Journal of Asian Earth Sciences 120 (2016) 17–28
Contents lists available at ScienceDirect
Journal of Asian Earth Sciences
journal homepage: www.elsevier.com/locate/jseaes
Imaging of magma intrusions beneath Harrat Al-Madinah
in Saudi Arabia
Mohamed F. Abdelwahed a,e,⇑, Nabil El-Masry a,f, Mohamed Rashad Moufti a, Catherine Lewis Kenedi b,
Dapeng Zhao c, Hani Zahran d, Jamal Shawali d
a
Geohazards Research Center, King Abdulaziz University, Saudi Arabia
School of Environment, University of Auckland, Auckland, New Zealand
Department of Geophysics, Tohoku University, Sendai, Japan
d
Saudi Geological Survey (SGS), Saudi Arabia
e
National Research Institute of Astronomy and Geophysics, NRIAG, Egypt
f
Geology Department, Faculty of Science, Suez Canal University, Ismailia 41522, Egypt
b
c
a r t i c l e
i n f o
Article history:
Received 10 December 2015
Received in revised form 24 January 2016
Accepted 26 January 2016
Available online 27 January 2016
Keywords:
Seismic tomography
Magma intrusions
Harrat Rahat
Al-Madinah
Saudi Arabia
a b s t r a c t
High-resolution tomographic images of the crust and upper mantle beneath Harrat Al-Madinah, Saudi
Arabia, are obtained by inverting high-quality arrival-time data of local earthquakes and teleseismic
events recorded by newly installed borehole seismic stations to investigate the AD 1256 volcanic eruption and the 1999 seismic swarm in the study region. Our tomographic images show the existence of
strong heterogeneities marked with low-velocity zones extending beneath the AD 1256 volcanic center
and the 1999 seismic swarm area. The low-velocity zone coinciding with the hypocenters of the 1999
seismic swarm suggests the presence of a shallow magma reservoir that is apparently originated from
a deeper source (60–100 km depths) and is possibly connected with another reservoir located further
north underneath the NNW-aligned scoria cones of the AD 1256 eruption. We suggest that the 1999 seismic swarm may represent an aborted volcanic eruption and that the magmatism along the western margin of Arabia is largely attributed to the uplifting and thinning of its lithosphere by the Red Sea rifting.
Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction
Large areas on the western margin of the Arabian plate are covered with intraplate basaltic volcanic fields that are commonly
related to the upwelling of the Afar mantle plume (30 Ma) and
the consequent formation of the East African, Gulf of Aden, and
Red Sea rift systems (Coleman et al., 1983; Bosworth et al., 2005;
Pallister et al., 2010; Chang et al., 2011). Thirteen of these volcanic
fields, known locally as the harrat, are located in western Saudi
Arabia (Coleman et al., 1983). One of these volcanic fields is Harrat
Rahat. Its importance stems from the fact that its northern extremities extend to the south-eastern and eastern suburbs of the holy
city of Al-Madinah Al-Munawarah.
Harrat Rahat forms a 50–75 km-wide, 300 km-long, N20–25°Wtrending plateau (Durozoy, 1970; Berthier et al., 1981; Coleman
et al., 1983). It comprises four coalesced smaller volcanic fields
⇑ Corresponding author at: Geohazards Research Center, King Abdulaziz
University, Saudi Arabia.
E-mail
addresses:
[email protected],
[email protected]
(M.F. Abdelwahed).
http://dx.doi.org/10.1016/j.jseaes.2016.01.023
1367-9120/Ó 2016 Elsevier Ltd. All rights reserved.
attaining an approximate area of 19,830 km2 and an estimated volume of 1999 km3 (Camp and Roobol, 1989). In previous studies, the
northern part of Harrat Rahat was referred to as Harrat Al-Madinah
(Coleman et al., 1983; Moufti, 1985; Moufti et al., 2012). In this
study, Harrat Al-Madinah is used to describe the volcanic terrains
of Harrat Rahat located north of latitude 24°N. It is mainly composed of olivine basalts and hawaiites with minor silicic differentiates of mugearites, benmoreites, and trachytes (Moufti et al.,
2012). The northern part of Harrat Al-Madinah (NHM), however,
is covered mainly with basaltic scoria cones, tephra fall deposits,
and lava flows (Camp and Roobol, 1989). On the contrary, the
southern part of Harrat Al-Madinah (SHM) is uniquely marked
with the development of trachytic lava domes, tuff rings, explosion
craters, and pyroclastic deposits (Moufti and Németh, 2013).
The most recent volcanic eruption took place in NHM in AD
1256, when an eruption lasted for 52 days, extruded 0.5 km3 of
alkali-olivine basalt from a 2.25 km-long fissure and produced at
least 7 scoria cones and a 23 km-long lava flow that came to within
8 km from Al-Madinah (Camp et al., 1987; Camp and Roobol, 1989;
Ambraseys et al., 1994; El-Masry et al., 2013; Murcia et al., 2014b).
In the SHM, an earthquake swarm of 500 events occurred at the
18
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
end of 1999. The recorded earthquakes with magnitudes ranged
between M0.5 and M3.0 went without any filed reports from the
locals. These events have raised the importance of understanding
the relationship between volcanicity and lithospheric structures
in Harrat Al-Madinah.
Although many researchers have extensively studied the lithospheric structure beneath the Arabian Plate and the Red Sea region
(e.g., Benoit et al., 2003; Nyblade et al., 2006; Park et al., 2007,
2008; Chang and Van der Lee, 2011; Hansen et al., 2007), quite
few have addressed this subject in the harrats (e.g., Harrat
Lunayyir, Pallister et al., 2010; Hansen et al., 2013; Koulakov
et al., 2015). In this work, the detailed 3-D structure beneath Harrat
Al-Madinah is investigated through the collaborative project of
Volcanic Risks in Saudi Arabia (VORiSA) between King Abdulaziz
University, Saudi Arabia, and the University of Auckland, New Zealand. Eight short-period seismic borehole stations were installed to
record local micro-seismic activity. Due to the sparsity of local seismicity in the area, we applied a tomography method (Zhao et al.,
1994, 2012) to invert abundant travel-time data of local and teleseismic events simultaneously. This method has been used successfully so far to image the deep structure of the Japan
subduction zone (e.g., Zhao et al., 1994, 2012; Abdelwahed and
Zhao, 2007; Huang et al., 2013; Liu et al., 2013, 2014), the Yellowstone and Cape Verde volcanic fields (Tian and Zhao, 2012; Liu and
Zhao, 2014), as well as many other regions in the world (see a
recent review by Zhao, 2015). After one-year of data collection,
and with the aid of data from the Saudi Geological Survey (SGS),
about 5800 arrival-time data were collected which enabled us to
image the crustal and upper mantle structures beneath Harrat
Al-Madinah.
2. Data and method
Our results are mainly derived from the analysis of two sets of
data. One is arrival-time data of local earthquakes and teleseismic
events recorded by the newly installed VORiSA Seismic Network
(VSN) (Abdelwahed, 2013; Kenedi et al., 2014). The other is the
1999 earthquake swarm data recorded by the Al-Madinah Seismic
Network (SGS-MSN) operated by the SGS (Endo et al., 2007).
The VSN consists of eight short-period borehole stations
deployed in the NHM area (Fig. 1). The sensors were installed in
the boreholes at the depth of 120 m and consisted of threecomponent 2 Hz velocity seismometers with a Reftek acquisition
system (DAS-130) recording the data at 250 sample/s. The technical specifications of the borehole sensor are shown in Table S1. This
network has been operated by the Geohazards Research Center
(GRC), King Abdulaziz University since April 2012. The SGS-MSN
has been operated by SGS since 1999. It consists of 16 threecomponent Nanometrics broadband seismic stations recording
the data at 100 sample/s. The distribution of the seismic stations
used in this study is shown in Fig. 1. Table S2 shows the coordinates, sensor types, and the installation dates of the seismic stations used in this study. The waveform data in this study are
analyzed using the SGRAPH software and a waveform modelling
method (Abdelwahed, 2012, 2013; Abdelwahed and Zhao, 2005,
2014).
2.1. Local earthquake data
A total of 4609 P-wave arrival times from 733 local earthquakes
are used in this study (Fig. 2a). This dataset was recorded by the
VSN during the period from April 2012 to November 2013 and
by the SGS-MSN during the period from November 1999 to December 1999. Fig. 2b shows the travel time–distance curve of the local
data used in this study. The first P-wave arrival times were picked
up manually by the GRC and the Saudi Geological Survey (SGS)
operators, respectively. The picking errors are estimated to be
0.1 s. Fig. S1 shows an example of the seismograms of a local
earthquake recorded by the VSN. The focal depths of the local
events range from 0 to 40 km, which caused some scattering in
the observed travel-time curve, particularly for larger distances.
2.2. Teleseismic data
The other dataset consists of 1179 P-wave relative travel-time
residuals from 151 teleseismic events recorded by the VSN. The
teleseismic travel-time data were firstly picked up manually before
applying the multi-channel cross-correlation technique (VanDecar
and Crosson, 1990) to improve the picking accuracy. Fig. S2 shows
an example of the seismograms of a teleseismic event recorded by
the VSN. The epicentral distances of the teleseismic events (M 5.5–
8.0) range from 30° to 90° (Fig. 2c). Most of the events occurred in
the subduction zones in the northern Pacific and Tonga, and eastern and southern Asia. Although the incident angles of the teleseismic rays are generally small (within 30° from the vertical
direction), the teleseismic rays crisscross in the upper mantle
beneath the study area. In the tomographic inversion, relative
travel-time residuals of the teleseismic events are used to minimize the effects of the hypocentral mislocations and the structural
heterogeneities outside the modelling space (Zhao et al., 1994,
2012). To calculate the relative travel-time residuals (RTTRs), we
first calculated the travel-time residuals (TTRs) by subtracting
the theoretical travel times (Tcal) that were calculated for the
iasp91 Earth model (Kennett and Engdahl, 1991) from the corresponding observed travel times (Tobs). The RTTRs for all the teleseismic events and stations were then calculated by removing
the average of the entire TTRs from the individual TTR (see Zhao
et al., 1994 for more details). Furthermore, the mean RTTR at each
station was calculated by averaging all the corresponding TTRs of
all the teleseismic events recorded by the corresponding station.
Fig. 3 shows the distribution of the RTTRs at the eight VSN stations
from the teleseismic events in four geographical quadrants. Positive residuals dominate the central and western parts of the study
region, suggesting the existence of significant low-velocity (low-V)
anomalies beneath those parts. Whereas negative residuals are
observed at stations VW02 and VW06 in the NE and SW quadrants,
respectively, indicating the existence of high-velocity (high-V)
anomalies beneath these stations. The patterns of the residual distribution for the four quadrants (Fig. 3a–d) are generally consistent
with that obtained from the data of all quadrants (Fig. 3e) with the
exception of the SW quadrant where a small number of teleseismic
events exist (see Fig. 2c). This confirms the reliability of the
observed features and suggests the existence of significant structural heterogeneities beneath the study region, which could
explain the observed volcanic and seismic activities in the area.
2.3. Method
In this study, the joint inversion method of Zhao et al. (1994,
2012) is used to analyze the teleseismic RTTRs together with the
local earthquake arrival times in tomographic inversions. This
method deals with a general velocity model in which complex
velocity discontinuities exist and velocity changes in three dimensions. The medium under the area is divided into layers, where
three-dimensional grid nodes are arranged. Velocity perturbations
at the grid nodes from a starting 1-D velocity model are taken to be
unknown parameters. The velocity perturbation at any point in the
model is calculated by linearly interpolating the velocity perturbations at the eight nodes surrounding that point. The 3-D ray tracing
technique (Zhao et al., 1992; Zhao and Lei, 2004) is used to trace rays
between hypocenters and receivers and to calculate theoretical
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
19
Fig. 1. Landsat satellite image of the study area showing the VORiSA borehole seismic stations (green triangles) and the broadband seismic stations (cyan triangles). The
yellow rectangle represents the area covered by the VORiSA seismic network. Trachytic volcanic centers (including domes, tuff rings and pyroclastic flows) are delineated
with white lines. The red triangles denote the scoria cones of the AD 1256 historical eruption. The yellow curved lines are surface faults. The NHM and SHM designate the
northern and southern parts of Harrat Al-Madinah, respectively. The inset map is after Moufti et al. (2012). (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)
travel times and the travel-time residuals. The LSQR algorithm
(Paige and Saunders, 1982) is used to solve the large and sparse
system of observation equations with damping and smoothing regularizations (Zhao, 2001, 2004). The local earthquakes are relocated in the inversion process using P-wave arrival times (Zhao
et al., 1992, 1994). After the relocation process, the travel-time
data become less scattered (Fig. 2b) and the event distribution is
correlated well with the lineament of the observed low-V anomalies (as shown in Fig. 8). The relative travel-time residuals of teleseismic events are employed to constrain the deeper structure of
the study area where teleseismic rays crisscross. The starting
model for the 3-D tomographic inversion is a local 1-D velocity
model that has been used for the earthquake location. The shallow
part of the model (0–40 km) consists of four layers representing
the surface layer, the upper crust, the lower crust, and the uppermost mantle. The P-wave velocities in the three crustal layers are
5.3, 6.0 and 6.6 km/s, respectively. Directly below the Moho discontinuity (40 km), the velocity is 7.75 km/s and gradually
increases downward according to the iasp91 Earth model
(Kennett and Engdahl, 1991).
For result optimizations, we further searched for the optimal
value of the damping parameter based on the trade-off curve
between the total RMS travel-time residuals and the norm of the
model for nine successive inversions. The optimal damping value
is found to be 4.0 taking in consideration the balance between
the solution norm and the RMS residual (Fig. S3).
3. Analysis and results
3.1. Checkerboard resolution test
The checkerboard resolution test (CRT) (Zhao et al., 1992;
Humphreys and Clayton, 1988) is used to assess the adequacy of
the ray coverage and to evaluate the resolution. The CRT is based
on estimating synthetic dataset corresponding to alternative positive and negative velocity perturbations assigned to the 3-D nodes.
The resolvability of the images depends on the degree of recovering the polarity and weight of the inserted pattern. In this study,
a checkerboard input model is constructed by assigning ±4% velocity perturbations to alternative grid nodes, starting with 6–15 km
grid intervals. The vertical grid interval varies from 1 to 2 km at
the top and increases gradually up to 20 km at 120 km depth
(Figs. S4–S7). The optimal grid interval is found to be 7 km in the
lateral direction and ranges from 2 to 20 km in depth (Fig. 4).
The input model of the CRT with a grid interval of 7 km is shown
in Fig. S8. The pattern is generally reconstructed in the swarm area
at 5–100 km depths. The resolution becomes significantly lower
20
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
Fig. 2. Local and teleseismic data used in this study. (a) Distribution of the 733 local earthquakes (red stars) and seismic stations (triangles). (b) P-wave travel-time curve of
the local earthquake data. (c) Epicentral distribution of the 151 teleseismic events (red dots). (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)
toward the boundaries and at shallow depths (2–4 km), which is
attributed to the sparsity of the seismic rays particularly in the
western side of the area. Beneath the historical eruption area, the
resolution is low below the 60 km depth.
3.2. Results
The primary features of interest in the tomographic images of
this study are the significant low-V and high-V anomalies in the
shallow and deeper parts of the Harrat Al-Madinah. These anomalies may reflect heterogeneous crust and upper mantle structures,
which may have resulted from episodes of different volcanic eruptions (e.g., basaltic and trachytic) and the complex tectonic history
of the area. The tomographic images show distinctive low-V and
high-V anomalies beneath the AD 1256 historical eruption and
the swarm areas at depth ranges of 2–20 km and 20–40 km,
respectively (Fig. 5). A low-V anomaly trending north–south is visible at the 2–10 km depths interval beneath the historical eruption
and the swarm area. This anomaly becomes incoherent at 10 km
depth and disappears downward. At 22 km depth, a large-sized
low-V anomaly is concentrated beneath the swarm area and
becomes less prominent and decreases in size with depth disappearing at the 46 km depth. Scattered low-V anomalies appear
eastward at the 46–55 km depths and northward at the 65–
100 km depths. On the other hand, significant high-V anomalies
emerge in many parts beneath the area (Fig. 5). They constitute
the predominant feature in the surface layer (0–2 km). Going deeper, a high-V anomaly appear in the SE part of the area at the 4 km
depth. South of the swarm area, high-V anomalies appear at the
16 km depth. Prominent high-V anomalies exist in the northern
part of the area at the 22–38 km depths and shift westward at
the 46–55 km depths (Fig. 5).
3.3. Vertical cross-sections
Eight vertical cross-sections are constructed to explain the current results (Fig. 6). Sections A and B pass from north to south
crossing the AD 1256 eruption and the swarm areas. Shallow,
low-V anomalies (SLV, 6–40 km depths) are visible beneath the
historical and trachytic domes with lateral offshoots dipping steeply to the south. Beneath the swarm area and the trachytic domes,
the low-V anomaly coincides with the location of 1999 seismic
swarm and appears wider in size (Section B). Another low-V anomaly is visible at 60–100 km depths in Section A. Due to the unresolved parts of the sections, it is not known whether this
anomaly extends laterally into the deeper parts of the sections.
High-V anomalies are also observed in these sections surrounding
the shallow low-V zones at 20 km depth and extending downward
to 60 km depth. Sections D and E are two parallel east–west sections passing through Al-Madinah city and the historical eruption.
A relatively small-sized, branched, low-V anomaly is revealed
beneath the historical eruption. This anomaly appears as a conelike structure which is narrow beneath the volcanic center and
becomes wider at 10–15 km depths. We suggest that these anomalies are sub-branches of the similar low-V zones appearing in Section A (just perpendicular to them). Beneath the aforementioned
SLV anomalies, deep low-V (DLV, >40 km depth) anomalies are
revealed (Fig. 6). The reliability of the SLV and DLV anomalies is
discussed in the synthetic test section.
Section H is a NE–SW trending section that passes through the
swarm area and the trachytic domes of the SHM. Beneath the
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
21
Fig. 3. Distribution of the average relative travel-time residuals (RTTRs) at the VORiSA seismic stations (triangles) calculated from the events located in the NW (a), NE (b), SW
(c), and SE (d) quadrants. The average of RTTRs from all the events is shown in (e). The red and blue colors denote delayed and early arrivals, respectively. The scale for the
RTTRs is shown below (d). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
swarm area, a wide SLV zone is visible occupying the entire crust
and coinciding at depth with the narrow zone of the 1999 earthquake swarm. A distinctive DLV anomaly is also observed in the
deeper parts of the section. The Sections F and G are constructed
further south parallel to Sections D and F to cover the swarm area
and the trachytic volcanic centers (Fig. 6). Prominent SLV and DLV
anomalies are also observed beneath the swarm area with a probable connection.
Prominent high-V anomalies are visible in different parts of the
sections surrounding the SLV zones beneath the historical eruption
22
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
Fig. 4. Results of a checkerboard resolution test with a lateral grid interval of 7 km. The vertical grid interval is 2 km for depths <10 km, 15 km at 30 km depth, and 20 km for
depths >30 km. These are found to be the optimal grid parameters for this study.
and the swarm areas. However, in deeper parts (>60 km), these
anomalies are elongated, which could be largely attributed to poor
resolution below this depth. The reliability of these anomalies is
discussed in the next section.
3.4. Synthetic test
Synthetic tests are conducted to evaluate the reliability of the
resulted images and to confirm the observed major features. The
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
23
Fig. 5. Map views of P-wave velocity tomography. The layer depth is shown beside each map. The red and blue colors denote low and high velocities, respectively. The
velocity perturbation scale (in %) is shown at the bottom. Areas hit by less than eight rays are masked out. (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)
test synthesizes a dataset corresponding to a model similar to the
resulted images and checks the image recovery after a tomographic
inversion. This test is applied to all the selected vertical crosssections shown in this study. All the sections are synthesized in a
way that the input model contains the typical high-V and low-V
features in the resulted image except that the velocity anomalies
greater than 1% are considered as 4% of velocity perturbation either
fast or slow (Fig. S9). The synthetic test simulates a dataset that
would have been given if these particular anomalies exist under
the study area. Fig. 7 shows the synthetic test results along the
eight cross-sections shown in Fig. 6.
In general, the synthetic test results show that most of the SLV
anomalies beneath the different types of volcanic centers are
recovered; therefore they are reliable features. The DLV anomalies
are only recovered in Sections A and D, which may suggest that the
existence of low-V anomalies at that depth may be reliable. Shallow high-V anomalies in all sections, except for Sections F and H,
do not exceed 40 km depth, which suggests that high-V anomalies
deeper than 40 km are not true structures. In Sections F and H
beneath the trachytic domes, the high-V anomalies are recovered
down to a depth of 120 km. However, the elongated shape of
the anomalies implies that there might be smearing effects in the
24
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
Fig. 6. Vertical cross-sections of P-wave velocity tomography beneath the study area. Locations of the profiles are shown on the inset map. The red and blue colours denote
low and high velocities, respectively. The velocity perturbation (in %) scale is shown below the map. The red and purple triangles denote the basaltic cones and trachytic lava
domes, respectively. Areas hit by less than eight rays are masked out. (For interpretation of the references to color in this figure legend, the reader is referred to the web
version of this article.)
deeper parts of the images (>60 km). This may be attributed to the
lack of crisscrossing teleseismic rays at the deeper parts that cannot be solved by the inversion. This implies that the dimensions
of the observed anomalies in deeper parts might not be true. In
spite of this deficiency in images, we believe that the resulted
images in the shallower parts (<60 km) are robust.
In the following sections, we discuss the tomographic results
emphasizing the role of the shallow lithosphere–asthenosphere
boundary (LAB) in magma generation, which helps understanding
the origin of the upper mantle thermal anomalies and its implication on volcanism in the area.
4. Discussion
Our tomographic images clearly demonstrate the presence of
low-V and high-V anomalies in the lithosphere underneath Harrat
Al-Madinah. These anomalies reflect the heterogeneous nature of
the lithosphere that could be attributed to the complexity of the
tectonic evolution of the Arabian lithosphere during the Precambrian and to episodes of active Cenozoic volcanism.
The Arabian Shield has a complex tectonic history reflected in
its highly heterogeneous and anisotropic structure. Most of
the Arabian Shield was generated during the Cryogenian
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
25
Fig. 7. Synthetic test results along eight vertical cross-sections. The figure description is the same as Fig. 6.
(850–630 Ma) at primitive arc systems, where collision between
continental fragments and tectonic escape occurred between
630–600 Ma and 600–550 Ma, respectively (Stern, 1994; Stern
and Johnson, 2010). Therefore, the Shield is a product of a cycle
of continental accretion that lasted about 300 million years
(Stern and Johnson, 2010). During the Neoproterozoic, the lithosphere of western Arabia became thinner (<80 km) compared to
that of eastern Arabia (150 km) (Stern and Johnson, 2010). This
was largely ascribed to lithospheric mantle removal by delamination (Kay and Kay, 1993) and/or Rayleigh–Taylor instability (dripping) processes (Gorczyk et al., 2012; Göğüsß and Pysklywec, 2008).
It was suggested that the Red Sea went through a two-stage rifting history, wherein an early stage of passive rifting was followed
by active rifting processes associated with a mantle upwelling
(McGuire and Bohannon, 1989; Camp and Roobol, 1992; Ebinger
and Sleep, 1998; Daradich et al., 2003; Hansen et al., 2007). Accordingly, the extension and erosion by flow in the underlying asthenosphere brought up significant variations in the LAB depth from the
Red Sea coasts eastwards underneath the Arabian Shield terranes
from 55 km to 100–110 km depths, respectively (Hansen et al.,
2007). Although the rise of the upper-mantle isotherms and the
thinning and stretching of the lithosphere may have produced a
broad band of high heat flow extending across the entire width
of the Red Sea (Bonatti, 1985), Chang and van der Lee (2011)
argued against a northwestward flow of low-velocity channels
from the Afar mantle plume beneath the entire Red Sea and suggested instead a channelled northward mantle flow from Afar
beneath Arabia. These factors could have contributed greatly to
the observed lithospheric heterogeneity depicted in our tomographic images.
26
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
Fig. 8. A 3-D image depicting the magma intrusion beneath Harrat Al-Madinah. The image shows the ±1% isovelocity perturbation surfaces of the obtained P-wave
tomography. The high and low velocities are shown in blue and orange, respectively. The 1999 seismic swarm is shown in black dots. The location of the AD 1256 eruption
and the trachyte areas are shown on the top. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Our observed high-V anomalies may reflect areas which were
not affected by faulting and large earthquakes (e.g., Abdelwahed
and Zhao, 2007; Zhao et al., 2004; Chen et al., 2014, 2015) and/or
not intruded by magma (e.g., Hansen et al., 2013; Asamori and
Zhao, 2015; Liu and Zhao, 2014, 2015), or may alternatively represent remnants of cumulates crystallized in the lower crust (e.g., AlMishwat and Nasir, 2004; Xia and Zhao, 2014; Xia et al., 2015),
which may in turn explain the presence of intrusive complexes
of gabbro in western Arabia (Stern and Johnson, 2010). The
observed high-V crustal heterogeneities may broadly signify relative variations in crustal ductility and/or composition underneath
Harrat Al-Madinah. In addition, fractionated solidified intrusions
(Chiarabba et al., 2000; Keranen et al., 2004; Daly et al., 2008), formerly acted as shallow reservoirs for the ascending patches of
magmatic material erupted to the Cenozoic volcanic fields of western Arabia, could be another source of heterogeneity in the crust.
The high-V anomalies in the upper mantle could be partly
inherited from the Proterozoic lithosphere formed by accreted
island arcs marked with recycled and dehydrated relics of subducting oceanic lithosphere (Stern and Johnson, 2010). The observed
thinning of the lithosphere in western Arabia (<80 km), whether
attributed to the Rayleigh–Taylor instability or to delamination
processes (Stern and Johnson, 2010), could also be considered as
another cause for the mantle heterogeneities in this rifted region.
Low-V structures observed in the tomographic images extend
beneath all types of volcanic edifices in Harrat Al-Madinah
(Fig. 6). However, the characteristics of these low-V zones are quite
different. A shallow low-V (SLV) zone is located beneath the
aligned cones of the historical eruption at shallow depths of 10–
20 km (Sections D, E, and the left halves of Sections A and B) and
becomes larger in size with depth. This SLV zone is possibly connected to a deeper low-V anomaly (DLV) at 60–120 km depths,
as observed in Sections A–E (Fig. 6).
Beneath the trachytic lava domes and the swarm area (Sections
H, F, and G, and the right halves of Sections A and B in Fig. 6), a SLV
zone is located at 20–40 km depths and is comparatively larger in
size. This low-V zone is apparently connected to its counterpart
beneath the basaltic cones of the AD 1256 eruption (Section A,
Fig. 6). This may suggest that both basaltic and trachytic magmas
have intricate and closely juxtaposed shallow reservoirs and pathways, wherein trachytic reservoirs tend to reside at a deeper depth.
In Sections B and F (Fig. 6), the low-V zone beneath the trachytic
lava domes and tuff rings of Matan, Mouteen, Gura-1, Gura-2,
and Gura-3, and Um-Junb is connected to the DLV zone (Sections
F and H, Fig. 6) and is coincidentally underlain by the loci of the
1999 seismic swarm. This may imply that the 1999 seismic swarm
was triggered when a shallow reservoir (the SLV zone at 20–40 km
depths) was recharged by an intruding magmatic material from a
deeper source, possibly the DLV zone. It appears that the ascending
magma moved aseismically below 45 km depth before it went
seismically in the crust until it stopped (aborted) and seismicity
died out at 10–15 km depths. This implies that the 1999 seismic
swarm may represent the precursors of an aborted volcanic eruption. Petrological studies have suggested that the source of magma
is located most possibly between 60 and 100 km depths, where
lithospheric and asthenospheric garnet peridotite mantle is governed by temperatures higher than 1200 °C and pressures between
2 and 3 GPa (Murcia et al., 2014a). Fig. 8 shows a 3-D tomographic
image of the ±1% isovelocity surface down to a depth of 60 km
beneath the historical eruption and the swarm area. A magma
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
reservoir appears clearly at the lowermost part of the crust. The
ascent of the magma toward shallower levels of the crust could
have initiated the 1999 seismic swarm. The cluster of the earthquake hypocenters is bounded by markedly low-V and high-V
anomalies. Incoherent patches of high-V anomaly of possibly solidified magma intrusions and/or cumulates can be identified from
60 km depth to the near surface. The low-V zone observed beneath
the historical eruption, particularly the shallow parts, may represent remnants of a drained shallow magma reservoir.
One can infer that the shallower LAB beneath western Arabia
has increased the geothermal gradient and henceforth triggered
primary magma generation in an already heterogeneous lithospheric mantle, which in turn reduced seismic velocity and
increased the velocity contrast across the LAB. Hence, we think that
the Red Sea rifting has significantly contributed to the magma generation and the lithospheric mantle heterogeneities in western
Arabia.
Although the present study provides crucial information on the
dynamics of volcanic eruption and magma intrusion in the Harrat
Al-Madinah volcanic field, it is not clearly evident whether there
will be enough magma for a near future eruption. More seismic
data and high-resolution S-wave and Poisson’s ratio images as well
as additional scientific evidence are essential for future studies.
5. Conclusions
A detailed three-dimensional P-wave velocity tomography
beneath the Al-Madinah Volcanic field, Saudi Arabia, is determined
down to a depth of 120 km by simultaneously inverting traveltime data of local and teleseismic events. We collected 4609 Pwave arrival times from 733 local earthquakes and 1179 arrivals
from 151 teleseismic events recorded by local seismic networks
in the Kingdom of Saudi Arabia. Main findings of this study are
summarized as follows:
(1) Our tomographic images reveal strong heterogeneities in the
lithosphere beneath Harrat Al-Madinah marked with low-V
and high-V anomalies which reflect the complexity of the
Arabian lithosphere evolution from the Precambrian to the
Cenozoic.
(2) A shallow low-V zone at 10–20 km depths is detected
beneath the AD 1256 eruption site, which may represent
remnants of a drained shallow magma reservoir.
(3) The 1999 seismic swarm may represent the precursors of a
failed volcanic eruption which was triggered when a rising
magma from a deeper source (>120 km depth) recharged a
shallow reservoir (20–40 km depths) before it stopped at
10–15 km depths.
(4) Our tomographic images also indicate that the low-V zones
beneath the AD 1256 eruption and the 1999 seismic swarm
areas are, to some extent, interconnected.
Acknowledgments
This work is financially supported by the King Abdulaziz
University as a part of the VORiSA (Volcanic Risk in Saudi Arabia)
collaborative project between King Abdulaziz University, Saudi
Arabia, and the University of Auckland, New Zealand. We thank
the Saudi Geological Survey for providing part of the seismic data
used in this study. We appreciate King Abdulaziz University
funding and technical support. Special thanks to the GHRC and
the University of Auckland staff for the scientific support. Two
anonymous referees provided very thoughtful review comments
and suggestions which have improved this paper.
27
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jseaes.2016.01.
023.
References
Abdelwahed, M.F., 2012. SGRAPH (SeismoGRAPHer): seismic waveform analysis
and integrated tools in seismology. Comput. Geosci. 40, 153–165.
Abdelwahed, M.F., 2013. Data analysis and waveform processing of VORiSA Seismic
Network Data. In: VORiSA Scientific Meeting, November 17–18, Jeddah, 2013,
King Abdulaziz University, Saudi Arabia.
Abdelwahed, M.F., Zhao, D., 2005. Waveform modelling of local earthquakes in
southwest Japan. Earth Planets Space 57, 1039–1054.
Abdelwahed, M.F., Zhao, D., 2007. Deep structure of the Japan subduction zone.
Phys. Earth Planet. Inter. 162, 32–52.
Abdelwahed, M.F., Zhao, D., 2014. Genetic waveform modeling for the crustal
structure in Northeast Japan. J. Asian Earth Sci. 89, 66–75.
Al-Mishwat, A.T., Nasir, S.J., 2004. Composition of the lower crust of the Arabian
Plate: a xenolith perspective. Lithos 72, 45–72.
Ambraseys, N.N., Melville, C.P., Adams, R.D., 1994. The Seismicity of Egypt, Arabia,
and the Red Sea: A Historical Review. Cambridge University Press, NY, 320 pp.
Asamori, K., Zhao, D., 2015. Teleseismic shear wave tomography of the Japan
subduction zone. Geophys. J. Int. 203, 1752–1772.
Benoit, M.H., Nyblade, A.A., VanDecar, J.C., Gurrola, H., 2003. Upper mantle P wave
velocity structure and transition zone thickness beneath the Arabian Shield.
Geophys. Res. Lett. 30 (10), 1531. http://dx.doi.org/10.1029/2002GL016436.
Berthier, F., Demange, J., Iundt, F., Verzier, P., 1981. Geothermal Resources of the
Kingdom of Saudi Arabia. Saudi Arabian Deputy Ministry for Mineral Resources,
Open-File Report, BRGM-OF-01-24, 116 pp.
Bonatti, E., 1985. Punctiform initiation of seafloor spreading in the Red Sea during
transition from a continental to oceanic rift. Nature 316, 33–37.
Bosworth, W., Huchon, P., McClay, K., 2005. The Red Sea and Gulf of Aden Basins. J.
Afr. Earth Sci. 43, 334–378.
Camp, V.E., Hooper, P.R., Roobol, M.J., White, D.L., 1987. The Madinah eruption,
Saudi Arabia: magma mixing and simultaneous extrusion of three basaltic
chemical types. Bull. Volcanol. 49, 489–508.
Camp, V.E., Roobol, M.J., 1989. The Arabian continental alkali basalt province: Part I.
Evolution of Harrat Rahat, Kingdom of Saudi Arabia. Geol. Soc. Am. Bull. 101,
71–95.
Camp, V.E., Roobol, M.J., 1992. Upwelling asthenosphere beneath western Arabia
and its regional implications. J. Geophys. Res. 97, 15255–15271.
Chang, S.-J., Merino, M., Van der Lee, S., Stein, S., Stein, C.A., 2011. Mantle flow
beneath Arabia offset from the opening Red Sea. Geophys. Res. Lett. 38, L04301.
http://dx.doi.org/10.1029/2010GL045852.
Chang, S.-J., Van der Lee, S., 2011. Mantle plumes and associated flow beneath
Arabia and East Africa. Earth Planet. Sci. Lett. 302, 448–454.
Chen, C., Zhao, D., Wu, S., 2014. Crust and upper mantle structure of New Madrid
Seismic Zone: insight into intraplate earthquakes. Phys. Earth Planet. Inter. 230,
1–14.
Chen, C., Zhao, D., Wu, S., 2015. Tomographic imaging of the Cascadia subduction
zone: constraints on the Juan de Fuca slab. Tectonophysics 647, 73–88.
Chiarabba, C.A., Amato, E., Boschi, E., Barbari, F., 2000. Recent seismicity and
tomography modeling of the Mount Etna plumbing system. J. Geophys. Res.
105, 10923–10938.
Coleman, R.G., Gregory, R.T., Brown, G.F., 1983. Cenozoic Volcanic Rocks of Saudi
Arabia. USGS Open-File Report: 83-788, 82 pp.
Daly, E., Keir, D., Ebinger, C.J., Stuart, G.W., Bastow, I.D., Ayele, A., 2008. Crustal
tomographic imaging of a transitional continental rift: the Ethiopian rift.
Geophys. J. Int. 172, 1033–1048.
Daradich, A., Mitrovica, J.X., Pysklywec, R.N., Willett, S.D., Forte, A.M., 2003. Mantle
flow, dynamic topography, and rift-flank uplift of Arabia. Geology 31, 901–904.
Durozoy, G., 1970. Groundwater Geology of the Basalt on the Harrat Rahat Plateau.
Bureau de Recherches Géologicques et Minères (BRGM), Open-File Report, 70
JED 24, 31 pp.
Ebinger, C.J., Sleep, N.H., 1998. Cenozoic magmatism throughout east Africa
resulting from impact of a single plume. Nature 395, 788–791.
El-Masry, N., Moufti, M.R., Murcia, H., Qaddah, A., Abdelwahed, M.F., 2013.
Historical accounts of the AD 1256 eruption near Al-Madinah. VORISA
Scientific Meeting, King Abdulaziz University, Jeddah, November 17–18, 2013,
pp. 9–13.
Endo, E., Zahran, H., Nofal, H., Hadidy, S., 2007. The Saudi national seismic network.
Seismol. Res. Lett. 78, 439–445.
Göğüsß, O.H., Pysklywec, R.N., 2008. Near-surface diagnostics of dripping or
delaminating lithosphere. J. Geophys. Res. 113, B11404. http://dx.doi.org/
10.1029/2007JB005123.
Gorczyk, W., Hobs, B., Gerya, T., 2012. Initiation of Rayleigh–Taylor instabilities in
intra-cratonic settings. Tectonophysics 514–517, 146–155.
Hansen, S.E., DeShon, H.R., Moore-Driskell, M.M., Al-Amri, A.M.S., 2013.
Investigating the P wave velocity structure beneath Harrat Lunayyir,
northwestern Saudi Arabia, using double-difference tomography and
earthquakes from the 2009 seismic swarm. J. Geophys. Res. 118, 4814–4826.
28
M.F. Abdelwahed et al. / Journal of Asian Earth Sciences 120 (2016) 17–28
Hansen, S.E., Rodgers, A.J., Schwartz, S.Y., Al-Amri, A.M.S., 2007. Imaging ruptured
lithosphere beneath the red sea and Arabian Peninsula. Earth Planet. Sci. Lett.
259, 256–265.
Huang, Z., Zhao, D., Hasegawa, A., Umino, N., Park, J., Kang, I., 2013. Aseismic deep
subduction of the Philippine Sea plate and slab window. J. Asian Earth Sci. 75,
82–94.
Humphreys, E., Clayton, R., 1988. Adaptation of back projection tomography to
seismic travel time problems. J. Geophys. Res. 93, 1073–1085.
Kay, R.W., Kay, S.M., 1993. Delamination and delamination magnetism.
Tectonophysics 219, 177–189.
Kenedi, C.L., Boese, C., Abdelwahed, M.F., Alvarez, M., Lindsay, J.M., Mokhtar, T.A.,
Moufti, M.R.H., 2014. Insights from the first borehole seismic network in the
volcanic fields (harrats) of the Kingdom of Saudi Arabia. In: Volcanic Risks in
Saudi Arabia (VORiSA) Project, Final Report. King Abdulaziz University, Part II:
Section 7, Jeddah.
Kennett, B., Engdahl, E.R., 1991. Travel time for global earthquake location and
phase identification. Geophys. J. Int. 105, 429–465.
Keranen, K., Klemperer, S.L., Gloaguen, R., Group, E.W., 2004. Three-dimensional
seismic imaging of a protoridge axis in the Main Ethiopian rift. Geology 32,
949–952.
Koulakov, I., El Khrepy, S., Al-Aifi, N., Kuznetsov, P., Kasatkina, E., 2015. Structural
cause of a missed eruption in Harrat Lunayyir basaltic field (Saudi Arabia) in
2009. Geology. http://dx.doi.org/10.1130/G36271.1.
Liu, X., Zhao, D., 2014. Seismic evidence for a mantle plume beneath the Cape Verde
hotspot. Int. Geol. Rev. 56, 1213–1225.
Liu, X., Zhao, D., 2015. Seismic attenuation tomography of the Southwest Japan arc:
new insight into subduction dynamics. Geophys. J. Int. 201, 135–156.
Liu, X., Zhao, D., Li, S., 2013. Seismic heterogeneity and anisotropy of the southern
Kuril arc: insight into megathrust earthquakes. Geophys. J. Int. 194, 1069–1090.
Liu, X., Zhao, D., Li, S., 2014. Seismic attenuation tomography of the Northeast Japan
arc: insight into the 2011 Tohoku earthquake (Mw 9.0) and subduction
dynamics. J. Geophys. Res. 119, 1094–1118.
McGuire, A.V., Bohannon, R.G., 1989. Timing of mantle upwelling: evidence for a
passive origin for the Red Sea rift. J. Geophys. Res. 94, 1677–1682.
Moufti, M.R., 1985. The Geology of Harrat Al Madinah Volcanic Field, Harrat Rahat,
Saudi Arabia. Unpublished Ph.D. Thesis, Department of Environmental Science,
University of Lancaster, United Kingdom, 476 pp.
Moufti, M.R., Moghazi, A.M., Ali, K.A., 2012. Geochemistry and Sr–Nd–Pb isotopic
composition of the Harrat Al-Madinah Volcanic Field, Saudi Arabia. Gondwana
Res. 21, 670–689.
Moufti, M.R., Németh, K., 2013. The intra-continental Al Madinah volcanic field,
Western Saudi Arabia: a proposal to establish Harrat Al Madinah as the first
volcanic geopark in the Kingdom of Saudi Arabia. Geoheritage 5, 185–206.
Murcia, H., Lindsay, J.M., Smith, I.E.M., Moufti, M.R., Németh, K., Cronin, S.J., ElMasry, N.N., Niedermann, S., 2014a. A melt source and magma ascent processes
revealed by recent basaltic eruptions in northern Harrat Rahat, Kingdom of
Saudi Arabia. In: Volcanic Risks in Saudi Arabia (VORiSA) Project, Final Report.
King Abdulaziz University, Part I: Section 4, Jeddah.
Murcia, H., Németh, K., Moufti, M.R., Lindsay, J.M., El-Masry, N., Cronin, S.J., Qaddah,
A., Smith, I.E.M., 2014b. Late Holocene lava flow morphotypes of northern
Harrat Rahat, Kingdom of Saudi Arabia: implications for the description of
continental lava fields. J. Asian Earth Sci. 84, 131–145.
Nyblade, A., Park, Y., Rodgers, A., Al-Amri, A., 2006. Seismic structure of the Arabian
Shield lithosphere and Red Sea margin. Margins Newslett. 17, 13–15.
Paige, C.C., Saunders, M.A., 1982. LSQR: an algorithm for sparse linear equations and
sparse least squares. ACM Trans. Math. Softw. 8, 43–71.
Pallister, J.S., McCausland, W.A., Jónsson, S., Lu, Z., Zahran, H.M., El Hadidy, S.,
Aburukbah, A., Stewart, I.C.F., Lundgren, P.R., White, R.A., Moufti, M.R.H., 2010.
Broad accommodation of rift-related extension recorded by dyke intrusion in
Saudi Arabia. Nat. Geosci. 3, 705–712.
Park, Y., Nyblade, A.A., Rodgers, A., Al-Amri, A.M., 2007. Upper mantle structure
beneath the Arabian Peninsula from regional body-wave tomography:
implications for the origin of Cenozoic uplift and volcanism in the Arabian
Shield. Geochem. Geophys. Geosyst. 8, Q06021. http://dx.doi.org/10.1029/
2006GC001566.
Park, Y., Nyblade, A., Rodgers, A., Al-Amri, A.M., 2008. S wave velocity structure of
the Arabian Shield upper mantle from Rayleigh wave tomography. Geochem.
Geophys. Geosyst. 9, Q07020. http://dx.doi.org/10.1029/2007GC001895.
Stern, R.J., 1994. Arc assembly and continental collision in the Neoproterozoic East
African Orogen: implications for the consolidations of Gondwanaland. Annu.
Rev. Earth Planet. Sci. 22, 319–351.
Stern, R.J., Johnson, P., 2010. Continental lithosphere of the Arabian Plate: a geologic,
petrologic, and geophysical synthesis. Earth Sci. Rev. 101, 29–67.
Tian, Y., Zhao, D., 2012. P-wave tomography of the western United States: insight
into the Yellowstone hotspot and the Juan de Fuca slab. Phys. Earth Planet. Inter.
200, 72–84.
VanDecar, J., Crosson, R., 1990. Determination of teleseismic relative phase arrival
times using multi-channel cross-correlation and least squares. Bull. Seismol.
Soc. Am. 80, 150–169.
Xia, S., Shen, Y., Zhao, D., Qiu, X., 2015. Lateral variation of crustal structure and
composition in the Cathaysia block of South China and its geodynamic
implications. J. Asian Earth Sci. 109, 20–28.
Xia, S., Zhao, D., 2014. Late Mesozoic magmatic plumbing system in the onshore–
offshore area of Hong Kang: insight from 3-D active-source seismic
tomography. J. Asian Earth Sci. 96, 46–58.
Zhao, D., 2001. Seismic structure and origin of hotspots and mantle plumes. Earth
Planet. Sci. Lett. 192, 251–265.
Zhao, D., 2004. Global tomographic images of mantle plumes and subducting slabs:
insight into deep Earth dynamics. Phys. Earth Planet. Inter. 146, 3–34.
Zhao, D., 2015. Multiscale Seismic Tomography. Springer, 304 pp.
Zhao, D., Hasegawa, A., Horiuchi, S., 1992. Tomographic imaging of P and S wave
velocity structure beneath northeastern Japan. J. Geophys. Res. 97, 19909–
19928.
Zhao, D., Hasegawa, A., Kanamori, H., 1994. Deep structure of Japan subduction zone
as derived from local, regional, and teleseismic events. J. Geophys. Res. 99,
22313–22329.
Zhao, D., Lei, J., 2004. Seismic ray path variations in a 3-D global velocity model.
Phys. Earth Planet. Inter. 141, 153–166.
Zhao, D., Tani, H., Mishra, O.P., 2004. Crustal heterogeneity in the 2000 western
Tottori earthquake region: effect of fluids from slab dehydration. Phys. Earth
Planet. Inter. 145, 161–177.
Zhao, D., Yanada, T., Hasegawa, A., Umino, N., Wei, W., 2012. Imaging the
subducting slabs and mantle upwelling under the Japan Islands. Geophys. J.
Int. 190, 816–828.