Fluid-controlled faulting process in the Asal Rift, Djibouti,
from 8 yr of radar interferometry observations
Cécile Doubre Earth and Space Science Department, University of California, Los Angeles, California 90095, USA
Gilles Peltzer Earth and Space Science Department, University of California, Los Angeles, California 90095, and
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91125, USA
ABSTRACT
The deformation in the Asal Rift (Djibouti) is characterized by magmatic inflation, diking, distributed extension, fissure opening, and normal faulting. An 8 yr time line of surface
displacement maps covering the rift, constructed using radar interferometry data acquired
by the Canadian satellite Radarsat between 1997 and 2005, reveals the aseismic behavior of
faults and its relation with bursts of microseismicity. The observed ground movements show
the asymmetric subsidence of the inner floor of the rift with respect to the bordering shoulders
accommodated by slip on three of the main active faults. Fault slip occurs both as steady creep
and during sudden slip events accompanied by an increase in the seismicity rate around the
slipping fault and the Fieale volcanic center. Slip distribution along fault strike shows triangular sections, a pattern not explained by simple elastic dislocation theory. These observations
suggest that the Asal Rift faults are in a critical failure state and respond instantly to small
pressure changes in fluid-filled fractures connected to the faults, reducing the effective normal
stress on their locked section at depth.
Keywords: Afar, Asal Rift, faulting process, radar interferometry.
INTRODUCTION
A dense network of northwest-southeast–
striking normal faults and fissures dissects
the basaltic surface of Afar (Manighetti et al.,
1998; Gupta and Scholz, 2000a). Between the
Ghoubbet Gulf and Asal Lake, the present-day
deformation is concentrated in the 10-km-wide
Asal Rift, bounded by fast-slipping normal
faults (Stein et al., 1991; Manighetti et al., 1998)
that accommodate most of the N45°E, 11–17
mm/yr diverging motion between Arabia and
Somalia (Ruegg et al., 1987; Vigny et al., 2006).
In November 1978, a seismo-volcanic crisis
produced the Ardukoba fissure eruption, involving 2 m of horizontal extension between the rift
shoulders, 0.7 m of subsidence of the inner floor,
and as much as 1 m of slip on large faults (Abdallah et al., 1979). The extension rate measured
across the activated part of the rift remained high
(60 ± 10 mm/yr) until 1987 (Kasser et al., 1987),
and then returned to a value of ~17 mm/yr (Cattin
et al., 2004), comparable to the long-term opening velocity (De Chabalier and Avouac, 1994).
Creep meters located across two faults showed
elevated, postcrisis slip rates following an exponential decay (Ruegg and Kasser, 1987; Doubre,
2004). The seismicity in the rift is characterized
by earthquakes of magnitude <3, concentrated
in crises of a few months in the Fieale caldera,
the Disa Le Mallo (DLM) subrift, and along the
northern bounding faults (Lépine and Hirn, 1992;
Doubre, 2004) (Figs. 1A, 1B).
In this paper we use 8 yr of synthetic aperture radar interferometry (InSAR) data to follow the evolution of the surface deformation
in the Asal Rift. We investigate the temporal
behavior of the faults in relation with episodes
of increased seismicity and possible changes in
crustal fluid pressure.
INSAR MEASUREMENTS
InSAR provides spatially continuous maps of
the surface displacement vector component along
the radar line of sight. Over the Asal Rift, the line
of sight of the Radarsat satellite on descending
passes has an azimuth of ~N283° and an incidence
angle of ~36° off nadir. The quasi-uninterrupted
sequence of data acquisitions every 24 days since
1997 (Fig. 1C) allowed us to construct a time line
of the surface displacement of the Asal Rift using
the small baseline subset approach (Berardino
et al., 2002). Because of the radar phase propagation delay in the troposphere (Zebker et al.,
1997), the phase evolution observed in the map
series shows a 12 month oscillation pattern of
large amplitude compared to the tectonic signal.
We worked around this problem by making two
kinds of observation. First, the linear trend of the
ground movement is obtained by computing the
mean line-of-sight velocity at each image pixel
over the 8 yr observation period, smoothing out
the effects of the seasonal phase oscillations
(Fig. 1A). Second, the slip history on a fault is
obtained by observing the time evolution of the
relative displacement between two nearby points
on either side of its trace. In this procedure, most
of the seasonal signal, which is spatially correlated, is cancelled.
MEAN SURFACE VELOCITY FIELD
The 8 yr average line-of-sight velocity map
provides a spatially continuous view of the
deformation field over the Asal Rift, highlighting the active structures and zones of concentrated deformation (Figs. 1A and 2). Slip rates
indicated here are line-of-sight components of
fault slip vectors. The displacement field bears a
strong asymmetry across the rift with a maximum
subsidence of the rift floor against fault γ and a
large uplift of the northern shoulder. This relative movement is accommodated at the surface
by ~2 mm/yr of slip along fault γ and distributed
deformation across the DLM subrift, probably
indicating faster slip (~7 mm/yr) on the deeper
section of the fault (Figs. 1A and 2). Near long
42.5°E, slip is transferred from fault γ to fault α
following a set of en echelon scarps (Fig. 1A).
The slip rate gradually decreases to the northwest along fault α, vanishing where the fault
splits into two branches. The subsidence of the
rift floor is accommodated to the southwest by
north-dipping faults, including β (~1.5 mm/yr),
A, and minor faults between A and α (Figs. 1A
and 2C). The 8 yr average slip rates on β and
γ are consistent with the Quaternary rates estimated from geological data on these two faults
(Stein et al., 1991).
South of the Fieale caldera, only fault E
appears to be slipping at the surface. With a
mean slip rate of ~2 mm/yr, this fault accommodates the uplift of the southern shoulder of
the active rift. The major faults F and H, which
form the southern border of the inner rift, do not
show any evidence of movement in the past 8 yr
(Fig. 2B). The creeping faults and the subsidence of the rift floor observed in the data appear
to be concentrated in the northeastern part of the
rift, where most of the surface deformation was
found after the 1978 crisis. This indicates that
the current stress conditions are still dominated
by the regime prevailing in 1978.
FAULT SLIP HISTORY AND SEISMICITY
Faults γ, β, and E show evidence of creep
near the surface in the average velocity map
(Fig. 1A). For each fault we compare the slip
evolution with the cumulative number of earthquakes and associated moment release occurring in the area of the creeping section (Fig. 3).
The slip evolution on fault γ shows two rapid
slip events, interrupting periods of slow and
steady creep (Fig. 3A). The first event occurred
in December 2000 with a movement of 6.5 ±
2.7 mm; the second occurred in March 2003
with a smaller slip increment of 3.5 ± 5.0 mm.
© 2007 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected].
GEOLOGY,
January
2007
Geology,
January
2007;
v. 35; no. 1; p. 69–72; doi: 10.1130/G23022A.1; 4 figures; Data Repository item 2007022.
69
A
13°N
RS Ara
AR
Lake
Asal
B
11°38′N
Ardukoba
11°N
α2
γ2
α1
P3
Fi
D
42°26′E
Ghoubbet
Gulf
F
11°34′N
0
2
Magnitude
-1 0 1 2 3
γ1
DL
M
E
H
G-K
P1
42°28′E
Fig 3g
42°30′E
42°32′E
42°30′E
42°32′E
2 km
2004
2002
1
GSA Data Repository item 2007022, Table DR1,
pairs of Radarsat orbits used in this study, is available online at www.geosociety.org/pubs/ft2007.htm,
or on request from [email protected] or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO
80301, USA.
70
2
0
1
2
3
4
5
6
7
8
Distance along profile P1 (km)
B
0
2000
1998 2000 2002 2004
Image1
Figure 1. Mean surface velocity and seismicity in Asal Rift between 1997 and 2005. A: Stack
of all interferograms used in this study. Colors represent average ground velocity toward
satellite. Intensity is radar backscatter image with illumination from east-southeast. Roman
and Greek letters refer to major faults (Manighetti et al., 1998; Stein et al., 1991). Solid lines
indicate location of profiles shown in Figure 2. DLM—Disa Le Mallo subrift; Fi—Fieale
caldera; G-K—Galaele Kôma lava field. Inset: Location map of studied area. Dashed box is
area covered by Radarsat data on descending track 324. Arrow points to study area (black
square) shown in A and B. Ara—Arabia; Som—Somalia; RS—Red Sea; AR—Aden Ridge.
B: 1997–2005 microseismicity recorded by Arta Geophysical Observatory permanent network and during 2000–2001 campaign (Doubre, 2004). Duration magnitudes of event range
from –1.8 to 3.0. Dark gray area shows 1978 Ardukoba eruption. Box shows area covered
by Figure 3G. C: Graph of image pairs analyzed in this study (black squares; Table DR11)
plotted as function of acquisition dates of first image (horizontal axis) and second image
(vertical axis).
Errors on slip rates are based on uncertainties on
linear regressions in Figure 3. Both events coincide in time with periods of increased seismic
activity near the creeping fault. A swarm of shallow (<2 km) earthquakes associated with fissure
opening occurred in the hanging wall of fault γ in
December 2000 (Doubre, 2004) (Fig. 3G). This
sequence was followed by an increase in seismicity below the Fieale caldera for a period of
six months. In March 2003, a similar sequence
of shallow earthquakes occurred on the southwestern side of the DLM subrift.
Fault β is a smaller, northeast-dipping fault
along the northeastern side of a narrow horst
structure, facing fault γ (Fig. 1A). The slip evolution on β shows that the fault was locked during the 8 yr period of observation, except for a
slip event of 9.3 ± 2.6 mm in March 2003, coincident with the second event on γ (Fig. 3C) and
Velocity
(mm/yr)
8
4
a swarm of microearthquakes located in the area
of maximum slip on β (Fig. 3G).
Fault E extends across a young lava field and
is associated with a topographic scarp of ~10 m
(Figs. 1A and 2B). The evolution of slip on E
shows a steady slip rate of 1.3 ± 0.1 mm/yr until
a slip event of 3.5 ± 1.7 mm in April–May 2002.
After this event, the fault slipped at a lower rate
of 0.6 ± 0.2 mm/yr. As for faults γ and β, the
event on E is associated with an increase in the
seismic activity. However, the coincident earthquakes appear to be distributed over a broad
area around the fault and below the Fieale caldera and not clustered in the fault hanging wall
(Fig. 3G).
SLIP DISTRIBUTION ALONG
FAULT LENGTH
Figure 4 shows the time evolution of the slip
distribution along the three previous faults.
On fault γ, the main creeping section extends
~2.2 km to the southeast from the transfer zone
with α (Fig. 4A). Subfault γ1 also shows fast
slip along its subaerial section before it enters
the Ghoubbet Gulf. To the northwest, the
movement transfers to fault α over a distance
0
C
Elevation (m)
200 SW
H
F
100
C
1998
42°28′E
β
NE
6
P2
11°32′N
42°26′E
D
0
km
b
γ1
F
δ
γ2
Fi
11°34′N
β
A
11°36′N
Som 11°36′N
42°E 43°E
δ
Elevation (m)
300 SW
H
200
100
ε
Image2
Lake Asal
11°38′N
A
Goubbet Gulf
Line-of-sight
velocity (mm/yr)
0
4
8
Elevation (m)
α β γ2 NE
200
100
E
A
Velocity
(mm/yr)
Velocity
(mm/yr)
4
4
2
0
2
-2
0
1
2
3
4 0 0
1
2
3
Distance along P2 (km) Distance along P3 (km)
Figure 2. Elevation and mean line-of-sight
velocity profiles across creeping faults in
Asal Rift. Locations of profiles P1 (A), P2
(B), and P3 (C) are indicated in Figure 1A.
Letters refer to fault names in Figure 1A and
text. Note evidence of surface slip on fault
γ 1 (P1), fault E (P2), and faults γ 2 and β (P3)
and distributed strain across Disa Le Mallo
subrift (P1 and P3; see also Fig. 1A).
of ~2 km. The entire α-γ 2-γ 1 system seems to
creep at a steady rate, while individual sections
show independent periods of acceleration.
An ~1-km-long section of γ 2 accelerated in
December 2000 with a triangular slip distribution centered at km 2.5 on the profile. The slip
did not propagate into the α-γ transfer zone,
but triggered delayed slip across the zone in the
following year with a rather flat slip distribution between km 0.6 and 2.1 (Fig. 4A). The subaerial section of γ1 accelerated at the same time
as γ 2. The March 2003 slip event on fault β
shows a clear triangular distribution extending
over an 800-m-long section. A small amount of
slip occurred at the same time along the 0.2–
0.6 km section of the fault (Fig. 4B). The slip
distribution on fault E exhibits a different overall shape from those on γ and β. The distribution
shape is acquired during the 2002 event on the
0.4–3.2 km profile section (Fig. 4C). This main
slip event was preceded by two smaller events,
3 yr and 2 yr earlier along the 0.2–0.6 km and
0.6–0.9 km sections of the profile, respectively
(Fig. 4C). Before the 2002 event, only the part
of the fault southeast of km 1.2 on the profile
seems to creep steadily. The slip distribution on
E resembles the double tip restricted first-order
shape as defined by Manighetti et al. (2001,
2004). The slipping section on E goes between
two volcanic cones along the fissure line of the
Galaele Kôma lava field, acting as mechanical
GEOLOGY, January 2007
100 B
G
25
15
50
5
0
Moment
(×1011 N.m)
Displ.(mm)
Number of
events
30
Fault γ2
20 A
10
0
–10 1997 1998 1999 2000 2001 2002 2003 2004
150
11°36′N
F
D
20
0
14
10
6
2
30
Fault E
20 E
10
0
1997 1998 1999 2000 2001 2002 2003 2004
–10
200
150 F
6
100
4
50
2
0
0
500
1000
1500
2000
2500
3000
Days since 1 January 1997
Number of
events
Displ.(mm)
11°34′N
1998 1999 2000 2001 2002 2003 2004
Moment
(×1011 N.m)
40
Fault β
1 km
42°29′E
42°30′E
42°31′E
42°32′E
×
Moment
(×1011 N.m)
Number of Displ.(mm)
events
30
20 C
10
0
–10 1997
60
Figure 3. Fault slip variations and cumulative seismicity as function of time between 1997
and 2005 for three creeping faults in Asal Rift (displ.—displacement). For faults γ 2, β, and
E, upper graphs show relative displacement between two points located on either side of
fault (dots) and linear fit by sections (gray lines); lower graphs show cumulative number
of earthquakes (black curve) and moment release (gray curve) in vicinity of fault. Seismic
moment M is derived from duration magnitude Md using log(M ) = 1.5Md + 9 (Kanamori,
1977). G: Shaded relief map showing earthquake epicenters. Dotted-line contours indicate
areas around faults where earthquakes are counted to construct graphs in B, D, and F. Open
circles are earthquakes that occurred in rift area during 1997–2005. Color-filled circles correspond to earthquakes during periods of accelerated creep on faults defined by epochs of
radar data acquisitions bracketing creep events: blue circles correspond to first event on
fault γ (6 November 2000–17 January 2001), yellow circles correspond to event on fault β
(24 February–20 March 2003), and green circles correspond to event on fault E (18 April–
5 June 2002). Diamonds indicate location of measurement for graphs A, C, and E. Double
solid lines along fault scarps show locations of profiles shown in Figure 4.
barriers, preventing lateral propagation of the
slipping zone (Manighetti et al., 2001) (Figs.
1A and 4C).
Assuming uniform slip down to a depth of
4 km (Ruegg, 1975), the observed slip distributions imply geodetic moments in the range of
1015–1016 Nm for the slip events on the three
faults, while the total seismic moment released
by earthquakes during these events is at least
three orders of magnitude lower (Fig. 3). This
comparison confirms the aseismic nature of the
observed slip on the Asal faults.
DISCUSSION
A notable coincidence exists between the
spatial and temporal distributions of microearthquakes in the Asal Rift and the location
and occurrence of slip events on three of the
main faults that were activated during the 1978
crisis. However, observations indicate that no
causal relationship exists between the fault activation and the earthquakes. First, the hypocen-
GEOLOGY, January 2007
ters are not located on the fault planes: the 2002
events were distributed over a broad region
around fault E and the 2000 and 2003 clusters
occurred in the narrow graben between faults γ
and β. Second, the 2000 swarm includes nondouble-couple mechanisms that are indicative
of crack opening, a mechanism inconsistent
with elastic relaxation in the hanging wall of a
slipping, normal fault (Doubre, 2004). No such
analysis could be done for the 2003 cluster in
the absence of a dense seismic network as used
in the 2000 campaign (Doubre, 2004). However,
the proximity and similar depth of both clusters
suggest that they are of same nature. We propose that both the occurrence of earthquake
bursts and the accelerated slip on the faults are
caused by fluid- pressure changes in the crust.
If the pressure of fluids filling the dikes and the
multitude of fissures that populate the base of
the seismogenic crust increases, it is conceivable that some fluids inject into the deep part of
a fault, resulting in the decrease of the effective
normal stress on the fault plane, hence inducing slip. This mechanism can be selective in
the way it activates certain faults and not others, depending on the complexity of the connection between the opening fissures and the
faults. Stress change induced by magma injection has been advocated to explain earthquake
triggering in Iceland (Sigmundsson et al., 1997).
Fluid injection can also be advocated to explain
non-double-couple earthquakes associated with
tensile crack opening in the shallow part of the
crust (Foulger and Long, 1984), such as those
observed in the 2000 seismic cluster (Doubre,
2004). The spatial proximity between the 2000
and 2003 earthquake clusters and the location
of maximum slip on faults γ and β during the
associated slip events (Fig. 3D) may suggest the
presence of conduits for fluid injection in this
part of the rift.
The fluid injection mechanism described here
may also explain the triangular shape of the slip
distribution observed on faults γ and β (Figs.
4A, 4B). Simple elastic theory predicts an elliptical distribution of slip on a fault with uniform
friction and responding to a uniform stress field
(Pollard and Segall, 1987; Scholz, 2002). However, the distribution of slip observed on subsections of faults γ and β appears to be remarkably
triangular (Fig. 4B). While triangular shapes of
scarp heights, representing the cumulative slip
on faults (Manighetti et al., 2001; Scholz, 2002),
have been successfully explained by off-fault,
anelastic deformation and damage around the
fault tips (Gupta and Scholz, 2000b; Manighetti
et al., 2004), the slip distribution in a small event
is more likely controlled by friction on the fault
surface (Bürgmann et al., 1994). Scarp height
profiles and short-term slip profiles observed
on faults γ and β do not have a common shape
(Figs. 4A, 4B), suggesting different controlling
processes. The injection of fluid on a fault can
reduce its effective friction, allowing more slip
around the injection point, hence modifying the
otherwise elliptical slip distribution into a triangular distribution.
We conclude that pressure changes in fluidfilled cracks in the crust provide a plausible
mechanism to explain both the selective triggered slip on faults during periods of increased
seismic activity and triangular slip distributions.
The shallow crust tensional stress regime in the
rift is controlled by both the steady plate motion
and pressure in the magmatic system at depth.
The overall inflation of the zone surrounding the
Asal Rift (Ballu et al., 2003; Ruegg and Kasser,
1987) and the extension rate across the rift
exceeding the far-field plate velocity indicate
that the magmatic system under the rift is currently overpressurized, and therefore prone to
force fluids into the array of fractures at intermediate and shallow depths, triggering earthquakes
and controlling fault slip.
71
γ2 - γ1
overlap
120
100
80
60
40
0
1
2
Fault β
NW
10
40
5
20
0
–5
0.0
0.5
1.0
Distance (km)
09 May 98
14 Apr 97
1.5
Distance (km)
C
SE
0
3
NW
15
4
5
Fault E
SE
10
60
5
40
0
20
–5
0
1
2
Distance (km)
Time Scale
22 May 00
12 May 02
18 Jun 04
04 May 99
20
0
17 May 01
07 May 03
Scarp height (m)
Fault γ1
split of γ2
B
LOS Displ (mm)
SE
3
0
Scarp height (m)
Fault α1
LOS Displ. (mm)
30
25
20
15
10
5
0
–5
Fault γ2
α−γ
transfer
zone
NW
Scarp height (m)
LOS Displ. (mm)
A
07 Jul 05
Figure 4. Profiles showing variations of fault scarp height (gray shade) and evolution of
fault slip with time along fault strike (LOS—line of sight; displ.—displacement). A: Fault γ.
B: Fault β. C: Fault E. Scarp height and slip profiles are constructed by difference between
along-fault profiles located on either side of fault (see location in Fig. 3G). Fault-slip profiles
are obtained from displacement maps computed for each epoch of radar data acquisition
between 1997 and 2005. Only one profile per year is represented for clarity. Colors correspond to epoch dates shown on time-scale bar. Note that apparent reverse slip on fault
β at beginning of sequence reflects distributed strain in hanging wall of fault γ. Apparent
increase of scarp height near ends of fault E is due to steep slope of volcanic cones at two
ends of creeping section.
ACKNOWLEDGMENTS
We thank J.B. De Chabalier, E. Ivin, E. Jacques,
G. King, I. Manighetti, P. Rosen, J.C. Ruegg, P. Tapponnier, and C. Vigny for numerous discussions,
and G. King and R. Bürgmann for reviewing the
paper. The National Aeronautics and Space Administration (NASA) Solid Earth and Natural Hazard
and RADARSAT’s Application Development and
Research Opportunity (ADRO) programs supported
this research (grant NAG51102). Peltzer’s contribution was done in part at the Jet Propulsion Laboratory, California Institute of Technology, under contract with NASA.
REFERENCES CITED
Abdallah, A., Courtillot, V., Kasser, M., Le Dain,
A.Y., Lépine, J.-C., Robineau, B., Ruegg,
J.-C., Tapponnier, P., and Tarantola, A.,
1979, Afar seismicity and volcanism: Relevance to the mechanics of accreting plate
boundaries: Nature, v. 282, p. 17–23, doi:
10.1038/282017a0.
Ballu, V., Diament, M., Briole, P., and Ruegg, J.-C.,
2003, 1985–1999 gravity field variations across
the Asal Rift: Insights on vertical movements
and mass transfer: Earth and Planetary Science
Letters, v. 208, p. 41–49.
Berardino, P., Fornano, G., Lanari, R., and Sansosti, E., 2002, A new algorithm for surface
deformation monitoring based on small baseline differential SAR interferograms: IEEE
Transactions on Geoscience and Remote
Sensing, v. 40, p. 2375–2383, doi: 10.1109/
TGRS.2002.803792.
Bürgmann, R., Pollard, D.D., and Martel, S.J., 1994,
Slip distributions on faults: Effects of stress
gradients, inelastic deformation, heterogeneous
host-rock stiffness, and fault interaction: Jour-
72
nal of Structural Geology, v. 16, p. 1675–1690,
doi: 10.1016/0191–8141(94)90134–1.
Cattin, R., Doubre, C., De Chabalier, J.-B., Vigny, C.,
King, G., Avouac, J.-P., and Ruegg, J.-C., 2004,
Numerical modelling of Quaternary deformation and post-rifting displacement in the AsalGhoubbet Rift (Djibouti, Aar): Earth and Planetary Science Letters, v. 239, p. 352–367.
De Chabalier, J.-B., and Avouac, J.-P., 1994, Kinematics of the Asal rift (Djibouti) determined
from the deformation of Fieale Volcano: Science, v. 265, p. 1677–1681.
Doubre, C., 2004, Structure de mécanismes des segments de rift volcano-tectoniques: Étude de
rift anciens (Ecosse, Islande) et d’un rift actif
(Asal-Ghoubbet) [Ph.D. thèse]: Le Mans, Université du Maine, 420 p.
Foulger, G., and Long, R.E., 1984, Anomalous
focal mechanisms: Tensile crack formation on
an accreting plate boundary: Nature, v. 310,
p. 43–45, doi: 10.1038/310043a0.
Gupta, A., and Scholz, C.H., 2000a, A brittle strain
regime transition in the Afar Depression:
Implications for the fault growth and seafloor
spreading: Geology, v. 28, p. 1087–1090, doi:
10.1130/0091–7613(2000)028<1087:BSRTIT>
2.3.CO;2.
Gupta, A., and Scholz, C.H., 2000b, A model of
normal fault interaction based on observations and theory: Journal of Structural Geology, v. 22, p. 865–879, doi: 10.1016/S0191–
8141(00)00011–0.
Kanamori, H., 1977, The energy release in great
earthquakes: Journal of Geophysical Research,
v. 82, p. 2981–2987.
Kasser, M., Ruegg, J.-C., and Lépine, J.-C., 1987,
Geodetic measurements on the Asal Rift (Djibouti). Twelve years of observations: Journal
of Geodynamics, v. 7, p. 221–226.
Lépine, J.-C., and Hirn, A., 1992, Seismotectonics
in the Republic of Djibouti linking the Afar
Depression and the Gulf of Aden: Tectonophysics, v. 209, p. 65–86, doi: 10.1016/0040–
1951(92)90011-T.
Manighetti, I., Tapponnier, P., Gillot, P.-Y., Jacques,
E., Courtillot, V., Armijo, R., Ruegg, J.-C.,
and King, G., 1998, Propagation of rifting
along the Arabia-Somalia plate boundary: Into
Afar: Journal of Geophysical Research, v. 103,
p. 4947–4974, doi: 10.1029/97JB02758.
Manighetti, I., King, G.C.P., Gaudemer, Y., Scholtz,
C.H., and Doubre, C., 2001, Slip accumulation and lateral propagation of active normal faults in Afar: Journal of Geophysical
Research, v. 106, p. 13,667–13,696, doi:
10.1029/2000JB900471.
Manighetti, I., King, G., and Sammis, C.G., 2004,
The role of off-fault damage in the evolution of
normal fault: Earth and Planetary Science Letters, v. 217, p. 399–408, doi: 10.1016/S0012–
821X(03)00601–0.
Pollard, D.D., and Segall, P., 1987, Theoretical displacements and stresses near fractures in rock:
With applications to faults, joints, veins, dikes,
and solution surfaces, in Atkinson, B.K., Fracture mechanics of rock: London, Academic
Press, p. 277–349.
Ruegg, J.-C., 1975, Structure profonde de la croûte
et du mateau supérieur du Sud-Est de l’Afar
d’après les données sismiques: Annals of Geophysics, v. 31, p. 329–360.
Ruegg, J.-C., and Kasser, M., 1987, Deformation
across the Asal-Ghoubbet rift, Djibouti, uplift
and crustal extension: Geophysical Research
Letters, v. 14, p. 745–748.
Ruegg, J.-C., Kasser, M., and Lépine, J.-C., 1987,
Strain accumulation across the Asal-Ghoubbet
rift, Djibouti: East Africa: Journal of Geophysical Research, v. 89, p. 6237–6246.
Scholz, C.H., 2002, The mechanics of earthquakes
and faulting: Cambridge, Cambridge University Press, 496 p.
Sigmundsson, F., Einarsson, P., Rognvaldsson, T.,
Foulger, G.R., Hodgkinson, K.M., and Thorbergsson, G., 1997, The 1994–1995 seismicity
and deformation at the Hengill triple junction,
Iceland: Triggering of earthquakes by minor
magma injection in a zone of horizontal shear
stress: Journal of Geophysical Research, v. 102,
p. 15,151–15,161, doi: 10.1029/97JB00892.
Stein, R., Briole, P., Ruegg, J.-C., Tapponnier, P.,
and Gasse, F., 1991, Contemporary, Holocene,
and Quaternary deformation of the Asal Rift,
Djibouti: Implications for the mechanics of
slow spreading ridges: Journal of Geophysical
Research, v. 96, p. 21,789–21,806.
Vigny, C., Huchon, P., Ruegg, J.-C., Khanbari, K.,
and Asfaw, L.M., 2006, Confirmation of Arabia
plate slow motion by new GPS data in Yemen:
Journal of Geophysical Research, v. 111,
p. B02402, doi: 10.1029/2004JB003229.
Zebker, H.A., Rosen, P.A., and Hensley, S., 1997,
Atmospheric effects in interferometric synthetic aperture radar surface deformation
and topographic maps: Journal of Geophysical Research, v. 102, p. 7547–7563, doi:
10.1029/96JB03804.
Manuscript received 22 May 2006
Revised manuscript received 22 August 2006
Manuscript accepted 30 August 2006
Printed in USA
GEOLOGY, January 2007