1

STATE OF ISRAEL ‫מדינת ישראל‬

THE MINISTRY of ENERGY AND WATER RESOURCES ‫משרד האנרגיה והמים‬
Office of the Chief Scientist ‫המדען הראשי‬

The use of P-wave spectra in the determination

of earthquake source parameters in Israel

Final report

Dr. A. Hofstetter and G. Ataev

THE GEOPHYSICAL INSTITUTE OF ISRAEL

December 2012 Publication No: ES-11-2012

GII Report No: 569/701/12
Contract Number 211-17-004
 2

Contents

List of Figures .................................................................................................................... 3

List of Tables...................................................................................................................... 4
List of abbreviations ......................................................................................................... 5
List of definitions .............................................................................................................. 5

‫ תקציר בעברית‬.............................................................................................................. 6

Abstract ...................................................................................................................... 8

1. Introduction......................................................................................................... 9

2. Data collection .................................................................................................. 11

3. Processing and calculations........................................................................ 18

3.1 Spectral analysis of seismograms and strong motion data ............................ 18
3.2 The time window determination........................................................................... 22
3.3 Ratio between spectral amplitudes of S- and P- waves .................................. 26
3.4 Regional attenuation of the low frequency spectral displacements............. 29
3.5 Distance effect on the measured corner frequency and ratio between
corner frequencies obtained from the analysis of S- and P-waves. ................ 32

4. Source parameters determination and analysis ............................... 34

4.1 Seismic moment and moment magnitude .......................................................... 34
4.2 Stress drop.................................................................................................................. 38
4.3 Source radius and source area ............................................................................... 41

5. Discussion and conclusions......................................................................... 42

Acknowledgements ............................................................................................... 44

References ................................................................................................................ 44
 3

List of Figures
Figure 1. Location map of earthquake foci and seismic stations used in this study. .............. 11
Figure 2. Distribution of seismic data according to earthquake magnitudes Md and epicentral
distances. Broadband, short period and accelerometric data are shown by red, black and blue
circles, respectively. ................................................................................................................. 12
Figure 3. Histogram of earthquake magnitudes used in this study (a) and corresponding
hypocentral depths (b).............................................................................................................. 13
Figure 4. Seismograms of an earthquake, which occurred on 07.09.2004 in the Dead Sea
region with Md=3.7. Also shown are the epicentral distances. ............................................... 18
Figure 5. (a) Example of a seismogram of an earthquake (07.09.2004, Dead Sea region,
Md=3.7) recorded by the station MZDA. The analyzed time window is marked in blue; (b)
The P-wave displacement spectrum marking the measured level of the low-frequency spectral
amplitude Ω(m*s) and the corner frequency f0(Hz). ............................................................... 19
Figure 6. Displacement spectra (m*s) calculated from seismograms of an earthquake
(07.09.2004, Dead Sea region, Md=3.7) recorded at stations located at distances of 26 km(a)
and 232 km(b) from the epicenter. Also marked are the measured f0 and Ω. ......................... 20
Figure 7. The P-wave displacement spectra of an earthquake (2008.02.15, Md=5.1, Lebanon)
calculated (a) Spectral displacement in m*s from the vertical component of station Sal'it and
(b) Spectral displacement in cm*s from the vertical component of accelerometer in the town
Petah Tiqva. ............................................................................................................................. 21
Figure 8. The P-wave displacement spectra of the earthquake (1997.03.26, Md=5.6, Roum
Fault), calculated in m*s from the vertical component of the broad band station JER (a) and
in cm*s from the vertical component of the accelerometer at KIT (b). ................................... 21
Figure 9. The P-wave displacement spectra of the earthquake (2012.03.22, Md=3.7, Dead
Sea), calculated in m*s from the vertical component of the short period station MZDA in
Mezada (a) and in cm*s from the vertical component of the accelerometer installed in ARAD
(b), located at distances 23 km and 20 km from the epicenter, respectively. .......................... 22
Figure 10. Short period vertical seismograms from (a) a local earthquake (2010.19.10,
Md=3.5 in the Arava), recorded by the station ZFRI and (b) a local earthquake (2011.01.01,
Md=4.0 in the Jordan Valley), recoded by the station BLGI. The selected time windows of
1.5 sec each are marked in blue. .............................................................................................. 23
Figure 11. Short period vertical seismograms from (a) a local earthquakes (2004.10.28,
Md=3.9 in East Sinai), recorded by station SLTI and (b) a local earthquake (1999.04.11,
Md=3.6 in the Hula-Kinneret region), recorded by station KMTI. The selected time windows
of 3.0sec and 4.0sec, respectively, are marked in blue. ........................................................... 24
Figure 12. The broad band vertical seismograms from (a) a regional earthquake (2012.07.10,
Md=3.7 in Lebanon), recorded by station MMLI and (b) a regional earthquake (2012.05.11,
Md=5.3 in Cyprus), recorded by station EIL. The selected time windows of 2.0sec and
6.0sec, respectfully, are marked in blue. .................................................................................. 25
Figure 13. The relation between the low frequency spectral displacements of S- and P-waves.
Also included is the measurement from the Gulf of Aqaba earthquake (22.11.1995, Mw=7.2).
Marked in red are data of the earthquakes that occurred in the Gulf of Aqaba and Saudi
Arabia region. .......................................................................................................................... 26
Figure11 . The ratio between the low frequency spectral displacement from S- and P-waves
for all studied events versus the epicentral distance (a) and duration magnitudes (b). The red
circles mark earthquakes that occurred in the Gulf of Aqaba and the Saudi Arabia region. ... 27
Figure 15. The relation between the low frequency spectral displacements of S- and P-waves.
Also shown is the log-log regression line. ............................................................................... 28
 4

Figure 16. Examples of the empirical attenuation functions of Ω(P) for some studied
earthquakes. Also shown are the duration magnitudes. ........................................................... 30
Figure 17. Distance correction D for the low frequency spectral displacement Ω(P), as a
function of distance R. ............................................................................................................. 31
Figure 18. Residuals of the parameter D=R*℮ 0.0042R against distance, R. ............................. 31
Figure 19. The ratio between f‫׳‬0 and f0 as a function of epicentral distance, R. ..................... 32
Figure 02 . Residuals of the regression ln (f0‫ ׳‬/f0)=0.00043R against the distance R............. 33
Figure 01 . Corner frequency determined from P-wave analysis, f0 (P) versus f0 determined
from S-wave analysis, f0 (S)..................................................................................................... 33
Figure 22. Seismic moment, M0 (P) versus seismic moment M0 (S). The straight line
represents M0 (P) = M0 (S). ..................................................................................................... 35
Figure 02. The relation between M0 and Md for all events used in this study. The red solid
lines represent the linear equations for two magnitude ranges; the green line represents the
polynomial quadratic equation for the whole magnitude range. .............................................. 37
Figure 01. The relation between Mw obtained from P-waves and Md. The solid line is the
least-squares fit to the data ....................................................................................................... 37
Figure 25. P-wave stress drop, ∆σ (P) versus S-wave stress drop, ∆σ (S). The straight line
represents ∆σ (P) = ∆σ (S). ...................................................................................................... 38
Figure 26. The stress drop, ∆σ (P) variation with the seismic moment M0 (P) for all events
used in this study. ..................................................................................................................... 39
Figure 27. The stress drop, ∆σ versus the seismic moment M0 obtained from P- (red circles)
and S-wave spectra. Also included is the Gulf of Aqaba earthquakes of 1995, (main shock
Mw=7.2) (blue circles). ........................................................................................................... 40
Figure02 . Relationship between the seismic moment, M0 and the corner frequency, f0
obtained from P-waves (red circles) and S-waves (blue circles). The solid lines indicate the
expected relation when the stress drop is constant. Note that the corner frequency divided by
a factor of 1.24. ........................................................................................................................ 40
Figure 29. Moment magnitude, Mw(P) versus the source are, S=πr2 in km2, where r is source
radius of earthquakes. .............................................................................................................. 41

List of Tables

Table 1 . Source parameters of earthquakes obtained from P-wave spectra used in this study.
.................................................................................................................................................. 14
Table 0 . M0 and Mw estimates obtained from waveform inversion and spectral analysis of
P- and S-waves. ........................................................................................................................ 36
 5

List of abbreviations

ISN -Israel Seismic Network

List of definitions

P-wave-primary seismic body waves,- produced by earthquakes.

S-wave –secondary seismic body waves, produced by earthquakes
Fourier Spectra – A representation of time-domain signal in the frequency domain,
which can be generated via Fourier transform of the signal.
Low frequency spectral amplitude- is constant amplitude level at low frequencies of
displacement spectrum.
Corner frequency- is boundary in a system’s frequency response at which energy
flowing through the system begins to be reduced (attenuated) rather than passing through and
defines as the intersection point of the low- and high-frequency asymptotes.
Seismic Moment -measure of quake rupture size related to leverage of forces
(couples) across area of fault slip, provides estimate of overall size of the seismic source.
Stress drop -the amount of stress released by an earthquake, stress drop considers
average of a highly variable slip over a fault.
 ‫‪6‬‬

‫תקציר בעברית‬

‫הערכת פרמטרי המקור של רעידות אדמה בישראל מספקטרום התנודות של גלי ‪.P‬‬

‫במחקר זה נבחנה האפשרות להעריך את פרמטרי המקור של רעידת אדמה מתוך ספקטרום קבוצת‬
‫גלי ‪ P‬המגיעים ראשונים לתחנה‪ .‬הפרמטרים שאותם מבקשים להעריך הם המומנט הסיסמי‪,Mo ,‬‬
‫המגניטודה ‪ , Mw‬הפרשי המאמץ‪, ∆σ ,‬רדיוס המקור‪ ,ro ,‬ושטח המקור ע"פ המודל של ברון‪.‬‬
‫בהתאם למודל של ברון )‪ (Brune, 1970, 1971‬ניתן להעריך את המומנט הסיסמי וממנו לחשב‬
‫את שאר הפרמטרים של המקור‪ ,‬ממדידה של האמפליטודה הספקטרית של ההעתקה (‪)Displacement‬‬
‫בתחום התדירויות הנמוכות ‪ Ω‬ומדידה של תדירות הקצה (‪ .)corner frequency f0‬המודל של ברון פותח‬
‫במקורו עבור אנליזה של גלי ‪ S‬ובוצעו מספר מחקרים בעבר‪ ,‬עבור הרשת הסיסמית של ישראל‪,ISN ,‬‬
‫שעניינם הערכת פרמטרי המקור מאנליזה ספקטרית של גלי ‪( S‬ראה ‪, Hofstetter and Ataev, 2011‬‬
‫‪ ) Shapira and Hofstetter, 1993‬מחקר זה הוא הראשון לבחינת הישימות של האנליזה הספקטרית של‬
‫השניות הראשונות של גלי ‪ .P‬במחקר זה נבחנו ‪ 0421‬סיסמוגרמות של ‪ 39‬רעידות אדמה מקומיות‬
‫ואזוריות בעוצמות ‪ , Md =2.7-5.6‬שנקלטו בתחנות סיסמיות ובמדי תאוצה ברחבי ישראל בטווח‬
‫המרחקים ‪ 8-551‬ק"מ‪ ,‬בין השנים ‪ . 0335-4104‬קטע הזמן שנבחר לאנליזה נקבע על פי המרחק ממוקד‬
‫הרעש ונע בין ‪ 0.5‬שניות בתחנות הקרובות ועד ‪ 6‬שניות בתחנות הרחוקות‪ .‬כשניה אחת יותר על כל ‪011‬‬
‫ק"מ כאשר תחילת הקטע ברגע הגעת גל ‪ P‬ראשון וסופו לאחר השיא של קבוצת גלי ‪ P‬הראשונה‪.‬‬
‫בשלב ראשון של המחקר נבחנה התלות במרחק האפיצנטרי ‪ R‬של ערכי המדידה ‪ Ω‬ו ‪ fo‬מגלי ‪P‬‬
‫דהיינו‪ Ω(P) ,‬ו )‪ .fo(P‬התברר שהתלות במרחק של תדירות הקצה )‪ fo)P‬וכן מקדם הניחות האנ‪-‬אלסטי‬
‫)‪ Ω(P‬כמעט זהים לאלו שהתקבלו עבור גלי ‪ . S‬פונקציות הניחות לספקטרום גלי ‪ P‬הן‬ ‫של‬
‫‪ D=R*℮0.0042R‬ו ‪. f0=f'0*℮0.00043R‬‬
‫משנקבעו פונקציות הניחות עבור מדידות )‪ Ω(P‬ו )‪ fo(P‬ניתן לחשב את ערכי ‪ Ω‬ו ‪ fo‬בקרבת‬
‫המקור (דהיינו‪ ,‬לאחר התיקון למרחק ‪ ) R‬ומכאן לחשב את ערכי הפרמטרים ; ‪ ∆σ , , Mw , Mo‬ו ‪ ro‬ע"פ‬
‫המודל של ברון‪ .‬ערכי הפרמטרים עבור כל רעידת אדמה נקבעו להיות הממוצע הגיאומטרי של הערכים‬
‫שהתקבלו מכל אחת מהתחנות שמדדה את רעידת האדמה‪ .‬תוצאות החישובים מוצגים בטבלה ‪.0‬‬
‫השוואה עם ערכי הפרמטרים של רעידות אדמה אלה כפי שהתקבלו מאנליזה של גלי ‪ S‬מראה‬
‫התאמה טובה מאד‪ .‬יש גם התאמה טובה על ערכי מומנט סיסמי שהתקבלו בשיטת האינברסיה של‬
‫הסיסמוגרמה‪.‬‬
‫כמו כן מצאנו את הקורלציה ‪ . Ω (S) = 3.06*Ω (P)0.95‬בבדיקת מתאם זה‪ ,‬התגלה כי ערכי ‪Ω‬‬
‫)‪ (P‬בתחנות המרוחקות ‪ 051‬ק"מ ממוקדי רעידות אדמה במפרץ אילת (ודרומה להם) הרבה יותר נמוכים‬
‫מאלה הנצפים מרעידות אדמה במקומות אחרים‪ .‬לפיכך‪ ,‬נתוני מדידה אלה לא שותפו במחקר‪.‬‬
 ‫‪7‬‬

‫בהמשך המחקר בחנו את המיתאם בין הפרמטרים הדינאמיים של המקור‪ ,‬כפי שחושבו ממדידות‬
‫של ספקטרום גלי ‪ P‬לבין המגניטודה ‪ Md‬שנקבעת לפי משך הסיגנל הסיסמי בתחנות רשת ישראל ‪ISN‬‬
‫מזה למעלה מ ‪ 91‬שנים‪ .‬המשוואות הניסיוניות הן‪:‬‬

‫‪log M0=1.01(±0.08) Md + 10.63(±0.30), σ=0.25‬‬ ‫‪for‬‬ ‫‪2.7 ≤ Md ≤4.0,‬‬

‫‪log M0=1.53(±0.11) Md + 8.51(±0.49), σ=0.31‬‬ ‫‪for‬‬ ‫‪4.0 ≤ Md ≤5.6,‬‬

‫‪log M0 =13.16-0.46 Md +0.21Md2 ,‬‬ ‫‪σ=0.27‬‬ ‫‪for‬‬ ‫‪2.7 ≤ Md ≤ 5.6‬‬

‫‪Mw = 0.82 (±0.03) Md+0.52(±0.13),‬‬ ‫‪σ=0.20‬‬ ‫‪for‬‬ ‫‪2.7 ≤ Md ≤ 5.6,‬‬

‫‪.‬כאשר ‪ M0‬ביחידות של ‪.Nm‬‬

‫הפרש המאמץ ע"פ מודל ברון ברעידות האדמה ששימשו במחקר הוא בין ‪ 1.0‬לבין ‪ 93‬מגה‪-‬פסקאל ‪ .‬ככלל‪,‬‬
‫המאמץ שהשתחרר ברעידת האדמה גדל ככל שעוצמת רעידת האדמה גדלה‪ .‬הנתונים לא מאפשרים לקבוע‬
‫בודאות אך נראה שיש גבול עליון שבו‪ ,‬עבור רעידות אדמה חזקות‪ ,‬הפרש המאמץ קבוע‪ .‬הערך הזה כנראה‬
‫קרוב ל ‪ 21‬מגה‪-‬פסקאל ובודאי גבוה יותר ממה שהוערך עד היום‪.‬‬

‫נתוני מחקר זה מצביעים על ישימות השיטה הספקטרית לקביעה מהירה של פרמטרים של רעידת האדמה‬
‫והם מאפשרים מעבר לקביעת שיגרתית של ערכי ‪ Mw‬כמדד לעוצמת רעידות אדמה שנקלטות ברשת‬
‫ישראל‪.‬‬
 8

Abstract
We investigated the applicability and reliability of determining earthquake source
parameters, such as the seismic moment, the stress drop and the source and slip dimensions,
by analyzing the spectra of the P wave-train recorded by the stations of the Israel Seismic
Network, ISN. The methodology follows the work of Hanks and Wyss, (1972), Baumbach
and Borman (1999, 2012) and others.
Applying Brune’s seismic source model (Brune, 1970, 1971) for P-wave, we have
analyzed 1240 digital records from 93 local and regional earthquakes in the magnitude range
Md=2.7-5.6, occurring in a distance range of 8 km to 550km during 1995-2012. The first step
of this study was to estimate for the empirical dependence of P-wave displacement spectrum
(the amplitude at low frequencies, Ω) and the corner frequency, f0, on the epicentral distance
R, similar to the approach of Shapira and Hofstetter (1993) applied to investigate the
attenuation of low frequency amplitudes and corner frequency of S-wave spectra. Using
nonlinear regression procedure, we obtained, from P-wave spectra measurements a distance
correction parameter, D=R*℮0.0042R for the low frequency spectral amplitude and
f0=f'0*℮0.00043R for the corner frequency. This attenuation model for P-wave spectra was
applied to determine the source spectrum of the ground displacement for each earthquake in
the database and then to determine the seismic moment, M0, the moment magnitude, Mw, the
source radius, r0 and the stress drop, ∆σ.
We obtained the seismic moments for earthquakes in the magnitude range 1.06*1013
≤M0 ≤ 1.25*1017 (Nm) and their corresponding moment magnitudes, Mw, from 2.6 to 5.3.
The corner frequencies of analyzed earthquakes vary between 1.0 to 5.6 Hz. Computed
values of stress drop, ∆σ are in range between 10-1 MPa and 39 MPa.
Comparing the kinematic source parameters estimated from P-wave analysis with
those obtained by Hofstetter and Ataev (2011) from spectra of S-waves shows that the values
are consistent. The empirical correlations between source parameters and the duration
magnitudes are in agreement with those obtained previously from the analysis of S-waves by
Shapira and Hofstetter (1993), Hofstetter (2003), Hofstetter et.al (2008) and Hofstetter and
Ataev (2011) for events along the Dead Sea Fault or the Eastern Mediterranean region. The
earthquake source parameters estimated by taking into account the regional conditions can be
used for a wide range of applications in Israel, including seismic hazard assessments and
developing earthquake scenarios. The obtained results also facilitate routine analysis of
seismic data of the Israel Seismic Networks.
 9

1. Introduction

Seismologists consider the seismic moment M0 and moment magnitude Mw

(Kanamori, 1977) as a more reliable measure of earthquake size that represents the total
deformation at the source and correlates with the seismic energy release. These dynamic
earthquake source parameters have been widely used for computing scaling laws applicable
for seismic hazard assessments. Therefore defining the seismic moment M 0 and other source
parameters and developing a reliable relationship between them and local magnitude Md that
is in use for more than 30 years is of fundamental importance. The empirical determination of
the dynamic source parameters using the spectra of S-waves and analysis of their
interrelationships for the active seismic zones in Israel was first performed by Shapira and
Hofstetter (1993). One of the important relationships obtained by Shapira and Hofstetter
(1993) is

log M0=16.0( ± 0.4) + 1.5( ± 0.1) Md, 0.4< Md <6.7, (1)

where M0 is the seismic moment in dyne*cm and Md is the coda magnitude determined from
the duration of the seismic signal (Shapira, 1988).

In an earlier study of the Dead Sea basin van Eck and Hofstetter (1989, 1990) used
mainly low-magnitude events (1.0< ML<4.0) and obtained a relationship

log M0=17.0 + 1.2 ML, (2)

In a study of the aftershock sequence of earthquakes occurring between November

1995 and December 1996 in the Gulf of Aqaba, Hofstetter (2003) obtained :

log M0=17.93(±0.07) + 0.87(±0.02) ML, 2.5<ML≤4.0 (3)

log M0=15.01(±0.25) + 1.59(±0.06) ML, 4.0<ML≤6.2 (4)

Begin and Steinitz (2005), using data only of local earthquakes 0 ≤ ML ≤ 5.2 which
occurred during 1984-2004, developed the relationship:

log M0=18.46+ 0.177 ML+0.160 ML2 , (5)

where the seismic moment is in dyne*cm.

A similar second order relationship with ML was found for earthquakes in the Carmel-
Tirtza fault (Hofstetter et al. 1996). Relationships between M0 and Md derived from
 11

recordings of the Dead Sea Mw=5.1earthquake, and its aftershocks on February 2004
(Hofstetter et al., 2008) are:

log M0=10.8(±0.5) + 0.9(±0.03) Md, 1.1≤Md≤3.5 (6)

log M0=8.01(±0.5) + 1.6(±0.3) Md, 3.5≤Md≤5.2 (7)

where M0 is in Nm. Hofstetter and Ataev (2011) estimated the source parameters of 100 local
and regional events occurred during 1995-2011 by analyzing the S-wave displacement
spectra obtained the empirical relations:

log M0= 9.84±(0.19)+1.18(±0.05) Md, 2.7 ≤ Md ≤4.0 (8)

log M0= 8.35(±0.39)+1.55(±0.09) Md , 4.0 ≤ Md ≤5.6 (9)

log M0 =12.27-0.20 Md +0.19Md2 , 2.7 ≤ Md ≤ 5.6, (10)

where M0 is in Nm.

The use of P-wave data can be of particular interest when the S-wave data are clipped
or when trying to estimate magnitudes in real time. In such cases, analysis of P-waves has
obvious advantage. Hanks and Wyss (1972), Baumbach and Borman (1999, 2012)
demonstrated the use of P-wave spectrum to estimates seismic moment M0, corner frequency
f0 and source radius r0. Watanabe et al., (1996) have investigated the source characteristics
from P- and S-wave for earthquakes with magnitudes ranging from 3.3 to 6.0 and confirmed
that the seismic moment determined from P-and S-waves are consistent. Ottemoller and
Havskov (2003) presented good results of the method to automatically determine the moment
magnitude from the source spectrum of P and S, or Lg waves of local and regional
earthquakes. Tusa and Gresta (2008) had used displacement spectra of P-wave to estimate
source parameters for events of magnitude less than 4.6 with the hypocentral distances in
range between 13 and 90 km. Harrington and Brodsky (2011) also used P- and S-wave
displacement spectra for estimating seismic moment and corner frequency.

We have investigated earthquakes source parameters and seismic relations using P-

wave spectra for regional and local earthquakes that occurred during 1995-2012, by applying
Brune’s seismic source model (1970, 1971). This is the first time in Israel that the source
characteristics have been obtained from P-wave and both of results from P- and S-wave are
compared.
 11

2. Data collection

In this study we used seismograms and some accelerograms of 93 local and regional
earthquakes of magnitude Md > 2.7 (See Figure1).

Earthquake
Short period station
Broad band station
Array

Figure 1. Location map of earthquake foci and seismic stations used in this study.
 12

The data include the short period vertical component seismograms of the Geotech S-
13 (1 Hz) seismometer, recorded with a sampling rate of 50 sps, and vertical component
seismograms of the STS-2 broad band seismometers.

Figure 2 shows the distribution of the used seismograms with respect to magnitudes
and epicentral distances. It should be noted that there is a wealth of data of small and
moderate earthquakes, but strong-motion data are few. We also included in the analysis, the
broad band data of the stations in Jerusalem and Eilat that recorded the Beirut earthquake of
Md =5.6 which occurred in 1997.

Figure 2. Distribution of seismic data according to earthquake magnitudes Md and

epicentral distances. Broadband, short period and accelerometric data are shown by red, black
and blue circles, respectively.

The selection of a waveform for analysis follows the condition that the earthquake
was recorded by a number of calibrated seismic stations. It is also required that the signals are
not clipped and are at least twice the background noise. We tried to obtain a balanced
coverage through a wide magnitude range (Figure 3,a) and a wide range of epicentral
distances. Most of these events occurred in the upper crust at depths of 3-20 km (see Figure
3, b).
 13

a)

93 events

b)

Figure 3. Histogram of earthquake magnitudes used in this study (a) and corresponding
hypocentral depths (b).

Table 1 presents a summary of the studied earthquakes and their kinematic source
parameters as obtained in this study. For 73 events we could determine, individually, the
attenuation functions. Other, marked by *, were not recorded by sufficient number of stations
to allow such an analysis or occurred in the Gulf of Aqaba or Saudi Arabia. These events
show different spectral characteristics, as will be discussed below. Nevertheless, the
measurements of those events with some restriction also were used to determine the source
parameters. Number of stations indicates the number of seismograms used in the analysis.
The number in brackets is the number of accelerograms (see Table1).
 14

Table 1 . Source parameters of earthquakes obtained from P-wave spectra used in this study.

Date and time Region Lat. Lon Depth Md M0 Mw f0 Stress r0 No of

(Year, month, day, (0N) (0E) (km) (Nm) (Hz) drop (km) stations
hour, minute) (MPa)
1 199511221512 East Sinai 29.96 34.40 13 4.3 2.23E+14 3.5 3.41 0.84 0.565 7
2 *199606011606 Aragonese-Deep 28.84 36.48 12 4.6 1.35E+16 4.7 2.70 25.22 0.713 1
3 *199606022123 Saudi-Arabia 28.79 34.91 10 4.1 5.91E+14 3.8 2.99 1.51 0.643 1
4 *199606251045 Eilat-Deep 29.16 34.72 4.3 5.66E+14 3.8 2.01 0.44 0.957 1
5 *199609041717 Aragonese-Deep 28.89 34.78 4.6 2.60E+15 4.2 2.17 2.53 0.887 1
6 *199701050008 Saudi-Arabia 28.81 34.94 8 4.1 3.45E+14 3.6 3.33 1.22 0.578 1
7 *199703260423 Beirut 33.78 35.62 10 5.6 9.03E+16 5.2 1.00 8.71 1.917 4(2)
8 199705290705 Golan Heights 33.35 35.62 8 3.7 3.00E+14 3.6 3.52 1.23 0.548 12
9 199804142129 Northern-Jordan 31.45 35.58 10 3.0 4.32E+13 3.0 3.59 0.19 0.537 10
10 199804250129 East Sinai 29.78 34.59 18 4.1 6.36E+14 3.8 3.91 3.41 0.493 7
11 199904111944 Hula-Kinneret 33.15 35.63 4 3.6 3.03E+14 3.6 4.11 1.89 0.469 11
12 *199910050500 Aragonese-Deep 29.00 34.84 5 4.0 1.26E+14 3.3 3.82 0.67 0.505 3
13 *199910050544 Aragonese-Deep 28.91 34.90 14 4.8 4.55E+15 4.4 2.21 4.45 0.886 2
14 199910281538 Negev 30.40 34.98 9 4.6 1.86E+15 4.1 2.84 4.04 0.679 12
15 *200003081422 Aragonese-Deep 28.78 34.71 5 5.0 1.90E+16 4.8 2.04 7.36 1.078 3
16 200003110151 Jordan-Valley 32.11 35.50 4 3.1 5.81E+13 3.1 3.78 0.30 0.509 11
17 *200003190949 Eilat-Deep 29.21 34.81 2 3.2 1.20E+14 3.3 2.80 0.25 0.687 5
18 200003250935 East Sinai 29.83 34.52 14 3.3 8.80E+13 3.2 4.40 0.72 0.437 10
19 200003280102 Dead-Sea-Basin 31.71 35.52 17 3.2 4.60E+13 3.1 4.07 0.29 0.473 11
20 *200003302016 Aragonese-Deep 28.80 34.69 3 3.0 9.01E+13 3.2 4.12 0.6 0.483 3
21 *200004060637 Aragonese-Deep 28.79 34.76 5 5.0 6.48E+15 4.5 1.80 3.34 1.064 3
 15

22 200004260103 East-Sinai 29.98 34.41 13 3.7 2.20E+14 3.5 3.98 1.32 0.484 11
23 200007050333 Dead-Sea-Basin 31.48 35.58 12 3.5 7.61E+13 3.2 3.32 0.26 0.579 11
24 *200010080258 Aragonese-Deep 28.84 34.67 5 4.6 5.53E+14 3.8 2.58 0.90 0.747 2
25 *200108041530 Aragonese-Deep 28.74 34.78 5 4.2 1.33E+15 4.0 2.46 1.80 0.795 1
26 200110081125 Arava 30.47 35.29 10 4.2 1.16E+15 4.0 3.67 4.86 0.524 12
27 200202240955 E.Shomron 32.07 35.47 2 3.6 1.60E+14 3.4 2.60 0.25 0.741 12
28 *200203110125 Aragonese-Deep 28.95 34.79 14 4.0 1.87E+14 3.5 3.39 0.69 0.568 2
29 *200307231636 Aragonese-Deep 28.82 34.79 2 4.4 1.04E+15 4.0 2.43 1.22 0.816 2
30 200310190529 Lebanon 33.30 35.43 1 3.1 1.09E+14 3.3 3.95 0.64 0.487 9
31 200310231438 E.Mediter.-Sea 33.88 34.70 25 3.7 9.72E+14 3.9 3.94 5.67 0.489 8
32 200312312044 Dead Sea Basin 31.50 35.51 16 3.7 3.98E+14 3.7 3.48 1.60 0.553 9
33 200402110812 Dead-Sea-Basin 31.70 35.56 17 5.2 5.70E+16 5.1 1.73 30.82 1.115 8
34 200402130702 Dead-Sea-Basin 31.69 35.55 18 3.7 2.48E+14 3.5 4.21 1.70 0.457 10
35 200402190547 Dead-Sea-Basin 31.72 35.49 17 3.0 2.89E+13 2.9 5.61 0.49 0.343 9
36 200402240211 Dead-Sea-Basin 31.72 35.49 17 3.5 1.20E+14 3.3 4.81 1.34 0.401 12
37 200403152349 Dead-Sea-Basin 31.36 35.57 9 4.3 1.47E+15 4.1 2.68 2.16 0.767 9
38 200406031318 Jordan-Valley 32.02 35.49 4 3.2 5.38E+13 3.1 3.49 0.22 0.553 9
39 *200406071345 Palmira 33.78 36.51 5 4.0 8.86E+14 3.9 2.53 1.39 0.761 6
40 200407071434 Jordan-Valley 31.97 35.55 14 4.8 9.74E+15 4.6 2.28 9.94 0.835 14(4)
41 200407091215 Dead-Sea-Basin 31.69 35.57 16 3.7 2.01E+14 3.5 4.80 2.11 0.401 10
42 200407202142 Carmel-Tirza 32.47 35.24 13 3.6 1.21E+14 3.3 4.92 1.39 0.390 10
43 200408081241 Carmel-Tirza 32.47 35.26 10 4.0 4.95E+14 3.7 2.51 0.67 0.768 12
44 200410280710 East-Sinai 29.95 34.42 13 3.9 3.07E+14 3.6 2.99 0.78 0.644 11
45 200410290351 Hula-Kineret 33.18 35.61 6 3.2 6.97E+13 3.2 4.22 0.51 0.456 13
46 200410290609 Lebanon 33.19 35.61 11 3.1 3.76E+13 3.0 4.43 0.31 0.435 10
 16

47 *200502120803 Saudi-Arabia 28.90 35.12 5 4.1 2.27E+14 3.5 2.70 0.43 0.713 2
48 *200502260057 Saudi-Arabia 28.87 35.19 5 4.2 6.15E+14 3.8 2.68 0.91 0.772 2
49 200506010326 Dead-Sea-Basin 31.29 35.41 8 3.7 2.25E+14 3.5 3.54 1.18 0.545 9
50 200506012130 Barak-fault 30.27 35.01 7 3.4 5.52E+13 3.1 3.80 0.29 0.508 9
51 200507030224 E. Mediter.-Sea 33.81 34.83 2 3.4 1.58E+14 3.4 4.41 1.20 0.439 7
52 200507180952 Dead-sea-Basin 31.23 35.38 11 3.5 4.09E+13 3.0 4.66 0.38 0.413 11
53 200509021017 Lebanon 33.37 35.49 4 3.1 4.18E+13 3.0 3.94 0.26 0.488 8
54 200509070835 Jordan-valley 32.18 35.56 9 3.7 1.80E+14 3.4 4.69 1.62 0.411 11
55 200510030405 Jordan-Valley 32.04 35.59 15 4.0 1.05E+15 4.0 3.82 6.0 0.504 9
56 200609090457 E.Shomron 32.02 35.48 4 4.4 2.35E+15 4.2 2.17 2.09 0.888 12
57 200609170822 E.Shomron 32.02 35.47 1 4.0 3.86E+14 3.7 2.41 0.51 0.798 9
58 200609212107 E.Shomron 32.04 35.47 2 3.5 1.22E+14 3.3 2.90 0.29 0.665 14
59 200702092211 Dead-Sea Basin 31.09 35.52 9 3.6 8.73E+14 3.9 2.69 1.72 0.715 8
60 200702092212 Dead-Sea Basin 31.09 35.53 10 4.5 4.81E+15 4.4 2.37 5.77 0.814 11(3)
61 200711200919 Dead-Sea Basin 31.69 35.55 10 4.5 2.40E+15 4.2 2.66 3.94 0.719 10
62 200711232219 Judea-Samaria 31.86 34.94 11 4.2 1.23E+15 4.0 3.73 6.09 0.516 8(1)
63 200712020737 Dead-Sea-Basin 31.67 35.55 9 4.0 9.44E+14 3.9 3.51 3.92 0.544 7
64 200802112346 Lebanon 33.26 35.42 1 4.3 2.43E+15 4.2 2.73 4.60 0.705 10(1)
65 200802151036 Lebanon 33.33 35.41 3 5.1 3.70E+16 5.0 1.60 14.19 1.205 14(1)
66 200803010547 Lebanon 33.26 35.43 3.4 1.48E+14 3.4 3.77 0.76 0.511 13
67 200803200229 Hula-Kinneret 33.24 35.44 4 2.7 2.89E+13 2.9 4.50 0.25 0.428 9
68 200806121257 Lebanon 33.28 35.46 1 3.9 3.91E+14 3.7 2.52 0.67 0.760 17
69 200806121542 Lebanon 33.32 35.44 3.9 9.70E+14 3.9 2.36 1.27 0.817 12
70 200806130528 Lebanon 33.32 35.45 3.9 3.54E+14 3.6 2.72 0.77 0.709 15
71 200810121255 Sinai 29.77 34.76 3 2.7 1.06E+13 2.6 4.75 0.12 0.406 8
 17

72 200901210640 E. Mediter.-Sea 32.51 34.46 10 3.6 2.60E+14 3.6 3.91 1.42 0.493 11
73 200904171103 Lebanon 33.34 35.43 4.2 7.83E+14 3.9 2.34 0.95 0.824 9
74 200904211318 Northern-Jordan 31.48 35.56 10 3.3 6.42E+13 3.1 4.41 0.56 0.435 9
75 201003090050 E. Mediter.-Sea 33.51 34.81 21 3.5 4.64E+14 3.7 3.75 2.07 0.511 10
76 201003201845 Hula-Kinneret 32.91 35.62 13 3.6 1.34E+14 3.4 3.59 0.67 0.536 8
77 *201007151125 Aragonese-Deep 28.97 34.76 9 4.2 9.28E+14 3.9 2.41 1.24 0.798 4
78 201008222223 Hula-Kinneret 33.10 35.46 2 3.1 8.37E+13 3.2 3.75 0.43 0.514 9
79 201010170800 Dead Sea Basin 31.09 35.47 10 2.9 5.31E+13 3.1 3.89 0.30 0.495 9
80 201010191646 Arava 30.40 35.30 10 3.5 1.08E+14 3.3 4.80 1.15 0.401 12
81 201101011629 Jordan-Valley 32.66 35.63 3 4.0 8.20E+14 3.9 2.94 1.60 0.656 11(1)
82 201101040536 Galilee 33.09 35.44 1 2.9 3.42E+13 3.0 3.70 0.18 0.514 7
83 201101121243 Hula-Kinneret 33.13 35.49 3 2.9 5.59E+13 3.1 4.07 0.36 0.473 9
84 201101202137 Hula-Kinneret 33.09 35.46 3 3.5 1.83E+14 3.5 3.00 0.48 0.641 10
85 201101311130 E. Mediter.-Sea 33.26 33.57 32 3.3 2.23E+14 3.5 3.89 1.28 0.491 8
86 201103020749 E. Mediter.-Sea 33.80 33.88 32 3.5 3.27E+14 3.6 3.98 2.04 0.484 8
87 201105291916 W.-Sirhan 33.01 35.88 7 3.4 5.58E+13 3.1 4.19 0.42 0.459 12
88 201108070851 E. Mediter.-Sea 32.57 34.47 30 4.2 1.01E+15 3.9 3.37 3.73 0.569 12
89 201112042054 Hula-Kinneret 33.10 35.60 5 3.9 3.94E+14 3.7 2.73 0.81 0.705 11(1)
90 201203220415 Dead Sea Basin 31.31 35.42 23 3.7 1.81E+14 3.5 4.84 1.91 0.398 11(1)
91 201205111847 Cyprus 34.25 34.17 32 5.3 1.25E+17 5.3 1.46 39.26 1.319 10
92 201207101733 Lebanon 33.30 35.37 1 3.7 2.03E+14 3.5 3.07 0.53 0.627 12
93 201208110359 Arava 29.97 35.08 12 3.6 1.14E+14 3.3 4.01 0.67 0.481 11
 18

3. Processing and calculations

3.1 Spectral analysis of seismograms and strong motion data

The analysis involved measuring the low-frequency spectral amplitudes of the

displacement of P-waves, Ω(P) and the observed corner frequency f0(P). These measured
values are used to determine the seismic moment, M0, the moment magnitude, Mw, the stress
drop, ∆σ, in accordance with the Brune’s circular source model (Brune, 1970, 1971). This
analysis was performed by using program JSTAR, developed by the Seismology Division of
GII (Polozov and Pinsky, 2007).Velocity and acceleration records are base-line corrected by
removal of the mean. These and broadband seismograms are band-pass filtered between 0.2
and 12 Hz. Example of records is shown in Figure 4.

P
26 km

P S 49 km

P
65 km

P 119 km

P
132 km

P S
159 km

P S
160 km

P 176 km

P S
194 km

P 232 km

Figure 4. Seismograms of an earthquake, which occurred on 07.09.2004 in the Dead

Sea region with Md=3.7. Also shown are the epicentral distances.
 19

In the case of small or distant events we analyzed only those recordings where the
signal-to-noise ratio exceeds 2. The displacement amplitude spectra are calculated from the
vertical component of the ground motions, starting at the time of arrival of the P-waves. The
low-frequency spectral displacement level Ω and corner frequency f0 are manually measured
from the computed displacement spectrum for each record (examples in Figures 5 and 6).

a)
Short period station MZDA

b)
4.6Hz; 5.0e-07 m*s

Figure 5. (a) Example of a seismogram of an earthquake (07.09.2004, Dead Sea

region, Md=3.7) recorded by the station MZDA. The analyzed time window is marked in
blue; (b) The P-wave displacement spectrum marking the measured level of the low-
frequency spectral amplitude Ω(m*s) and the corner frequency f0(Hz).
 21

Short period station DRGI Broad band station EIL

a b

Figure 6. Displacement spectra (m*s) calculated from seismograms of an earthquake

(07.09.2004, Dead Sea region, Md=3.7) recorded at stations located at distances of 26 km(a)
and 232 km(b) from the epicenter. Also marked are the measured f0 and Ω.

Figure 6 also demonstrates the decrease of the amplitude of the displacement spectrum
and change of the corner frequency when increasing distance. It should be noted that the
displacement spectra are obtained from the records of the short period station (DRGI) where
the analyzed time window is 1.5sec and 3sec in the broad band station (EIL)

In addition to the data obtained from seismic stations of the Israel Seismic Network we
have examined the applicability of similar spectral analysis using data from strong motion
accelerometers. The spectral analysis of accelerometers data was performed by using program
SEISPECT, developed at the Seismology Division of GII (Perelman and Zaslavsky, 2001).
For reason not yet understood some of the accelerograms are inconsistent with measurements
at other stations or missing the P-wave onset. Consequently, only few accelerograms were
included in this study (Figure 2 and Table1). Figures 7, 8 and 9 show examples of P-wave
spectra from accelerometers together with the displacement spectra obtained from
seismometers for comparison.

Figure 7 shows displacement spectra from earthquake on 15.02.2008, Md=5.1 in

Lebanon, obtained by the station Salit and by the accelerometer installed in Petah Tiqva,
located at a distance of 125km and 147 km from epicenter, respectively.
 21

Short period station SLTI Accelerometric station PET

1.57Hz, 3.0e-05 1.6Hz, 2.0e-03

2.0e-3

a b

Figure 7. The P-wave displacement spectra of an earthquake (2008.02.15, Md=5.1,

Lebanon) calculated (a) Spectral displacement in m*s from the vertical component of station
Sal'it and (b) Spectral displacement in cm*s from the vertical component of accelerometer in
the town Petah Tiqva.

Figure 8 demonstrates the P-wave displacement spectra from Beirut’s earthquake that
occurred on 1997.03.26 on the Roum Fault with Md=5.6. Spectra are calculated from the
recording of the broad band station (JER) in Jerusalem (R=233km) and from the
accelerogram from the town Kiryat Shemona (KIT) (R=75km).

Broad band station JER Accelerometric station KIT

a b

Figure 8. The P-wave displacement spectra of the earthquake (1997.03.26, Md=5.6,

Roum Fault), calculated in m*s from the vertical component of the broad band station JER
(a) and in cm*s from the vertical component of the accelerometer at KIT (b).
 22

Short period station MZDA Accelerometric station ARAD

5.48Hz, 6.8e-07 5.8Hz, 8.0e-05

a b

Figure 9. The P-wave displacement spectra of the earthquake (2012.03.22, Md=3.7,

Dead Sea), calculated in m*s from the vertical component of the short period station MZDA
in Mezada (a) and in cm*s from the vertical component of the accelerometer installed in
ARAD (b), located at distances 23 km and 20 km from the epicenter, respectively.

3.2 The time window determination

We studied the effect of the width of the P-waves time window to define the optimal
length of the time window always starting at the P-onset, hence, disregarding the type of the
first arriving P phase. The recommended length of the P-window is different by Different
authors, e.g., Watanabeet al. (1996), Tusa and Gresta (2008), Harrington and Brodsky (2011)
recommend using different windows of the P-wave train. Times vary from 1.02 sec to 2.5sec
for distances 30-250km.

In this study the time window starts from the first P onset and includes the first group
of the P phase (e.g., Pg, Pb, Pn) and excluding S-wave. Examples are shown in Figures 5a, 9,
10, 11 and 12. When the epicentral distance is short, the duration is limited by the arrival of
the S-waves. The length of time window is in a range from 1.5sec to 6.0sec thus increases
with distance, adding 1 sec per each 100km.
 23

a) R=24km

1.5sec

R=42km
b)

1.5sec

Figure 10. Short period vertical seismograms from (a) a local earthquake (2010.19.10,
Md=3.5 in the Arava), recorded by the station ZFRI and (b) a local earthquake (2011.01.01,
Md=4.0 in the Jordan Valley), recoded by the station BLGI. The selected time windows of 1.5
sec each are marked in blue.
 24

a) R=262km

3.0sec

b) R=349km

4.0sec

Figure11 . Short period vertical seismograms from (a) a local earthquakes

(2004.10.28, Md=3.9 in East Sinai), recorded by station SLTI and (b) a local earthquake
(1999.04.11, Md=3.6 in the Hula-Kinneret region), recorded by station KMTI. The selected
time windows of 3.0sec and 4.0sec, respectively, are marked in blue.
 25

a) R=95km

2.0sec

b)
R=520km

6.0sec

Figure 12. The broad band vertical seismograms from (a) a regional earthquake
(2012.07.10, Md=3.7 in Lebanon), recorded by station MMLI and (b) a regional earthquake
(2012.05.11, Md=5.3 in Cyprus), recorded by station EIL. The selected time windows of
2.0sec and 6.0sec, respectfully, are marked in blue.
 26

3.3 Ratio between spectral amplitudes of S- and P- waves

We have investigated the relations between the low frequency spectral displacements
of P- and S-waves, arriving to the same station (Figure 13) and its dependence on distance and
magnitude. Measured low frequency spectral amplitudes of S-waves are copied from
Hofstetter and Ataev (2011).

Figure 13. The relation between the low frequency spectral displacements of S- and P-
waves. Also included is the measurement from the Gulf of Aqaba earthquake (22.11.1995,
Mw=7.2). Marked in red are data of the earthquakes that occurred in the Gulf of Aqaba and
Saudi Arabia region.

As seen in Figure13, spectra of the first P-waves from earthquakes that occurred in
Gulf of Aqaba and the Saudi Arabia region, have, in general, much smaller spectral
displacements, thus yielding higher Ω(S)/Ω(P) ratios. Figure 14 shows the observed
Ω(S)/Ω(P) ratios as a function of the epicentral distance and the duration magnitudes. On
 27

these figures we marked (points in red) the data that corresponds to measurements of
earthquakes located in the Gulf of Aqaba. It is evident that events that occurred in the Gulf
and recorded at stations as far as 150 km or more, show significantly higher Ω(S)/Ω(P) ratios.
All other spectral ratios are roughly of the same order. A careful check shows that the
relatively high Ω(S)/Ω(P) ratios are due to significantly low Ω(P) observations. Evidently, this
anomaly will be observed mainly on seismograms of higher magnitude earthquakes that are
recorded at greater distances.

Figure10 . The ratio between the low frequency spectral displacement from S- and P-
waves for all studied events versus the epicentral distance (a) and duration magnitudes (b).
The red circles mark earthquakes that occurred in the Gulf of Aqaba and the Saudi Arabia
region.
 28

The reason for the low Ω(P) observations is not really clear. We may expect that it is a
propagation path effect. However, it should be further investigated in a separate study.
Consequently, we have excluded Ω(P) observations obtained at epicentral distances greater
than 150 km from earthquakes in the Gulf of Aqaba and neighboring Saudi Arabia.

Based on the above, and as shown in Figure 15, we derive the relationship:

Ω (S)= 3.06*Ω (P)0.95, (11)

Ω (S) = 3.06*Ω (P)0.95

R2=0.92
Number of stations used= 702

Figure 15. The relation between the low frequency spectral displacements of S- and P-
waves. Also shown is the log-log regression line.
 29

3.4 Regional attenuation of the low frequency spectral displacements

To obtain the source spectrum from the displacement spectrum of P-waves, we study
the distance effect on the low frequency spectral amplitude Ω (P). We adhere to the procedure
presented by Shapira and Hofstetter (1993) to determine the coefficients of the equation:

ln(Ωij) = Cj-α*ln(Ri)- δ*Ri, (12)

Cj =a0+a1*Mj (13)

where Ωij –is the low-frequency spectral displacement level at stations i from event j
of magnitude Mj. In that procedure we first used individual earthquakes to determine the best
values of α and δ common to all individual attenuation functions ( see Figure 16) and then,
for these α and δ values, we derived the best values a0 and a1 to yield the minimum standard
deviation of the attenuation function (Eq. 12).

From a total of suggested 93 events we selected a set of 73 earthquakes that are

recorded by at least 6 stations (usually 8-11 stations) having in total-754 measurements.
Shapira and Hofstetter (1993) described the attenuation D of S-wave in terms of geometrical
spreading Rα and the inelastic attenuation ℮δR , i.e.,

D= Rα℮δR, (14)

where R denotes epicentral distance. The updated function for S-waves yield: α 5/6
and δ 3.65*10-3km-1 (Hofstetter and Ataev, 2011)

In this study, using P-waves, we fixed α =1 for the geometrical attenuation and then
estimated δ by searching the lowest sum of squared discrepancy between observed and
predicted Ω(P) values. The obtained optimal value of δ is δ=0.0042.

Figure 16 displays some examples of attenuation curves of the form ln (Ωij)

=a0+a1*Mj-α*ln(Ri)-δ*Ri , where α =1.

In Figure17 we show the distance corrections, D= Rα℮δR versus the distance R, where
α =1 and δ= 0.0042km-1. The regression coefficient is R2=0.87. It should be noted that the
inelastic attenuation with distance is roughly the same for P- and S-waves.

The plot in Figure 18 shows the residuals (ln(observed)-ln(calculated)), for this

regression Eq. 12 against distance.
 31

-8

Md=5.2
Md=5.1

Md=4.5 Md=5.3
-10 Md=4.8
Md=4.6
ln low frequency spectral amplitude 

Md=4.2
-12

Md=4.0
Md=3.2
Md=3.6
-14

Md=3.1

Md=3.0
-16

-18

5 10 20 30 50 100 200 300 500

Distance R, km

Figure 16. Examples of the empirical attenuation functions of Ω(P) for some studied
earthquakes. Also shown are the duration magnitudes.
 31

7
Ln distance correction

1
0 50 100 150 200 250 300 350 400 450 500 550
Distance, km

Figure 17. Distance correction D for the low frequency spectral displacement Ω(P), as
a function of distance R.

1
log residual

-1

5 10 20 50 100 200 500

Distance,km

Figure 18. Residuals of the parameter D=R*℮ 0.0042R against distance, R.

 32

3.5 Distance effect on the measured corner frequency and ratio between corner
frequencies obtained from the analysis of S- and P-waves.

To estimate the corner frequency at the source from the P-wave displacement
spectrum at distance R, we need to estimate the distance scaling parameters. The distance
effect for corner frequency of S-wave was formulated by Shapira and Hofstetter (1993) by the
equation:

f0 f‫׳‬0*℮γR, (15)

where f0 is corrected corner frequency at the source and f‫׳‬0 is the corner frequency at the
station at distance R. For S-wave Hofstetter and Ataev (2011) obtained γ =4.36*10-4 km-1.

The corner frequencies measured from displacement spectra of P-waves on 714

stations that recorded 74 earthquakes, were used in this study. Analysis, applying the search
technique yield γ=4.3*10-4 km-1 (see Figure 19). Again, almost the same scaling as obtained
by Hofstetter and Ataev (2011) from S-waves. The distribution of the ln(f0‫ ׳‬/f0) residuals, as a
function of the epicentral distance is presented in Figure 20.

f0= f4‫׳‬℮4.3*10-4*R
R2=0.67
Number of used stations =714

Figure 19. The ratio between f‫׳‬0 and f0 as a function of epicentral distance, R.
 33

Figure 04 . Residuals of the regression ln (f‫׳‬0 /f0)=0.00043R against the distance R.

Defining the events corner frequency f0 to be the geometrical average of f0 estimates

from available seismograms of the event with standard deviations between 0.05 to 0.6 Hz
and average σ= 0.2 Hz for all 93 events.

Following Hanks and Wyss (1972), Molnar, Tucker and Brune, (1973), later by
Watanabe et al (1996) and by Tusa and Gresta (2008), it is expected that f0 (P) > f0 (S). Hanks
and Wyss (1972) claimed that f0 (P) should be shifted by a factor of α/β relative to f0(S),
where α and β are P-wave and S-wave velocities, respectively but, as they suggested, the
validity of that assumption has to be checked empirically. Our data set, based on 74
earthquakes show an average ratio: f0(P)/ f0(S)=1.24(±0.14). A comparison between corner
frequencies estimated from P- and from S- waves is presented in Fig.21.
6
Corner frequency for P-wave, Hz

93events

1 2 3 4 5 6
Corner frequency for S-wave, Hz
Figure 01 . Corner frequency determined from P-wave analysis, f0 (P) versus f0
determined from S-wave analysis, f0 (S).
 34

4. Source parameters determination and analysis

4.1 Seismic moment and moment magnitude

Seismologists consider the seismic moment M0 and moment magnitude Mw

(Kanamori, 1977) as a more reliable measure of earthquake size that represents the total
deformation at the source and correlates with the seismic energy release. These earthquake
source parameters have been widely used for computing scaling laws exploitable for seismic
hazard assessments. Therefore defining the seismic moment M0 and moment magnitude Mw
and developing a reliable relationship between them and the local magnitude Md is of
fundamental importance. Here we used P-wave spectrum to provide estimates of the seismic
moment M0 and moment magnitude, Mw, and compared these values with those obtained
by Hofstetter and Ataev ( 2011) from the spectra of S-waves..

In agreement with the source model developed by Brune (1970 1971).The seismic
moment, M0 for P-wave is:

M0= Ω (P)4πρDα3/cF, (16)

where Ω(P) is the measured low frequency spectral level of the ground vertical
displacement (m*s), ρ is the density (2700kg/m3); α is the average P-wave velocity (6200
m/s), D=Rα℮δR is the distance correction, c is the free surface amplification, assumed to be
equal to 2, F is the P-wave radiation-pattern correction factor, i.e., F=0.64, (Baumbach and
Bormann, 2012).

Kanamori, (1977) and Hanks and Kanamori (1979) defined the moment magnitude,
Mw, and we followed the same definition:

Mw=2/3(logM0-9.1). (17)

M0 is given in Nm (kg*m2*s-2).

M0 and Mw for each event were estimated by geometrical averaging of all moment
and magnitude determined by the individual stations that recorded the event. The standard
deviation for determining of moment magnitude, Mw was defined for each event, then as
average for all events, σ=0.1.

The seismic moment, M0 and moment magnitude, Mw of 93 earthquakes were

determined and are in ranges 1.06*1013≤ M0 ≤ 1.25*1017 Nm and 2.6≤ Mw ≤5.3 (see Table
1).
 35

In Figure 22 we compare the seismic moment estimated from P-waves, M0(P), and
that from S-waves, M0(S). The good agreement demonstrates the reliability of M0(P)
estimations. The relation between M0(P) and M0(S) is:

log M0(P)=log M0(S)+0.16, σ=0.20.

.
1E+018

1E+017
Seismic moment, M0(P)

1E+016

1E+015

1E+014

93events
1E+013
1E+013 1E+014 1E+015 1E+016 1E+017 1E+018
Seismic moment, M0 (S)

Figure 22. Seismic moment, M0 (P) versus seismic moment M0 (S). The straight line
represents M0 (P) = M0 (S).

Table 2 presents estimates of M0 and Mw obtained in this study from P-wave spectra
analysis, the estimates obtained from waveform inversions (Hofstetter, 2008, Meirova and
Hofstetter, 2011), GFZ CMT solution of the Cypriot earthquake of 11.05.2012, and the
corresponding estimates obtained from S-wave spectral analysis (Hofstetter and Ataev ,
2011). All these estimations are in agreement with each other.
 36

Table 0 . M0 and Mw estimates obtained from waveform inversion and spectral

analysis of P- and S-waves.
Waveform inversion,
Events P-wave S-wave
CMT
Data Region Md M0 Mw M0 Mw M0 Mw
31.12.2003 Dead Sea 3.7 2.17*1014 3.5 3.98*1014 3.7 1.86*1014 3.5
14 15 14
15.03.2004 Dead Sea 4.3 5.81*10 3.8 1.47*10 4.1 6.83*10 3.8
11.02.2004 Dead Sea 5.2 5.94*1016 5.1 5.70*1016 5.1 4.0*1016 5.0
Jordan
07.07.2004 Valley
4.8 3.78*1015 4.4 9.74*1015 4.6 6.0*1015 4.5

15.02.2008 Lebanon 5.1 4.3*1016 5.1 3.7*1016 5.0 2.78*1016 4.9

11.02.2008 Lebanon 4.3 1.76*1015 4.1 2.43*1015 4.2 2.15*1015 4.1
16 17 16
11.05.2012 Cyprus 5.3 6.5*10 5.1 1.25*10 5.3 8.28*10 5.2

Regression analysis is applied to correlate the seismic moment M0 with the duration
magnitude Md, determined for the Israel seismic Network (see Figure 23):

log M0=1.23(±0.05) Md + 9.88(±0.20), σ=0.29 for 2.7 ≤ Md ≤5.6 (18)

log M0=1.01(±0.08) Md + 10.63(±0.30), σ=0.25 for 2.7 ≤ Md ≤4.0 (19)

log M0=1.53(±0.11) Md + 8.51(±0.49), σ=0.31 for 4.0 ≤ Md ≤5.6, (20)

where M0 is in Nm .

In general, these relations are in agreement with those obtained from S-wave
displacement spectra by Hofstetter and Ataev (2011) for 100 local and regional events which
occurred during 1995-2011 in and around Israel, and with those developed previously by Van
Eck and Hofstetter (1989), Shapira and Hofstetter (1993), Hofstetter et al. (1996), Hofstetter
(2003), Hofstetter et.al (2008) for events along the Dead Sea Fault or the Eastern
Mediterranean region.

For whole magnitude range, we obtained the orthogonal polynomial quadratic

equation with R2=0.88 (Figure23):

log M0 =13.16-0.46 Md +0.21Md2 , σ=0.27 for 2.7 ≤ Md ≤ 5.6 (21)

The correlations between moment magnitudes, Mw defined from P-waves and

duration magnitudes, Md copied from the ISN catalogue is shown on Figure 24 and given by
Eq. 22.

Mw = 0.82 (±0.03) Md+0.52(±0.13), σ=0.20 for 2.7 ≤ Md ≤ 5.6 (22)

 37

93event
s

Figure 02 . The relation between M0 and Md for all events used in this study. The red
solid lines represent the linear equations for two magnitude ranges; the green line represents
the polynomial quadratic equation for the whole magnitude range.

Mw=4.80(±4.42)*Md +4.50(±4.12)

R2=0.87 93events

Figure 00 . The relation between Mw obtained from P-waves and Md. The solid line is
the least-squares fit to the data.
 38

4.2 Stress drop

The seismic moment, M0 and the corner frequencies f0 obtained in this study were
used to assess the Brune stress drop, ∆σ (1970, 1971), i. e.:

∆σ=8.47 M0 (f0/ β)³ , (23)

where β is the S-wave velocity, β=3600 m/s, f0 is the correct corner frequency (corrected for
the distance) and divided by a factor of 1.24 (see discussion in paragraph 3.5). The ∆σ is in
Pa (kg*m-1*s-2).

The event's stress drop, ∆σ, given in Table 1 is the geometrical average of individual
determinations of ∆σ by the stations that recorded the event. In Figure 25 we compare the
stress drop estimated from P-wave spectra and that from S-wave spectra (Hofstetter and
Ataev, 2011). We find a good agreement between ∆σ (P) and ∆σ (S) estimates.

100
Stress drop (P), MPa

10

93events
0.1
0.1 1 10 100
Stress drop (S), MPa
Figure 25. P-wave stress drop, ∆σ (P) versus S-wave stress drop, ∆σ (S). The straight
line represents ∆σ (P) = ∆σ (S).

The stress drop, ∆σ of this study vary from 10-1 MPa to 39 MPa and are consistent
with stress drop values determined by Shapira and Hofstetter (1993), by Kanamori et al
(1993), Abercrombie (1995), Mayeda and Walter (1996). Shapira and Hofstetter (1993),
 39

Hofstetter (2003) suggested a constant stress drop of 88-90 bar (8-9MPa) for events in the
range 1015-1016 Nm. The constant stress drop level is reach asymptotically at high magnitude
events. This has been pointed by Hanks and Thatcher (1972, 1973), Kanamori and Anderson
(1975), Ide and Beroza (2001), Kanamori and Brodsky, 2004, Atkinson (1993, 2004),
Imanishi and Ellsworth, 2006, Allman and Shearer (2009). It should be noted that some
studies find several anomalous events with high or low stress drops (Baltay, et.al, 2011) or
observe a systematic increase of stress drop with seismic moment or magnitude (Takahashi et.
al, 2005, Walter et al.,2006, Hofstetter et.al, 2008, Tusa and Gresta, 2008, Drouet, et.al,
2011). Mayeda, et.al (2005) show significant differences in the spectral shapes of main
shocks, larger aftershocks and small events due to the existing of complexities in the rupture
process and suggest that earthquakes scaling are not self-similar.

The variation of the stress drop, ∆σ with the seismic moment, M0 is shown in Figures
26, 27.
100
Stress drop (MPa)

10

93events (this study)

0.1
1E+013 1E+014 1E+015 1E+016 1E+017 1E+018
Seismic moment, M0 (Nm)

Figure 26. The stress drop, ∆σ (P) variation with the seismic moment M0 (P) for all
events used in this study.

Based on the data presented in this study, we may suggest that there is a clear trend of
increasing stress drop with increasing seismic moment in the range 1013-1016 Nm (Hofstetter
and Ataev, 2011). Events with higher seismic moments may reach a stress drop almost
40MPa. This is also demonstrated in Figure 28.
 41

100
Stress drop (MPa)

10

0.1
1E+013 1E+014 1E+015 1E+016 1E+017 1E+018 1E+019 1E+020
Seismic moment, M0 (Nm)
Figure 27. The stress drop, ∆σ versus the seismic moment M0 obtained from P- (red
circles) and S-wave spectra. Also included is the Gulf of Aqaba earthquakes of 1995, (main
shock Mw=7.2) (blue circles).

1E+020

1E+019 10
MP
a
Seismic moment, M0 (Nm)

1M
1E+018 Pa
0.1
MP
1E+017 a

1E+016

1E+015

1E+014

1E+013

1E+012
0.2 0.4 0.6 0.8 2 4 6 8
0.1 1 10
Corner frequency (Hz)
Figure 08 . Relationship between the seismic moment, M0 and the corner frequency,
f0 obtained from P-waves (red circles) and S-waves (blue circles). The solid lines indicate
the expected relation when the stress drop is constant. Note that the corner frequency
divided by a factor of 1.24.
 41

4.3 Source radius and source area

The source radius, r0 was estimated from P-wave records on basis of kinematic
rupture model of Brune (1970, 1971). Baumbach and Borman (1999, 2012) suggested:

r0 (P)=KP* β /2πf0(P) (24)

where KP=3.36, β=3.6 km/s, f0(P) is corner frequency measured on the P-wave
displacement spectrum.

For the analyzed earthquakes, the calculated source radius range from 0.34 km to
1.92 km (see Table 1). Correspondingly, the source areas, πr2, vary from 0.4km2 to 12 km2.
These results are in consistent with values obtained from S-wave investigations (Hofstetter
and Ataev, 2011). Figure 29 shows moment magnitude, Mw(P) versus the source area,
S=πr2 (km2). There is an obvious increase of the source area with increasing Mw.

6
Moment magnitude, Mw

1 10

Source area, km2

Figure 29. Moment magnitude, Mw(P) versus the source are, S=πr2 in km2, where r
is source radius of earthquakes.
 42

5. Discussion and conclusions

We have analyzed 1240 digital records from 93 local and regional earthquakes in the
magnitude range Md=2.7-5.6, occurring in a distance range of 8 km to 550km during 1995-
2012. The main goal of this study was to develop and check the applicability and reliability of
estimating source parameters from the first few seconds of the arriving P-waves to the stations
of the Israel Seismic Network, ISN. The analysis was performed with short period and broad
band vertical component seismograms, and vertical component accelerograms.

The optimal P-wave time window used in the analysis is a function of the epicentral
distance. The time window increase with distance, varying from 1.5sec to 6.0sec, starting at
the onset of the P arrival and include the maximum of the first arriving P phase.

It has been observed that for earthquakes that occurred in the Gulf Aqaba and farther
south in the Saudi Arabia region, the spectral displacement amplitudes, Ω of the P-waves
arriving to stations at epicentral distances greater than 150km, have, in general much lower
values (by a factor of 10 and more) in comparison with the spectral amplitudes of other
earthquakes recorded at the same distances. Consequently, we had to exclude these data from
further analysis.

In a previous study, Hofstetter and Ataev (2011) studied the displacement spectra of S-
waves and thus we obtained an empirical ratio between low frequency spectral displacement
amplitudes of P- and of S-waves:

Ω (S) = 3.06*Ω (P)0.95,

Based on these observations we derived the attenuation of Ω (P) with distance to be of

the form D= R*℮δR, where δ= 0.0042 km-1 i.e., almost the same as previously observed for Ω
(S).

On the P-wave displacement spectra we also measured the corner frequency and
determined the dependence to be of the form f0= f'0*℮γR , where γ=4.3*10-4 km-1, again,
almost identical to the findings of Shapira and Hofstetter (1993) and Hofstetter and Ataev
(2011) for S-wave spectra. Our data show that the corner frequencies measured on P-wave
spectra are 1.24 time greater (on the average) than those measured from S- wave spectra.
 43

Once the distance dependence of Ω and of f0 are determined, following Brune's model
we have computed spectral levels and corner frequencies at the source from which the
average seismic moment and the associated corner frequency of the event define the event's
source parameters M0, Mw, f0, ∆σ, source radius and source area. The determined values for
the analyzed earthquakes are shown in Table 1. These parameters, obtained from spectral
analysis of the first few seconds of the arriving P-waves, are consistent with those obtained
from analyzing S-waves (Hofstetter and Ataev, 2011). It should be noted, however, that the
source parameters determined from analyzing P-waves show a significantly higher scatter.

Empirical correlations between the dynamic source parameters and the duration
magnitudes used by the ISN are:

log M0=1.01(±0.08) Md + 10.63(±0.30), σ=0.25 for 2.7 ≤ Md ≤4.0,

log M0=1.53(±0.11) Md + 8.51(±0.49), σ=0.31 for 4.0 ≤ Md ≤5.6,

log M0 =13.16-0.46 Md +0.21Md2 , σ=0.27 for 2.7 ≤ Md ≤ 5.6

Mw = 0.82 (±0.03) Md+0.52(±0.13), σ=0.20 for 2.7 ≤ Md ≤ 5.6,

where M0 is in Nm.

Again, similar correlations were obtained previously by Shapira and Hofstetter (1993),
Hofstetter (2003) Hofstetter et.al (2008), Hofstetter and Ataev (2011) for events along Dead
Sea Fault and the Eastern Mediterranean region.

Local and regional earthquakes in and around Israel have a wide range of stress drop;
from 10-1 MPa to 39 MPa. The observations show that the stress drop increase with the
increasing seismic moments, it is possible that the stress drop becomes constant at high
magnitude earthquakes, reaching a maximum value of about 40 MPa. This value is more than
4 times higher than previously assumed.

This study confirms that spectral analysis of the first few seconds of the arriving P-
waves to the ISN stations can be used to reliably determine the earthquake source parameters.
We should however pay attention to the observed anomaly regarding earthquakes that occur
in the Gulf of Aqaba and recorded at distances great than 150 km. There might be other
situations which would not follow the empirical relationships described in this report, possibly
due to certain, currently not identified, path effects.

The correlations with Md are of special importance. Md magnitudes of 5 or more start

saturating and become unreliable. Mw magnitudes, as already observed by Hanks and
 44

Kanamori (1979), Hanks and Boore, (1984), Ottemoller and Havskov (2003) and others, are
stable also at high magnitudes. Magnitude determinations, md, of past events can now be
converted into Mw magnitudes that better correlated with seismic hazard parameters.

This study may now open new directions in real time monitoring by the ISN and may
also be integrated in early warning systems.

Acknowledgements

This study was financed by The Ministry of Energy and Water Resources. Special
thanks to Dr. Avi Shapira for the fruitful discussion, support and constructive comments.
Thanks are also due to Andre Polozov and Vladimir Pinsky for help and comments and to
Nahum Perelman for some data processing.

References

Abercrombie, R.E., 1995. Earthquake source scaling relationships from 1 to 5 ML

using seismograms recorded at 2.5 –km depth. J. Geophys. Res., 100, 24015-24036.
Allmann, B., Shearer, P., 2009. Global variations of stress drop for moderate to larger
earthquakes. J. Geophys. Res., 114, B01310.
Atkinson, G. M., 1993. Earthquake source spectra in Eastern North America. Bull.
Seism. Soc. Amer., 83, 1778-1798.
Atkinson, G. M., 2004. Empirical Attenuation of Ground- Motion Spectral Amplitudes
in Southeastern Canada and the Northeastern United States. Bull. Seism. Soc. Amer., 94,
1079-1095.
Baltay, A., Ide, S., Priet, G., Beroza, G., 2011. Variability in earthquake stress drop
and apparent stress. Geophys. Res. Lett., 38, L06303.
Baumbach, M. Borman, P.,1999. Determination of source parameters from seismic
spectra. GeoForschugsZentrum Potsdam. International Training Courses on Seismology and
Seismic Risk assessment for Developing Countries.
Baumbach, M., Borman, P., 2012. Determination of source parameters from seismic
spectra., New Manual of Seismological Observatory practice 2 (NMSOP-2) (pp.1-7).
Potsdam: Deutsches Geo Forschugs Zentrum GFZ.
Begin, Z.B., Steinitz, G, 2005. Temporal and spatial variations of microearthquake
activity along the Dead Sea Fault, 1984-2004. Earth Sciences, 54, 1-14.
 45

Brune, J.N., 1970. Tectonic stress and the spectra of seismic shear waves from
earthquake. J. Geophys. Res., 75, 4997-5009.
Brune, J.N., 1971 Correction. J. Geophys. Res., 76: 5002.
Drouet, S., Bouin, M., Cotton, F.,2011.New moment magnitude scale, evidence of
stress drop magnitude scaling and stochastic ground motion model for the French West
Indies. Geophysical J. Int., 187, 1625-1644.
Hanks, T., Boore, D.,1984. Moment magnitude relations in theory and practice.
J.Geophys.Res.89, 6229-6235.
Hanks, T.and Kanamori, H., 1979. A moment Magnitude Scale. J. Geophys. Res.,
84:2348-2350.
Hanks, T., Thatcher, W.,1972. A graphical representation of seismic source
parameters. J. Geophys. Res. 77, 4393-4405.
Hanks, T and Wyss, M.,1972. The use of body-wave spectra in the determination of
seismic-source parameters. Bul. Seism. Soc. Amer., 62, 2, 561-589.
Harrington, R., Brodsky, E., 2011. Energy and Rupture Dynamics are Different for
Earthquakes on Mature Faults vs. Immature Faults. University of California, Dept. Earth
Sciences.
Hofstetter A., Ataev G., 2011. Re-examination of correlation coefficients of
earthquake source parameter, GII, Report No 556/638/11.
Hofstetter, T. van Eck, Shapira, A., Shapira, A., 1996. Seismic activity along fault
branches of the Dead Sea-Jordan transform system: The Carmel-Tirtza fault system,
Tectonophysics, 267: 317-330.
Hofstetter, A., 2003. Seismic observations of the 22/11/1995 Gulf of Aqaba
earthquake sequence. Tectonophysics, 369: 21-36.
Hofstetter, A., Gitterman, Y., Pinsky, V., Kraeva, N., Feldman, L., 2008.
Seismological observation of the northern Dead Sea basin earthquake on 11 February 2004
and its associated activity, Isr. J. Earth-Sci,57, 2, 101-124.
Ide, S. Beroza, G., 2001. Does apparent stress vary with earthquake size? Geophys.
Res. Lett., 28, No.17, 3349-3352.
Imanishi, K., and Ellsworth, W., 2006. Source scaling relationships of
microearthquakes at Parkfield, CA determined using the SAFOD pilot hole array. In
Earthquakes: Radiated Energy and the Physics of earthquake Faulting, R. Abercrombie, A
McGarr, H. Kanamori and G. Toro. American Geophysical Monograph 170, 81-90.
Kanamori, H., 1977. The energy release in great earthquakes. J. Geophys. Res., 82:
2981-2987.
Kanamori, H., Anderson, D.,1975. Theoretical basis of some empirical relations in
seismology. Bull. Seism. Soc. Amer., 65, 1073-1095.
Kanamori, H., Mori, J. Hauksson, E., Heaton, T., Hutton, L., Jones, L., 1993.
Determination of earthquake energy-release and M (L) using Terrascope. Bull. Seism. Soc.
America, 83, 330-346.
Kanamori, H. and Brodsky E., 2004. The physics of earthquakes, Report On Progress
in Physics, 67, 1429-1496.
 46

Mayeda, K., and Walter, W., 1996. Moment, energy, stress drop, andsourse spectra of
western United States earthquakes from regional coda envelopes. J. Geophys. Res., 101: 195-
208.
Mayeda, K., Gok, R., Walter, W., Hofstetter, A., 2005. Evidence for non-constant
energy/moment scaling from coda- derived source spectra. Geophys. Res. Lett., 32, L1036.
Meirova T., Hofstetter R., 2011. Observations of seismic activity in Southern
Lebanon. Annual Meeting, Geological Society of Israel, Ashkelon, 2011.
Molnar, P., Tucker, B., Brune, J., 1973. Corner frequencies of P and S waves models
of earthquake sources. Bull. Seism. Soc. Amer., 63, 2091-2104.
Ottemoller, L., Havskov, J., 2003. Moment Magnitude Determination for Local and
Regional Earthquakes Based on Source spectra. Bull. Seism. Soc. Amer., 93, 1:203-214.
Perelman, N.; Zaslavsky, Y. Analysis of seismic signals in spectral domain,
(SEISPECT). GII Report 569/345/01. 2001.
Polozov, A., Pinsky, V., 2007. New software for seismic network and array data
processing and joint seismological data base. GII report 546/307/07.
Shapira, A., 1988. Magnitude scales for regional earthquakes monitored in Israel. Isr.
J. Earth-Sci, 37, 17-22.
Shapira, A. and van Eck, T., 1993.Synthetic uniform hazard site specific response
spectrum, Natural Hazard, 8: 201-205.
Shapira, A., and Hofstetter, A., 1993. Source parameters and scaling relationship of
earthquakes in Israel, Teconophysics: 217-226.
Takahashi, T., Sato, H., Ohtake, M., Obara, K.,2005. Scale dependence of apparent
stress for earthquake along the subducting pacific plate in northeastern Honshu, Japan. Bul.
Seism. Soc. Amer., 95, 1334-1345.
Tusa, G and Gresta S., 2008. Frequency-Dependent Attenuation of P waves and
Estimation of Earthquake Source Parameters in Southeastern Sicily, Italy. Bul. Seis. Soc.
Amer., 98, 6, 2772-2794.
Van Eck, T. and Hofstetter, A., 1989. Microearthquake activity in the Dead Sea
region. Geophys. J. Int., 99: 605-620.
Van Eck, T. and Hofstetter, A., 1990. Fault geometry and spatial clustering of
microearthquake along the Dead Sea-Jordan rift fault zone. Tectonophysics, 180:15-27
Walter, W. Mayeda, R., Gok, R. and Hofstetter, A., 2006. The scaling of seismic
energy with moment: Simple models compared with observations, in Earthquakes: Radiated
Energy and the Physics of Faulting, Geophys. Monogr. Ser., 170, 25-41.
Watanabe, K., Sato, H., Kinoshita, S, Ohtake, M.,1996. Source characteristics of
Small to Moderate Earthquakes in the Kanto Region, Japan: Application of a New Definition
of the S-Wave Time Window Length. Bull. Seism. Soc. Amer., 86, 5:1284-1291.
 47

Publication No ES_11_2012 2. 3. Recipient Accession No.

4. Title and Subtitle 5. Publication Date
The use of P-wave spectra in the December, 2012
determination of earthquake source parameters 6. Performing Organiz. Code
in Israel.
7. Author 8. Performing Organiz. Rep. No
Dr Hofstetter A., Ataev G. 559/701/12

9. Performing Organization Name and 10. Project / Task / Work Unit No.
Address
The Geophysical Institute of Israel
P.O.B. 182, Lod 71100, Israel 11. Contract.
211-17-004
12. Sponsoring Organization(s) Name and 13. Type of report and period covered
Address
Office of the Chief Scientist
The Ministry of Energy and Water Resources 14. Sponsoring Organiz. Code

15. Supplementary Notes

16. Abstract (Limit 200 Words)
The source parameters were estimated from P-wave spectra using Brune’s seismic
source model (Brune 1970,1971).We have analyzed 1240 digital records from 93 local
and regional earthquakes in the magnitude range Md=2.7-5.6, occurring in a distance
range of 8 km to 550 km during 1995-2012. We obtained a distance correction parameter,
D=R*℮0.0042R for the low frequency amplitude and f0 f'0*℮0.00043R for corner frequency,
almost identical to the findings of Shapira and Hofstetter (1993) and Hofstetter and Ataev
(2011) for S-wave spectra. This attenuation model was used to determine the source
spectrum of the ground displacement, the seismic moment, M0, moment magnitude, Mw,
source radius, r0, and stress drop, ∆σ for each earthquake.
The analyzed earthquakes have the seismic moment between 1.06*1013 and
1.25*1017 (Nm), moment magnitudes- 2.6-5.3, the corner frequencies- 1.0-5.6Hz, and
stress drop between 10-1MPa and 39MPa. The observations show that the stress drop
increase with the increasing seismic moments, it is possible that the stress drop becomes
constant at high magnitude earthquakes, reaching a maximum value of about 40MPa. This
study confirms that spectral analysis of the first few seconds of the arriving P-waves to the
ISN stations can be used to reliably determine the earthquake source parameters.

17. Keywords
Earthquake, seismograms, P-wave, S-wave, source parameters, distance
corrections, seismic moment, stress drop, empirical correlations.
‫ צניחת‬,‫ מומנט סיסמי‬,P ‫ אמפליטודה ספקטרית של גלי‬,‫ פרמטרים של המקור‬, ‫רעידות אדמה‬
. ‫ סיכון סיסמי‬, ‫ בחינה סטטיסטית‬,‫המאמץ‬
18. Availability Statement 19. Security Class 20. Security Class
The Geophysical Institute of Israel, the (This Report) (This Page)
library. 21. No of Pages 22. Price
46 79,090 NIS