Tectonophysics 460 (2008) 237–247

Contents lists available at ScienceDirect

Tectonophysics

j ourna l homepage: www.e lsev ie r.com/ locate / tecto

Source parameters of March 10 and September 13, 2007, United ArabEmirates earthquakes

Y. Al Marzooqi a, K.M. Abou Elenean b,⁎, A.S. Megahed b, I. El-Hussain c, A.J. Rodgers d, E. Al Khatibi a

a Dubai Municipality, Survey Department, P.O. Box: 67 Dubai, UAEb Seismology Department, National Research Institute of Astronomy and Geophysics (NRIAG), Helwan, Egyptc Earthquake Monitoring Center, Sultan Qaboos University, P.O. Box 50, Al-Khoudh, PC 123, Muscat, Omand Atmospheric, Earth and Energy Division, Lawrence Livermore National Laboratory, L-046, Livermore, CA, 94551 USA

⁎ Corresponding author.E-mail addresses: yamarzooqi@dm.gov.ae (Y. Al Marz

(K.M. Abou Elenean), megahed_ali@yahoo.com (A.S. Me(I. El-Hussain), rodgers7@llnl.gov (A.J. Rodgers), eakhatib

0040-1951/$ – see front matter © 2008 Elsevier B.V. Aldoi:10.1016/j.tecto.2008.08.017

a b s t r a c t

a r t i c l e i n f o

Article history:

On March 10 and Septembe Received 11 March 2008Received in revised form 14 August 2008Accepted 22 August 2008Available online 5 September 2008

Keywords:United Arab EmiratesArabian–Eurasian collisionRegional waveform inversionSource parameters

r 13, 2007 two earthquakes with moment magnitudes 3.66 and 3.94, respectively,occurred in the eastern part of the United Arab Emirates (UAE). The two events were widely felt in thenorthern Emirates and Oman and were accompanied by a few aftershocks. Ground motions from theseevents were well recorded by the broadband stations of Dubai (UAE) and Oman seismological networks andprovide an excellent opportunity to study the tectonic process and present day stress field acting in this area.In this study, we report the focal mechanisms of the two main shocks by two methods: first motion polaritiesand regional waveform moment tensor inversion. Our results indicate nearly pure normal faultingmechanisms with a slight strike slip component. We associated the fault plane trending NNE–SSW with asuggested fault along the extension of the faults bounded Bani Hamid area. The seismicity distributionbetween two earthquake sequences reveals a noticeable gap that may be a site of a future event. The sourceparameters (seismic moment, moment magnitude, fault radius, stress drop and displacement across thefault) were also estimated from displacement spectra. The moment magnitudes were very consistent withwaveform inversion. The recent deployment of seismic networks in Dubai and Oman reveals tectonic activityin the northern Oman Mountains that was previously unknown. Continued observation and analysis willallow for characterization of seismicity and assessment of seismic hazard in the region.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

The Arabian plate is surrounded by diverse plate boundaries. To theeast and north, the Arabian Plate is colliding with the Eurasian platealong the Zagros and Bitlis sutures (Mckenzie, 1976) and presents oneof the most seismically active continental regions on the Earth (Fig. 1).To the southwest and south the Arabian Plate is bounded by seafloorspreading along the Red Sea, the Gulf of Aden and the Arabian Sea. TheArabian Plate is bounded along the northwestern side by the majorleft lateral strike-slip motion on the Dead Sea Fault. The eastern part ofthe Arabian Plate was affected by numerous intraplate tectonismthroughout Mesozoic and Cenozoic time (Brew et al., 2001).

The Late Cretaceous Semail ophiolite is exposed in northern Omanand northeastern part of the United Arab Emirates (UAE), where itforms the world's largest exposure of oceanic crust and upper mantleemplaced onto continental crust (e.g. Glennie et al., 1974). Highpressure rocks, including carpholite-bearing meta-sediments, garnet-blue-schists and eclogites of continental crustal origin are exposed in

ooqi), kamal_atiya@yahoo.comgahed), elhussain@squ.edu.om@dm.gov.ae (E. Al Khatibi).

l rights reserved.

north-eastern Oman, structurally beneath the ophiolite (e.g. Lippard,1983). Despite almost 100% exposure, factors including variable andcomplex stratigraphy, multiple episodes of deformation and rough,often inaccessible terrane have complicated the geological interpreta-tion of the region. This has led to numerous tectonic models andmuchdebate about the evolution of the Arabian margin in general and thehigh pressure terrane in particular (e.g. Searle et al., 1994, 2004;Gregory et al., 1998; Gray et al., 2000; El-Shazly et al., 2001; Bretonet al., 2004; Warren and Miller, 2007). Most structural and tectonicmodels proposed for the emplacement of the ophiolite and underlyingthrust sheets have involved the NE-directed subduction away from thepassive continental margin of the Arabian plate (e.g. Le Métour et al.,1990; Hanna, 1990; Searle et al., 1994).

Following erosion and subsidence of the obducted mass, a periodof quiet shallowwater carbonate shelf deposition prevailed during theEocene (Alsharhan and Nairn, 1997). A second compressional eventaffected the northeastern and northern margin of the Arabian Plate inthe Oligocene–Miocene as a result of the final closure of themain tractof Neo-Tethys (Glennie et al., 1973). This event continues to thepresent day as a slowcontinent–continent collision responsible for thevast Alpine–Himalayan ranges of which the Zagros Mountains are onepart (Sengör, 1987). The Alpine event produced the SW-vergingthrusts of the Zagros andwest-verging thrusts and associated huge N–

238 Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

S trending folds in the Tertiary limestone cover rocks in the Emiratesand northern Oman (Searle, 1985; Warrak, 1996).

Unfortunately little research has been carried out in the northernOman Mountains on neotectonics. There are no detailed field surveysof the Tertiary faults or assessment of their seismicity. These faultstructures include the Dibba Line (Glennie et al., 1990), the WadiShimal and Wadi Ham Faults (Gnos and Nicolas, 1996; Husky et al.,2005). These and associated faults lie within the Dibba–Masafi–Fujairah area of the northern UAE (Figs. 2, 3).

The local seismic activity of UAE is low. Historically there were noreports of any destructive earthquakes in the country. This could be aresult of poor catalogue completeness with respect to MN5 earth-quakes or their long recurrence time. On March 11, 2002 a moderate(Mw~5) earthquake occurred in Masafi area (Fig. 3) which was

Fig. 1. Tectonic boundaries of the Arabian plate. Seismicity data was compiled from ISC (1964–Harvard CMT solutions for events with MW≥6.0. Solutions are presented with lower hemisphby the square. Major tectonic features are indicated as: AG, Arabian Gulf; AS, Arabian Sea; BSMS, Mediterranean Sea; RS, Red Sea; ZB, Zagros Belt. Dashed red, thin black and heavy blueOphiolite exposures are shown by violet colors. (For interpretation of the references to colo

recorded by the worldwide seismic networks. The uncertainty inUSGS-PDE location for this event was 10–20 km because no localseismic networks were operating at this time (Rodgers et al., 2006a).The event was felt throughout the northern emirates and wasaccompanied by smaller events before and after the March 11 mainshock. A report on these events and accompanying damage wasprovided by Othman (2002). Rodgers et al., (2006a) studied the sourcemechanism of Masafi 2002 event and reported a normal mechanismwith a slight right-lateral strike-slip component consistent with thelarge-scale tectonics. The focal mechanism provided for northeasttrending steeply southeast-dipping normal faults similarly in orienta-tion to an important pair of faults bounded Bani Hamid metamorphicrocks that coalesce northwards and continue north-east through theKhor Fakkan ophiolite block (Gnos and Nicolas, 1996; Musson et al.,

2005) and NEIC (2005–2007) for earthquakes with mb ≥4. Focal mechanisms representere projection and dark quadrant denotes compression. The area of interest is enclosed, Bitlis Suture; CY, Cyprean Arc; DSF, Dead Sea Fault; GA, Gulf of Aden; IO, Indian Ocean;lines represent the rifting, transform and subduction–collision boundaries, respectively.ur in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. Seismicity of the Oman Mountains in the UAE and Oman recoded by Dubai and Oman seismic networks from June 2006 to February 2008. Blue lines are known surface faultscrossing the area. Solid stars represent the gas filed in Dubai near Wadi Nazwa. Major tectonic features are indicated: MK, Makkran; ZFZ, Zenden Fault zone; AS, Arabian Sea; ZB,Zagros Belt. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

239Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

2006). Since 2003, the Earthquake Monitoring Center (EMC) of Omanhas published a yearly bulletin of earthquakes in the region.Information gathered from the EMC bulletin for local earthquakes inUAE shows that there is moderate seismic activity in the northeasternpart of UAE (Al Khatibi et al., 2007). This activity was scattered due tolocation errors as a result of the limited azimuth cover of the Omanistations. By June, 2006 Dubai Municipality installed four broadbandstations in Dubai Emirate, UAE. Dubai and Oman local seismicnetworks are exchanging real-time data which improves the detec-tions and accuracy of earthquake locations in the region. The observedseismic activity by these two networks from June 2006 to February,2008 indicates a clustering of activity in the northern part of UAEalong Masafi-Bani Hamid area and near Wadi Nazwa gas filed, shownin Figs. 2, 3.

UAE has witnessed rapid economic development and high-risebuilding construction in recent years. The majority of the populationoccupies the flat land near the coast that is covered by thicksedimentary deposits. The occurrences of moderate (MW 4–5) earth-quakes at shorter distances (e.g. the 2002 Masafi earthquake)constitute another ground motion hazard besides larger (MWN5)earthquakes along Zagros belt (southwestern Iran). The seismic wavesfrom these earthquakes includewaves in awide range of wavelengths,and soil conditions significantly amplify the waves at periods near thedominant soil periods (Safak, 2008).

On March 10 and September 13, 2007, two felt earthquakesoccurred in the northeastern part of UAE to the northeast of Masafi2002 earthquake (Fig. 3). The estimated average local magnitudes

from the Dubai local network were 4.0 and 4.4 for the two events,respectively. The maximum observed intensity for the two eventsranges from III–IV MM. No injuries were reported but the shaking wasstrong enough to frighten many residents of northern UAE and Oman.Such reactions suggest that amplification of ground motion due tosedimentary structure played a role in these events; because suchmoderate events typically do not cause so much shaking (Safak, 2008;Rodgers et al., 2006b).

In this article,we present analysis of regionalwaveforms recordedbytheDubai andOmannetworks to estimate the sourcemechanisms, focaldepths and other source parameters of March 10 and September 13,2007 UAE earthquakes. The mechanisms and source parameters areuseful for hazard evaluation in theUAEwhichhas developed rapidly andwhose exposure to seismic groundmotion is only nowbeing revealedbythe joint analysis of seismic data from the Dubai and Oman networks.Furthermore, the estimation of the focal mechanisms in this poorlystudied area reveals the nature of faulting and the state of stress in thecrust.

2. Location and polarities focal mechanisms

The two investigated events were located using arrival time picksfrom the Dubai, Oman and Iran seismic networks, utilizing theHYPOCENTER location algorithm (Lienert et al., 1988). The digitalwaveform data were band-limited between 0.8 Hz to 20 Hz beforearrival timeswere picked. In total, 23 (18 P and 5 S) and 17 (11 P and 6 S)crustal phases from all stations within maximum epicentral distances

Fig. 3. Topographic map of the study area. Focal mechanisms of the two studiedearthquakes of March and September, 2007. The location (red star) and focalmechanism of March 2002 are also shown (Rodgers et al., 2006a). Local seismicityrecorded by Dubai and Oman networks from June 2006 to February 2008. Dashed linerepresents the suggested extension of the active NNE–SSW Bani Hamid fault.

Table 2Arabian platform crustal model (Rodgers et al., 1999)

Thickness VP VS ρ

km km/s km/s gm/cm3

4 4.0 2.31 2.616 6.2 3.64 2.820 6.4 3.70 3.0∞ 8.1 4.55 3.2

240 Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

of 1000 kmwere used in the locations, respectively. The closest stationwas ~50 km from their epicenters, and the greatest azimuthal gaps instation coverage were 126° and 180°, respectively. The locationparameters are listed in Table 1. The RMS travel time residuals in thehypocentral estimate for the studied events were 0.6 s and 0.4 s,respectively. The two events were located on Khor Fakkan block alongthe suggested fault branching to northeast from Wadi Ham faultbounding Bani Hamid metamorphic rocks and extending northeast(Fig. 3). These events were followed by few aftershocks that do notclarify any fault geometry due to the limited number of seismicstations, high noise level and poor azimuth coverage.

Seismic locations obtained from merging seismic phase arrivaltime data from the Dubai and Oman seismic networks suggests anearly NE–SW trend starting north of Bulaydah–Bani Hamid andtrending towards the north-east and probably extending offshore intothe Gulf of Oman (Fig. 3). The seismicity distribution along this trendshows a noticeable gap that may be a site of a future event. A recentincrease of activity in February, 2008 (10 events with local magnitudesranging from 2.3 to 3.8) suggests that large events are more likely tohappen than what was previously believed.

Polarities of first arriving P-waves compiled for 10 and 12 localstations were used to determine the focal mechanism for these twoevents, respectively. A double-couple, fault plane solution was fit to

Table 1Hypocentral parameters of the two studied events

Date Origin Time Location

Day Mo. Yr. H. Mn. Sec. Lat.

10 03 2007 18 15 14.3 25.30314.4 25.222

13 09 2007 15 47 07.0 25.46908.7 25.460

⁎, ⁎⁎ and ⁎⁎⁎: locations of this study, NEIC and CSEM, respectively. f: fixed focal depth.

the observations using FOCMEC (Snoke, 2003). The 1-D velocity modelTable 2 and locations in Table 1 were used to determine takeoff angles.The resulting solutions (Fig. 4) indicate normal faulting for both eventstrending NNE–SSW and NE–SW. Of the two nodal planes, ourpreferred fault plane is the NNE–SSW south-easterly dipping whichis consistent with surface faults and large-scale tectonics. The faultparameters of the focal mechanism based on polarities solution arelisted in Table 3.

3. Regional waveform inversion

Because the source region is of a relatively low seismicity, themajority of earthquakes recorded by local and regional seismicnetworks along northeastern UAE are too small to be included inthe Harvard CMT. The occurrences of these earthquakes areparticularly important for characterizing regional tectonics andconstraining stress orientation. Focal mechanisms based on short-period P-wave polarities are possibly biased by errors in the crustalstructure, focal depth and poor azimuth coverage. Regional waveforminversion (e.g. Ichinose et al., 2003; Horton et al., 2004; Rodgers et al.,2006a; Abou Elenean and Hussein, 2007) rely on both P- and S-waveradiation and constrain the moment and depth as well as theorientation of faulting.

Generally, the waveform inversion is done in the lower frequencybandbetween0.01 to 0.12Hz. The use of seismicwaves at longer periodsimproves the estimation of earthquake source parameters because theyare relatively insensitive to the effects of lateral velocity and densityheterogeneities (Ritsema and Lay, 1995). Regional waveform inversionstypically provide good results for closer broadband stations (b1000 km)with good signal to noise ratio (Abou Elenean and Hussein, 2007).

The two investigated events are recorded by the broadbandseismographic stations in UAE and Oman. Five broadband (three-component) stations (Fig. 2) in the distance range from 50 to 107 kmare used to determine the focal mechanism and focal depth. Choice ofthe records is mainly based on clarity of the records (signal to noisespectral ratios N9). Instrument responses are removed and the twohorizontal components, NS and EW, are converted to R (radial) and T(transverse). The Arabian Platform 1-D crustal model (Rodgers et al.,1999, Table 1) is used to compute Green's functions in this study. Thequality factor Q is assigned according to Swanger's law “Qs=Vs/10”where Vs is the shear velocity in m/s and QP≈9/4 Qs (Mancilla, 2001).

The matrix inversion method (following Ichinose, 2006) is used toinvert for the point source moment tensor using Green's functionscomputed for a 1-D velocitymodel. The 1-D velocitymodelmay not beperfect and small time shift is required for maximum correlation

H Er. (km) ML Ref.

Long. (km) X Y Z

56.191 4.2 04 03 05 4.0 ⁎

56.074 10 f 6.6 5.7 – 4.1 ⁎⁎

56.297 8.2 03 2.2 04 4.4 ⁎

56.200 20 f 12.5 9.9 – 4.5 ⁎⁎⁎

Fig. 4. Focal mechanisms based on the polarities of P-wave of two studied earthquakes(A: 10/03/2007 and B: 13/09/2007). Lower hemisphere equal area projection is used.

Table 3The focal mechanism and source parameters for March 10, 2007 and September 13,2007 main shocks using polarities and regional waveform inversion

Eventindex

Fault plane Mw Mo× 1021 dyne-cm H Method

St.° Dip° Rake° (km)

10 03 2007 41° 51° −103° – – 06 POL37° 39° −108° 3.66 3.79 06 INV

13 09 2007 25° 65° −101° – – 08 POL13° 63° −115° 3.94 9.98 08 INV

POL: focal mechanism using polarities method.INV: focal mechanism and source parameters using regional waveform inversion.

241Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

between the observed and synthetics. In the matrix inversiontechnique, the shift is not known and it becomes a nonlinear modelparameter which has to be inverted along with the moment tensorand optimal source depth. To account for the horizontal mislocation,the synthetics were shifted relative to the observed by changing theorigin time few seconds before/after location origin time during ourinversion. The inversion is done imposing a deviatoric moment tensorwithout isotropic component. The Green's function is computed usinga fast reflectivity and frequency–wave number (f–k) summationtechnique (Zeng and Anderson, 1995). The details of the methodologywere described in Ichinose and Zeng (2000) and Ichinose et al. (2003).The preferred solution was obtained by a simple grid search over thefocal depths between 1 and 30 km with 1 km step and also over theorigin time between 2-s before/after the location origin time, since theorigin time trades off with focal depth. The solution that has a largepercentage of variance reduction and double couple component isselected. The percentage of variance reduction is found by:

VR ¼ 1−∑i

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffidatai−synthið Þ2

qffiffiffiffiffiffiffiffiffiffiffiffidata2i

q264

375T100 ð1Þ

Where data and synth are the data and synthetic time seriesrespectively and the summation is performed for all stations andcomponents. Fig. 5 shows the comparison between the observed and

synthetic waveforms for the preferred solution with high percentageof variance reduction and double couple component of March 10,2007 earthquake. The fit to the observed waveforms is quite good(variance reduction N70%). The optimal origin time from our inversion(14.0 s) is consistent with the location origin time seconds (14.3 s)indicating the adequacy of the 1-D model for the lower frequencyband of our inversion. The focal mechanism was stable over widerange of focal depths and represented a normal fault with a slightstrike slip component in good agreement with the first motionpolarity focal mechanism solution. The focal mechanism parametersare listed in Table 3. Fig. 6 illustrates the changes in variance reductionand source mechanism as a function of focal depth.

Fig. 7 shows a comparison between the observed and syntheticwaveforms for the preferred solutionwith high percentage of variancereduction and double couple component of September 13, 2007earthquake. The optimal origin time seconds (7 s) is also consistentwith the location origin time seconds and the fit is satisfactory. Thefocal mechanism of this event was also stable over a wide range offocal depths and the preferred solution also showed normal faulttrending NNE with a slight strike slip component. Fig. 8 illustrates thechanges in variance reduction (expressed in percentage) and sourcemechanisms as a function of focal depth. The best-fitting focalmechanism parameters are listed in Table 3. The focal mechanismsestimated from these two recent events are similar with that of theMarch 11, 2002 event (Rodgers et al., 2006a).

4. Source parameters

Source parameters are useful in micro-zonation and the assess-ment of seismic hazard. The seismic moment (Mo), fault radius (ro),average displacement across the fault (do), stress drop (Δσ) and themoment magnitude (Mw) were determined for the two events usingthe P-wave far-field amplitude displacement spectra. The analyzeddata consist of four broadband stations that overlie hard rock sites andhave good signal to noise ratio. We applied our analysis to the verticalcomponents only and we ignore the UAE stations that overliesedimentary cover to avoid possible site effects and spectral complex-ity. The analyzed stations are located within the epicentral distancerange 50 km to 135 km. A cosine taper was applied to the instrumentcorrected, de-meaned signal window and the Fourier amplitudespectrum was computed. Signal windows of varying length weretested in order to select a length that would avoid contamination fromother phases and maintain the resolution and stability of the spectra.A selected signal windows ranges from 6 to 8 s.

FollowingBrune (1970,1971), the farfielddisplacement spectrad(f) is:

d fð Þ ¼ G r; hð ÞTD fð ÞTMoTK

4πρv3cT

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1þ f

fo

� �2γr ð2Þ

where G(r, h) is geometrical spreading which is a function of bothepicentral distance (r) and focal depth (h). D(f) is the diminutionfunction due to anelastic attenuation that includes path and near

Fig. 5. Comparison between observed (solid lines) and synthetic (dashed lines) displacement waveforms for the preferred focal mechanism deduced from the regional waveforminversion of 10/03/2007 main shock. The epicentral distance (km) and azimuth (degree) of each station is written above their trace. Z, R, T indicates vertical, radial and transversecomponents consequently. Time scale in s and amplitude scale in μm are located to the bottom and left of the T component.

242 Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

surface attenuation, f is the frequency, fo is the corner frequency and γis the source spectral fall off. While ρ is the density, Vc is the velocity ofeither shear wave or compressional wave and K is a factor to correctfor the free surface effect and radiation pattern. In our analysis, thegeometrical spreading has been defined according to Herrmann andKijko (1983) relation. Meanwhile, the anelastic attenuation e−πft/Qc(f)with Qc(f) being the quality factor, t is the travel time was taken intoaccount by a quality factor QP of 648 that presents the average alongthe path.

Fig. 6. Percent of variance reduction and source mechanism as a function of focal depthsresulted from regional waveform inversion for 10/03/2007main shock. Mechanisms areplotted each 2 km only except the first 5 km. Preferred solution is labeled above by itsmoment magnitude. Dark arrow indicates the preferred focal depth.

The attenuation of seismic wave near the site is commonlyaccounted for by e−πfk (Singh et al., 1982). Assuming an omega-squareBrune's source model, the k parameter, low frequency spectralamplitude Ωo and corner frequency fo are estimated using thenonlinear least square inversion techniques. The estimated value ofk parameter ranges from 0.01 to 0.03. For circular source model theseismic moment, fault radius (ro), displacement across the fault andthe moment magnitude can be derived from the P-wave displacementspectra following Brune's (1970, 1971), Hanks and Wyss (1972) andKanamori (1977) relations:

Mo ¼ 4πρv3PFTRθ/TG r;hð Þ ð3Þ

ro ¼ 2:34VP

2πfoð4Þ

Δσ ¼ 7Mo

16r3oð5Þ

do¼ Mo

πρV2S r

2o

ð6Þ

Mw ¼ 2=3log10 Moð Þ−10:73 ð7Þ

where ρ, VP and VS are picked from Table 2 according to the sourcedepth. F, free surface effect is estimated for individual station based onVP/VS and the emergence angles of P and S-wave using the focmecpackage (Snoke, 2003). For the radiation pattern effect an averagevalue of Rθϕ of 0.52 was assumed (Boore and Boatwright, 1984). Thesource parameters obtained for the two recent events are listed atTables 4 and 5. The fits to the observed displacement spectra at somestations for both events are plotted in Figs. 9 and 10, respectively.

ΩoT

Fig. 7. Comparison between observed (solid lines) and synthetic (dashed lines) displacement waveforms for the preferred focal mechanism deduced from the regional waveforminversion of 13/09/2007 main shock. The epicentral distance (km) and azimuth (degree) of each station is written above their trace. Z, R, T indicates vertical, Radial and transversecomponents consequently. Time scale in s and amplitude scale in μm are located to the bottom and left of the T component.

243Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

For each event, the average values ⟨x⟩were computed for seismicmoment, source radius, stress drop, average displacement andmoment magnitude Table 6. Calculation was made following Arch-uleta et al. (1982):

hxi ¼ anti log1Ns

∑Ns

i¼1logxi

!ð8Þ

where Ns is the number of stations used. This procedure gives equalweight to each observation. In the case of simple arithmetic average,the result would be biased toward larger values. Another reason is that

Fig. 8. Percent of variance reduction and source mechanism as a function of focal depthsresulted from regional waveform inversion for 13/09/2007main shock. Mechanisms areplotted each 2 km only and the preferred solution is labeled above by its momentmagnitude. Dark arrow indicates the preferred focal depth.

the errors associated with Ωo and ro are log-normally distributed(Garcia-Garcia et al., 1996). The standard deviation of the logarithm,SD[log⟨x⟩], and a multiplicative error factor, Ex, were calculated as:

SD logbxNð Þ ¼ 1Ns−1

∑Ns

i¼1logxi−loghxið Þ2

" #12

ð9Þ

Ex ¼ anti log SD loghxi½ �ð Þ ð10Þ

The calculated source parameters of the first event of March 10,2007 are relatively consistent which indicates a homogenous rupture.Meanwhile the parameters of the second event of September 13, 2007show a slight scatter that may be related to a directivity effect ofrupture towards NNE (corner frequency and stress drop are relativelylarger at BAN station compared with other stations). The Mw at ASUstation to the south is also larger than the average probably due to aslight site effect relative to the others.

5. Discussion

Estimation of accurate hypocentral parameters and focal mechan-ism can provide important information about the slip, fault structureat depth and the stress field in seismically active areas. In this study,we used broadband records from the Dubai and Oman seismicnetworks to obtain a precise location and focal mechanism of two

Table 4Source parameters of March 10, 2007 main shock deduced from P-wave spectra

St.Code

Δ(km)

Az° fo(Hz)

Ωo cm⁎s Mo

(dyne-cm)ro(m)

Δσ(bar)

do(cm)

Mw

BAN 69.70 9 2.72 4.41e-05 4.00×1021 843 2.91 0.53 3.67HAT 53.40 186 3.03 6.20e-05 4.20×1021 755 4.13 0.70 3.68ASH 70.10 191 2.94 4.13e-05 3.85×1021 780 3.55 0.60 3.66ASU 114.9 229 3.79 2.03e-05 3.66×1021 604 7.25 0.95 3.64

Table 5Source parameters of September 13, 2007 main shock deduced from P-wave spectra

St.Code

Δ(km)

Az° fo(Hz)

Ωo cm⁎s Mo

(dyne-cm)ro(m)

Δσ(bar)

do(cm)

Mw

BAN 50.50 0 3.60 1.75e-04 13.4×1021 635 22.90 3.16 4.02HAT 73.40 193 2.55 9.26e-05 8.82×1021 896 5.36 1.04 3.90ASH 90.50 195 2.29 4.83e-05 5.73×1021 998 2.52 0.54 3.77ASU 135.06 226.5 2.77 9.67e-05 19.6×1021 826 15.22 2.73 4.13

244 Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

moderate size but felt events of March 10 and September 13, 2007. Thehypocenters of the two events were located in the Khor Fakkan blockbetween Wadi Shimal Fault and pair of faults branching to thenortheast from the Wadi Ham Fault (Gnos and Nicolas, 1996).

The focal mechanisms of the two investigated events determinedby two methods (polarities of P-wave and regional waveforminversion) indicate normal faulting mechanisms with slight strikeslip component. Because the dominant surface faults crossing the areaare trending NNE and south-easterly dipping, we argue the NNE planewith slight right lateral strike slip movement from our solution as thefault plane. The right-lateral strike-slip component of the mechanismis consistent with right-lateral motion along the Oman Line in Iran(e.g. Kadinsky-Cade and Barazangi, 1982; Rodgers et al., 2006a). The

Fig. 9. Displacement spectra from four selected stations for 10/03/2007 main shock. The staspectrum while the red indicates the fitted omega-square source model curve. (For interpreversion of this article.)

Oman Line defines the boundary between continental collision of theArabian and Eurasian Plates along the Zagros Thrust and oceanic–continental convergence in the Makkran region. Convergence alongthe Zagros Thrust is much faster than along the Makkran, leading todextral motion along the Zendan Fault Zone (see Fig. 2, inset). Oursolutions are in good agreement with the solutions of March 11, 2002Masafi (Rodgers et al., 2006a, Fig. 3). Relying upon the similarities offocal mechanisms of the three events and uncertainties on location of2002 earthquakewe argue that they are related to the same suggestedNNE–SSW fault (Fig. 3). The normal component is dominant and isconsistent with brittle extension of the Khor Fakkan Block, a massif ofmainly Semail peridotite bounded to the west and northwest by theWadi Shimal Fault, and bounded to the southwest by the Wadi HamFault (Rodgers et al., 2006a). Searle and Cox (1999) represent the twonortheast trending faults at Bulaydah as reverse faults. Meanwhile,Searle (1988), Boote et al. (1990) and Rodgers et al. (2006a) suggesteda reactivation of the NNE reverse faults as normal faulting in thepresent day stress field.

Harvard CMTsolutions for larger earthquakes (Mw ≥6, Fig.1) to thenorth of UAE along the Zagros Belt show a dominant NNW to NWthrust faults with right-lateral strike slip component. The P-axis ofthese solutions generally trends ~NNE. In the present stress field,where the maximum horizontal stress is oriented ~NNE, roughly

tion code, the distance in km and azimuth in degrees are given. Blue line indicates thetation of the references to colour in this figure legend, the reader is referred to the web

Fig. 10. Displacement spectra from four selected stations for 13/09/2007 main shock. The station code, the distance in km and azimuth in degrees are given. Blue line indicates thespectrum while the red indicates the fitted omega-square source model curve. (For interpretation of the references to colour in this figure legend, the reader is referred to the webversion of this article.)

245Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

parallel faults will be essentially normal, whereas roughly perpendi-cular ones are thrust faults. Other faults in intermediate directions willentail predominant strike-slip movements.

6. Conclusions

We estimated the focal mechanisms of two felt events thatoccurred on March 10 and September 13, 2007 using P-wave polarityand regional waveform inversion techniques. Results indicate anormal faulting mechanism with slight right-lateral strike-slipcomponent. The normal faulting mechanism is consistent with

Table 6Mean value (MV), standard deviation (SD) and multiplicative error factors (ME) of theestimated seismic moment, Mo, the fault radius, r, the stress drop, Δσ, averagedisplacement, do and moment magnitude, Mw

Main shock of March 10, 2007 Main shock of September 13, 2007

Mo

(dyne-cm)ro(m)

Δσ(bar)

do

(cm)Mw Mo

(dyne-cm)ro(m)

Δσ(bar)

do

(cm)Mw

MV 3.92×1021 739 4.19 0.67 3.67 10.73×1021 827 8.28 1.48 3.95SD 0.02 0.06 0.17 0.10 0.00 0.23 0.08 0.43 0.36 0.01ME 1.06 1.15 1.48 1.28 1.00 1.69 1.21 2.72 2.30 1.04

relaxation of the Khor Fakkan Block. The slight right-lateral strike-slip component is consistent with dextral motion across the Omanline. The event provides constraints on active tectonics in therelatively aseismic northern UAE and Oman Mountains.

The seismic activity revealed by Dubai and Oman networksconstitute a nearly NE–SW trends starting north of Bulaydah–BaniHamid toward north-east and probably extending offshore into theGulf of Oman and showing a noticeable gap (~10 km length) that maybe a site of a future event. Assuming a rupture depth ~10 km, this gapcould result in ~Mw=6.0 earthquake. This magnitude value isestimated according to Wells and Coppersmith (1994) equation:

Mw ¼ 4:07þ 0:98log10a km2� � ð11Þ

Where a is the fault area in km squared. A similar value is obtainedaccording to Ambraseys and Jackson (1998) equation which relate thefault length with surface wave magnitude. Based on the clustering ofsmaller earthquakes, NE fault plane of the focal mechanisms and theelevation difference (Fig. 3) we confirm the NE extension of this faultwhich was previously suggested by Musson et al. (2006) based onsatellite and total magnetic intensity maps. The recent increase ofactivity in February, 2008 (10 events with local magnitudes rangesfrom 2.3 to 3.8) suggests that the region is more active and capable oflarger events than previously believed.

246 Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

The seismic moment and moment magnitude of March 10earthquake based on the regional waveform inversion are 3.79×1021

dyne-cm and 3.66, respectively, which are nearly similar to the valuesderived from the displacement spectra 3.92×1021 dyne-cm and 3.67,respectively. The estimated fault length, displacement across the faultand stress drop based on displacement spectra are 739m, 0.67 cm and4.19 bar, respectively.

Meanwhile, the seismic moment and moment magnitude ofSeptember 13 earthquake based on the regional waveform inversionare 9.98×1021 dyne-cm and 3.94, respectively, which have goodagreement with the values derived from the displacement spectra10.73×1021 dyne-cm and 3.95, respectively. The estimated faultlength, displacement across the fault and stress drop for this eventare 827 m, 1.48 cm and 8.28 bar, respectively.

Our stress drop values for the studied events are smaller than usualvalues of type III intraplate earthquakes (midplate) but similar to II(intraplate, plate boundary related) type of Scholz (1990). Thisobservation may be related to the tectonic situation of UAE andOman closer to themajor Arabian–Eurasian plate boundary that mightreactivate previous weakness zones. These earthquakes causedconsiderable alarm in the northern Emirates and highlight the factthat damaging earthquakes can occur in this region. The deploymentof modern seismic networks in the UAE and Oman allow detailedstudy of earthquakes in the region and enable improvement ofknowledge of tectonic processes and seismic hazard.

Acknowledgments

The authors would like to thank the staff of EarthquakeMonitoringCenter of Oman and Dubai Municipality for providing us with thedigital records of the two events investigated. We are also grateful toProf. G. Ichinose, MTC/AFTAC Technology for offering us the regionalmoment tensor inversion codes.

References

Abou Elenean, K.M., Hussein, H.M., 2007. Source mechanism and source parameters ofMay 28, 1998 earthquake. Egypt. J. Seismol. 11, 259–274.

Al Khatibi, E., Al Marzooqi, Y., Megahed, A., Mathias, F., Mallalah, A., 2007. Preliminaryseismicity of United Arab Emirates deduced from Dubai Seismic Network, FourthGulf Seismic Forum 24–27 March, 2007, Kuwait.

Alsharhan, A.S., Nairn, A.E., 1997. Sedimentary Basins and Petroleum Geology of theMiddle East. Elsevier Science, Amsterdam. 940 pp.

Ambraseys, N.N., Jackson, J.A., 1998. Faulting associated with historical and recentearthquakes in the Eastern Mediterranean region. Geophys. J. Int. 133, 390–406.

Archuleta, R.J., Cranswinck, E.,Mueller, C.H., Spudich, P.,1982. Sourceparametersof the 1980Mammoth lakes, California, earthquakes sequence. J. Geophys. Res. 87, 4595–4607.

Boore, D.M., Boatwright, J., 1984. Average body-wave radiation coefficients. Bull.Seismol. Soc. Am. 74 (5), 1615–1621.

Boote, D.R., Mou, D., Waite, R.I., 1990. Structural evolution of the Suneinah Foreland,central Oman Mountains. In: Robertson, A.H.F., Searle, M.P., Ries, A.C. (Eds.), TheGeology and Tectonics of the Oman Region. Geological Society Special Publication,vol. 49, pp. 397–418.

Breton, J.P., Béchennec, F., Le Métour, J., Moen-Maurel, L., Razin, P., 2004. Eoalpine(Cretaceous) evolution of the Oman Tethyan continental margin: insights from astructural field study in Jebel Akhdar, Oman Mountains. GeoArabia 9, 41–58.

Brew, G., Barazangi, M., Al-Maleh, A.K., Sawaf, T., 2001. Tectonic and geologic evolutionof Syria. GeoArabia 6, 573–616.

Brune, J.N., 1970. Tectonic stress and the spectra of seismic shear waves from earthquakes.J. Geophys. Res. 75, 4997–5009.

Brune, J.N., 1971. Correction. J. Geophys. Res. 76, 5002.El-Shazly, A.K., Brocker, M., Hacker, B., Calvert, A., 2001. Formation and exhumation of

blueschists and eclogites fromNEOman: new perspectives fromRbeSr and 40Ar/39Ardating. J. Metamorph. Geol. 19, 233–248.

Garcia-Garcia, J.M., Vidal, F., Romacho, M.D., Martin-Marfil, J.M., Posadas, A., Luzon, F.,1996. Seismic source parameters for microearthquakes of the Granada basin(southern Spain). Tectonophysics 261, 51–66.

Glennie, K.W., Boeuf, M.G.A., Hughes Clarke, M.W., Moody-Stuart, M., Pilaar, W.F.H.,Reinhardt, B.M., 1973. Late Cretaceous nappes in Oman Mountains and theirgeologic evolution. Bull. Am. Assoc. Pet. Geol. 57, 5–27.

Glennie, K.W., Boeuf, M.G., Hughes-Clarke, M.H.W., Moody-Stuart, M., Pilaar, W.F.,Reinhardt, B.M., 1974. Geology of the Oman Mountains. Verhandelingen KoninklijkNederlandsGeologischMijnbouwkundidgGenootschap Transactions, vol. 31, p. 423.

Glennie, K.W., Hughes Clarke, M.W., Boeuf, M.G.A., Pilaar, W.F.H., Reinhardt, B.M., 1990.Inter-relationship ofMakran–OmanMountains belts of convergence. In: Robertson, A.

H.F., Searle, M.P., Ries, A.C. (Eds.), The Geology and Tectonics of the Oman Region.Geological Society Special Publication, vol. 49, pp. 773–786.

Gnos, E., Nicolas, A., 1996. Structural evolution of the northern end of the OmanOphiolite and enclosed granulites. Tectonophysics 254, 111–137.

Gray, D.R., Gregory, R.T., Miller, J.M., 2000. A new structural profile along theMuscat-Ibratransect, Oman: implications for emplacement of the Samail ophiolite. In: Dilek, Y.,Moores, E.M., Elthon, D., Nicholas, A. (Eds.), Ophiolites and Oceanic Crust: NewInsights from Field Studies and the Ocean Drilling Program. Special Paper, vol. 349.Geological Society, America, pp. 513–523.

Gregory, R.T., Gray, D.R., Miller, J.M., 1998. Tectonics of the Arabian margin associatedwith the formation and exhumation of high-pressure rocks, Sultanate of Oman.Tectonics 17 (5), 657–670.

Hanks, T.C., Wyss, M., 1972. The use of body-wave spectra in the determination ofseismic-source parameters. Bull. Seismol. Soc. Am. 62, 561–589.

Hanna, S., 1990. The alpine deformation of the Central Oman Mountains. In: Robertson,A.F.H., Searle, M.P., Ries, A.C. (Eds.), The Geology and Tectonics of the Oman Region.Special Publication, vol. 49. Geological Society, London, pp. 341–359.

Herrmann, R., Kijko, A.,1983.Modelling some empirical vertical component Lg relations.Bull. Seism. Soc. Am. 73, 157–171.

Horton, S., Kim, W.-Y., Withers, M., 2004. The 6 June 2003 Bardwell, Kentucky,earthquake sequence: evidence for a locally perturbed stress in the Mississippiembayment. Bull. Seism. Soc. Am. 95, 431–445.

Husky, T., Robinson, C., El-Baz, F., 2005. Tertiary-Quaternary faulting and uplift innorthern Oman Hajar Mountains. J. Geol. Soc. Lond. 162, 871–888.

Ichinose, G., 2006. Moment Tensor Inversion Toolkit (MTINV) Version 0.9, ReleasedNovember, 2006. http://www.seismo.unr.edu/htdocs/students/Ichinose/.

Ichinose, G., Zeng, Y., 2000. Waveform Modeling and Moment Tensor Inversion Tools.http://www.seismo.unr.edu/htdocs/students/Ichinose/WaveformTools.

Ichinose, G., Anderson, J., Smith, K., Zeng, Y., 2003. Source parameters of EasternCalifornia andWestern Nevada Earthquake from regional moment tensor inversion.Bull. Seism. Soc. Am. 93, 61–84.

Kadinsky-Cade, K., Barazangi, M., 1982. Seismotectonics of southern Iran: the Omanline. Tectonics 1, 389–412.

Kanamori, H.,1977. Theenergy release ingreat earthquakes. J. Geophys. Res. 82,1981–1987.Le Métour, J., Rabu, D., Tegyey, M., Bechennec, F., Beurrier, M., Villey, M., 1990.

Subduction and obduction: two stages in the Eo-Alpine tectonometamorphicevolution of the Oman mountains. In: Robertson, A.F.H., Searle, M.P., Ries, A.C.(Eds.), The Geology and Tectonics of the Oman Region. Special Publication, vol. 49.Geological Society of London, pp. 327–339.

Lienert, B., Berg, E., Frazer, L., 1988. HYPOCENTER, an earthquake location methodusing centered, scaled and adaptively least squares. Bull. Seismol. Soc. Am. 76,771–783.

Lippard, S.J., 1983. Cretaceous high pressure metamorphism in NE Oman and itsrelationship to subduction and ophiolite nappe emplacement. J. Geol. Soc. Lond.140, 97–104.

Mancilla, P., 2001: Surface wave dispersion about the NewMadrid Region, M. Sc. Thesis,Faculty of the Graduate School of Saint Louis University, USA.

McKenzie, D., 1976. The East Anatolian fault; a major structure in eastern Turkey. EarthPlanet. Sci. Lett. 29, 189–193.

Musson, R.W., Northmore, K.J., Sargeant, S.L., Phillips, E.R., Boon, D., Long, D., McCue, K.,Ambraseys, N.N., 2006. The geology and geophysics of the United Arab Emirates.Geological Hazards, vol. 4. Ministry of Energy, UAE, p. 239.

Othman, A., 2002. Study of the effects of the Fujairah earthquake, March 2002. UnitedArab Emirates University, Al-Ain, UAE. Technical Report (in Arabic). 58 pp.

Ritsema, J., Lay, T., 1995. Long period regional waves moment tensor inversion forearthquakes in the western United States. J. Geophys. Res. 100, 9853–9864.

Rodgers, A., Walter, W., Mellors, R., Al-Amri, A., Zhang, Y., 1999. Lithospheric structure ofthe Arabian Shield and Platform from complete regional waveform modeling andsurface wave group velocities. Geophys. J. Int. 138, 871–878.

Rodgers, A., Fowler, A., Al-Amri, A., Al-Enezi, A., 2006a. The March 11, 2002 Masafi,United Arab Emirates earthquake: insights into the seismotectonics of the northernOman Mountains. Teconophysics 415, 57–64.

Rodgers, A., Tkalcic, H., McCallen, D., Larsen, S., Snelson, C., 2006b. Seismic groundmotion and site response in Las Vegas Valley, Nevada from NTS explosions. PureAppl. Geophys. 163, 55–80.

Safak, E., 2008. Assessment of seismic hazard in Dubai, United Arab Emirates, projectsp323, internal report, Dubai Municipality, Survey Department, UAE.

Scholz, C.H., 1990. The mechanics of earthquake faulting. Cambridge University Press,pp. 326–329.

Searle, M.P., 1985. Sequence of thrusting and the origin of culminations in the northernand central Oman Mountains. J. Struct. Geol. 7, 129–143.

Searle, M.P., 1988. Thrust tectonics of the Dibba zone and the structural evolution of theArabian continental margin along the Musandam Mountains (Oman and UnitedArab Emirates). J. Geol. Soc. Lond. 145, 43–53.

Searle, M., Cox, J., 1999. Tectonic setting, origin and obduction of the Oman ophiolite.Bull. Geol. Soc. Am. 111, 104–122.

Searle, M.P., Waters, D.J., Martin, H.N., Rex, D.C., 1994. Structure and metamorphism ofblueschist-eclogite facies rocks from the northeastern Oman mountains. J. Geol.Soc. Lond. 151, 555–576.

Searle, M.P., Warren, C.J., Parrish, R.R., Waters, D.J., 2004. Structural evolution,metamorphism and restoration of the Arabian continental margin, Saih Hatatregion, Oman Mountains. J. Struct. Geol. 26, 451–473.

Sengör, A.M.C., 1987. Tectonics of the Tethy sides: orogenic collage development in acollisional setting. Annu. Rev. Earth Planet. Sci. 15, 213–244.

Singh, S.K., Apsel, R.J., Fried, J., Brune, J.N., 1982. Spectral attenuation of SH-waves alongthe Imperial fault. Bull. Seismol. Soc. Am. 72, 2003–2016.

247Y. Al Marzooqi et al. / Tectonophysics 460 (2008) 237–247

Snoke, A., 2003. Focal mechanism determination software (FOCMEC package). http://www.geol.vt.edu/outreach/vtso/focmec.

Warrak, M., 1996. Origin of the Hafit structure: implications for timing the Tertiarydeformation in the northern Oman Mountains. J. Struct. Geol. 18, 803–818.

Warren, C.J., Miller, J. Mcl., 2007. Structural and stratigraphic controls on the origin andtectonic history of a subducted continental margin, Oman. J. Struct. Geol. 29,541–558.

Wells, D.L., Coppersmith, K.J., 1994. New empirical relationships among magnitude,rupture length, rupture width, rupture area, and surface displacement. Seismol.Soc. Am. Bull. 84 (4), 974–1002.

Zeng, Y., Anderson, J., 1995. A method for direct computation of the differentialseismogram with respect to the velocity change in a layered elastic solid. Bull.Seismol. Soc. Am. 85, 300–307.