See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/226901483

The 1992 Cairo earthquake: A case study of a small destructive event

Article in Journal of Seismology · December 1998

DOI: 10.1023/A:1009717023043

CITATIONS READS

17 3,428

3 authors, including:

Ronald Arvidsson Ota Kulhánek

Geological Survey of Sweden Uppsala University
38 PUBLICATIONS 1,251 CITATIONS 63 PUBLICATIONS 1,161 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Seismology View project

MinLand H2020 View project

All content following this page was uploaded by Ronald Arvidsson on 04 March 2016.

The user has requested enhancement of the downloaded file.

 Journal of Seismology 2: 293–302, 1998.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.
293

The 1992 Cairo earthquake: A case study of a small destructive event

Attia El-Sayed, Ronald Arvidsson & Ota Kulhánek

Seismology, Department of Earth Sciences, Uppsala University, Villavagen 16, S-752 36 Uppsala, Sweden,
e-mail: [email protected]

Received 28 January 1997; accepted in revised form 5 May 1998

Key words: Egypt, Cairo, Red Sea, synthetic seismogram, focal mechanism, seismic hazard, site effect and peak
ground acceleration

Abstract
On 12 October 1992, an earthquake, magnitude mb = 5.9 and Ms = 5.2, hit the City of Cairo, Egypt. It was this
century’s largest earthquake in northern Egypt with related destruction in the City of Cairo, the Nile Valley and the
Nile Delta areas.
Our source parameter determinations show that the 1992 earthquake had a normal faulting mechanism, seismic
moment Mo = 5.2 × 1017 Nm, centroid depth of 23 km and a source time function duration of 3 seconds. The
mechanism is compared with those corresponding to two other events that occurred in the northern Red Sea. The
similarity between the mechanisms as well as the spatial distribution of the geological faults around Cairo suggest
seismic activity along the extension of the stress field of the Red Sea rift system to the area around the City of
Cairo. This situation affects the level of seismic hazard in the Cairo area.
The 1992 earthquake belongs to an unusual class of relatively small, Mw < 6.0, yet destructive earthquakes.
The damage caused by these events is usually attributed to their shallow focal depth, ≤ 5 km, and to amplification
of seismic waves in the local soil beneath the damaged structures. However, the Cairo earthquake deviates from
other events of this class since the focal depth was determined to be 23 km. We calculated synthetic accelerograms
for the 1992 earthquake with the loose sediments observed in the Nile Valley, and show that this enhanced the
amplitude of the acceleration by a factor of two. However, the determined accelerations, about 0.5 m s−2 , cannot
alone explain the relatively large amount of damaged structures. Hence, a major cause to the destruction is likely
the poor state of construction of the Cairo buildings.

Introduction of the earthquake to densely populated areas, (2) local

site effects, i.e., the possible amplification of ground
shaking due to the loose sedimentary layer that fills the
On 12 October 1992, an earthquake, with magni-
Nile Valley graben and overlays the bedrock, (3) poor
tude mb = 5.9, Ms = 5.2 by PDE (Preliminary Deter-
quality of construction and (4) poor preparedness, i.e.,
mination of Earthquake), ML = 5.3 by HLW (Helwan
the majority loss of life was caused by panic.
observatory), occurred beneath south-western Cairo,
According to PDE and reports from Helwan Obser-
northern Egypt. It was the largest earthquake in north-
vatory, the earthquake was felt in large parts of Egypt,
ern Egypt in the past hundred years. Numerous after-
from Aswan to Alexandria (Figure 1a), and practically
shocks with magnitudes of up to 4.5 were recorded
all over Israel. Maximum damages were reported from
(Ibrahem, 1995).
Cairo, El-Giza and El-Faiyum which lie near the epi-
In spite of its relatively moderate magnitude, dam-
centre and within the Nile Valley graben. As published
age caused by this earthquake was large. A prelimi-
in Al-Aharam (13–20 October 1992), Al-Gomoheria
nary report listed total losses in the range of one billion
(13–26 October 1992) and PDE, at least 541 people
US$. Factors which most likely contributed to the ex-
were killed and about 6 500 were injured.
tensive destruction were: (1) proximity of the focus

Article: jose33 Pips nr. 180045 (josekap:bio2fam) v.1.1

jose33.tex; 2/12/1998; 0:50; p.1
294

Totally, 8500 buildings were damaged in the Cairo find a reasonable explanation to the destructive cause
area alone. An official investigation revealed that of the event.
1343 schools were damaged beyond any repair, 2544
needed major repair and 2248 needed repair mainte-
nance. On the other hand, ancient historical monu- Data
ments suffered only minor damages, e.g., two stone
slabs fell from the roof of Kom Ombo temple, six More than 100 stations responded to our request
Coptic churches and 140 Islamic buildings, including and contributed to data collection with seismograms,
mosques, suffered some damage. One stone, 1500 kg and/or bulletins. The majority of these stations pro-
of weight, moved and fell down from the Khefren vided copies of three-component, short- and long-
pyramid. period, analog and/or digital records. The distribution
A number of surface distortions associated with of contributing stations around the epicentre provided
the shock were noticed in the area just after the earth- a good coverage from all directions except the az-
quake. The most common feature was surface fissures imuth range between 180 to 270 degrees. This gap
observed at epicentral distances of up to tens of kilo- corresponds to northern and central African countries
meters with displacements of up to about 150 cm where only few stations are working.
(Badawi and Mourad, 1994) visible in the asphalt Polarities of first P-onsets were identified by
roads between Cairo and El-Faiyum. present authors in most of the available seismograms.
The Cairo earthquake belongs to a rare group of Some of the records reveal a double P-pulse with less
relatively weak earthquakes, 5 < m < 6, which caused than 1 sec delay. In some cases this phenomenon in-
surprisingly severe destruction within the respective troduces difficulties in the polarity reading. Since a
epicentral area. Other shocks with similar character- large number of records are available, only sharp clear
istics are, e.g., Skopje 1956 (Poceski, 1963), Agadir polarities were picked. Less pronounced onsets were
1960 (Duvergé, 1969), and San Salvador 1986. All ignored. Altogether, 43 polarities, well distributed
these shocks were shallow crustal events taking place around the source, were considered.
in areas with unconsolidated sedimentary covers. Waveforms of digital records usually show good
As follows from available geological maps quality, only a few stations have high seismic noise.
(Egyptian Geological Survey, 1981) and the geolog- Analog seismograms, especially from nearby stations,
ical study of Said (1981), the epicentral area of the caused some problems, e.g., high skew and truncated
1992 event is covered by Miocene-Oligocene-Eocene waveforms which introduced certain difficulties in the
sediments of thickness of about 900 m (El-Gamili, data analysis. In total, 9 stations with good-quality
1982). As yet, no surface tectonic lineaments have records have been used in the waveform analysis. For
been observed in the epicentral area (El-Shazly et al., teleseismic body wave inversion, digital data from the
1980), probably due to the thick loose sedimentary world wide networks were received from IRIS data
cover. However, on both sides of and close to the Nile centre, Seattle, Washington, U.S.A.
river, a large number of NW–SE and NE–SW faults
have been observed (Figure 1b). Trends of the new
observed lineaments (Arab et al., 1994; Badawi and Hypocentre location
Maurad, 1994) are in good agreement with the surface
structure in the surrounding areas (Figure 1b). The As mentioned above, the geology and tectonics of
NW–SE and NE–SW faults are assumed to be part of, Egypt is complicated. The crustal structure is rather
respectively, the Red Sea rifting system and of the Gulf heterogeneous and could change abruptly from area
of Aqaba-Levant system (Bayoumi, 1983; Hempton, to area, especially near the Nile Valley. Within the
1987). These two systems are considered to be the Nile Valley graben, loose sediments, e.g., silt and
main factors behind the seismicity of Egypt (Kulhánek sand are thicker (900–1000 m) and more water satu-
et al., 1992). rated than outside the graben where they are almost
The main objectives of the present work are the fol- absent (thickness of 1–2 m). As one departs from
lowing: To deduce the source parameters (mechanism, the graben, water saturation and thickness of the
source time function, seismic moment, focal depth) loose sedimentary layers decrease rapidly (El-Gamili,
for the 1992 Cairo earthquake; to suggest a plausible 1982). Consequently, dramatic changes in the velocity
tectonic model for the occurrence of the event; and to of seismic waves may be expected. Velocity models

jose33.tex; 2/12/1998; 0:50; p.2

 295

Figure 1. (a) Distribution of seismic stations (crosses) used in locating the 12 October 1992 earthquake (open circle). See the insert (Figure 1b)
for details. (b) Fault pattern around the City of Cairo, main faults (solid line), secondary faults (dashed line), solid circle – Cairo earthquake of
12 October 1992. The surface is covered by quaternary deposits, except for the doted area where Tertiary – Cretaceous sediment are represented.
Collected and modified from El-Shazly (1980) and Arab et al. (1994).

jose33.tex; 2/12/1998; 0:50; p.3

296
Table 1. Crustal model used in this study

Layer Layer P velocity S velocity Density

Number thickness km sec−1 km sec−1 g cm−3

1 0–1 km 3.30 1.90 1.9

2 1–2 km 4.50 2.57 2.2
3 2–10 km 6.00 3.40 2.53
4 10–30 km 6.35 3.60 2.64
5 30– km 8.00 4.57 2.95
Layer 1 and 2 correspond to Eonile Canyon deposits which were
inserted to replace the uppermost 2 km in the first layer in the original
profile of Gaulier et al. (1988).

deduced from various seismic profiles surrounding the Earthquake source parameters
epicentral region of the 1992 Cairo earthquake were
examined to find the appropriate crustal structure to The earthquake source parameters, focal mechanism,
be used in location and waveform modelling. A crustal seismic moment, centroid depth and source time func-
model modified from profile number 7 (Gaulier et al., tion were determined through a three-step procedure.
1988) for the Gulf of Suez and adjacent area was P-polarities and forward modelling were used to con-
used (Table 1). This profile agrees with the results strain the mechanism style, whereas the inversion of
given by Hoseny (1985). A minor modification for teleseismic body waves yielded the final results, in-
the uppermost sedimentary layer was made by making cluding seismic moment, source time function and
use of results from gravity and seismic studies of the earthquake centroid depth.
Nile Valley (El-Gamili, 1982; Said, 1981). In profile
number 7, the first layer (layer 3 in Table 1) extends Polarity mechanism
from 0 to 10 km. Our modification was made by in-
serting two layers (first and second layer in Table 1), To determine the fault-plane solution from P-wave
corresponding to the Eonile Canyon deposits (Said, polarity readings, a modified version of the program
1981) beneath the Cairo area. We are fully aware of FOCMEC (Snoke et al., 1984) was employed. Polari-
the fact that, by using such a distant profile (profile 7 ties were weighted as full weight for sharp polarities
mentioned above), an error could appear in our conclu- and half weight for emergent ones. 43 azimuthally
sions. However, profile number 7 is the closest profile well distributed polarities were used to derive the
to the epicentral area currently available. mechanism for the Cairo earthquake. 40 correspond
The 1992 earthquake has been located by using the to sharp polarities and were given full weight and 3 to
program HYPOINVERSE (Klein, 1978) which makes emergent polarities were given half weight. Under the
use of the first P and/or S arrival times as well as condition of the smallest polarity error, two families
of the crustal model (Table 1) of the studied area. of solutions were obtained (Figure 2), one showing a
Arrival-time readings are firstly weighted according normal faulting mechanism with a strike-slip compo-
to the record quality, then the program uses an auto- nent and the other indicating pure strike slip. In terms
matic weighting with respect to the epicentral distance of stress axes, the two families reveal that the T- and P-
and travel-time residual. Arrival-time data from sta- axes are oriented approximately SSW–NNE and E–W,
tions deployed in Egypt, Israel and Saudi Arabia are respectively.
available (Figure 1a). Altogether, 21 P-arrivals as well
as 5 S-arrivals have been used in locating the earth- Forward modelling
quake. From our calculation the hypocenter is located
at 29.79◦ N and 30.68◦ E at a depth of 20 km. This To resolve the ambiguity in the P-wave polarity mech-
location is in fairly good agreement with those given anism, forward modelling was performed for waves
by PDE and Helwan. recorded at regional distance of up to 16◦ . The synthet-
ics were calculated by using a discretized wavenumber
(Bouchon, 1980) propagator matrix wavenumber in

jose33.tex; 2/12/1998; 0:50; p.4

 297

Figure 2. Lower-hemisphere projections of the focal sphere for the 1992 Cairo earthquake. (1): distribution of possible nodal planes from
FOCMEC solutions, (2): corresponding maximum (P) and minimum (T) deviatoric stress axes, (3): distribution of P-wave polarities; dilatation
denoted by rectangulars and compression by crosses. Small crosses refer to polarities taken from ISC. (4): preferred polarity solution.

tegration scheme (Herrmann and Wang, 1985) which

is valid for distances of up to 2000 km. The source,
was assumed to be a parabolic pulse with a duration
of 3 seconds. Observed seismograms were converted
to ground velocities by removal of the instrumental
response in the frequency domain. A frequency pass
band of 0.025–0.01 Hz was applied to clean the wave-
forms from possible influences of 3-D heterogeneities
in the higher frequencies. The main emphasis was put
onto modelling of Love and Rayleigh waves, i.e., the
horizontal components were rotated into correspond-
ing radial and transversal directions. However, the
long-period body-waves were considered as well.
The structural model we used was PREM upper
mantle with the aforementioned local crustal struc-
ture. There were only small deviations in the phase
velocity from this model when we compared observed
data with synthetics. Altogether, eight long-period
records from RYD (Riyadh), ANTO (Ankara) and
Figure 3. Observed (solid line) and synthetic (dashed line) seismo-
ATH (Athens) were employed which allowed a plausi- grams of the 1992 earthquake at the ANTO (Ankara) station at an
ble modelling of the source. The analog seismograms epicetral distance of about 1040 km. The synthetic seismograms are
from RYD and ATH were digitized and converted derived through the propagator matrix method. Z – Vertical com-
from unequal to equal sampling rate of 1 Hz. ponent, R – Radial component, T – Transverse component, P – P
wave, R – Rayleigh wave, L – Love wave.
Different focal depths were examined but, due to
the long period band used, we were not able to de-
rive an unambiguous focal depth. We tested different Teleseismic body-wave inversion
likely solutions deduced from the polarity approach
(Figure 2) and found that the normal-faulting style We made use of teleseismic P and SH waves to deduce
derived from polarity readings is to be preferred be- the earthquake source parameters, final focal mecha-
fore the strike-slip mechanism. The Harvard solution nism, moment tensor, centroid depth and source time
(Dziewonski et al., 1993) is in good agreement with function, by employing a body-wave inversion scheme
our result. Even though there is a discrepancy between (Nabelek, 1984, McCaffrey et al., 1991). The P and
the relative amplitudes of the Love and Rayleigh SH waves were derived from long-period digital seis-
waves at ATH, probably due to the proximity of mograms that were retrieved from the IRIS data centre
the station to the nodal plane in the radiation pat- in Seattle, Washington. Altogether, 14 seismogram
tern, the overall correlation between observed and components were used. This type of inversion can be
synthetic seismograms was good. As an example, Fig- dependent upon starting source parameters. However,
ure 3 shows observed and synthetic traces for ANTO by making use of the mechanism derived in the pre-
station. vious section we were able to achieve a stable result

jose33.tex; 2/12/1998; 0:50; p.5

298
Table 2. Fault parameters of the northern Red Sea events of 1969 (1) and of 1972 (2) and of the Cairo
earthquake of 1992 (3)

Event Date Location Focal Fault plane P-axis T-axis

Lat Lon. Depth parameters t p t p
N E km s d r

1 19690331 27.61◦ 33.91◦ 6 294 37 271 113 52 221 13

2 19720628 27.70 33.80 6 288 40 260 75 81 205 05
3 19921012 29.79 30.68 23 159 50 303 117 65 26 0
∗ " 29.74 30.63 22 136 42 –75 145 80 36 3
∗∗ " 29.78 31.14 22 190 62 –53 150 56 254 10
Parameters of events 1 and 2 are from Huang and Solomon (1987). 3 is our solution, ∗ and ∗∗ are from
the Harvard group (Dziewonski et al., 1993) and NEIC, respectively. s = strike, d = dip, r = rake, t = trend
and p = plunge.

(Table 2). The derived fault plane solution (Figure 4) are presently no observations of ground acceleration
is close to that presented by Harvard and NEIC (Ta- available from Egypt. Therefore, ground acceleration
ble 2). Our moment, 5.2 × 1017 Nm, is within 10% of values given for different areas in Egypt are only esti-
the Harvard CMT moment. In general, the synthetic mates calculated by using various empirical formulas
P waves fit the data better than the SH waves which (Maamoun et al., 1984). These formulas often pay
is a common observation when inverting teleseismic small attention to local geologic conditions, in spite
body phases. The derived source time function of 3 s of the fact that these strongly affect the acceleration.
duration and the centroid depth of 23 ± 3 km satisfy Hence, estimates of ground acceleration at a specific
the observed data well. The depth was well resolved site made by using an empirical formula and at the
due to the clear separation between the direct body same time ignoring local conditions can easily lead to
waves and the pP, sP and sS phases. This makes the large errors. It has been shown that loose sediments
source time function estimate more reliable since the indeed amplify seismic waves (Aki, 1982). We inves-
tradeoff which occurs when the depth phases interfere tigated specifically how the soft sedimentary deposits
with direct phases is smaller when pP, sP and direct in the Nile Valley affect the peak ground acceleration
P and S phases are separated. A difference between associated with the Cairo earthquake.
the observed and synthetic SH seismograms is obvious We used the same technique as described in mod-
for the Harvard station, HRV. It can be explained by a elling regional seismograms above to synthetisize ac-
proximity to the node of the radiation pattern as well celerograms to investigate the general effect of the
as by heterogeneities of the Earth structure. known sedimentary top layers. The accelerograms
In general, source parameters deduced through two were synthesized using our fault orientation (Table 2),
different modelling techniques are in good agreement source size, source time function, source depth and
with the available Harvard solution and also compare crustal model parameters.
well with those of the northern Red Sea quakes of The effect of the loose sedimentary cover was
1969 and 1972 (Table 2). determined by comparing our results considering sed-
imentary cover, (i.e., layer, 1 in Table 1), and by
ignoring this layer. The computed synthetics indicate
Ground acceleration that the presence of the soft soil cover amplifies the
ground acceleration almost by a factor of 2. How-
The destruction in Cairo may have been due too poor ever, the obtained accelerations, about 0.5 m s−2 , were
quality of a part of the buildings in northern Egypt but too small to explain the destruction. The cause of
obviously the duration and peak ground acceleration the destruction is probably to be found in the state
also have a certain effect. Peak ground acceleration is of the damaged buildings themselves, even though
one of the most important factors to consider in the we should not discount effects from a more complex
design of structures like dams, nuclear power plants, 3-dimensional crustal structure as well as local soil
buildings, etc. To the best of our knowledge, there conditions beneath a given construction.

jose33.tex; 2/12/1998; 0:50; p.6

 299

Figure 4. Top: Lower hemisphere projection of teleseismic inversion focal mechanism. Observed (solid line) and synthetic (dashed line) P-wave
seismograms of the 1992 earthquake, Mo = 5.2 × 1017 Nm, h = 23 ± 3 km. STF – source time function. 50 s – time scale of seismograms in
seconds. Station codes are given to the left of seismograms, d – dilatation. Bottom: Same as top, except for SH waves.

Discussion dicting this view. Using data from Ambraseys et al.

(1994) we find that the area has a history of being re-
The Cairo earthquake of 12 October 1992 and its loca- peatedly shaken by similar-size earthquakes (Table 3).
tion in particular, came as a surprise for the Egyptian The last reported destructive earthquake occurred in
geological community. It occurred in an area con- 1920. Using data from Ambraseys et al. (1994, see
sidered stable with an apparently simple geological Table 3) a mean return period of about 80 years for
setting and with no observable active surface tectonics. destructive earthquakes is indicated for the time period
However, historical seismological evidence is contra- since 1754.

jose33.tex; 2/12/1998; 0:50; p.7

300
Table 3. Historical seismicity in the El-Faiyum-Cairo area as reported in historical Arab docu-
ments

Date Location MF Intensity Remarks

Y M D Lat. Lon. MSK
Scale

885 11 06 30.0◦ N 31.2◦ E – VII 1000 people were killed and

many houses and mosques
have collapsed. (Ambraseys,
et al., 1994).

950 07 25 30.0 31.2 VII-IX Most of the houses in old

Cairo were destroyed and a
portion of the old mosque fell
down.

1111 05 26 31.0 31.0 – VII Church of St Michael in the

island of Ruda, on the orders
of the wazir al-Afdal was de-
stroyed.

1264 03 20 29.0 31.0 VII-IX Many houses were destroyed

in a number of places.

1754 10 29.6 32.2 6.6 VII-IX A great loss of life was re-
ported. Qarafa, Bulaq and
part of new Cairo were de-
stroyed.

1847 08 07 29.7 30.8 5.8 VII-VIII 274 people killed and thou-
sands injured and heavy
wide destruction were re-
ported practically on all
places.

1920 10 01 29.5 31.3 4.9 VI Three houses were destroyed

in El-Faiyum.
Destruction reports, MF and intensity were taken from Ambraseys et al. (1994). These are in fairly
good agreement with parameters found in the other sources mentioned in the text. MF corresponds
to surface-wave magnitude.

shape. We should, however, not discount the local

However, the area has also been suffering damage surface geology which may amplify the ground mo-
from distant earthquakes. The latest such earthquakes tion values, since the City of Cairo is located partially
are the events of 1995 (Gulf of Aqaba) and of 1996 on the tip of Nile Delta graben and partially within
(Cyprus). A review of reports given by Ambraseys the major Nile Valley graben. According to El-Gamili
et al. (1994) shows that for the majority of the large (1982) the width of the Nile Valley graben is up to
earthquakes which occurred in Egypt or in adjacent several kilometers and filled by sedimentary materials.
areas, the City of Cairo was affected more than other The upper most part is silty to fine-sand water satu-
sites in Egypt. This, together with the non-extreme rated in some areas and sandy partially saturated in
acceleration values calculated above, suggests that a other areas. When this type of soil is strongly shaken,
large part of the Cairo building stock is in very poor it is highly susceptible to the liquefaction phenomena.

jose33.tex; 2/12/1998; 0:50; p.8

 301

Destruction due to liquefaction could be observed in ments with no surface lineaments. Nevertheless, near
the Nile flood plain areas (Badawi and Mourad, 1994). to the 1992 epicentre, NW–SE and NE–SW fault sys-
Earthquake damage in northern Egypt with con- tems (Figure 1b) were reported by El-Shazly et al.
centrations in the City of Cairo, Nile Valley and Nile (1980). These are probably due to the complexity of
Delta has been reported from earthquakes which oc- the region which is affected by the northern Red sea
curred as far as Turkey, Gulf of Aqaba, Cyprus or rifting system and by the north-east trending systems
Crete. The most recent example of severe destruc- such as the Gulf of Aqaba-Levant fault system or the
tion in Cairo city due to a distant earthquake is the eastern Mediterranean – Cairo – Faiyum system. It is
case of Mw = 7.2, Gulf of Aqaba, 22 November 1995 possible that the stress fields of the two systems affect
earthquake and the Mw = 6.8, Cyprus, 9 October 1996, north-eastern Egypt (Kulhánek et al., 1992; El-Sayed
earthquake. These earthquakes were located at dis- et al., 1996). The proposed normal-faulting mech-
tances of about 400 and 600 km, respectively, from anism and the NW–SE oriented strike of the Cairo
the City of Cairo. Even if these earthquakes caused earthquake imply that the event occurred along the ex-
relatively minor damages, this nevertheless indicates tension of the stress field of the Red Sea rifting system.
that at least some of the buildings in the Cairo are in A comparison between the mechanism of this event
a poor state since normally destruction would not be and the northern Red Sea events of 1969 and 1972
observed at these distances. shows a high degree of similarity. These similarities
The 1992 earthquake was followed by a large num- also support the proposed location of the 1992 event,
ber of aftershocks but no foreshock was recorded. i.e., along the extension of the Red Sea fault system
The absence of microseismic activity before the Cairo (see Figure 1b). As shown in Table 2, the only differ-
earthquake as well as the low b-value (approximately ence among the three mechanisms is an increase of the
0.7) given for the area by El-Sayed and Wahlström strike-slip component from the 1969 event to the event
(1996) manifest a period of seismic quiescence and of 1992, i.e., a likely decrease of the effect of the Red
stress accumulation before a major earthquake. In Sea rift stress field as we depart from the rifting axis.
view of the historical seismicity (Table 3), the maxi- The Cairo earthquake of 12 October 1992 rings an
mum experienced earthquake magnitude is close to 6.6 alarm for the future urban plans in Egypt. To mitigate
(see Ambraseys et al., 1994). Based on the historical the effects of similar future disastrous earthquakes,
and recent seismic activity, the Cairo area will proba- seismic risk factors should be considered. In the fur-
bly experience another damaging shock in the coming ther planning, one should also realize that building
hundred years. codes are different for different localities in Egypt.
The polarity mechanism exhibited in Figure 2 Economically speaking, it is cheaper to use suitable
shows two families of solutions. Both forward mod- building codes for different parts of the country, which
elling of regional waves as well as teleseismic body- will become available through detailed studies of local
wave inversion (Figures 3 and 4) support the normal- geology.
faulting mechanism. This solution is similar to that
given by the Harvard group and by PDE (Table 2).
Field observations, such as displacements in the cov-
ering sediments and roads also indicate the normal-
faulting mechanism (Badawi and Mourad, 1994). Acknowledgements
However, from a large-scale tectonic point of view, re-
verse or strike slip faulting is the predicted style for the
predominantly compressional stress regime in Egypt This work was carried out at the Seismological De-
(Mousa, 1989). Nevertheless, this conclusion may be partment, Uppsala University, Sweden. Financial sup-
an oversimplification of the true situation. The tec- port was provided partly by the Swedish Institute,
tonic stress regime in north-eastern Egypt is influenced University of Uppsala, Swedish Foundation for In-
by an interaction of at least three plates, namely the ternationalization (contract 720-22195 (G10940-300))
African, Arabian and Eurasian plate. Our results indi- and Swedish Natural Science Research Council (con-
cate that, within areas close to the Red Sea, extension tract G-GU 3164-309, 312). The authors are very
is likely to be the prevailing regime. grateful to all station directors who provided the data
As mentioned above, the surface geology of the required. We are greateful to the anomymous reviewer
epicentral area is covered by loose quaternary sedi- who revealed a numerical error in the text.

jose33.tex; 2/12/1998; 0:50; p.9

 302

References Gaulier, J. M., Le-Pichon, X., Lyberis, N., Avedik, F., Geli, L., Mor-
retti, I., Deschamps, A. and Hafez, S., 1988, Seismic study of the
Aki, K., 1982, Strong motion prediction using mathematical tech- crust of the northern Red Sea and Gulf of Suez, Tectonophysics
niques, Bull. Seism. Soc. Am. 72, S29–S41. 153, 55–88.
Arab, N., Kazi, A. and Rieke, H., 1994, Fractal geometry of the Ibrahem, I., 1995, Seismicity and seismic hazard in Egypt the earth-
fault in relation to 12 October 1992 Cairo earthquake, Natural quake of Dahshour of 12, 10, 1992. Published abstract in XXIV
Hazards 10, 221–233. General Assembly, September 19–24 1994, Athens, Greece.
Ambraseys, N. N, Melville, C. P. and Adam, R. D., 1994, The Seis- Hempton, M. R., 1987, Constraints on Arabian plate motion and
micity of Egypt, Arabia and the Red Sea a Historical Review, extension history of the Red Sea, Tectonics 6, 687–705.
Cambridge Univ. Press, 181 pp. Herrmann, R. B. and Wang, C. Y., 1985, A comparison of synthetic
Badawi, H. S. and Mourad, S. A., 1994, Observation from 12 Oc- seismograms, Bull. Seism. Soc. Am. 75, 41–56.
tober 1992 Dahshour earthquake in Egypt, Natural Hazards 10, Hosney, H. M., 1985, Geophysical investigations in the Gulf of Suez
261–274. – Consequence for the tectonic evolution of the northern Red
Bayoumi, A. I., 1983, Tectonic origin of the Gulf of Suez Egypt as Sea, PhD Thesis, Univ. Kiel, 188 pp.
deduced from Gravity data, Handbook, Geophysics Exploration Huang, P. Y. and Solomon, S. C., 1987, Centroid depths and mecha-
at Sea, CRC Press Inc., U.S.A., pp. 417–431. nisms of Mid Ocean Ridge earthquakes in the Indian Ocean, Gulf
Bouchon, M., 1980, The motion of the ground during an earthquake, of Aden and Red Sea, J. Geophys. Res. 92, 1361–1382.
the case of dip slip fault, J. Geophys. Res. 85, 367–375. Klein, R. W., 1978, Hypocentre location program: HYPOIN-
Duvergé, J., 1969, Seismicity of Morocco; Agadir earthquake and VERSE, part 1: Users guide, Open file report 78?694, U.S.
paraseismic protection (In French), Service de le Météorologie Geological Survey, Menlo Park, California, 113 pp.
Nationale, Morocco, 110 pp. Kulhánek, O., Korrat, I. and El-Sayed, A., 1992, Connection of
Dziewonski, A. M., Ekström G. and Salgnik, M. P., 1993, Centriod the Seismicity in the Red Sea and Egypt. Published abstract in
moment tensor solution for October–December, 1992, Phys. the 10th annual meeting of Egyptian Geophysical Society, 1–3
Earth Planet. Interiors 80, 89–103. March 1992.
Egyptian Geological Survey and Mining Authority, 1981, Geo- Maamoun, M., Megahed, A. and Allam, A., 1984, Seismicity of
logic map of Egypt: Egyptian Geological Survey and Mining Egypt, Bull. HIAG. Vol. IV Ser. B, 109–160.
Authority, Cairo, Scale 1: 2 000 000. McCaffrey, R., Abers, G. and Zwick, P., 1991, Inversion of
El-Gamili, M., 1982, A geophysical interpretation of a part of the teleseismic body waves, IASPEI Software library.
Nile Valley, Egypt based on gravity data, J. Geol. (Special Vol., Mousa, H.: 1989, Earthquakes activities in Egypt and Adjacent re-
part 2), 101–120. gions and its relation to geotectonic features in A.R.E. Egypt. M
El-Sayed, A. and Wahlström, R., 1996, Distribution of the energy Sc Thesis Mansoura University Egypt, 170 pp.
release, b-value and seismic risk in Egypt, Natural Hazards 13, Nabelek, J., 1984, Determination of earthquake source parameters
133–150. from inversion of body waves, PhD Thesis, Mass. Inst. of Tech.
El-Sayed, A., Wahlström, R. and Kulhánek, O., 1996, Triggering Poceski, A., 1963, The ground effect of Skopje, 26 July 1963
sources for seismicity and distribution of ground acceleration in earthquake, Bull Seism. Soc. Am. 59, 1–22.
Egypt, 10th International Seminar on Earthquake Prognostics, Said, R., 1981, The Geological evolution of the River Nile, Elsevier
Jan 5–10, Cairo, Egypt P 59. Publ. Co., Amsterdam & New York, 180 pp.
El Shazly, E. M., Abdel Hady, A. M., El Ghawaby, A. M., El Kassas, Sieberg, A., 1932, Erdbebengeographie. In: B. Gutenberg (ed.),
I. A., Khawasik, S. M., El Rakaiby, M. M. and Aassy, I. E., 1980, Handbuch der Geophysics, Band IV, Abschnitt VI, Berlin.
Structural lineament map of Egypt based on landsat satellite im- Snoke, J. A., Munsey, J. W., Teague, A. G. and Bollinger, G. A.,
age interpretation field and laboratory investigation. Acad. Sci. 1984, A program for focal mechanism determination by com-
Res. Technol., Cairo. bined use of polarity and SV-P amplitude ratio data, Earthquake
Notes, 55–3, 15.

jose33.tex; 2/12/1998; 0:50; p.10

View publication stats