spectra of the Beirut earthquake at station ZEF,
EW (a) and NS (b) component, for different damping values. The building code
413 curve for S2 soil conditions is also presented.
Figure 52 Accelerograms (a),
spectra (b) and spectral H/V ratios (c) at the KIT station for the 4.8.97 earthquake (t = 3 sec).
1. Introduction
The
study presented in this report was conducted after the occurrence of strong
earthquakes in the region: November 1995, The Gulf of Eilat/Aqaba (Mw=7.1)
and October 1996, Cyprus (Mw=6.8), which were felt throughout Israel and
triggered the Israel strong motion network. Most of the results were obtained
during 1996 and presented at the Annual Conference of the Israel Geological
Society (Gitterman et al., 1996a; Gitterman and Shapira, 1997a), the XXV
ESC Assembly in Iceland (Gitterman et al., 1996b) and the 29th
IASPEI Assembly in Greece (Gitterman and Shapira, 1997b).
The
main reason for initiating this study was the enhanced Peak Ground Acceleration
(PGA) values (as compared with previously known attenuation relationships)
observed. During the Gulf of Aqaba earthquake, almost all nine stations of the
Israel Accelerograph Network, from Eilat to Haifa, that were triggered by the
main shock show enhanced PGA values. We suspect that this phenomenon is
associated with the amplification caused by sediment layers under the stations.
In order to verify this hypothesis we attempted to evaluate local site effects
using the three-component strong motion records.
2. Method
of Seismic Response Estimation
In
order to estimate the spectral seismic response function for a specific site,
we calculated and analyzed spectral ratios of horizontal to vertical components
(H/V ratio) of the most intense part of the strong motion records (i.e. S-wave
and S-coda). In actual fact this is a modification or application to earthquake
data of the Nakamura method for microtremors (Nakamura, 1989) first applied by
Lermo and Chavez-Garcia (1993) to weak and strong motion records of
earthquakes.
The
empirical single station spectral ratio method was applied to accelerograms of
strong earthquakes under the same assumption, which is common in microtremor
investigations:
-
reliable estimation of the fundamental resonance
frequency fres;
-
approximate evaluation of amplification levels
The results obtained were
verified, wherever possible, by comparison with:
a)
The average site response H/V ratio estimations using weak
motion observations of microtremor and small events (Nakamura technique).
b)
An analytical transfer function computed by using:
*
a linear 1-D SH model
*
a non-linear model (Joyner, 1997)
*
Rayleigh wave spectral ratio (ellipticity) (Gitterman
et al., 1996c)
The calculation procedure
includes:
i)
Selection of a processing time window (usually the most
intense part of S and S-coda waves on an accelerogram) denoted by arrows on the
accelerograms presented in the attached figures.
ii)
Calculation of amplitude spectra followed by H/V spectral
ratios and their smoothing in a frequency window, usually 0.7-1.0 Hz.
iii) Calculation
of the average H/V ratio (over both horizontal components and all available
recordings of several earthquakes at a station).
3. Data
Used in the Study
All available
digital strong motion records were used. Initially, we processed the data of
the two recent earthquakes and then included in the analysis accelerograms of
several previous events which occurred during 1984-1993 and two recent felt
earthquakes from 1997. A list of the triggered accelerograph stations is
presented in Table 1. Parameters of recorded earthquakes are presented in
Table 2. Digital accelerographs A-700 and A-800 (Teledyne Geotech) were
used. The instrument trigger threshold in all cases was 0.006g. Note that all
accelerograms may be divided roughly into two groups:
c)
Records of the strongest (ML>6) and farthest
(r>100 km) earthquakes (Gulf of Aqaba 22.11.95 and Cyprus 9.10.96).
d)
All other records obtained from weaker (ML<5.5) and closer
(r<100 km) events.
In order to verify the
estimation method we used seismograms of two weak events near Eilat (see
Table 2a) and microtremors recorded by the Eilat accelerograph station
(data obtained by Dr. Y. Zaslavsky). Fig. 1 shows earthquake epicenters
and locations of the stations.
4. The
Gulf of Aqaba Earthquake
The Gulf of
Eilat/Aqaba earthquake (22.11.95, ML=6.2, Mw=7.1), which caused minor damage in
the town of Eilat, is the strongest event ever instrumentally recorded in
Israel. Ground fractures and cracking of road surfaces were noted; liquefaction
and sand boils were observed close to the coast (Osman and Ghobarah, 1996). The
main shock was followed by many aftershocks, two of which were also recorded by
accelerographs (see Table 2). PGA values measured at various distances
were compared with known attenuation relationships (Joyner and Boore, 1988;
Shapira et al., 1994) and showed anomalous high acceleration values at
almost all recording sites (see Table 3 and Fig. 2).
It should be noted
that the enhanced amplitudes may also be associated with the rupture
directivity phenomenon (Gitterman and Shamir, 1994). The directivity effect is
expected because of the 35-40 km rupture of the fault in a direction
approximately 15°
from the north. A rough estimation of the effect is shown on Fig. 3
(Gitterman et al., 1996a).
4.1 Eilat Station (EIL)
Fig. 4
shows accelerograms of the Eilat station from the main shock and two
aftershocks on the same scale (Fig. 41) for amplitude comparison and at different
scales (Fig. 4b) for waveform comparison. The accelerograms were found to
be similar for the main shock and the closer aftershock-2 of 26.12.95.
Aftershock-1 of 23.11.95, located on the eastern bank of the Gulf (see
Fig. 1) demonstrated very weak P-waves at all components. For the main
shock and aftershock-2, maximum amplitudes in the S-wave windows are
significantly larger at the EW and NS components than at the vertical
component, indicating a possible site effect.
The
acceleration response spectra, calculated by the SMA program, from the
horizontal EW component of the main shock (at different damping values) and an
average (±s)
obtained from both horizontal components of the main shock and aftershock-1
(for the 5% damping) are presented in Fig. 5. A comparison with Building
Code 413 response spectrum for S2 soil conditions (bold line on
Fig. 5b) shows a clear enhancement in the range of 0.2-0.5 sec. For
this station, located on an alluvial fan (about 50cm thick) overlying granite,
the ratios for the Gulf events demonstrated a strong, clear amplification
factor of about 6 at a frequency of around 2.2 Hz for the main shock
(Fig. 6). (It is possible that owing to a trough at this frequency in the
vertical component spectrum on Fig. 6a, the real amplification factor in
this case is smaller). A factor of about 3-4 in the frequency range
2.0-3.2 Hz was obtained for the aftershocks (Figs. 7 and 8). All
spectral ratios are plotted on Fig. 9. The ratios for aftershock-2 show
some differences from the main shock and aftershock-1 caused by the different
waveforms mentioned above. Averaging over three events gives an amplification
of about 4 at 2.2-2.3 Hz.
Using a
subsurface structure model (Table 4) inferred from observed microtremors
using the Aki spatial correlation method (Gitterman et al., 1996c) and the
EW main shock accelerogram as input for the SHAKE program (Schnabel
et al., 1972), we generated ground motions at the outcropped rock basement
and calculated the non-linear site response and the PGA on rock (Fig. 10).
The estimated PGA on hardrock is lower by a factor of 2 as compared to the
observed surface PGA and corresponds better to the attenuation curves. This
result agrees with the observations of the horizontal PGA at two Jordanian
stations founded on sandy soil (0.10g) and on hardrock (0.045g) in Aqaba when
the difference was explained by the effect of local site conditions (Osman and
Ghobarah, 1996).
4.2 OTHER STATIONS
Site response
estimates for other stations are presented on Figs. 11-20. They all more
or less demonstrate a site amplification factor of 3-7 at various frequencies
from 0.5 to 6 Hz. GMK station shows a strong amplification, which is
consistent for the horizontal components and for both recorded events (main
shock and aftershock-1, Figs. 12 and 13). H/V ratios for the most remote
station GOS (r=502 km) are not calculated because the vertical component
signal is very weak and close to the noise level (Fig. 20). Nevertheless,
relatively large horizontal component signals provide evidence of a site effect
at the station.
4.3
comparison with weak motion estimates at Eilat station
Correlation
was obtained with evaluations based on microtremor measurements taken a year
before the earthquake at the same site (compare Figs. 6 and 21) (Gitterman
et al., 1996). PGA of about 0.1g were observed at the Eilat station
(located on alluvium, Vs»300 m/s), the strongest ever recorded in Israel.
The ratios for both horizontal components, obtained from a 15 sec window
W1 (Fig. 22) of dominant S-waves from the main shock, demonstrated a
distinct amplification of factor A=5.8 at the frequency of f0=2.24 Hz,
whereas estimated based on lower amplitudes (about 0.01g) from S-coda window W2
(Fig. 23a) and aftershock‑2 (Fig. 23b) (same length window as W1)
showed a clear shift of f0 to 2.6 Hz and A»4.
The latest frequency estimate agrees with evaluations obtained from
microtremors (f0~2.8 Hz) (Fig. 21) and weak earthquakes
recorded at the same site (f0=2.7) (Figs. 24 and 25) as well as
with the analytical evaluations (f0=2.6) based on the subsurface
model (Table 4): 1-D SH transfer function (Fig. 28b) and Rayleigh
spectral ratio (ellipticity) (Fig. 26).
A
non-linearity effect could explain the fundamental frequency shift. A
discrepancy in amplification factors (increased values, contradictory to the
non-linear theory) could refer to the known inaccuracy of the H/V ratio method
in estimating amplification factors.
An
additional observation which could be related to evidence of non-linear
behavior, is provided by the anomaly shift of the dominant frequency to the
right (higher values) for later phases in spectra of short windows
(~6-7 sec) of S-waves from the main shock: 17-23 sec, 23-29 sec
and 26-43 sec for the EW component and 18-24.5 sec, 24.5-32 sec
and 38-45 sec for the NS component (Fig. 27a). This type of shift is
not observed for the main shock vertical component or for the horizontal
components of aftershock-2 (Fig. 27b). Usually, in the linear case, the
opposite effect is observed on recordings when late arrivals have lower
frequencies due to longer pathways and high frequency energy attenuation. This
is confirmed by analysis of aftershock-2 and weak earthquake spectra
(Fig. 26c) demonstrating a normal shift of the dominant frequency for
several successive windows.
The
proposed hypothesis is supported by calculation of non-linear transfer
functions (Joyner, 1997) using accelerograms of the main shock and aftershock
as input for the Joyner program (Fig. 28a): the same shift of f0
from 2.3 to 2.6 Hz was obtained (Fig. 28b). According to expert
opinion, an acceleration level of ~0.1g is large enough to cause a detectable
non-linear effect (Dr. I. Beresnev, personal communication). All
available estimations of site response at the Eilat station obtained using
different methods are presented in Table 5.
5. The
Cyprus Earthquake
6.
Results
of Site Response Estimations for Previous (1984-1993) and Recent (1997)
Earthquakes
Here
we analyze much weaker events (ML<5.5) than the Gulf of Aqaba and
Cyprus earthquakes. However these events occurred at closer distances
(r<100 km) and, therefore, some stations located on hardrock were
triggered although, as may be expected, no site effect was observed at these
stations (see Table 8). The measured PGA values are presented in
Table 7 (these data were published in part in Arieh et al., 1982 and
Gitterman et al., 1994). In most cases one or both horizontal values are
significantly larger than the vertical, indicating possible site amplification.
Spectral H/V ratio estimates are shown on Figs. 36-52. The 1979 earthquake
was not used for estimation since the data were analog recordings and due to
the absence of a vertical component. In some cases the ratios were not
calculated owing to unreliable weak signals (such as for station DA2 for the
27.04.87 earthquake, Fig. 38).
All
stations located on sediments show a more of less noticeable site effect. Most
stations installed at hardrock sites, such as the Haifa Technion (HAT),
Jerusalem (JER), Mifal (MIF) and Tunnel Gate (TUG), Dead Sea do not show any
stable amplification. There are several stations, located on consolidated rocks
(per available data presented in Table 1) such as Alonim (ALN), Ashqelon
(ASQ), Hadera (HAD), Izreel Valley (ISR), Mizpe Shalem (MIZ), Reading IRDG) and
Zefat (ZEF), which do nevertheless show an enhancement in a specific frequency
range. Some stations, such as Bet Shean (BET), demonstrate a very stable site
effect and similar ratios for two earthquakes with similar magnitudes and
locations (Figs. 43-44). In other cases, stations Almog (ALM)
(Figs. 16 and 46) and Qiryat Shemone (KIT) (Figs. 48 and 52) show
different ratios for two different events and horizontal components. In both
cases, the two recorded earthquakes were of very different magnitudes,
distances and azimuth. The EW component at station ALM showed a remarkable
spectral amplification corresponding well with the enhanced amplitudes on EW
accelerograms as compared with the vertical component. Nevertheless, this fact
can hardly be related to the site effect because it was observed at different
frequencies for the two earthquakes.
An
interesting example of different site responses at two neighboring stations MIF
and NET for the Moav earthquake of 28.09.91 is presented on Fig. 45.
Station MIF with hardrock conditions did not show any significant spectral
amplification (corresponding to almost equal amplitudes at all three components
of the accelerogram), whereas an enhancement was obtained for the NET station
installed on soft sediments though in a broad frequency range (large amplitudes
at horizontal accelerograms can be observed as compared with the vertical
component).
Response
spectra for both horizontal components of the Beirut earthquake recorded in
Zefat (ZEF) are presented in Fig. 51. Comparison with the response
spectrum of Code 413 for S2 soil conditions demonstrates a clear ground
motion enhancement at the ranges 0.2-0.4 and 0.06-0.8 sec.
7. Conclusions
Principal
parameters of the site response (amplification factors and fundamental
frequency or frequency range) of all triggered stations are presented in
Table 8 (stations with clear, stable and remarkable site amplifications
are marked in bold; stations which did not display any noticeable site effect
are marked in italics). We suggest that, due to local site effects, the
earthquake ground motions exceeded the instrument triggering threshold (0.006G)
and, therefore, several Israeli accelerographs were probably triggered. If this
is indeed the case, then it has major consequences on the evaluation of
earthquake hazard.
The
H/V spectral ratio estimates of site response based on strong motion records
are consistent with those obtained from weak motion seismograms of microtremor
and small events with analytic functions. An observed reduction of 15-20% in
the resonance frequency for PGA»0.1g compared to all numerous weak motion estimates, suggests
a possible non-linearity of site response at the Eilat station. If so, this is
one of the few cases for non-linear effect associated with PGA<0.2g.
Nevertheless, for engineering applications, the differences in resonance
frequency and amplification factor estimates obtained from strong and weak
motions are not significant and, hence, the estimates can be averaged.
Reliable
confirmation of site amplification effects should be carried out using detailed
microtremor and weak earthquake motion measurements and analytical estimates,
based on detailed local geotechnical and geological information followed by a
detailed comparative analysis of different methods of site effect evaluation as
realized for the Eilat accelerograph station.
Acknowledgments
Thanks
are due to Dr. Avi Shapira for his review of the manuscript and
numerous valuable comments. Uri Peled assisted by supplementing strong
motion data and station geology conditions. This study was supported by the
Center of Absorption in Science of the Ministry of Absorption and the Earth
Sciences Research Administration of the Ministry of National Infrastructures.
REFERENCES
Arieh, E.,
Rotstein, Y. and Peled, U., 1982. The Dead Sea Earthquake of
23 April 1979, Bull. Seism. Soc. Am., 1982, 72:1627-1634.
Gitterman, Y., Zaslavsky, Y. and Shapira, A., 1994.
Analysis of strong ground records in Israel, in Earthquake Engineering,
A. Rutenberg (Ed.), Balkema Publ., Rotterdam, 109-118.
Gitterman, Y. and Shamir, G., 1994. Application of the
stochastic model of rupture directivity to some felt events along the Dead Sea
plate boundary, XXIV ESC General Assembly, Athens, Greece, Abstracts.
Gitterman, Y., Shapira, A. and Peled, U., 1996a. Analysis
of strong motion records of the 22.11.95 Nuweiba earthquake and its aftershocks,
Israel Geological Society Annual Meeting, Eilat, Abstracts, p.33.
Gitterman, Y., Shapira, A., Zaslavsky, Y. and
Peled, U., 1996b. Preliminary analysis of strong motion records of the
Nuweiba (Egypt) 22.11.95 earthquake and the site-effects in Eilat, Israel XXV
ESC General Assembly, Reykjavik, Iceland, Abstracts, p.96.
Gitterman, Y., Zaslavsky, Y., Shapira, A. and
Shtivelman, V., 1996c. Empirical site response evaluations: case studies
in Israel, Soil Dynamics and Earthquake Engineering, 15:447-463.
Gitterman, Y. and Shapira, A. 1997a. Evaluation of
site-response from Israeli strong motion records, Israel Geological Society
Annual Meeting, Kefar Giladi, Abstracts, p.36.
Gitterman, Y. and Shapira, A., 1997b. A study of Site-Response
from Strong motion records, 29th General Assembly of the International
Association of Seismology and Physics of the Earth’s Interior, Thessaloniki,
Greece, Abstracts, p.319.
Joyner, W.B. and Boore, D.M., 1988. Measurement,
characterization and prediction of strong ground motion, in: Proc. Conf. on
Earthquake Eng. and Soil Dynamics II, GT Div/ASCE, Park City, Utah, 43-102.
Joyner, W.B., 1997. A Fortran program for calculating nonlinear
seismic response, U.S. Geological Survey, Open File Report No. 77-671.
Lermo, J. and Chavez-Garcia, F.J.
Site effect evaluation using spectral ratios with only one station, Bull.
Seism. Soc. Am., 1993, 83:1574-1594.
Osman, A. and
Ghobarah, A., 1996. The Aqaba earthquake of November 22, 1995,
EERI Special Earthquake Report.
Nakamura, Y., 1989. A
method for dynamic characteristics estimation of subsurface using microtremor
on the ground surface, QR of RTRI, 30, No.1;25-33.
Schnabel, P.B.,
Lysmer, J. and Seed, H.B., 1972. SHAKE - a computer program
for earthquake response analysis of horizontally layered sites, Earthquake
Engineering Research Center, Rep. EERC-72--12, Berkeley, California.
Shapira, A., Gitterman, Y. and Zaslavsky, Y., 1994.
Semi-empirical attenuation functions for seismic ground motions in Israel,
Israel Geological Society Annual Meeting, Nof Ginosar, Abstracts, p.33.
Figure
1. Earthquake epicenters and location of accelerograph stations.
