You are on page 1of 17

Arabian Journal of Geosciences (2017) 10:540

https://doi.org/10.1007/s12517-017-3325-1

ORIGINAL PAPER

Seismic hazard assessment and rheological implications: a case study


selected for cities of Saudi Arabia along the eastern coast of Red Sea
Faisal Rehman 1 & Abdullah M. Alamri 2 & Sherif M. El-Hady 3,4 & Hussein M. Harbi 3 & Ali H. Atef 3

Received: 19 January 2017 / Accepted: 30 November 2017


# Saudi Society for Geosciences 2017

Abstract
A probabilistic approach is used to evaluate the seismic hazard for 12 strategic cities in Saudi Arabia along the eastern
coast of Red Sea. The focal depth variations controlled by rheological characteristics are taken into account for hazard
calculations, and its creditability is tested through sensitivity analysis for hazard results. This study presents a neo-
probabilistic seismic hazard assessment methodology in which the focal depth distribution of earthquakes within
seismogenic layer is divided into three depth slices. These depth slices are based upon rheological characteristic of
seismogenic layer. The hazard results are obtained using this depth-slice methodology and conventional approach in
which uniform distribution of seismicity within seismogenic layer is assumed. The sensitivity analysis culminated in
underestimation of hazard values in higher frequencies for uniform distribution of seismicity within seismogenic layer.
Foregoing the observations recorded above, it can be concluded that the exploitation of depth-slices biased by the
rheology to calculate hazard is relatively preferable in the situations demanding safety measures.

Keywords Seismic hazard assessment . Rheological implications . Saudi Arabia

Introduction exceedance probability of ground motion from specified


level at the site of interest by integrating the probability
Probabilistic seismic hazard analysis (PSHA) is a quanti- theory with prototype of seismological sources and
fied methodology which provides a relationship among ground motion attributes (Anbazhagan et al., 2009;
potential seismic sources, corresponding ground motion Mihalić et al., 2011; Muço et al., 2012). For assessment
parameters, and associated earthquake occurrence proba- of ground motion, we must take into account the earth-
bilities (Lin and Baker, 2011). PSHA reckons the quake occurrence frequency for various magnitudes, the
ambiguity of the epicenter locations, and the attenuation
of ground motion together with its uncertainty (Wong, and
District, U.S.A.C.o.E.J., Corporation, U, 2004).
Electronic supplementary material The online version of this article
(https://doi.org/10.1007/s12517-017-3325-1) contains supplementary Rheology is the understanding of deformation processes in
material, which is available to authorized users. the crust and upper mantle. Brittle processes control mainly
deformation in the shallow crust (Emmerson et al., 2006). The
* Faisal Rehman deformation at greater depths is dominated by plastic process-
mail.faisalrehman@gmail.com es as result of the higher temperatures (Handy and Brun,
2004). However, during the episodes of deformation, such
1
Department of Earth Sciences, University of Sargodha, as associated with earthquakes, semi-plastic deformation be-
Sargodha, Pakistan havior will penetrate deeper in the crust (Scholz, 2002). The
2
Department of Geology & Geophysics, King Saud University, strength of continental lithosphere may likely to be contained
Riyadh 11451, Saudi Arabia within the seismogenic layer (Maggi et al., 2000). These rhe-
3
Geophysics Department, Faculty of Earth Sciences, King Abdulaziz ological changes control earthquake occurrence with respect
University, Jeddah, Saudi Arabia to depth in seismogenic layer (Albaric et al. 2009). A very
4
Earthquake Department, National Research Institute of Astronomy little attention has been given to the subject of seismic hazard
and Geophysics, Helwan, Egypt assessment and rheological implications.
540 Page 2 of 17 Arab J Geosci (2017) 10:540

Throughout the human history, urban settlement hap- Regional geology and tectonic setting
pened in coastal areas, which are actually the link be-
tween shore and offshore. The coastal areas provide trans- The Arabian Plate is one of the youngest fragments of Earth
portation sources, defensible strategic positions, and tour- lithosphere (Sadek, 2004). It is bordered in all direction by
ism places. In the situations demanding safety measures, active plate margins. The drifting from NE Africa furnished
seismic hazard assessment for coastal cities is unavoid- the origin of Arabian Plate ~ 25 Ma ago (Stern and Johnson,
able. Previous studies related to PSHA in the western 2010). The Arabian Plate is subdivided into two major regions,
coastal province of Saudi Arabia are limited (e.g., Al- Arabian Platform (eastern part) and Arabian Shield (western
Amri et al., 2003, Al-Amri and Rodgers (2013), Al-Arifi part). The Arabian Platform is characterized by sedimentary
et al., (2013) Almadani et al., (2015) and Zahran et al., cover thickening towards Arabian Gulf. Igneous and metamor-
(2015 & 2016). These studies did not incorporate rheo- phic rocks are mainly exposed in Arabian Shield area and are
logical constraints. The main aim of this research is to devoid of sedimentary cover (Al-Amri and Rodgers, 2013).
identify the effect of rheology attribute on hazard estima- The main tectonic provinces that affect our study area are
tions besides uncertainty and sensitivity analysis. PSHA is summarized in Fig. 2. The Arabian Shield separated from
carried out for 12 cities of Saudi Arabia (Fig. 1). These Nubian Shield in early Tertiary age by the Red Sea Rifting.
cities are rapidly growing major coastal settlements. The continental lithosphere rifting, beginning in Oligocene,

Fig. 1 Location map of selected


cities and town circled in red
Arab J Geosci (2017) 10:540 Page 3 of 17 540

Fig. 2 Major tectonic provinces


affecting the study area modified
after USGS (1998)

leads to break and generation of the sea floor by approximate- shields (Ghebreab, 1998). The Red Sea tectonics indicates that
ly 5 Ma in the southern part of the Red Sea (Cochran, 1981). initially rift-normal extension started which lately switched to
On the basis of geology and tectonics point of view, the oblique rifting. The Red Sea continental rifting transformed
Arabian Shield can be divided into five major terrains (Fig. 2): into oceanic rift having no connection to world mid-oceanic
Assir Terrain, Hijaz Terrain, Madyan Terrain, Afif Continental ridge system (Bosworth et al., 2005).
Terrain, and ArRayn Terrain (Al-Shanti, 2009). These terrains
are dissected and crossed by a group of different strike-slip
faults. Proterozoic age strike-slip faults are major structural Seismicity and Seismotectonics
elements extending in the northeastern Arabian Peninsula
from the southeast having 120-km length and 300-km width The Arabian Shield and Red Sea are examples of active tec-
(Agar, 1987; Al-Shanti, 2009; Stern, 1985). tonics regimes in the world experiencing continental rifting
Red Sea is elongated and is somewhat twisting escarpment- and new oceanic lithosphere formation (Al Amri et al.,
bound basin trending, having 250- to 450-km width, 2017). Youssef (2015) discussed the assumption that
surrounded by uplifted edges of the Arabian and African Arabian Shield was relatively stable and has low seismic
540 Page 4 of 17 Arab J Geosci (2017) 10:540

activity character of Saudi Arabia. However, ongoing tectonic associated with existing transform faulting and normal
activity and occurrence of felt earthquakes changed the as- faulting, which is a result of spreading. In Yemen, the
sumption especially for western part of Arabian Shield. The focal mechanism solution indicates pure normal faulting
seismicity map of the Arabian Shield and its surrounding for associated with seismic activity. In the Red Sea southern
period from 1900 to 2016 for magnitude 4 and above is pre- part towards central, the faulting trend changes from
pared from various national and international available seis- strike-slip to normal. The normal faulting can be attribut-
mological data sources (Fig. 3). The majority of seismic ac- ed to sea floor spreading or extension while at the south-
tivities are restricted along the divergent boundaries and the ern part the counter-clockwise movement Arabian Plate
spreading centers of the Red Sea and the Gulf of Aden and the produces strike-slip behavior. Gulf of Aqaba indicates
Dead Sea strike-slip fault zone in the northwest. The overall two styles of faulting; strike-slip and normal.
paucity of seismic activity in the interior part of the Arabian
promontory indicates that presently a very little deformation
of the Arabian Plate is taking place. Seismic hazard assessment
The focal mechanism solution map is prepared using
Harvard Centroid Moment Tensor solution (Fig. 4). The The effective way to minimize the seismic disasters loss is
East African rift system indicates pure normal faulting. to evaluate the seismic hazard and publicize these pieces
The tension axis nearly oriented NNE-SSW direction. of information to improve building construction style and
The Gulf of Aden illustrates two types of faulting re- structural design (El-Hussain et al., 2012). The seismic
gimes, strike-slip and normal faults. Strike-slip faults hazard assessment outputs the ground motion attributes

Fig. 3 Instrumental seismicity


map of the Arabian shield and its
surrounding for the period from
1900 to 2016 for magnitude 4 and
above
Arab J Geosci (2017) 10:540 Page 5 of 17 540

Fig. 4 Harvard centroid moment


tensor solution for earthquakes
greater than 5 magnitudes

for earthquake resistant design, to be implemented for the are well determined when the catalogue has longer cover-
site of interest (Hashemi et al., 2013). Seismic hazard age (Woessner and Wiemer, 2005).
assessment study involves utilization of seismological da- All available seismicity data from national and internation-
ta mainly comprised of seismic activity, attenuation, and al sources is gathered and compiled into one comprehensive
site response studies (Dowrick, 2009). catalogue. The catalogue is built by merging historical and
In the current study, seismic hazard assessment followed instrumental records. The historical catalogue is assigned
the generation of homogenized earthquake catalogue, selec- three quality schemes: Qa, quality for references; Qb, quality
tion of suitable and comprehensive seismotectonic model, cal- for location; and Qc, quality for magnitude.
culations of Mmax and recurrence parameters for seismogenic On the basis of quality control for historical events, the
source zones, and selection of appropriate ground motion pre- historical earthquake of 1121 AD is considered very low
dictions equation. in quality and is excluded in catalogue compilation
(Fig. 5). The priority given to consideration of magnitudes
is in the following succession respectively: the moment
Earthquake catalog magnitude (Mw), the surface wave magnitude (Ms), the
body-wave magnitude (mb), and the local magnitude
Earthquake catalogue, which provide a broad data set, is (ML), respectively. The earthquake catalogue is compiled
being used for various investigations related to earthquake for a spatial region extending from 15° to 35° N and 29°
physics and geohazard analysis. The hazard parameters to 47° E and includes events from magnitude 3 and above.
540 Page 6 of 17 Arab J Geosci (2017) 10:540

Fig. 5 Historical seismicity from


827 to 1899 AD site area and its
vicinity

The catalogue covers a period of time from 827 AD to Zhuang et al. (2002), Hainzl et al. (2006), and Barani
August 2016. et al. (2007).
Earthquake magnitudes reported are not uniform, so In this study, catalogue is declustered using the Gardner
the different earthquake magnitudes are converted to mo- and Knopoff (1974) technique. This technique is simple, ap-
ment magnitude. The conversation of different magnitude plicable to both historical and instrumental catalogue and to
into Mw is done by using formulae of Scordilis (2006) and catalogue with fewer parameters, time, date, and magnitude
Al Kathery (2010) . (Barani et al., 2007).
Subsequent to magnitude conversation, declustering is ap-
plied, to the catalogue to sort out main shocks, foreshocks. and Seismotectonic source model
aftershocks. This methodology is used for seismic hazard as-
sessment and earthquake predictions models (Van Stiphout The delineation of seismic source plays an important role
et al., 2012). The statistical analyses for seismicity are ob- in the evalution of hazard (Vipin and Sitharam, 2013).
scured by earthquake clusters, particularly in the case of back- The historical events, seismicity pattern, active faults,
ground seismicity (Zhuang et al., 2002). and their seismogenic potential play dynamic role in
Several techniques have been proposed for catalogue source characterization (Meletti et al., 2008). The
declustering, such as Knopoff (1964), Gardner and Rehman et al. (2016) seismogenic source model is utilized
Knopoff (1974), Öncel and Alptekin (1999), in the current study (Fig. 6). The seismogenic source
Reasenberg (1985), Reasenberg and Jones (1989), model comprised of total 23 area sources.
Arab J Geosci (2017) 10:540 Page 7 of 17 540

Fig. 6 Seismogenic source model


by Rehman et al. (2016)

Recurrence parameters and Mmax Ground motion prediction equations

The Gutenberg Richter (1944) equation describes the relation- Ground motion prediction equations (GMPE) recounts the
ship between the number of earthquakes of specific magnitude attributes of ground motion as a function of magnitude of
(M) or larger per year and the activity rate (Gutenberg and the earthquake, distance from the source to the site, and local
Richter, 1944). Kijko and Sellevoll (1989, 1992) expanded site conditions (Atkinson and Boore, 1995; Peruš and Fajfar,
the Gutenberg Richter (1944) relationship for heterogeneous 2009; Yazdani and Kowsari, 2013). The GMPE basically pre-
and varying quality seismicity data containing large historical sents empirical relationships that are obtained from seismo-
events and recent observation. logical data. GMPE model is obtained by stochastic tech-
Beside recurrence parameters, the choice of Mmax has an niques when a fewer number of seismological data is avail-
influence on seismic hazard especially for long return period able. The stochastic simulation method is applied by Boore
and short distance. Different techniques have been utilized for (1983, 2003), Toro et al. (1997), and Raoof et al. (1999).
estimation of Mmax in earlier hazard studies, e.g., global statis- Douglas (2007) suggested a better approach to utilize well-
tical models (Kagan, 1990), regression strain based (DeMets develped models from alternative regions rather to use equa-
et al., 1990), and observed seismicity-based numerical ap- tions from local and often poorly constrained models.
proach (Kijko, 2004; Kijko and Graham, 1998; Kijko et al., Ground motion attenuation relationships in the current
2001). In the current study, recurrence parameters and Mmax study are selected under the guidance provided by Cotton
are calculated using Kikjo (2004) and presented in Table 1. et al. (2006). Six different ground motions scaling
540 Page 8 of 17 Arab J Geosci (2017) 10:540

Table 1 Recurrence parameters and Mmax calculated for each seismogenic source zone

Sr. No Zone Name Mmin Mmax B values Beta Lembda

Kijko Error Obs Kijko Error Kijko Error Kijko Error

1 Gulf of Aden 3.6 6.62 0.50 6.60 0.94 0.02 2.16 0.05 12.01 0.60
2 Southern Red Sea 3.6 6.66 0.50 6.60 1.06 0.03 2.43 0.08 4.13 0.36
3 Afar Region 3.6 6.95 0.52 6.80 1.13 0.02 2.60 0.05 17.59 0 .73
4 Yemen Jezan Zone 3.6 6.45 0.50 6.40 0.99 0 .03 2.27 0.08 3.01 0.32
5 Central Southern Red Sea 3.6 6.79 0.22 6.70 0.75 0.03 1.74 0.06 10.17 0.63
6 Baha to Abha 3.6 6.00 0.54 5.50 0 .94 0.07 2.16 0.16 0.37 0.13
7 Central Red Sea 3.6 5.47 0.26 5.30 1.01 0.05 2.34 0.11 2.38 0.37
8 Western Coast Seismicity 3.6 6.30 0.54 5.80 0.85 0.07 1.96 0.16 0.49 0.13
9 Eastern Background Seismicity 3.6 5.30 0.54 4.80 0.98 0.07 2.27 0.16 0.10 0.05
10 Harrat Ad Damn Zone 3.6 5.53 0.50 5.00 0.99 0.07 2.28 0.15 0.87 0.19
11 Egypt 1 3.6 4.72 0.23 4.60 1.11 0.06 2.55 0.13 1.61 0.30
12 Egypt 2 3.6 5.35 0.32 5.10 0.97 0.05 2.24 0.11 1.62 0.21
13 Western Background Seismicity 3.6 4.99 0.53 4.90 0.99 0.07 2.29 0.16 0.66 0.20
14 Hijaz Zone 3.6 5.84 0.52 5.70 0.98 0.07 2.26 0.15 0 .37 0.11
15 Eastern Coast Seismicity 3.6 4.40 0.22 4.30 1.04 0.07 2.39 0.16 1.10 0.22
16 Egypt 3 3.6 4.86 0.26 4.70 1.13 0.06 2.60 0.13 1.72 0.27
17 Yanbu Suture Zone 3.6 6.20 0.54 5.70 1.00 0.06 2.31 0.15 3.36 0.49
18 Egypt 4 3.6 5.19 0.35 4.90 1.11 0.06 2.55 0.13 1.19 0.21
19 Northern Red Sea 3.6 5.87 0.26 5.70 1.27 0.03 2.92 0.08 8.42 0.50
20 Afif Terrain 3.6 5.52 0.50 5.00 1.06 0.07 2.45 0.15 0.93 0.20
21 Tabuk Zone 3.6 7.32 0.59 7.00 1.18 0.04 2.71 0.10 3.07 0.31
22 Gulf of Suez 3.6 6.69 0.51 6.60 1.19 0.03 2.73 0.07 4.91 0.30
23 Gulf of Aqba 3.6 7.33 0.50 7.20 1.13 0.03 2.61 0.07 8.02 0.45

relationships are chosen to meet the tectonic conditions of the The deviation in elastic characters and focal depth distribu-
study area which include Abrahamson and Campbell and tion of earthquakes has been used to portray the lithospheric
Bozorgnia (2003), Abrahamson and Silva (1997), Sadigh rheological character and its deformation (Zhang et al., 2013).
et al. (1997), Boore et al. (1997), Atkinson and Boore The direct interpretation for variation in rheological properties
(1995), and Campbell and Bozorgnia (2008) (next-generation with depth can be made over seismicity data gathered over
attenuation equation). many years for a region (Bodri, 1996). The changes in the
geological framework are clearly illustrated by variation in
depth distribution in seismicity. Variety of earthquake cutoff
Rheology and focal depth distribution depths and seismicity peaks are direct an indication of signif-
icant changes in rheological behavior (Albaric et al., 2009).
The crustal rheology describes the inter-relationship between The cutoff seismicity depth is a clear indicator of the
different lithologies, their mechanical properties, existence of seismogenic layer thickness. Ito and Nakamura (1998) de-
different fluids, and their temperature present in the crust. The scribed that the focal depth cumulative frequency distribution
deformation style and strength of the lithosphere are largely curve could be used to estimate cutoff depth of seismogenic
controlled by rhology (Searle and Escartin, 2004). Yield layer, the 90% of the cumulative frequency curve of indicator
Strength Envelope (YSE) curve is used to portray dispersal of lower cut off depth of seismogenic layer. This methodology
of various stress regimes with respect to depth in crust. The is used by many researchers (e.g., Ito 1999; Motohashi et al.
lithosphere’s mechanical strength estimation has been carried 2004).
out by utilizing this technique (Albaric et al., 2009; In the current study, an important parameter focal depth
Déverchère et al., 2001). YSE and vertical depth distribution distribution which effects the ground motion acceleration is
of earthquakes can be used to predicate wet or dry character of considered and its effect on hazard assessment is tested. Ito
the earth crust and mantle (Konstantinou, 2010). and Nakamura (1998) methodology is used to estimate cutoff
Arab J Geosci (2017) 10:540 Page 9 of 17 540

Fig. 7 Estimation of cutoff


seismicity and division of
seismogenic layer into three depth
slices. (SGB, base of the
seismogenic layer)

seismicity depth. A focal depth distribution curve (FDDC) is Hazard calculations


constructed as a result of application of the Ito and Nakamura
(1998) technique. Depending upon average error in focal The hazard calculations generate ground motions curves for
depth, which less than 5 km, a bin size of 5 km is chosen for selected return periods. The seismogenic source zones, re-
FDDC. The base of the seismogenic layer is marked at the currence parameters, Mmax and GMPE, and effect of
observation of 90% of total seismicity. After that, on the basis rheology that have been discussed above are all included
of percentage of seismicity observed the in the FDDC, the into the hazard calculations. The incorporation of these
seismogenic layer is divided into three different depth slices. elements is carried out through formulation the of the
These three slices actually present vertical distribution of seis- logic tree. Kulkarni et al. (1984) introduced the logic tree
micity. Figure 7 displays some examples for seismogenic cut- as a tool to quantify and capture the uncertainties in PSHA.
off depths estimated by technique mentioned above and gen- In hazard calculation, uncertainties are a result of inade-
eration of three depth slices for different seismic source zones. quate understanding of the complex process of earthquake
540 Page 10 of 17 Arab J Geosci (2017) 10:540

Seismotectonic
Magnitude Depth Attenuation Models
Source Model

0-200km

Abra.-Silva (1997) Rock =0.12


M1= D1=0.6 Boore-Joyner-Fumal (1997) = 0.12
M max Calc Campbell-Bozorgnia (2003) =0.32
Campbell-Bozorgnia (2008) NGA=0.32
0.5
Sadigh (1997) Rock =0.06
Atkinson-Boore (1995) =0.06

200-600km

Rehman et al., D2=0.2 Abra.-Silva (1997) Rock= 0.22


2016 Boore-Joyner-Fumal (1997) = 0.22
Campbell-Bozorgnia (2003) = 0.22
Atkinson-Boore (1995) = 0.34

> 600 km

M2 =
M maxObs Abra.-Silva (1997) Rock 0.25
Boore-Joyner-Fumal (1997) = 0.25,
0.50
Campbell-Bozorgnia (2003) = 0.25
D3=0.2 Atkinson-Boore (1995) = 0.25

Fig. 8 Systematic layout for logic tree generated

generation and seismic wave propagation (Lombardi et al., for maximum observed seismicity part of the crust while the
2005). It is essentially critical to differentiate between un- equal probability of 20% is given for the rest of the remaining
certainty in knowledge (epistemic uncertainty) and random- two other depth zones (Fig. 7). The ground motion equations
ness in the process (aleatory uncertainty) (McGuire and are given weight according to the distance to the site. The
Toro 1986, Atkinson 2004). Aleatory variability is basically source to site distance is divided into three ranges (0–200,
the natural randomness in any process which results by 200–600, 600–1000 km). The weight assigned to GMPE ac-
simplified modeling of a complex process, parameterized cording to distance followed by a scheme presented in Fig. 8.
by probability density functions. The scientific uncertainty
present in simplified model is epistemic uncertainty
(Abrahamson, 2006). Results and discussion

Logic tree Seismic hazard assessment is conducted for return periods of


475 and 2475 years which correspond to 10 and 2% probabil-
The logic tree incorporates a succession of branches which ity of exceedance in 50 years, respectively. The hazard curves
depicts the alternate models and parameter. The relative con- illustrate the variation of strong motions parameters with prob-
fidence for each parameter or model is illustrated by weights ability of exceedance for specific return period. Figure 9 illus-
assigned to them (Abrahamson and Bommer, 2005; Bommer, trates the hazard curves for 12 cities in the western coast of
2002; Bommer and Scherbaum, 2008). In the current study, Saudi Arabia. These curves show ground motion acceleration
logic tree incorporates one seismogenic source model, two at various spectral periods.
scenarios for Mmax, three depth distribution slices of seismic-
ity, and six ground motion prediction equations. Deaggregation analysis
The magnitude is given for two different probabilities.
Mmax calculated is termed M1. The second probability M2 is Deaggregation technique is a significant and essential proce-
basically maximum observed magnitude in the source zone. dure for understanding of seismic hazard results. This process
The three depths D1, D2, and D3 are assigned different has become more basic practice for answering the question
weights. The maximum weight is given to the depth range related to sources, which significantly contribute to the hazard
where highest frequency of earthquakes is observed. In most (Jaradat 2008). It enables to identify the attributes of seismic-
of the seismogenic model zones, a weight of 60% is adopted ity accountable for seismic hazard in term of distance to site
Arab J Geosci (2017) 10:540 Page 11 of 17 540

Fig. 9 Uniform hazard curves for


475- and 2475-year return periods

and earthquake magnitude (Fauzi and Fauzi 2013; Bazzurro period in four cities (Jeddah, Yanbu, Jizan, and Tabuk) to
and Cornell 2004). Normally, seismic hazard is deaggregated determine the contribution of sources to hazard level.
in provisions of two parameters, source to site distance and In Jeddah and Yanbu, the largest contribution to hazard
magnitude (McGuire and Shedlock. 1981; Stepp et al. 1993). originates from very near sources at PGA, 0.1 s and 0.2 s
The PSHA results are deaggregated for a 475-year return spectral periods (Fig. 10 and Supplementary material). Both
540 Page 12 of 17 Arab J Geosci (2017) 10:540

PGA

Fig. 10 Deaggregation results for Jeddah at PGA, 0.1 s, 0.2 s, 1 s, and 2 s spectral periods
Arab J Geosci (2017) 10:540 Page 13 of 17 540

Fig. 11 Sensitivity analysis for


Jeddah at return periods of
475years (a) and 2475 years (b)

cities are surrounded by low seismicity. However, at 1 s and northern part. The abovementioned scenarios for both cities
2 s spectral periods for Jeddah, the hazard contribution origi- indicate that seismic sources very close to the two cities are of
nates from distant earthquakes (187–300 km) with magnitude low activity rates. The high ground-motion levels for short
6.3–6.4 Mw. For Yanbu, the magnitude of earthquakes range distance events are not likely to occur. However, for distant
from 6 to 6.2 Mw and distance to site from 375 to 544 km at 1 s earthquakes, higher motions can be expected at lower
and 2 s spectral periods possible linked with sources in frequencies.
540 Page 14 of 17 Arab J Geosci (2017) 10:540

The seismicity affecting Jizan and Tabuk is higher and very sensitivity analysis are carried out for Jeddah, Yanbu, Jizan,
close to both cities as compared to Jeddah and Yanbu and Tabuk.
(Supplementary material). The major contribution for seismic The deaggregation of hazard results for Jeddah and Yanbu
hazard to Jizan at PGA, 0.1 s, 0.2 s, 1 s, and 2 s spectral period revealed that short distant earthquakes are not expected to
originates from Yemen Jizan Zone (4), Southern Red Sea (2), produce highest ground motions at PGA, 0.1 s, and 0.2 s spec-
and Central Southern Red Sea (5) (Fig. 6). The controlling tral periods. However, the earthquake events at large distances
sources for Tabuk are Tabuk Zone (21), Northern Red Sea can produce higher accelerations at higher spectral periods
(19), and Gulf of Aqaba (23) (Fig. 6). The hazard for both like 1 and 2 s. Jizan and Tabuk are expected to have alarge
cities is dominated by moderate to large earthquakes at smaller contribution to hazard by both near and far distance earth-
and larger spectral periods (Supplementary material). This quakes at all spectral periods.
means high ground motion levels can be expected for shorter The main objective of the study is to find out the effect of
as well as larger distance event. crust rheological characteristics on hazard estimation. This
effort leads us to divide the focal depth distribution into three
depth slices. The maximum weight is assigned to the depth
Sensitivity analysis
slice that had the highest frequency of earthquakes occur-
rence. The sensitivity analysis carried out for three depth slices
Sensitivity analysis (SA) is an indispensable step in seis-
and for a uniform distribution for focal depths. The results
mic hazard analysis to determine the significance of input
obtained from the uniform distribution of seismicity within
parameters and uncertainty to the outcome (Beauval and
seismogenic layer are lower predicted than that of three depth
Scotti, 2004; Marin et al., 2004; Sabetta et al., 2005;
slices. This underestimation of hazard results is observed at
Sokolov et al., 2009). The SA points out the major contrib-
higher frequencies for return period 475 and 2475 years. The
utors in the variability of seismic hazard (Hassaballa et al.,
effect of focal depth distribution is negligible at lower
2011; Lombardi et al., 2005). The basic goal of SA is to
frequencies.
recognize the input parameters that encompass the maxi-
The epistemic uncertainty and aleatory variability are
mum impact on risk estimation and its uncertainty (Cramer
topics of great interest in the hazard analysis. The depth dis-
et al., 2002; Giner et al., 2002).
tribution is epistemic variability which directly comes from
The creditability of depth bin distribution of seismicity
stress field distribution and real tectonic conditions. There is
within seismogenic layer into three depth slices is evaluated
a clear observation that seismic hazard in a uniform distribu-
for its effect on hazard calculation. The hazard calculation is
tion of seismicity leads to underestimation of the hazard. This
carried out again by using a prototype of focal depth distribu-
concludes that it is suitable to use depth slices for safety pur-
tion. This approach divides seismicity in seismogenic layer
pose. Any future work for seismic hazard studies should take
uniformly. The results indicate that uniform distribution of
the stress field distribution and rheological characteristics of
focal depth in seismogenic layer leads to underestimation of
the crust into their consideration.
hazard at PGA and higher frequencies 0.1 and 0.2 s for the
return period of 475 and 2475 years (Fig. 11 and
Supplementary material). The lower frequency spectrum is Acknowledgements The authors gratefully appreciate the support by
unaffected for both 475- and 2475-year return periods. The Deanship of Graduate Studies and Department of Geophysics, Faculty
underestimation for hazard calculation increases in weight as of Earth Sciences, King Abdulaziz University. The authors would like to
ground acceleration values for that area increases. thank Earthquake Monitoring Center, Sultan Qaboos University, Oman
who give us the permission to run EZ-FRISK software for academic
purposes and scientific cooperation.

Conclusions

The hazard analyses, conducted in this study, are based on References


homogenized earthquake catalogue, updated source model,
Abrahamson N (2006) Seismic hazard assessment: problems with
and next-generation attenuation equations with rheological
current practice and future developments. First European
constraints. The instrumental seismicity catalogue used is gen- Conference on Earthquake Engineering and Seismology,
erated by cross-referencing other international seismological Geneva, Switzerland, pp 3–8
centers data. The ground-motion prediction equations used in Abrahamson NA, Bommer JJ (2005) Probability and uncertainty in seis-
this study include NGA model. The logic tree is built to take mic hazard analysis. Earthquake Spectra 21(2):603–607. https://doi.
org/10.1193/1.1899158
consideration of uncertainties. The PGA and spectral acceler- Abrahamson N A, Silva W J (1997) Empirical response spectral attenu-
ation values for 475- and 2475-year return period are calcu- ation relations for shallow crustal earthquakes. Seismol Res Lett
lated for 12 cities. Furthermore, deaggregation of hazard and 68(1):94–127
Arab J Geosci (2017) 10:540 Page 15 of 17 540

Agar R (1987) The Najd fault system revisited; a two-way strike-slip Boore DM, Joyner WB, Fumal TE (1997) Equations for estimating hor-
orogen in the Saudi Arabian Shield. J Struct Geol 9(1):41–48. izontal response spectra and peak acceleration from western North
https://doi.org/10.1016/0191-8141(87)90042-3 American earthquakes: a summary of recent work. Seismol Res Lett
Al Amri A, Abdelrahman K, Andreae MO, & Al-Dabbagh M (2017) 68(1):128–153. https://doi.org/10.1785/gssrl.68.1.128
Crustal and upper mantle structures beneath the Arabian Shield Bosworth W, Huchon P, McClay K (2005) The Red Sea and Gulf of Aden
and Red Sea. In Lithosphere dynamics and sedimentary basins of basins. J Afr Earth Sci 43(1-3):334–378. https://doi.org/10.1016/j.
the Arabian Plate and surrounding areas (pp. 3-29). Springer jafrearsci.2005.07.020
International Publishing Campbell KW, Bozorgnia Y (2003) Updated near-source ground-motion
AlKathery AM (2010) Short-term and long-term seismic hazard assess- (attenuation) relations for the horizontal and vertical components of
ment, NW Arabian Peninsula. MSc thesis, Geology Department, peak ground acceleration and acceleration response spectra. Bull
College of Science, King Saud University, p. 179 Seismol Soc Am 93(1):314–331. https://doi.org/10.1785/
Al-Amri A M, Rodgers A J (2013) Improvement of seismicity parameters 0120020029
in the Arabian shield and platform using earthquake location and Campbell KW, Bozorgnia Y (2008) NGA ground motion model for the
magnitude calibration. In: Lithosphere dynamics and sedimentary geometric mean horizontal component of PGA, PGV, PGD and 5%
basins: The Arabian plate and analogues. Springer, Berlin, damped linear elastic response spectra for periods ranging from 0.01
Heidelberg, pp. 281–293 to 10 s. Earthquake Spectra 24(1):139–171. https://doi.org/10.1193/
Al-Amri A, Punsalan BT, Khalil A, Uy EA (2003) Seismic hazard as- 1.2857546
sessment of western Saudi Arabia and the Red Sea region. IISEE, Cochran JR (1981) The Gulf of Aden: structure and evolution of a young
Japan, pp 95–112 ocean basin and continental margin. J Geophys Res: Solid Earth
Al-Arifi NS, Fat-Helbary RE, Khalil AR, Lashin AA (2013) A new (1978–2012) 86(B1):263–287. https://doi.org/10.1029/
evaluation of seismic hazard for the northwestern part of Saudi JB086iB01p00263
Arabia. Nat Hazards 69(3):1435–1457 Cotton F, Scherbaum F, Bommer JJ, Bungum H (2006) Criteria for
Albaric J, Déverchère J, Petit C, Perrot J, Le Gall B (2009) Crustal rhe- selecting and adjusting ground-motion models for specific target
ology and depth distribution of earthquakes: insights from the cen- regions: application to central Europe and rock sites. J Seismol
tral and southern East African Rift System. Tectonophysics 468(1- 10(2):137–156. https://doi.org/10.1007/s10950-005-9006-7
4):28–41. https://doi.org/10.1016/j.tecto.2008.05.021 Cramer CH, Wheeler RL, Mueller CS (2002) Uncertainty analysis for
Almadani S, Al-Amri A, Fnais M, Abdelrahman K, Ibrahim E, seismic hazard in the southern Illinois basin. Seismol Res Lett 73(5):
Abdelmoneim E (2015) Seismic hazard assessment for Yanbu met- 792–805. https://doi.org/10.1785/gssrl.73.5.792
ropolitan area, western Saudi Arabia. Arab J Geosci 8(11):9945– DeMets C, Gordon RG, Argus DF, Stein S (1990) Current plate motions.
9958 Geophys J Int 101(2):425–478. https://doi.org/10.1111/j.1365-
Al-Shanti A (2009) Geology of Arabian Shield of Saudi Arabia. King 246X.1990.tb06579.x
Abdulaziz Universiyt Press, Jeddah Déverchère J, Petit C, Gileva N, Radziminovitch N, Melnikova V,
Anbazhagan P, Vinod JS, Sitharam TG (2009) Probabilistic seismic haz- San'Kov V (2001) Depth distribution of earthquakes in the Baikal
ard analysis for Bangalore. Nat Hazards 48(2):145–166. https://doi. rift system and its implications for the rheology of the lithosphere.
org/10.1007/s11069-008-9253-3 Geophys J Int 146(3):714–730. https://doi.org/10.1046/j.0956-
Atkinson GM (2004). An overview of developments in seismic hazard 540x.2001.1484.484.x
analysis. In 13th World Conference on Earthquake Engineering (pp. Douglas J (2007) On the regional dependence of earthquake response
1-6). Vancouver, B.C., Canada August 1-6, 2004 spectra. ISET J Earthq Technol 44:71–99
Atkinson GM, Boore DM (1995) Ground-motion relations for eastern Dowrick D J (2009) Earthquake resistant design and risk reduction. John
North America. Bull Seismol Soc Am 85:17–30 Wiley & Sons, Hoboken
Barani S, Ferretti G, Massa M, Spallarossa D (2007) The waveform El-Hussain I, Deif A, Al-Jabri K, Toksoz N, El-Hady S, Al-Hashmi S, Al-
similarity approach to identify dependent events in instrumental Toubi K, Al-Shijbi Y, Al-Saifi M, Kuleli S (2012) Probabilistic
seismic catalogues. Geophys J Int 168(1):100–108. https://doi.org/ seismic hazard maps for the sultanate of Oman. Nat Hazards
10.1111/j.1365-246X.2006.03207.x 64(1):173–210. https://doi.org/10.1007/s11069-012-0232-3
Bazzurro P, Cornell CA (2004) Nonlinear soil-site effects in probabilistic Emmerson B, Jackson J, McKenzie D, Priestley K (2006) Seismicity,
seismic-hazard analysis. Bull Seismol Soc Am 94(6):2110–2123 structure and rheology of the lithosphere in the Lake Baikal region.
Beauval C, Scotti O (2004) Quantifying sensitivities of PSHA for France Geophys J Int 167(3):1233–1272. https://doi.org/10.1111/j.1365-
to earthquake catalog uncertainties, truncation of ground-motion 246X.2006.03075.x
variability, and magnitude limits. Bull Seismol Soc Am 94(5): Fauzi A, Fauzi UJ (2013) Deaggregation of new national seismic hazard
1579–1594. https://doi.org/10.1785/012003246 maps for Indonesia. In proceedings of 10 International Conference
Bodri B (1996) Thermal state, rheology and seismicity in the Pannonian on Urban Earthquake Engineering, Tokyo, Japan
basin, Hungary. J Geodyn 21(4):309–328. https://doi.org/10.1016/ Gardner J, Knopoff L (1974) Is the sequence of earthquakes in southern
0264-3707(96)00002-6 California, with aftershocks removed. Poissonian Bull Seismol Soc
Bommer JJ (2002) Deterministic vs. probabilistic seismic hazard assess- Am 64:1363–1367
ment: an exaggerated and obstructive dichotomy. J Earthq Eng Ghebreab W (1998) Tectonics of the Red Sea region reassessed. Earth Sci
6(sup001):43–73. https://doi.org/10.1080/13632460209350432 Rev 45(1-2):1–44. https://doi.org/10.1016/S0012-8252(98)00036-1
Bommer JJ, Scherbaum F (2008) The use and misuse of logic trees in Giner J, Molina S, Jauregui P (2002) Advantages of using sensitivity
probabilistic seismic hazard analysis. Earthquake Spectra 24(4): analysis in seismic hazard assessment: a case study of sites in south-
997–1009. https://doi.org/10.1193/1.2977755 ern and eastern Spain. Bull Seismol Soc Am 92(2):543–554. https://
Boore DM (1983) Stochastic simulation of high-frequency ground mo- doi.org/10.1785/0120000299
tions based on seismological models of the radiated spectra. Bull Gutenberg B, Richter CF (1944) Frequency of earthquakes in California.
Seismol Soc Am 73:1865–1894 Bull Seismol Soc Am 34:185–188
Boore DM (2003) Simulation of ground motion using the stochastic Hainzl S, Kraft T, Wassermann J, Igel H, Schmedes E (2006) Evidence
method. Pure Appl Geophys 160(3):635–676. https://doi.org/10. for rainfall-triggered earthquake activity. Geophys Res Lett 33(19).
1007/PL00012553 https://doi.org/10.1029/2006GL027642
540 Page 16 of 17 Arab J Geosci (2017) 10:540

Handy MR, Brun J-P (2004) Seismicity, structure and strength of the seismic hazard assessment of the Italian territory. Tectonophysics
continental lithosphere. Earth Planet Sci Lett 223(3-4):427–441. 450(1-4):85–108. https://doi.org/10.1016/j.tecto.2008.01.003
https://doi.org/10.1016/j.epsl.2004.04.021 Mihali S, Maja O, Krka M (2011) Seismic microzonation: A review of
Hashemi M, Alesheikh AA, Zolfaghari MR (2013) A spatio-temporal principles and practice. Geofizika 28(1):5-20
model for probabilistic seismic hazard zonation of Tehran. Comput Motohashi S, Ebisawa K, Sakagami M (2004). Evaluation of the
Geosci 58:8–18. https://doi.org/10.1016/j.cageo.2013.04.005 Seismogenic Layer Depth in Japan Using the JMA Catalogue. In
Hassaballa A, Mohamed ARE, Sobaih M (2011) Sensitivity analysis of proceedings of OECD/NEA CSNI Workshop, Tokyo, Japan
parameters for probabilistic seismic hazard for Sudan. J Sci Technol Muço B, Alexiev G, Aliaj S, Elezi Z, Grecu B, Mandrescu N, Milutinovic
12:02 Z, Radulian M, Ranguelov B, Shkupi D (2012) Geohazards assess-
Ito K (1999) Seismogenic layer, reflective lower crust, surface heat flow ment and mapping of some Balkan countries. Nat Hazards 64(2):
and large inland earthquakes. Tectonophysics 306(3-4):423–433 943–981. https://doi.org/10.1007/s11069-012-0185-6
Ito K, Nakamura S (1998) Variation in thickness of the Seismogenic layer Öncel A, Alptekin Ö (1999) Effect of aftershocks on earthquake hazard
in southwestern Japan and their relation to large inland earthquake. estimation: an example from the North Anatolian fault zone. Nat
Annuals Disas Prev Res Inst 41(B-1):27–35 Hazards 19(1):1–11. https://doi.org/10.1023/A:1008139802609
Kagan Y (1990) Random stress and earthquake statistics: spatial depen- Peruš I, Fajfar P (2009) How reliable are the ground motion prediction
dence. Geophys J Int 102(3):573–583. https://doi.org/10.1111/j. equations. In: Proceedings of the 20th international conference on
1365-246X.1990.tb04584.x structural mechanics in reactor technology (SMiRT 20), Espoo,
Kijko A (2004) Estimation of the maximum earthquake magnitude, M Paper, vol. 9
max. Pure Appl Geophys 161(8):1655–1681. https://doi.org/10. Raoof M, Herrmann R, Malagnini L (1999) Attenuation and excitation of
1007/s00024-004-2531-4 three-component ground motion in southern California. Bull
Kijko A, Graham G (1998) Parametric-historic procedure for probabilis- Seismol Soc Am 89:888–902
tic seismic hazard analysis part I: estimation of maximum regional Reasenberg P (1985) Second-order moment of central California seismic-
magnitude mmax. Pure Appl Geophys 152(3):413–442. https://doi. ity, 1969–1982. J Geophys Res: Solid Earth (1978–2012) 90(B7):
org/10.1007/s000240050161 5479–5495. https://doi.org/10.1029/JB090iB07p05479
Kijko A, Sellevoll M (1989) Estimation of earthquake hazard parameters Reasenberg PA, Jones LM (1989) Earthquake hazard after a mainshock in
from incomplete data files. Part I. Utilization of extreme and com- California. Science 243(4895):1173–1176. https://doi.org/10.1126/
plete catalogs with different threshold magnitudes. Bull Seismol Soc science.243.4895.1173
Am 79:645–654 Rehman F, El-Hady SM, Atef AH, Harbi HM (2016) Seismic hazard
Kijko A, Sellevoll MA (1992) Estimation of earthquake hazard parame- assessment of western Coastal Province of Saudi Arabia: determin-
ters from incomplete data files. Part II. Incorporation of magnitude istic approach. Earthq Sci 29(5):299–309
heterogeneity. Bull Seismol Soc Am 82:120–134 Sabetta F, Lucantoni A, Bungum H, Bommer JJ (2005) Sensitivity of
Kijko A, Lasocki S, Graham G (2001) Non-parametric seismic hazard in PSHA results to ground motion prediction relations and logic-tree
mines. Pure Appl Geophys 158(9):1655–1675. https://doi.org/10. weights. Soil Dyn Earthq Eng 25(4):317–329. https://doi.org/10.
1007/PL00001238 1016/j.soildyn.2005.02.002
Knopoff L (1964) The statistics of earthquakes in Southern California. Sadek A (2004) Seismic map for the state of Kuwait. Emirates J Eng Res
Bull Seismol Soc Am 54:1871–1873 9:53–58
Konstantinou K (2010) Crustal rheology of the Santorini–Amorgos zone: Sadigh K, Chang C-Y, Egan J, Makdisi F, Youngs R (1997) Attenuation
implications for the nucleation depth and rupture extent of the 9 relationships for shallow crustal earthquakes based on California
July 1956 Amorgos earthquake, southern Aegean. J Geodyn strong motion data. Seismol Res Lett 68(1):180–189. https://doi.
50(5):400–409. https://doi.org/10.1016/j.jog.2010.05.002 org/10.1785/gssrl.68.1.180
Kulkarni R B, Youngs R R, Coppersmith K J (1984) Assessment of Scholz CH (2002) The mechanics of earthquakes and faulting, 2nd edn.
confidence intervals for results of seismic hazard analysis. In: Cambridge University Press, UK, p 496. https://doi.org/10.1017/
Proceedings of the eighth world conference on earthquake engineer- CBO9780511818516
ing, vol 1. pp. 263-270 Scordilis E (2006) Empirical global relations converting M S and m b to
Lin T, Baker J (2011) Probabilistic seismic hazard deaggregation of moment magnitude. J Seismol 10(2):225–236. https://doi.org/10.
ground motion prediction models. In: 5th international conference 1007/s10950-006-9012-4
on earthquake geotechnical engineering, Santiago, Chile. pp. 10-13 Searle R, Escartin J (2004) The rheology and morphology of oceanic
Lombardi AM, Akinci A, Malagnini L, Mueller CS (2005) Uncertainty lithosphere and mid-ocean ridges. Geophys Monogr Ser 148:63–93
analysis for seismic hazard in Northern and Central Italy. Ann Sokolov VY, Wenzel F, Mohindra R (2009) Probabilistic seismic hazard
Geophys assessment for Romania and sensitivity analysis: a case of joint
Maggi A, Jackson J, Mckenzie D, Priestley K (2000) Earthquake focal consideration of intermediate-depth (Vrancea) and shallow
depths, effective elastic thickness, and the strength of the continental (crustal) seismicity. Soil Dyn Earthq Eng 29(2):364–381. https://
lithosphere. Geology 28(6):495–498. https://doi.org/10.1130/0091- doi.org/10.1016/j.soildyn.2008.04.004
7613(2000)28<495:EFDEET>2.0.CO;2 Stepp JC, Silva WJ, McGuire RK, Sewell RW (1993) Determination of
Marin S, Avouac J-P, Nicolas M, Schlupp A (2004) A probabilistic ap- earthquake design loads for a high level nuclear waste repository
proach to seismic hazard in metropolitan France. Bull Seismol Soc facility (No. CONF-9310102–VOL. 2). In: Proceedings of the
Am 94(6):2137–2163. https://doi.org/10.1785/0120030232 Natural Phenomena hazards Mitigation Conference, Vol 2. pp 651-
McGuire RK, Shedlock KM (1981) Statistical uncertainties in seismic 657, Oct. 19-22, Atlanta GA
hazard evaluations in the United States. Bull Seismol Soc Am Stern RJ (1985) The Najd Fault System, Saudi Arabia and Egypt: a Late
71(4):1287–1308 Precambrian rift-related transform system? Tectonics 4(5):497–511.
McGuire RK, Toro G (1986) Methods of earthquake ground motion https://doi.org/10.1029/TC004i005p00497
estimation for the eastern United States. Electric Power Research Stern RJ, Johnson P (2010) Continental lithosphere of the Arabian Plate:
Institute Research Project No. RP2556-16, prepared by Risk a geologic, petrologic, and geophysical synthesis. Earth Sci Rev
Engineering, Inc., Acton 101(1-2):29–67. https://doi.org/10.1016/j.earscirev.2010.01.002
Meletti C, Galadini F, Valensise G, Stucchi M, Basili R, Barba S, Toro GR, Abrahamson NA, Schneider JF (1997) Model of strong ground
Vannucci G, Boschi E (2008) A seismic source zone model for the motions from earthquakes in central and eastern North America:
Arab J Geosci (2017) 10:540 Page 17 of 17 540

best estimates and uncertainties. Seismol Res Lett 68(1):41–57. Youssef SEH (2015) Seismicity and seismotectonic setting of the Red Sea
https://doi.org/10.1785/gssrl.68.1.41 and adjacent areas. In: The Red Sea. Springer, Berlin Heidelberg, pp
Van Stiphout T, Zhuang J, Marsan D (2012) Seismicity declustering, 151–159
community online resource for statistical seismicity analysis. Zahran HM, Sokolov V, Youssef SEH, Alraddadi WW (2015)
https://doi.org/10.5078/corssa-52382934 Preliminary probabilistic seismic hazard assessment for the
Vipin K, Sitharam T (2013) Delineation of seismic source zones based on Kingdom of Saudi Arabia based on combined areal source model:
seismicity parameters and probabilistic evaluation of seismic hazard Monte Carlo approach and sensitivity analyses. Soil Dyn Earthq
using logic tree approach. J Earth Syst Sci 122(3):661–676. https:// Eng 77:453–468. https://doi.org/10.1016/j.soildyn.2015.06.011
doi.org/10.1007/s12040-013-0300-4 Zahran HM, Sokolov V, Roobol MJ, Stewart, IC, Youssef SEH, El-
Woessner J, Wiemer S (2005) Assessing the quality of earthquake cata- Hadidy M (2016) On the development of a seismic source zonation
logues: estimating the magnitude of completeness and its uncertain- model for seismic hazard assessment in western Saudi Arabia. J
ty. Bull Seismol Soc Am 95(2):684–698. https://doi.org/10.1785/ Seismol 20(3):747–769
0120040007 Zhang Z, Deng Y, Chen L, Wu J, Teng J, Panza G (2013) Seismic struc-
Wong IG, District, U.S.A.C.o.E.J, Corporation, U (2004) Deterministic ture and rheology of the crust under mainland China. Gondwana Res
and probabilistic seismic hazard analyses: Portuguese dam. URS 23(4):1455–1483. https://doi.org/10.1016/j.gr.2012.07.010
Corporation, Puerto Rico Zhuang J, Ogata Y, Vere-Jones D (2002) Stochastic declustering of space-
Yazdani A, Kowsari M (2013) Earthquake ground-motion prediction time earthquake occurrences. J Am Stat Assoc 97(458):369–380.
equations for northern Iran. Nat Hazards 69(3):1877–1894. https:// https://doi.org/10.1198/016214502760046925
doi.org/10.1007/s11069-013-0778-8

You might also like