Geophysical Journal International

Geophys. J. Int. (2015) 202, 1407–1418 doi: 10.1093/gji/ggv229

GJI Seismology

No thermal anomalies in the mantle transition zone beneath

an incipient continental rift: evidence from the first receiver
function study across the Okavango Rift Zone, Botswana

Y. Yu,1 K.H. Liu,1 M. Moidaki,2 C.A. Reed1 and S.S. Gao1

1 Geology and Geophysics Program, Missouri University of Science and Technology, Rolla, MO 65409, USA. E-mail: [email protected]

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

2 Department of Physics, University of Botswana, Gaborone, Botswana

Accepted 2015 May 29. Received 2015 May 3; in original form 2014 December 20

SUMMARY
Mechanisms leading to the initiation and early-stage development of continental rifts re-
main enigmatic, in spite of numerous studies. Among the various rifting models, which were
developed mostly based on studies of mature rifts, far-field stresses originating from plate
interactions (passive rifting) and nearby active mantle upwelling (active rifting) are commonly
used to explain rift dynamics. Situated atop of the hypothesized African Superplume, the
incipient Okavango Rift Zone (ORZ) of northern Botswana is ideal to investigate the role of
mantle plumes in rift initiation and development, as well as the interaction between the upper
and lower mantle. The ORZ developed within the Neoproterozoic Damara belt between the
Congo Craton to the northwest and the Kalahari Craton to the southeast. Mantle structure and
thermal status beneath the ORZ are poorly known, mostly due to a complete paucity of broad-
band seismic stations in the area. As a component of an interdisciplinary project funded by
the United States National Science Foundation, a broad-band seismic array was deployed over
a 2-yr period between mid-2012 and mid-2014 along a profile 756 km in length. Using P-to-S
receiver functions (RFs) recorded by the stations, the 410 and 660 km discontinuities bordering
the mantle transition zone (MTZ) are imaged for the first time. When a standard Earth model is
used for the stacking of RFs, the apparent depths of both discontinuities beneath the Kalahari
Craton are about 15 km shallower than those beneath the Congo Craton. Using teleseismic P-
and S-wave traveltime residuals obtained by this study and lithospheric thickness estimated
by previous studies, we conclude that the apparent shallowing is the result of a 100–150 km
difference in the thickness of the lithosphere between the two cratons. Relative to the adjacent
tectonically stable areas, no significant anomalies in the depth of the MTZ discontinuities or in
teleseismic P- and S-wave traveltime residuals are found beneath the ORZ. These observations
imply an absence of significant thermal anomalies in the MTZ and in the upper mantle beneath
the incipient rift, ruling out the role of mantle plumes in the initiation of the ORZ. We propose
that the initiation and development of the ORZ were the consequences of relative movements
between the South African block and the rest of the African plate along a zone of lithospheric
weakness between the Congo and Kalahari cratons. An area of thinner-than-normal MTZ is
found at the SW corner of the study area. This anomaly, if confirmed by future studies, could
suggest significant transferring of heat from the lower to the upper mantle.
Key words: Wave propagation; Cratons; Continental tectonics: extensional.

asthenosphere followed by extensive magmatic diking and regional

1 I N T RO D U C T I O N
uplift, while passive rifting demonstrates crustal extension ambigu-
The mechanisms controlling the initiation and development of con- ously accompanied by volcanism or uplift, and is commonly caused
tinental rifting can be generally categorized into active and passive by far-field plate driving forces (Sengor & Burke 1978; White &
processes, wherein the principle distinction depends upon the driv- McKenzie 1989; Schmeling 2010). Observations on the state of
ing forces that promote lithospheric extension (Sengor & Burke rifting and the geodynamic controls governing their evolution were
1978). Active rifting is broadly induced by a hot, actively upwelling primarily obtained, however, from relatively mature rift zones such


C The Authors 2015. Published by Oxford University Press on behalf of The Royal Astronomical Society. 1407
1408 Y. Yu et al.

as the Main Ethiopian and Kenyan Rifts of the northernmost East Surface geological observations and crustal geophysical studies
African Rift System (EARS). Meanwhile, the characterization of indicate that the ORZ is in its earliest recognizable stage of devel-
the initiation and early-stage evolution of continental rifting, ex- opment, with initiation of rifting estimated to be between 40 and
emplified by the amagmatic southern and southwestern EARS rift 27 ka (Modisi et al. 2000; Kinabo et al. 2008; Miensopust et al.
segments such as the Malawi (Ebinger et al. 1984) and Okavango 2011; Bufford et al. 2012). High heat flux has been recorded within
rift zones, is largely underdeveloped. a 60 km wide NE-trending zone within the ORZ indicative of sub-
The Okavango Rift Zone (ORZ), located in northern Botswana, stantial fluid migration along rift-related fault zones, suggesting
is believed to be an incipient rift at the terminus of the southwestern pervasive lithospheric-scale faulting despite the immaturity of the
branch of the EARS (Reeves 1972; Modisi et al. 2000). The ORZ rift zone (Leseane et al. 2015). Due to a complete paucity of broad-
is mainly situated within the Neoproterozoic Damara belt and the band seismic stations in the vicinity of the ORZ, the structure and
reworked Paleo-Mesoproterozoic crust of the Rehoboth Province, thermal state of the upper mantle and the mantle transition zone
which are sandwiched between the Neoproterozoic Congo Craton to (MTZ) have not been investigated with sufficient spatial resolution

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

the northwest and the composite Archean Kalahari Craton (which, beneath the incipient rift.
in the study area, includes the Kaapvaal and Zimbabwe cratons and Numerous previous studies suggest that the MTZ, which is con-
the Limpopo belt) to the east–southeast (Hanson 2003; Kinabo et al. fined by seismic discontinuities at approximate depths of 410 km
2008; Begg et al. 2009; Fig. 1). The Damara belt is a Pan-African and 660 km (hereafter referred to as d410 and d660, respectively), is
suture formed during east and west Gondwanaland collision, or sensitive to variations in mantle composition and temperature in the
is a collapsed intracontinental basin (Stern 1994). It is a major vicinity of the discontinuities (Helffrich 2000). Mineral physics and
zone of lithospheric weakness in southern Africa, extending from seismological investigations (Ringwood 1975) suggest that d410 is
Namibia on the southwestern African coast northeastwards into associated with the transition from olivine to wadsleyite, and d660
Zambia (Begg et al. 2009). represents the transition from ringwoodite to perovskite. Relatively

Figure 1. A topographic relief map of the study area showing the SAFARI (red triangles) and SASE (white triangles) seismic stations used in the study. Crosses
are ray-piercing points above the depth of 535 km. Red lines show active faults (Kinabo et al. 2008) and grey lines delineate boundaries of major tectonic units
modified from Hanson (2003) and McCourt et al. (2013). LP: Limpopo Belt, MB: Magondi Belt, ORZ: Okavango Rift Zone. The fan-shaped area enclosed by
the white lines is the Okavango Delta. The inset in the upper-right corner shows the study area which is indicated by the black rectangle. The blue lines in the
inset represent the rift axes of the East African Rift System. The inset in the lower-left corner is an azimuthal equidistant projection map centred at the study
area, showing the distribution of earthquakes (red dots) used for the receiver function study.
 Okavango mantle discontinuities 1409

colder and warmer regions are expected to have thicker and thin-
ner MTZ, respectively, due to their characteristic Clapeyron slopes
(about +1.5 to +3.0 MPa K−1 for d410 and −4.0 to −0.4 MPa
K−1 for d660; Ringwood 1975; Bina & Helffrich 1994; Litasov
et al. 2005; note the significant differences in the values deter-
mined by previous studies). Global and regional studies suggest
that the observed thick MTZ beneath subduction zones is associ-
ated with thermal anomalies ranging between −100 and −300 K,
and the deepening d410 beneath hotspots corresponds to thermal
anomalies between +100 and +300 K (Tauzin et al. 2008). A +200
to +300 K increase in MTZ temperature was proposed to explain
the observed depression of d410 beneath the mature sections of

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

the EARS (Nyblade et al. 2000; Owens et al. 2000; Huerta et al.
2009; Mulibo & Nyblade 2013). The presence of water in the MTZ
can have the same effect as decreasing temperature and will sub-
sequently thicken the MTZ (Litasov et al. 2005). The existence
of water has been proposed to explain the observed thicker-than-
normal MTZ beneath southern Africa (Blum & Shen 2004), the
Nubian and Arabian plates adjacent to the Red Sea (Mohamed et al.
2014), and the Afar Depression (Thompson et al. 2015). Hence,
variations in MTZ thickness and depths of the MTZ discontinuities
can behave as a proxy with which to detect the existence of ther-
mal and compositional anomalies and help distinguish the mode of
rifting.
Such a tool has been applied to other regions of continental rift- Figure 2. Number of RFs in radius = 1◦ bins and seismic stations (triangles)
ing. For instance, the Baikal rift was found to have a cold MTZ based used in the study. The size of the triangles is proportional to the number of
on an observed uplifted d410 (Liu & Gao 2006), and studies in the RFs recorded by the corresponding stations, and the colour of the circles
Rio Grande rift suggested the lack of a large-scale thermal anomaly represents the number of RFs per bin.
in the MTZ (Wilson et al. 2005; Gao & Liu 2014a). Beneath the
Kenya and Tanzania segments of the EARS, receiver function (RF)
studies revealed a thin MTZ, which was interpreted to be associ-
2.1 Receiver function stacking
ated with the African Superplume (Huerta et al. 2009; Mulibo &
Nyblade 2013). In contrast, beneath southern Africa, Gao et al. The original seismograms are windowed 20 s before and 260 s
(2002) observed a normal MTZ thickness and concluded that the after the first P arrival and are bandpass filtered with a four-pole,
African Superplume, if it exists, is constrained in the lower mantle two-pass filter in the frequency range of 0.02–0.2 Hz. The filtered
beneath the study area. This conclusion is supported by the seismic three-component seismograms with signal-to-noise ratios (SNR)
tomography study of James et al. (2001). The present study repre- exceeding 4 on the vertical component are converted into radial
sents the first seismological investigations of the upper mantle and RFs by deconvolving the vertical from the radial components using
mantle transition zone beneath the ORZ using broad-band seismic the procedure of Ammon (1991). Before the computation of the RFs,
data recorded in the vicinity of the rift. a set of exponential functions with a half width of 30 s and centred
at the PP-phase are adopted to minimize the degenerating effects of
the PP arrivals on the RFs (Gao & Liu 2014b). The resulting RFs
are subsequently subjected to an SNR-based procedure to select the
2 D ATA A N D M E T H O D S high-quality RFs (see Gao & Liu 2014a for more information about
the specific definition of the SNR and details of the procedure). A
In the summer of 2012, a total of 17 broad-band seismic stations total of 1064 high-quality RFs from 22 seismic stations and 159
were deployed in northern and central Botswana along a NW–SE teleseismic events are obtained (Fig. 1).
profile with a total length of 756 km (Fig. 1), as part of the Seismic We next employ the technique developed by Gao & Liu (2014b)
Arrays for African Rift Initiation (SAFARI) project (Gao et al. to conduct moveout correction and stacking of P-to-S converted
2013; Yu et al. 2013). The 17 SAFARI broad-band stations, which phases generated from the MTZ discontinuities. The procedure is
were equipped with Quanterra Q330 digitizers and Guralp CMG-3T similar to the common-conversion-point (CCP) method (Dueker
120 s sensors, recorded continuously with a sampling rate of 50 Hz & Sheehan 1997) but was developed under the non-plane wave-
until the summer of 2014. Additionally, data from 5 broad-band front assumption, in which the ray-parameters for the direct P wave
seismic stations from the 1997–1999 Southern African Seismic and the converted S wave are different, leading to more accurate
Experiment (SASE; Gao et al. 2002) are used to enhance coverage discontinuity depths and more sharply focused arrivals relative to
and resolution of our data set (Fig. 1). techniques based on the plane-wave assumption. The geographic
Data from events in the epicentral distance range of 30–100◦ are coordinates of the ray-piercing points for each of the selected RFs
selected with a cut-off magnitude of Mc as defined by Mc = 5.2 + are computed at approximately the middle of the MTZ (535 km)
( − 30.0)/(180.0 − 30.0) − D/700.0, where  is the epicentral using the IASP91 Earth model. Based on the locations of the ray-
distance in degree and D is the focal depth in km (Liu & Gao 2010). piercing points (Fig. 1), we group the RFs into bins with a radius
Such an empirical formula was designed to balance the quantity and of 1◦ (Fig. 2). The bins are aligned along the latitudinal and longi-
quality of the data to be processed. tudinal lines with a separation of one geographic degree between
1410 Y. Yu et al.

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

Figure 3. Depth series from stacking of RFs in radius = 1◦ bins along nine latitudinal profiles. The thick red lines show the mean depth series averaged over
all the 50 bootstrap iterations, and the bordering thin black lines show the mean ± two standard deviations. The circles indicate the average depths of the d410
and d660, and the error bars show two standard deviations of the depths. The number on the top of each trace represents the number of RFs in the bin.

adjacent bins. The RFs in a given bin are then moveout corrected and Figs 3 and 4 show all of the resulting depth series along latitudinal
stacked for the depth range of 350–700 km with a vertical resolution and longitudinal profiles, respectively. We also stack the RFs using
of 1 km (eq. 2 in Gao & Liu 2014b). Because a 1-D Earth model is data recorded by each of the stations (Fig. 5) for the purpose of
used for the migration and stacking of the RFs, the observed MTZ investigating data quality variations between the stations. Note that
discontinuity depths are apparent rather than true depths. A boot- the existence of lateral variations in the velocity structure can reduce
strap resampling approach with 50 resampling iterations (Efron & the stacking amplitude and broaden the peak at the estimated MTZ
Tibshirani 1986) is applied to obtain the mean and stan- discontinuity depths. In addition, the stacking results near the edges
dard deviation (SD) of the MTZ discontinuity depths of the study area are not as reliable as those in the interior due to
(Liu et al. 2003). the lower number of RFs.
 Okavango mantle discontinuities 1411

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

Figure 4. Same as Fig. 3, but for 10 longitudinal profiles. Note plot (i) contains two profiles.

2.2 Traveltime residuals about ±0.1 s, that is, within five data points from the true value.
The mean residual from the event is removed to obtain relative trav-
Both thermal anomalies in the vicinity of and velocity undulations
eltime residuals. The use of relative traveltime residuals minimizes
above the MTZ discontinuities can lead to observed variations in the
the effects of errors of hypocentral parameters and lateral velocity
depths of the discontinuities. In order to provide constraints on the
heterogeneities outside the study area. Events recorded by less than
interpretation of the resulting MTZ discontinuity depths, we also
five stations are not used.
measure P- and S-wave traveltime residuals. P- and S-wave arrivals
We use the following equation to correct for traveltimes due to
relative to the IASP91 Earth model are manually picked on the
variations in station elevation (Nolet 2008)
vertical and transverse components, respectively. The accuracy of
the hand-picked residuals is dependent on the signal strength rela-
tive to the background noise, as well as the sharpness of arrivals. hi
δticj = δti j − , (1)
For most of the seismograms, the uncertainty is estimated to be v × cos[a sin(Ri j × v)]
1412 Y. Yu et al.

coherency of the d410 and d660 arrivals are higher on the bin-
specific depth series (Figs 3 and 4) than on the station-specific ones
(Fig. 5). To produce spatially continuous images for the observed
depths and MTZ thickness at the bins, we fit the observations using
a continuous curvature surface gridding algorithm, with a tension
factor of 0.5 (Smith & Wessel 1990; Fig. 6). For the majority of the
study area, the SD of the resulting MTZ thickness is less than 5 km
(Fig. 6d).
The mean apparent depth of d410 for the entire study area is
398 ± 6 km, that of d660 is 646 ± 7 km and the average MTZ
thickness is 248 ± 6 km which is almost identical to the normal
thickness of 250 km in the IASP91 Earth model and is statistically

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

consistent with the global average from previous studies (Lawrence
& Shearer 2006; Tauzin et al. 2008). The apparent depths of both
d410 and d660 decrease systematically towards the southeast, from
approximately 410 km and 660 km beneath the NW end to 395 km
and 645 km near the SE extreme of the study area (Figs 6a and b).
The depths beneath the former area (Congo Craton) are identical
to those in the IASP91 Earth model, and those beneath the latter
(Kalahari Craton in southern Botswana) are about 15 km shallower,
an observation that is consistent with the study of Gao et al. (2002)
for southern Botswana.
The results show that, relative to the immediate adjacent areas,
the ORZ displays no observable anomalies in either the apparent
depths of the discontinuities or the resulting MTZ thickness. An
area of anomalously thin MTZ is observed on the SW corner of the
study area, approximately centred at (22◦ E, 22◦ S). The anomaly is
caused by both a depressed d410 and an uplifted d660, relative to
the immediate surrounding areas (Figs 3g and 4b).
Figure 5. Depth series for each of the stations from moveout correction and
stacking of the RFs. The number above each trace is the number of RFs from 3.2 Body-wave traveltime residuals
the station, and the tectonic region that the stations belong to is labelled on
the top of the plot. The bottom row shows station names. RP: Rehoboth We have hand-picked 1832 P- and 1356 S-wave residuals from
Province; MB: Magondi Belt. 172 teleseismic events. The residuals are displayed using different
approaches. In Figs 7(a) and (b), station averages with standard
where hi is the elevation (in km) for the ith station, δtij is the original deviations are displayed along a profile orthogonal to the strike
residual (in seconds) observed at the ith station from the jth event, of the ORZ. Such an approach implies that the anomalies in the
Rij is the ray parameter (in s km−1 ), and v is the average velocity (in residuals originate near the surface, which is unlikely. To more
km s−1 ) in the layer above sea-level. In this study we use 5.5 km s−1 reasonably represent the observations, the residuals from individual
√ station-event pairs are displayed at the ray-piercing point at the depth
for P-wave, and 5.5/ 3 km s−1 for S-wave velocities.
of 100 km, which is approximately the centre of the lithosphere in
the study area (Muller et al. 2009). The residuals are then projected
3 R E S U LT S to an NW–SE oriented profile (Figs 7a and b) and are also plotted
as images (Figs 7c and d).
3.1 Apparent discontinuity depths and MTZ thickness Both the P- and S-wave traveltime residuals decrease towards the
SE. The spatially averaged residuals have a peak-to-peak magnitude
A total of 53 bins with clearly observable d410 or d660 arrivals of about 0.6 s for P and 1.8 s for S waves (Fig. 7). The SE boundary
are obtained from our data set. Among these bins, 53 of them pos- of the ORZ marks the NW end of a linear increase of the traveltime
sess a strong arrival in the depth range of 350–450 km, and 52 residuals (especially for S waves). Nevertheless, significant travel-
have a strong arrival in the depth range of 600–700 km. The arrival time delays are not observed in the ORZ, which is in sharp contrast
with the maximum amplitude in the above ranges is considered with other continental rift zones. For instance, across the Baikal
as the P-to-S phase from d410 and d660, respectively (Figs 3 and rift zone, which was similarly developed along the edge of a stable
4). All of the 53 depth series plotted along nine latitudinal lines craton (the Siberian Craton), a P-wave traveltime delay as large as
(from 16◦ S to 24◦ S with an increment of 1◦ ) are shown in Fig. 3, 1 s relative to surrounding areas was observed (Gao et al. 2003).
and those along 10 longitudinal lines are plotted in Fig. 4. The
observed discontinuity depths and the MTZ thickness for each of
4 DISCUSSION
the bins are summarized in Table 1. To the first-order, the station-
specific results (Fig. 5) are similar to those obtained from the 1◦ bins
4.1 Causes of apparent variations in MTZ discontinuity
(Figs 3 and 4). However, due to off-profile variations in the discon-
depth
tinuity depths and the fact that the RFs recorded by a station sample
an area that is mostly larger than the 1◦ size of the bins especially To the first-order, the apparent depths of d410 and d660 (Fig. 6)
at the d660 depth, the overall strength (relative to noise) and spatial as well as the P- and S-wave traveltime residuals (Fig. 7) decrease
 Okavango mantle discontinuities 1413

Table 1. Results of receiver function stacking for each of the bins. clon and clat: longitude and
latitude of the bin’s center; d410: depth of the d410; sig4: standard deviation of the depth of the
d410; d660: depth of the d660; sig6: standard deviation of the depth of the d660; MTZ: mantle
transition zone thickness; sigm: standard deviation of MTZ thickness; a4: stacking amplitude of the
d410; a6: stacking amplitude of the d660; NRFs: number of receiver functions in the bin.
clon clat d410 sig4 d660 sig6 MTZ sigm a4 a6 NRFs
(deg.) (deg.) (km) (km) (km) (km) (km) (km)
20.0 −18.0 402 4.1 651 8.5 248 8.2 0.0215 0.0151 29
20.0 −19.0 405 1.8 – – – – 0.0313 0.0000 36
21.0 −17.0 404 4.6 661 4.0 256 5.9 0.0336 0.0268 12
21.0 −18.0 405 7.1 645 8.8 239 11.1 0.0200 0.0150 41
21.0 −19.0 401 3.1 658 7.5 257 8.4 0.0176 0.0189 117
21.0 −20.0 399 0.9 658 3.8 258 3.6 0.0218 0.0265 146

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

21.0 −21.0 402 0.8 653 4.8 250 4.7 0.0208 0.0197 88
21.0 −22.0 404 1.9 646 3.3 242 3.7 0.0269 0.0309 20
22.0 −16.0 404 5.5 657 2.5 253 5.1 0.0292 0.0280 15
22.0 −17.0 409 2.6 657 1.3 247 2.8 0.0175 0.0259 55
22.0 −18.0 410 1.7 655 7.6 244 7.8 0.0207 0.0093 45
22.0 −19.0 398 2.9 661 5.8 263 6.1 0.0171 0.0218 85
22.0 −20.0 401 0.9 654 2.0 253 2.1 0.0245 0.0212 165
22.0 −21.0 400 1.8 648 4.6 247 5.1 0.0167 0.0172 114
22.0 −22.0 408 2.1 636 4.7 227 5.4 0.0277 0.0356 37
22.0 −23.0 407 1.5 641 8.7 233 8.8 0.0464 0.0264 10
23.0 −17.0 415 4.6 657 2.4 241 4.5 0.0148 0.0198 75
23.0 −18.0 399 7.0 655 1.8 255 7.0 0.0132 0.0147 133
23.0 −19.0 399 1.5 648 2.4 249 2.9 0.0264 0.0170 99
23.0 −20.0 399 1.5 646 2.2 246 2.7 0.0222 0.0243 101
23.0 −21.0 396 2.4 645 2.9 249 4.0 0.0189 0.0244 110
23.0 −22.0 394 3.9 633 3.7 239 4.7 0.0165 0.0241 61
24.0 −18.0 394 2.6 652 2.7 258 4.4 0.0258 0.0213 105
24.0 −19.0 396 1.3 648 2.7 251 2.7 0.0319 0.0183 143
24.0 −20.0 396 1.8 645 5.1 249 5.1 0.0266 0.0179 97
24.0 −21.0 394 1.3 641 1.8 246 1.8 0.0215 0.0267 86
24.0 −22.0 390 1.1 636 1.6 245 2.1 0.0289 0.0251 68
24.0 −23.0 390 2.0 634 2.4 244 3.2 0.0266 0.0298 29
25.0 −18.0 399 7.2 650 2.3 250 7.7 0.0207 0.0329 44
25.0 −19.0 395 2.1 642 3.6 246 4.0 0.0237 0.0173 112
25.0 −20.0 393 2.6 636 4.0 242 4.8 0.0227 0.0173 98
25.0 −21.0 397 2.3 644 3.8 246 4.1 0.0246 0.0194 48
25.0 −22.0 387 2.2 637 1.9 249 2.6 0.0193 0.0254 67
25.0 −23.0 390 1.5 637 1.6 247 1.9 0.0246 0.0285 61
26.0 −19.0 391 5.9 638 2.1 247 6.4 0.0172 0.0207 58
26.0 −20.0 395 1.9 639 2.2 243 2.4 0.0195 0.0214 92
26.0 −21.0 397 2.8 647 2.0 250 3.1 0.0222 0.0247 68
26.0 −22.0 394 6.2 643 4.6 249 8.2 0.0171 0.0188 65
26.0 −23.0 394 1.7 639 2.2 245 2.6 0.0205 0.0196 81
26.0 −24.0 394 1.5 642 2.8 247 3.0 0.0250 0.0166 59
27.0 −19.0 394 4.9 646 2.1 252 5.4 0.0299 0.0307 9
27.0 −20.0 395 4.0 645 3.5 250 5.4 0.0239 0.0213 43
27.0 −21.0 396 3.7 643 3.5 246 4.4 0.0155 0.0257 67
27.0 −22.0 392 3.8 645 3.7 252 6.0 0.0159 0.0210 65
27.0 −23.0 407 9.7 642 3.1 235 10.3 0.0114 0.0185 40
27.0 −24.0 395 2.2 643 3.9 248 5.1 0.0309 0.0244 25
28.0 −20.0 390 3.6 640 2.4 250 3.6 0.0128 0.0270 20
28.0 −21.0 392 1.6 647 8.2 255 8.3 0.0249 0.0181 60
28.0 −22.0 391 2.5 641 6.1 250 6.3 0.0226 0.0173 55
28.0 −23.0 397 1.6 634 4.6 237 4.4 0.0133 0.0303 24
29.0 −21.0 395 4.8 649 10.5 253 12.5 0.0183 0.0192 19
29.0 −22.0 391 3.8 649 7.4 258 9.1 0.0254 0.0190 29
29.0 −23.0 401 3.1 642 10.4 241 11.0 0.0222 0.0172 17

from the Congo Craton at the NW part of the profile towards the velocity variations in the upper mantle, probably by variations in the
Kalahari Craton at the SE. The fact that the MTZ thickness remains thickness of the lithosphere. In the following, we quantify this using
the same throughout most of the study area suggests that the appar- our RF results and lithospheric thickness estimated from previous
ent variations of the observed depths are mainly caused by lateral studies (James et al. 2001; Muller et al. 2009).
1414 Y. Yu et al.

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

Figure 6. (a) Smoothed spatial distribution of resulting d410 depth; (b) same as (a) but for d660; (c) MTZ thickness; (d) standard deviations of the MTZ
thickness measurements. Circles indicate the centre of circular areas with high-quality measurements. Note that in (a)–(c), only measurements that are 5 or
more km different from the corresponding global average are coloured.

As discussed above, relative to the Congo Craton, both d410 γ = dln(Vs )/dln(Vp ), is required. Under the assumption of γ = 1.7,
and d660 beneath the Kalahari Craton are uplifted by about 15 km. which is the same as that found for the stable central and eastern
Under the assumption that this apparent uplift is purely caused by United States (Gao & Liu 2014a), the S-wave velocity anomaly
velocity anomalies, we estimate that a 15 km uplift corresponds to a is about +2 per cent and the predicted S-wave traveltime residual
Vp anomaly of +1.2 per cent in the depth range of 0–410 km using is −1.8 s, which is once again consistent with the observed value
the approach of Gao & Liu (2014b). If the anomalous zone is limited (Fig. 7).
to a 100 km thick layer, the required anomaly is about 5 per cent, We next explore the possibility that the observed velocity anoma-
and this value reduces to about 3 per cent if the thickness is 150 km. lies originate from lateral variations of lithospheric thickness
In addition, the velocity anomalies resulting in a 15 km apparent beneath the study area. Seismic tomography studies indicated that
uplift of both d410 and d660 can produce a P-wave traveltime the Kalahari Cratonic root extends to at least 250 km and perhaps
residual of about −0.6 s, which is almost exactly what was observed as deep as 300 km (James et al. 2001). On the other hand, Muller
(Fig. 7). To estimate the predicted S-wave traveltime residual using et al. (2009) found the lithospheric thickness in the vicinity of
the apparent depression of the discontinuities, the γ value, which is the Damara belt in the ORZ to be approximately 160 km. These
the ratio between S- and P-wave relative velocity anomalies, that is, studies suggest that beneath the Kalahari Craton, the lithosphere is
 Okavango mantle discontinuities 1415

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

Figure 7. Teleseismic P- and S-wave traveltime residuals displayed above the ray-piercing points at 100 km depth. (a) P-wave residuals projected on to the
profile shown in (c). Dots are individual event values, and circles with error bars are station-averaged values. (b) Same as (a) but for S-wave residuals. (c)
Spatial distribution of P-wave residuals. (d) Spatial distribution of S-wave residuals. The numbers along the profile in (c) and (d) indicate the distance in km
from the SE boundary fault of the ORZ.

approximately 100–150 km thicker than that beneath the northwest- Kalahari Craton imply higher-than-normal velocities and are con-
ern part of the study area. If we assume that this 100–150 km ex- sistent with seismic tomographic studies (James et al. 2001). We
cessive lithospheric thickness is solely responsible for the observed note that considerable discrepancies in the experimentally deter-
15 km apparent uplift of the MTZ discontinuities and variations in mined Clapeyron slopes were found among previous studies (see
traveltime residuals, the required P-wave velocity contrast is 3–5 Litasov et al. 2005 and Ghosh et al. 2013 for some examples),
per cent between the lithosphere and the asthenosphere. Velocity making it difficult to reliably convert undulations in the observed
contrasts of similar magnitude are commonly found in numerous MTZ discontinuity depths into temperature anomalies. Using the
previous seismic tomography and RF studies (e.g. Gao et al. 2003; Clapeyron slopes of +2.9 MPa K−1 for d410 and −2.1 MPa K−1
Li et al. 2003; Rychert et al. 2007; Eaton et al. 2009). for d660 (Bina & Helffrich 1994), the resulting MTZ thickness of
The normal apparent discontinuity depths beneath the Congo 248 ± 6 km corresponds to a thermal anomaly ranging between
Craton and the ORZ suggest normal (relative to the IASP91 Earth −34 K and +68 K.
model) upper mantle P- and S-wave velocities. This is drastically The anomalously thin MTZ observed on the SW corner of the
different from the mature segments of the EARS beneath which study area (Fig. 6c) has an estimated thickness of 235 km, corre-
greater-than-normal apparent discontinuity depths (e.g. Thompson sponding to an estimated thermal anomaly of +127 K. This abnor-
et al. 2015) and lower-than-normal velocities (Bastow 2012) have mally thin MTZ, if confirmed, could suggest significant transfer-
been reported. Similarly, the shallow d410 and d660 beneath the ring of heat from the lower to the upper mantle beneath western
1416 Y. Yu et al.

Botswana. Such a transfer might be related to the African Super- varying far-field stress from continental collision, and non-uniform
plume, which is thought to be confined in the lower mantle beneath influence of mantle upwelling or downwelling (e.g. Sabadini &
southern Africa (James et al. 2001; Gao et al. 2002). Unfortunately, Guinchi 1992; Doglioni 1993). We speculate that for the ORZ,
this area is located on the edge of the region sampled by a limited variations of lithospheric thickness combined with the presence
number of RFs (Figs 1 and 2). of previously ruptured lithosphere give rise to one or more of the
The above discussions assume anhydrous conditions. Some pre- above heterogeneities, leading to rift initiation. Such a passive rift-
vious MTZ studies involved the presence of water in the MTZ to ing model is consistent with conclusions from the study of Leseane
explain the observed larger-than-normal MTZ thickness (Blum & et al. (2015) that the ORZ could be initiated through fluid-assisted
Shen 2004; Cao & Levander 2010; Mohamed et al. 2014; Thomson lithospheric weakening without asthenospheric involvement. Pas-
et al. 2015), and one of them (Blum & Shen 2004) investigated the sive stretching of the lithosphere may eventually lead to active up-
MTZ beneath the Kalahari Craton. Water can uplift the d410 and welling of the asthenosphere and resulting in the transition from
depress the d660, resulting in a thicker MTZ (Litasov et al. 2005). passive to active rifting (Huismans et al. 2001). This model could

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

The addition of water also reduces seismic velocities in the MTZ explain why most of the world’s continental rifts developed along
(Inoue et al. 1998) and thus leads to an apparent MTZ thickening. the edges of cratonic blocks, where rapid changes of lithospheric
Therefore, in order to produce the observed normal apparent MTZ thickness are commonly found (Keller et al. 1991).
thickness, higher-than-normal temperature in the MTZ is required
to reduce the thickness. While it is possible that the effects of water
and temperature can cancel each other and result in a normal MTZ 5 C O N C LU S I O N S
thickness, as observed, the chance for this to happen is not high.
This first receiver function study of the incipient ORZ reveals a nor-
This is because it requires a specific combination of the amplitude
mal MTZ thickness, arguing against a strong thermal anomaly in the
of the temperature anomaly and the amount of water through out
upper mantle and MTZ beneath most of the study area. The appar-
most of the study area where normal MTZ thickness is observed.
ently shallower-than-normal MTZ discontinuities observed beneath
Additionally, some studies suggested that excessive amount of wa-
the Kalahari Craton can be explained by the existence of a litho-
ter can increase the pulse width of d410 and decrease its amplitude
sphere that is 100–150 km thicker than that beneath the Congo Cra-
(Smyth & Frost 2002; Van der Meijde et al. 2003). This is contra-
ton and the Damara belt. In addition, the ORZ is not associated with
dictory to the sharp arrivals associated with both the d410 and d660
significant delays in teleseismic P- and S-wave traveltime residu-
discontinuities (Figs 3 and 4). Therefore, while the pervasive exis-
als, implying the absence of significant mantle anomalies beneath
tence of water in the study area cannot be completely ruled out, an
the ORZ. These observations are inconsistent with the existence
anhydrous MTZ with nearly normal temperature can satisfactorily
of mantle plumes or a broadly upwarped asthenosphere beneath the
explain the observations.
study area. Instead, they provide unprecedented supporting evidence
for a model of rift initiation involving lithospheric fracturing due to
differential movements between tectonic blocks along pre-existing
4.2 Implications for rifting mechanisms zones of weakness such as edges of cratonic blocks.
The observed normal MTZ thickness, lack of significant traveltime
residuals and the absence of elevated mantle conductivity (Khoza AC K N OW L E D G E M E N T S
et al. 2013) associated with the incipient ORZ suggest the absence
of high-temperature anomalies in both the MTZ and the upper man- This study is part of the passive seismic component of an inter-
tle, ruling out contributions of one or more mantle plumes in rift disciplinary project funded by the United States National Science
initiation and favouring a ‘passive’ origin. Foundation (NSF) Continental Dynamics Program. We appreciate
A recent geodetic study (Malservisi et al. 2013) suggested that discussions with and encouragements from other participants of the
with respect to the African continent, the rigid South African block project. Careful reviews by three anonymous reviewers and Edi-
rotates clockwise, probably along the Damara belt. Such differential tor G.R. Keller significantly improved the manuscript. We thank
movements between relatively rigid tectonic blocks can rupture the the Program for Array Seismic Studies of the Continental Litho-
pre-existing weak zones along edges of the blocks. The fractures can sphere (PASSCAL) Instrument Center for equipment and logistical
then propagate downwards through the lithosphere without mantle support, and the Incorporated Research Institutions for Seismology
upwelling anomalies, and rifting processes can continue as long as (IRIS) Data Management Center for achieving both the SAFARI
the regional dominant stresses are extensional or transtensional. A and SASE data sets used in the study. Field assistance provided
likely environment for reactivation of shear zones adjacent to the by Angela M. Reusch from PASSCAL and Keletso Kaisara from
ORZ is the transcontinental Mwembeshi dislocation zone which the University of Botswana is greatly appreciated. YY was partially
experiences left-lateral transtension as the consequence of differen- supported by the China Scholarship Council under contract No.
tial extension between Nubia and Somalia (Daly et al. 1989). The 2011645027 and CR was partly supported by the Missouri Univer-
Mwembeshi shear zone, which intersects the Luangwa and Malawi sity of Science and Technology Chancellor’s Fellowship. The field
rift zones farther to the northeast of the ORZ, has been suggested and data analysis tasks related to this study were supported by NSF
to exert geometrical control on rift basin development within the grant EAR-1009946 to SG and KL.
EARS (Versfelt & Rosendahl 1989). Similar lithospheric-scale Pro-
terozoic suture shear zones have been cited as a possible origin for
the Kenya Rift (Birt et al. 1997). Differential movements between REFERENCES
tectonic blocks in a relatively stable tectonic plate can originate from Ammon, C.J., 1991. The isolation of receiver effects from teleseismic P-
many factors, such as differences in the direction and/or strength waveforms, Bull. seism. Soc. Am., 81, 2504–2510.
of mantle flow (either as driving or resistant forces), uneven distri- Bastow, I.D., 2012. Relative arrival-time upper-mantle tomography and the
bution of ridge push and slab pull forces across the plate, spatially elusive background mean, Geophys. J. Int., 190, 1271–1278.
 Okavango mantle discontinuities 1417

Begg, G.C. et al., 2009. The lithospheric architecture of Africa: Seismic Inoue, T., Weidner, D.J., Northrup, P.A. & Parise, J.B., 1998. Elastic prop-
tomography, mantle petrology, and tectonic evolution, Geosphere, 5, erties of hydrous ringwoodite (Q-phase) in Mg2 SiO4 , Earth planet. Sci.
23–50. Lett., 160, 107–113.
Bina, C.R. & Helffrich, G., 1994. Phase transition Clapeyron slopes and James, D.E., Fouch, M.J., VanDecar, J.C. & Van Der Lee, S., 2001. Tecto-
transition zone seismic discontinuity topography, J. geophys. Res., 99, spheric structure beneath southern Africa, Geophys. Res. Lett., 28, 2485–
15 853–15 860. 2488.
Birt, C.S., Maguire, P.K.H., Khan, M.A., Thybo, H., Keller, G.R. & Patel, Keller, G.R., Khan, M.A., Morgan, P., Wendlandt, R.F., Baldridge,
J., 1997. The influence of pre-existing structures on the evolution of the W.S., Olsen, K.H., Prodehl, C. & Braile, L.W., 1991. A comparative
southern Kenya Rift Valley-evidence from seismic and gravity studies, study of the Rio Grande and Kenya rifts, Tectonophysics, 197, 355–
Tectonophysics, 278, 211–242. 371.
Blum, J. & Shen, Y., 2004. Thermal, hydrous, and mechanical states of the Khoza, T.D., Jones, A.G., Muller, M.R., Evans, R.L., Miensopust, M.P. &
mantle transition zone beneath southern Africa, Earth planet. Sci. Lett., Webb, S.J., 2013. Lithospheric structure of an Archean craton and adjacent
217, 367–378. mobile belt revealed from 2-D and 3-D inversion of magnetotelluric data:
Bufford, K.M. et al., 2012. Geometry and faults tectonic activity of the Oka- example from southern Congo craton in northern Namibia, J. geophys.

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

vango Rift Zone, Botswana: evidence from magnetotelluric and electrical Res., 118, 4378–4397.
resistivity tomography imaging, J. Afr. Earth. Sci., 65, 61–71. Kinabo, B.D., Hogan, J.P., Atekwana, E.A., Abdelsalam, M.G. &
Cao, A. & Levander, A., 2010. High-resolution transition zone struc- Modisi, M.P., 2008. Fault growth and propagation during incipient
tures of the Gorda slab beneath the western United States: impli- continental rifting: insights from a combined aeromagnetic and Shut-
cation for deep water subduction, J. geophys. Res., 115, B07301, tle Radar Topography Mission digital elevation model investigation of
doi:10.1029/2009JB006876. the Okavango Rift Zone, northwest Botswana, Tectonics, 27, TC3013,
Daly, M.C., Chorowicz, J. & Fairhead, J.D., 1989. Rift basin evolution doi:10.1029/2007TC002154.
in Africa: the influence of reactivated steep basement shear zones, in Lawrence, J.F. & Shearer, P.M., 2006. A global study of transition zone
Inversion Tectonics, Vol. 44, pp. 309–334, eds. Cooper, M.A. & Williams, thickness using receiver functions, J. geophys. Res., 111, B06307,
G.D., Geol. Soc. Spec. Pub. doi:10.1029/2005JB003973.
Doglioni, C., 1993. Geological evidence for a global tectonic polarity, Leseane, K., Atekwana, E.A., Mickus, K.L., Abdelsalam, M.G., Shemang,
J. geol. Soc. Lond., 150, 991–1002. E.M. & Atekwana, E.A., 2015. Thermal perturbations beneath the in-
Dueker, K.G. & Sheehan, A.F., 1997. Mantle discontinuity structure from cipient Okavango Rift Zone, northwest Botswana, J. geophys. Res., 120,
midpoint stacks of converted P to S waves across the Yellowstone hotspot 1210–1228.
track, J. geophys. Res., 102, 8313–8327. Li, A., Forsyth, D.W. & Fischer, K.M., 2003. Shear velocity structure and
Eaton, D.W., Darbyshire, F.A., Evans, R.L., Grutter, H., Jones, A.G. & Yuan, azimuthal anisotropy beneath eastern North America from Rayleigh wave
X., 2009. The elusive lithosphere-asthenosphere boundary (LAB) beneath inversion, J. geophys. Res., 108, doi:10.1029/2002JB02259.
cratons, Lithos, 109, 1–22. Litasov, K.D., Ohtani, E., Suzuki, A. & Funakoshi, K., 2005. Wet sub-
Ebinger, C.J., Crow, M.J., Rosendahl, B.R., Livingstone, D.A. & LeFournier, duction versus cold subduction, Geophys. Res. Lett., 32, L13312,
J., 1984. Structural evolution of the Malawi rift, Africa, Nature, 308, doi:10.1029/2005GL022921.
627–629. Liu, K.H. & Gao, S.S., 2006. Mantle transition zone discontinuities be-
Efron, B. & Tibshirani, R., 1986. Bootstrap methods for standard errors, neath the Baikal rift and adjacent areas, J. geophys. Res., 111, B11301,
confidence intervals, and other measures of statistical accuracy, Stat. Sci., doi:10.1029/2005JB004099.
1, 54–75. Liu, K.H. & Gao, S.S., 2010. Spatial variations of crustal characteristics
Gao, S.S. & Liu, K.H., 2014a. Mantle transition zone discontinu- beneath the Hoggar swell, Algeria revealed by systematic analyses of
ities beneath the contiguous United States, J. geophys. Res., 119, receiver functions from a single seismic station, Geochem. Geophys.
6452–6468. Geosys., 11, Q08011, doi:10.1029/2010GC003091.
Gao, S.S. & Liu, K.H., 2014b. Imaging mantle discontinuities using Liu, K.H., Gao, S.S., Silver, P.G. & Zhang, Y., 2003. Mantle lay-
multiply-reflected P-to-S conversions, Earth planet. Sci. Lett., 402, ering across central South America, J. geophys. Res., 108, 2510,
99–106. doi:10.1029/2002JB002208.
Gao, S.S., Silver, P.G. & Liu, K.H., 2002. Mantle discontinuities beneath Malservisi, R., Hugentobler, U., Wonnacott, R. & Hackl, M., 2013. How
southern Africa, Geophys. Res. Lett., 29, 1291–1294. rigid is a rigid plate? Geodetic constraint from the TrigNet CGPS network,
Gao, S.S., Liu, K.H., Davis, P.M., Slack, P.D., Zorin, Y.A., Mordvinova, V.V. South Africa, Geophys. J. Int., 192, 918–928.
& Kozhevnikov, V.M., 2003. Evidence for small-scale mantle convection McCourt, S., Armstrong, R.A., Jelsma, H. & Mapeo, R.B.M., 2013.
in the upper mantle beneath the Baikal rift zone, J. geophys. Res., 108(B4), New U-Pb SHRIMP ages from the Lubango region, SW Angola: in-
2194, doi:10.1029/2002JB002039. sights into the Palaeoproterozoic evolution of the Angolan Shield,
Gao, S.S. et al., 2013. Seismic Arrays to Study African Rift Initiation, EOS, southern Congo Craton, Africa, J. geol. Soc. Lond, 170, 353–
Trans. Am. geophys. Un., 94, 213–214. 363.
Ghosh, S., Ohtani, E., Litasov, K.D., Suzuki, A., Dobson, D. & Funakoshi, Miensopust, M.P., Jones, A.G., Muller, M.R., Garcia, X. & Evans, R.L.,
K., 2013. Effect of water in depleted mantle on post-spinel transition 2011. Lithospheric structures and precambrian terrane boundaries in
and implication for 660 km seismic discontinuity, Earth planet. Sci. Lett., northeastern Botswana revealed through magnetotelluric profiling as part
371–372, 103–111. of the southern African magnetotelluric experiment, J. geophys. Res., 116,
Hanson, R.E., 2003. Proterozoic geochronology and tectonic evolution of B02401, doi:10.1029/2010JB007740.
southern Africa, Geol. Soc. Spec. Pub., 206, 427–463. Modisi, M.P., Atekwana, E.A., Kampunzu, A.B. & Ngwisanyi, T.H., 2000.
Helffrich, G., 2000. Topography of the transition zone seismic discontinu- Rift kinematics during the incipient stages of continental extension: evi-
ities, Rev. Geophys., 38, 141–158. dence from the nascent Okavango rift basin, northwest Botswana, Geol-
Huerta, A.D., Nyblade, A.A. & Reusch, A.M., 2009. Mantle transition ogy, 28, 939–942.
zone structure beneath Kenya and Tanzania: more evidence for a deep- Mohamed, A.A., Gao, S.S., Elsheikh, A.A., Liu, K.H., Yu, Y. & Fat-Helbary,
seated thermal upwelling in the mantle, Geophys. J. Int., 177, 1249– R.E., 2014. Seismic imaging of mantle transition zone discontinuities
1255. beneath the northern Red Sea and adjacent areas, Geophys. J. Int., 199,
Huismans, R.S., Podladchikov, Y.Y. & Cloetingh, S., 2001. Transition from 648–657.
passive to active rifting: relative importance of asthenospheric doming Mulibo, G.D. & Nyblade, A.A., 2013. Mantle transition zone thinning be-
and passive extension of the lithosphere, J. geophys. Res., 106, 11 271– neath eastern Africa: evidence for a whole-mantle superplume structure,
11 291. Geophys. Res. Lett., 40, 3562–3566.
1418 Y. Yu et al.

Muller, M.R. et al., 2009. Lithospheric structure, evolution and diamond Stern, R.J., 1994. Arc-assembly and continental collision in the Neoprotero-
prospectivity of the Rehoboth Terrane and Western Kaapvaal Craton, zoic African orogen: implications for the consolidation of Gondwanaland,
Southern Africa: constraints from broadband magnetotellurics, Lithos, Annu. Rev. Earth planet. Sci., 22, 319–351.
112, 93–105. Tauzin, B., Debayle, E. & Wittlinger, G., 2008. The mantle transi-
Nolet, G., 2008. A Breviary of Seismic Tomography, Cambridge Univ. Press. tion zone as seen by global Pds phases: no clear evidence for a
Nyblade, A.A., Owens, T.J., Gurrola, H., Ritsema, J. & Langston, C.A., 2000. thin transition zone beneath hotspots, J. geophys. Res., 113, B08309,
Seismic evidence for a deep upper mantle thermal anomaly beneath east doi:10.1029/2007JB005364.
Africa, Geology, 28, 599–602. Thompson, D.A., Hammond, J.O.S., Kendall, J.-M., Stuart, G.W., Helffrich,
Owens, T.J., Nyblade, A.A., Gurrola, H. & Langston, C.A., 2000. Mantle G., Keir, D., Ayele, A. & Goitom, B., 2015. Hydrous upwelling across
transition zone structure beneath Tanzania, East Africa, Geophys. Res. the mantle transition zone beneath the Afar Triple Junction, Geochem.
Lett., 27, 827–830. Geophys. Geosyst., 16, 834–846.
Reeves, C.V., 1972. Rifting in the Kalahari? Nature, 237, 95–96. van der Meijde, M., Marone, F., Giardini, D. & van der Lee, V., 2003.
Ringwood, A.E., 1975. Composition and Petrology of the Earth’s Mantle, Seismic evidence for water deep in Earth’s upper mantle, Science, 300,
McGraw-Hill. 1556–1558.

Downloaded from https://academic.oup.com/gji/article/202/2/1407/596566 by University of Botswana user on 17 May 2023

Rychert, C.A., Rondenay, S. & Fischer, K.M., 2007. P-to-S and S-to-P imag- Versfelt, J. & Rosendahl, B.R., 1989. Relationships between pre-rift structure
ing of a sharp lithosphere-asthenosphere boundary beneath eastern North and rift architecture in Lakes Tanganyika and Malawi, East Africa, Nature,
America, J. geophys. Res., 112, B08314, doi:10.1029/2006JB004619. 337, 354–357.
Sabadini, R. & Giunchi, G., 1992. Plate motion and dragging of the upper White, R. & McKenzie, D., 1989. Magmatism at rift zones: the generation
mantle: lateral variations of lithospheric thickness and their implications of volcanic continental margins and flood basalts, J. geophys. Res., 94,
for intraplate deformation, Geophys. Res. Lett., 19, 749–752. 7685–7729.
Schmeling, H., 2010. Dynamic models of continental rifting with melt gen- Wilson, D., Aster, R., Ni, J., Grand, S., West, M., Gao, W., Baldridge,
eration, Tectonophysics, 480, 33–47. W.S. & Semken, S., 2005. Imaging the seismic structure of the crust
Sengor, A.M. & Burke, K., 1978. Relative timing of rifting and volcanism and upper mantle beneath the Great Plains, Rio Grande Rift, and Col-
on Earth and its tectonic implications, Geophys. Res. Lett., 5, 419–421. orado Plateau using receiver functions, J. geophys. Res., 110, B05306,
Smith, W.H.F. & Wessel, P., 1990. Gridding with continuous curvature doi:10.1029/2004JB003492.
splines in tension, Geophysics, 55, 293–305. Yu, Y. et al., 2013. Crustal and mantle structure and anisotropy beneath the
Smyth, J.R. & Frost, D.J., 2002. The effect of water on the 410-km discon- incipient segments of the East African Rift System: preliminary results
tinuity: an experimental study, Geophys. Res. Lett., 29, 1–4. from the ongoing SAFARI, AGU Fall Meet. Abstracts, 1, 2542.