doi:10.1144/0016-76492007-093 2008; v. 165; p. 535-547 Journal of the Geological Society

 S. Tesfaye, M.G. Rowan, K. Mueller, B.D. Trudgill and D.J. Harding  

Ethiopia and DjiboutiRelay and accommodation zones in the Dobe and Hanle grabens, central Afar, 

Journal of the Geological Society

serviceEmail alerting to receive free email alerts when new articles cite this article click here

requestPermission to seek permission to re-use all or part of this article click here

Subscribe to subscribe to Journal of the Geological Society or the Lyell Collection click here

Notes  

Downloaded by on 13 February 2008

© 2008 Geological Society of London

Journal of the Geological Society, London, Vol. 165, 2008, pp. 535–547. Printed in Great Britain.

535

Relay and accommodation zones in the Dobe and Hanle grabens, central Afar,

Ethiopia and Djibouti

S. TESFAYE 1, M. G. ROWAN 2, K. MUELLER 3, B. D. TRUDGILL 4 & D. J. HARDING 5

1Cooperative Research Programs, Lincoln University, 104 Foster Hall, 820 Chestnut Street, Jefferson City, MO 65102,

USA (e-mail: tesfayes@lincolnu.edu)2Rowan Consulting Inc., 850 8th Street, Boulder, CO 80302, USA

3Department of Geological Sciences, University of Colorado, Campus Box 399, Boulder, CO 80309, USA4Department of Geology and Geological Engineering, Colorado School of Mines, 1516 Illinois Street, Golden,

CO 80401, USA5NASA, Goddard Space Flight Center, Planetary Geodynamics Laboratory, Greenbelt, MD 20771, USA

Abstract: The right-stepping Dobe and Hanle grabens display a variety of structures that serve to transfer

extensional displacement. These structures range from relay zones between overlapping fault segments to

accommodation zones between interacting rift segments. The study reveals the presence of three examples of

displacement transfer structures: a breached relay zone, a large-scale accommodation zone that is partially

breached, and a composite zone that combines elements of both. All three examples exhibit common structural

elements. First, dipping ramps develop between horizontal horst blocks and graben floors. Second, these ramps

are cut by numerous faults, most of which are antithetic to the ramps and the graben boundary faults. The

antithetic faults bound elongate blocks that are rotated into the grabens. Third, crosscutting faults partially or

completely link the en echelon or overlapping graben boundary faults. The identification of precursory

structures (mode I fractures) at the leading edge of the Dobe–Hanle accommodation zone breaching faults

suggest that the breaching process may be continuing. The spatial alignment from north to south of the

crosscutting faults, open fractures and lineaments indicates that the breaching process is progressing from the

Dobe graben towards the Hanle graben.

Segmentation of extensional fault systems has been identified at

different scales. First, the segmentation of grabens and graben-

bounding faults has been documented in continental rift systems

from both surface mapping and subsurface geophysical investiga-

tions (Rosendahl et al. 1986; Ebinger 1989a, b; Morley et al.

1990; Nelson et al. 1992; Hayward & Ebinger 1996). The areas

between adjacent or overlapping rift segments have been termed

transfer zones (Morley et al. 1990; Nelson et al. 1992) or

accommodation zones (Bosworth 1985; Rosendahl et al. 1986;

Ebinger 1989a, b). In these regions of interaction between

boundary faults, complex arrays of smaller-scale faults conserve

extensional strain as displacement is transferred from one graben

to the other.

At a smaller scale, surface mapping and modern seismic data

show that faults are composed of linked segments (Peacock &

Sanderson 1991, 1994; Childs et al. 1995; Peacock 2002). These

may be hard-linked or soft-linked. Walsh & Watterson (1991)

defined hard-linked faults as those where the fault surfaces are

linked at the scale of the map or cross-section in use. Soft-linked

faults are isolated from one another at the scale of the map or

cross-section in use, with mechanical and geometric continuity

achieved by ductile strain of the intervening rock volume. The

zones of complex deformation between overlapping fault seg-

ments are referred to as relay ramps (Larsen 1988; Peacock &

Sanderson 1994; Trudgill & Cartwright 1994; Peacock et al.

2000), transfer zones (Morley et al. 1990) or relay zones (Childs

et al. 1995; Huggins et al. 1995; Densmore et al. 2003). These

structures accommodate the transfer of displacement between

interacting fault segments (Larsen 1988; Peacock & Sanderson

1991, 1994; Trudgill & Cartwright 1994; Childs et al. 1995;

Huggins et al. 1995; Peacock et al. 2000; Peacock 2002;

Densmore et al. 2003).

In this paper we use ‘displacement transfer’ to describe the

variable distribution of extensional strain that occurs between

interacting segments of faults or grabens. To describe the

geometries, we adopt ‘accommodation zone’ for the region

between adjacent or overlapping rift segments and ‘relay zone’

for the area between interacting en echelon or overlapping fault

segments.

We examine the right-stepping Dobe and Hanle grabens of

central Afar (Fig. 1) to determine the geometry and evolution of

displacement transfer in overlapping fault segments and interact-

ing rift segments. The study demonstrates that the processes that

accommodate displacement transfer are similar both in relay and

accommodation zones. Typical deformation includes ramp

development, normal faulting of the ramp, complex block

rotation on both synthetic and antithetic faults within the ramp,

and development of breaching faults that link the interacting

graben-bounding faults. We illustrate these geometries with three

examples: a breached relay zone; an accommodation zone that is

partially breached; and an intermediate structure that combines

elements of both relay and accommodation zones.

Our study in central Afar enriches the prevailing understand-

ing of relay and accommodation zone evolution. Relay and

accommodation zones are known to play various roles in

structural and tectonic studies. First, relay and accommodation

zones are important in the migration and trapping of hydro-

carbons (Larsen 1988; Morley et al. 1990; Nelson et al. 1992;

Peacock & Sanderson 1994; Peacock 2002). Second, relay

structures play a key role in understanding the growth and

scaling of faults (Dawers & Anders 1995; Cartwright et al. 1995;

Peacock 2002). Third, they are critical in determining fault

segment length, and thereby constrain maximum likely earth-

quake magnitudes on these structures (Jackson & Blenkinsop

1997; Ferrill et al. 1999).

The Dobe–Hanle graben area was chosen in this investigation

to seek an explanation for the anomalously high density of

normal fault occurrences adjacent to the least deformed fault

blocks in the area. The study is based on field investigations,

analysis of Advanced Spaceborne Thermal Emission Reflection

Radiometer (ASTER) imagery, aerial photographs, and digital

topographic data. The Dobe–Hanle graben area is ideal for using

remotely sensed images to investigate structures because of the

extremely good exposure of outcrops, except in the graben floors,

owing to the absence of vegetation and soil cover in the region.

Geological framework

The diffuse Afar triple junction, where the Red Sea, Gulf of

Aden and Ethiopian Rifts converge (Fig. 1, inset), is character-

ized by volcanic and seismic activity. Evidence for active rifting

is manifest as well-preserved fault scarps, voluminous volcanic

rocks, volcanic centres, and shallow earthquakes (CNR–CNRS

(Afar team) 1973; Manighetti et al. 1998; Beyene & Abdelsalam

2005). Magmatic activity commenced in the Afar region (south-

ern Red Sea rift) between 31 and 29 Ma (Hofmann et al. 1997;

Ukstins et al. 2002). Rifting in Afar, the result of divergence

between the African (Nubian) and Arabian plates that are

separated by the anticlockwise rotating Danakil microplate, has

been active since c. 25 Ma (Barberi et al. 1975; Ukstins et al.

2002; Beyene & Abdelsalam 2005).

Fig. 1. Generalized geological map of central Afar (modified from Abbate et al. 1995).

Fig. 2. Topographic image of the Dobe–Hanle graben area generated

from digital terrain elevation data. Elevation ranges from c. 100 m (dark)

to c. 1400 m (white). Diagonal lines are locations of topographic profiles

shown in Figure 4. Fault relief along major graben-bounding faults was

measured along X–X9 (Fig. 7b) and Y–Y9 (Fig. 14). The location of the

map is shown in Figure 1.

S . TESFAYE ET AL.536

Fig. 3. Georeferenced, Advanced

Spaceborne Thermal Emission and

Reflection Radiometer (ASTER) imagery of

the Dobe–Hanle graben area. Black areas

are water bodies, white represents evaporite

and lacustrine sediments and grey is Afar

stratoid volcanic rocks (alluvium is also

grey). The image is a black and white

reproduction of a false colour composite of

the visible and near IR bands of ASTER

data at 15 m spatial resolution. The location

of the image is shown in Figure 1.

Fig. 4. Topographic profiles extracted from

the digital terrain elevation data. The

profiles show a change in the geometry of

the grabens as a result of changes in the

graben-bounding fault geometries. Grabens

bounded by single escarpments show abrupt

elevation drops between graben shoulders

and floors, whereas gradual elevation

changes are observed in graben segments

bounded by relay or accommodation zones.

The location of the profiles is indicated in

Figure 2. Vertical exaggeration is

approximately 3:1 and is uniform in all

profiles.

RELAY AND ACCOMMODATION ZONES, DOBE – HANLE GRABENS 537

Central Afar, a graben- and horst-dominated topography, is

floored by volcanic and sedimentary sequences. Three stages of

volcanism are distinguished in this region. The first group

includes the Miocene to Early Pliocene alkali basalts and more

silicic rocks (Barberi et al. 1975) that are at present exposed

around the periphery of the region (Fig. 1, Lower Extrusive

Complex). The second group, Plio-Pleistocene flood basalts with

subordinate silicic interlayers and volcanic centres, are collec-

tively referred to as the Afar stratoid series (CNR–CNRS (Afar

team) 1973; Varet & Gasse 1978; Barberi & Santacroce 1980).

Available age data indicate that the stratoid series was emplaced

between 4.0 and 1.0 Ma and occupy the floor of central Afar

(Barberi et al. 1975; Barberi & Santacroce 1980; Lahitte et al.

2003). The third group encompasses the ,1 Ma axial volcanic

ranges (primarily basaltic) exposed along discrete zones that

trend NW–SE in the north and WNW–ESE farther south (Fig.

1, inset). The axial volcanic ranges are interpreted to represent

the subaerial expression of oceanic spreading ridges that link the

Gulf of Aden with the Red Sea rift (Barberi & Varet 1977;

Barberi & Santacroce 1980; Barberi et al. 1980; Manighetti et

Fig. 5. Aerial photography interpretation of

the Middle Dobe relay zone. A–A9 and

B–B9 are the field traverses illustrated in

Figure 6. The location of the map is shown

in Figure 2. The scale is approximate.

Fig. 6. Topographic profiles of the Middle

Dobe relay zone constructed from field

survey data. The relay ramp consists of

rotated fault blocks, with faults antithetic to

the graben-bounding faults being most

common. The relief of the relay ramp faults

is mostly less than 100 m whereas the

interacting graben-bounding faults show

relief in excess of 500 m. Vertical

exaggeration is 2.4:1 and is uniform in both

profiles. Profile locations are shown in

Figure 5.

S . TESFAYE ET AL.538

al. 2001). The tectonic basins are filled with detrital and

chemical sediments (Fig. 1) that range in age from Pleistocene to

Present (CNR–CNRS (Afar team) 1973; Barberi & Varet 1977).

The distinct graben and horst topography and rotation of

blocks in central Afar inspired the introduction of rigid micro-

blocks in the modelling of its tectonic evolution (Tapponnier et

al. 1990). The mode of deformation in central Afar, however, is

a contentious issue. Proposed models include rift propagation

induced ‘bookshelf’ style of fault rotations characterized by left

lateral displacements between rigid microblocks (Tapponnier et

al. 1990; Manighetti et al. 1998, 2001). Sigmundsson (1992)

modified the bookshelf model by introducing rift-normal exten-

sion in the grabens in addition to the fault-parallel, left lateral

displacements along the boundary of the rigid blocks. Acton et

al. (1991) favoured a rift-normal extension between the rigid

blocks. Souriot & Brun (1992) proposed a ‘crank-arm’ model

that suggests that external boundary conditions, primarily the 108

anticlockwise rotation of the Danakil horst, account for the

observed fault patterns and clockwise rotation of fault blocks in

central Afar.

Faults are ubiquitous in central Afar and affect rocks of all

ages (Abbate et al. 1995). The dominant fault trend varies from

NNW–SSE to WNW–ESE. North–south and east–west fault

trends are less common. Normal faulting is dominant, although

strike-slip motions along some faults (same strike as normal

faults) have been reported (Dakin et al. 1971; Souriot & Brun

1992; Abbate et al. 1995). Prevailing models of extensional

faults in the Afar region assume that they maintain a steeply

Fig. 7. (a) A perspective view of the

Middle Dobe relay zone generated from the

ASTER image and digital terrain elevation

data. (b) Plot of vertical relief (proxy for

throw) v. strike distance in the Middle Dobe

relay zone measured from digital terrain

elevation data between points X and X9

shown in Figures 2 and 7a. Only the

graben-bounding faults are shown.

RELAY AND ACCOMMODATION ZONES, DOBE – HANLE GRABENS 539

dipping, planar geometry through the entire seismogenic crust

(Morton & Black 1975; Vellutini 1990; Abbate et al. 1995;

Hayward & Ebinger 1996). The base of the seismogenic crust in

central Afar occurs between 3 and 10 km depth beneath axial

volcanic ranges (Gouin 1979; Lepine & Hirn 1992) and between

8 and 15 km elsewhere, including the Dobe–Hanle graben region

(Lepine & Hirn 1992).

Methods

A combination of remote sensing images and field surveys was used to

determine the geometry of the fault and graben systems. Fault patterns in

map view were obtained from digital terrain elevation data (DTED; Fig.

2), ASTER imagery (Fig. 3), and 1:60 000 aerial photographs. The

DTED, obtained from the National Aeronautics and Space Administration

(NASA), were instrumental in determining throw across faults, and

extracting topographic profiles (Fig. 4). The DTED are in the form of a

raster topographic image with 3 arc second (c. 90 m) x–y (horizontal)

resolution. Elevation values measured with respect to mean sea level are

recorded at 1 m intervals.

We use topographic profiles obtained from the digital elevation data as

a proxy for structural cross-sections because of the minimal amount of

erosion in this region. The ground surface generally represents the same

stratigraphic level of youngest lava flows, which cap the stratoid series.

Exceptions include the floors of the major grabens, where there is

alluvial, lacustrine and evaporite fill, and along fault scarps, where there

has been degradation of the original fault surfaces. We also gathered

high-resolution topographic data across fault scarps and grabens along

three field traverses. Surveying was carried out with a Wild Total Station

(electronic theodolite and distance measuring device). Continuous meas-

urements were taken along the width of fault blocks, at scarp top and

bottom, and intervening graben filling sediments. In addition, the

orientation of faults and fault block tilts were measured along the

profiles.

Geometry of displacement transfer

The NW–SE-striking Dobe–Hanle graben system, with a com-

bined length of c. 125 km and maximum width of c. 20 km

(scarp to scarp), is one of the prominent tectonic features of

central Afar (Fig. 2). The grabens are bounded by normal faults

with dips up to 808 that displace the Afar stratoid volcanic rocks

(Mohr 1971; Tesfaye 2005). There is significant elevation

difference between graben shoulders and graben floors, with the

lowest point (84 m above sea level) occurring in the middle of

Dobe graben and the highest point (1376 m above sea level)

found in Kadda Gamarri (Fig. 2), NW of the Hanle graben. The

floors of the grabens are filled with alluvial, lacustrine, and

evaporite deposits of unknown thickness (Fig. 3) (Varet & Gasse

1978). Recent volcanism (,1 Ma) is markedly absent in the

grabens (Varet & Gasse 1978; Abbate et al. 1995).

Anomalously high density of faulting marks the region be-

tween the right-stepping Dobe and Hanle grabens (Figs 2 and 3).

The area between the two grabens exhibits a geometry inter-

mediate between those of rift jumps and rift offsets (e.g. Nelson

et al. 1992). The symmetric nature of the grabens is evident in

the topographic profiles (Fig. 4). Along the central portions of

the grabens, the boundaries are defined by single escarpments

with sharp elevation drops between graben shoulders and floors

(Fig. 4a, f and g, both sides; Fig. 4b, NE side; Fig. 4c, SW side).

The graben shoulders (Adaghilu, Dida, and Kadda Gamarri

horsts) display one of the least deformed blocks in central Afar,

with nearly horizontal surfaces and insignificant faulting (Figs

2–4).

In contrast to the symmetric, unrotated nature of the major

horst blocks, the regions of displacement transfer are defined by

closely spaced, tilted and rotated fault blocks marked by a

gradual drop in elevation (Fig. 4b, d and e, SW sides). These

areas are the focus of this paper. We first examine a relay zone

along the western margin of middle Dobe graben and then the

accommodation zone between the two grabens. We then analyse

the composite relay–accommodation zone where the southern-

most boundary fault of the Dobe graben and the eastern

boundary fault of the Hanle graben interact. Measured topo-

graphic relief is used in all cases as a proxy for fault throw.

Middle Dobe relay zone

The Middle Dobe relay zone developed between two overlap-

ping, synthetic fault segments with a maximum separation of c.

4.5 km (Fig. 5). The NW–SE- to west–east-trending, right-

stepping, graben-bounding faults are linked by a connecting fault

(ramp-breaching fault) that strikes north–south. Whereas the

relief on the graben-bounding faults exceeds 500 m, maximum

relief on the connecting fault is c. 80 m. The ramp region

Fig. 8. (a) An aerial photograph of the Middle Dobe relay zone. The

location of the aerial photograph is shown in Figure 5. (b) Field

photograph of the Middle Dobe relay zone breaching fault. The figures

show that the NW–SE-striking faults (both ramp and graben bounding)

are offset by the north–south-striking relay zone breaching fault. The

arrow in both figures points to the same fault scarp along the breaching

fault.

S . TESFAYE ET AL.540

between the two interacting fault segments has an overall dip of

c. 98 towards the SE. Near the connecting fault, however, the

ramp dips 68 towards the south. The ramp is cut by a dozen or so

synthetic and antithetic normal faults with scarps generally less

than 100 m long (Fig. 6). Short fault segments and small fault

blocks resembling box faults (Griffiths 1980) are clustered

around the hinge zone near the connecting fault, whereas

elongate fault blocks parallel to the main graben-bounding faults

are more common elsewhere on the ramp. These blocks are

variably rotated about a horizontal axis, with dips of up to 308

towards the graben centre (Figs 5 and 6). The relief on the

overlapping graben-bounding faults is typical of that of interact-

ing faults in a relay zone (Peacock & Sanderson 1994; Huggins

et al. 1995); when the relief (throw) on one of the interacting

faults diminishes the relief on the other interacting fault in-

creases, with a deficit of throw occurring in the overlap zone

(Fig. 7a and b).

Based on field observations and the morphology of fault scarps

on the 1:60 000 scale aerial photographs, the NW–SE-striking

faults (both ramp and graben bounding) are offset by the north–

south-striking, connecting fault that links the right-stepping

escarpment of the Dobe graben (Fig. 8a and b). This observation,

the relay zone geometry, and the distribution of relief measured

along the strike of graben-bounding faults are all compatible

with published models of fault growth by segment linkage

(Peacock & Sanderson 1991, 1994; Trudgill & Cartwright 1994;

Cartwright et al. 1995; Childs et al. 1995). We infer that the

overlapping boundary faults were originally separated by a

dipping relay ramp. Only when net displacement increased was

the relay ramp breached by the connecting fault. The fact that

Fig. 9. Aerial photography interpretation of

part of the Dobe–Hanle accommodation

zone. The location of the map is shown in

Figure 2. The scale is approximate.

RELAY AND ACCOMMODATION ZONES, DOBE – HANLE GRABENS 541

Fig. 10. Shaded relief image of the Dobe–

Hanle graben area. NE-facing fault scarps

are in the shadow and SW-facing scarps

appear bright. The relief on the

accommodation zone breaching faults is

small compared with the major graben-

bounding faults, as indicated by the size of

the shadow cast. The dashed black line

(south–central part of figure) marks the

boundary between the rotated and unrotated

fault blocks in the accommodation zone

ramp. The white squares indicate the

epicentres of the 1989 Dobe earthquake

sequence (Sigmundsson 1992). The focal

plane solutions for some of the events are

obtained from the Harvard Centroid

Moment Tensor Catalog (http://

www.seismology.harvard.edu/

CMTsearch.html).

Fig. 11. ASTER image of a portion of the

Dobe–Hanle accommodation zone. It

shows the spatial relationship of the

interpreted structural elements, which

include accommodation zone breaching

faults, open fractures, lineaments, and

rotated and unrotated fault blocks. The

location of the image is shown in Figure 3.

S . TESFAYE ET AL.542

the connecting fault extends south of the intersection with the

outer Dobe fault (Figs 5 and 8a) suggests that linkage was not

accomplished by lateral propagation of one of the main faults,

but by development of a new, separate segment.

Dobe–Hanle accommodation zone

The accommodation zone between the right-stepping Dobe and

Hanle grabens is a c. 16 km wide region that is cut by closely

spaced faults. In this section, we focus on the part of the

accommodation zone where the southwestern footwall of the

Dobe graben, called the Adaghilu horst, plunges to the SE into

the centre of the Hanle graben (Figs 2 and 3). This densely

faulted zone displays variable topography that ranges from an

average elevation of c. 880 m above sea level in the NW to c.

100 m in the floor of Hanle graben in the SE.

The accommodation zone is cut by two distinct, north–south-

to NNW–SSE- and NW–SE- to WNW–ESE-trending, fault sets

(Fig. 9). The prominent, NW–SE- to WNW–ESE-trending,

normal faults can be further divided into two subsets. In the

northeastern quadrant of Figure 9, the faults trend NW–SE, dip

mostly to the SW, and bound blocks that are rotated toward the

NE. In the southern half of Figure 9, faults trend WNW–ESE

and dip both to the south and north, forming a series of narrow

(,1 km) horsts and grabens with no rotation discernible on aerial

photographs.

The dominant (NW–SE- to WNW–ESE-trending) fault set is

cut by swarm of NNW–SSE-trending, mainly east-dipping

normal faults, tens to hundreds of metres long (Fig. 9). They

have relatively small relief compared with the major graben-

bounding faults, as evidenced from the shadow cast by the scarps

on the shaded relief image of Figure 10. These faults are younger

Fig. 12. Aerial photography interpretation

of the South Dobe composite zone. A, B, C

and D are faults that make up the

southwestern boundary of the Dobe graben.

K–K9 is the topographic profile illustrated

in Figure 16. The location of the map is

shown in Figure 2. The scale is

approximate.

RELAY AND ACCOMMODATION ZONES, DOBE – HANLE GRABENS 543

than the dominant set (NW–SE to WNW–ESE) and terminate to

the south near the boundary between the rotated and unrotated

NW–SE-trending fault blocks (Figs 9 and 10). Two open

fractures with a north–south trend occur south of the terminus of

the crosscutting faults (Figs 9 and 11). These fractures are

narrow, up to tens of metres wide, and hundreds of metres long,

and no vertical separation can be discerned on the 1:60 000 aerial

photographs or on the DTED. Further south, close to the west

Hanle graben-bounding fault, lineaments with similar crosscut-

ting trend that have no identifiable opening or displacements are

observed on the ASTER image (Fig. 11).

The NNW–SSE-trending, crosscutting faults are linked, at the

northern end, with the southern tip of the southernmost boundary

fault system of the Dobe graben (Fig. 10). This, combined with

the crosscutting nature of the north–south-trending faults, sug-

gests that the Dobe–Hanle accommodation zone is in the process

of being breached. The presence of open fractures at the leading

edge of the NNW–SSE-trending crosscutting faults signifies

(Fig. 11) the importance of precursory structures (mode I

fractures) in the initial stages of normal faulting (e.g. Crider &

Peacock 2004). Continued slip on the crosscutting faults and

development of the precursory structures to through-going faults

could result in the linking of the western boundaries of the Dobe

and Hanle grabens. The spatial alignment from north to south of

the crosscutting faults, open fractures and lineaments could

indicate that the breaching process is progressing from the Dobe

graben towards the Hanle graben.

South Dobe composite zone

The South Dobe composite zone is so named because it

combines elements of both accommodation and relay zones. It is

part of the Dobe–Hanle accommodation zone, occurring between

the terminations of the southernmost boundary fault of the Dobe

graben and the eastern boundary fault of the Hanle graben (Figs

12 and 13). It also acts in part as a relay zone where the south

Dobe graben boundary fault steps southward to an en echelon

segment of the southernmost Dobe graben boundary fault (fault

D, Fig. 12). This fault (fault D) has a maximum relief of over

400 m (Fig. 14) and is linked to the south Dobe graben boundary

fault by three fault segments that trend north–south to NNW–

SSE (faults A, B and C, Fig. 12) with maximum relief of less

than 200 m (Fig. 14). The southern termination of fault D

coincides with the beginning of the NNW–SSE-trending, cross-

cutting faults that are in the process of breaching the Dobe–

Hanle accommodation zone (Figs 10, 12 and 13).

The South Dobe composite zone is dominated by a relay ramp

that dips 8–108 to the SE and fault blocks that are variably

rotated to the NE (Figs 12 and 15a, b). The relay ramp is cut by

numerous faults that are antithetic to the graben-bounding faults

(Figs 12, 15a, b and 16). Most of the fault blocks that make up

the ramp strike WNW–ESE and are rotated toward the graben

centre, similar to the northeastern boundary fault of the Hanle

graben, which extends into the ramp (Figs 12 and 13). The

faulting pattern in the ramp becomes more chaotic close to faults

A and B (Figs 12 and 15a), where fault blocks are shorter, more

highly rotated, and strike north–south to NNW–SSE. Some of

the WNW–ESE-striking faults veer towards a NNW–SSE

orientation as they approach the graben boundary faults (faults

A, B and C), and others are cut by NNW–SSE-trending faults.

The presence of an old stream drainage network that runs across

four of the rotated fault blocks (Figs 12 and 15a) indicates that

initial basinward tilting of the relay ramp occurred prior to the

antithetic faulting. No lateral offset is observed in the truncated

stream drainage, indicating dip-slip movement.

The width and tilt of the antithetic fault blocks in the South

Dobe composite zone show a systematic variation. The width of

fault blocks along profile K–K9 (Fig. 12) ranges from 22 m to

1173 m (Table 1) and generally decreases from NE to SW as fault

D is approached (Figs 16 and 17a). Fault block dip ranges from 48

to 258 (Table 1) and increases to the SW. In general, we observe

an inverse relation between fault block widths and dip (Fig. 17b).

Fig. 13. A schematic illustration of the interacting Dobe and Hanle

grabens interpreted from the ASTER image. It shows the spatial

relationship of the structural elements that play a role in the evolution of

the relay and accommodation zones in the area, including relay and

accommodation zone breaching faults, open fractures, and lineaments

(possibly fractures).

Fig. 14. Plot of vertical relief v. strike distance of graben-bounding faults

(SW Dobe and east Hanle) in the South Dobe composite zone. Relief is

measured from the digital terrain elevation data between points Y and Y9

in Figure 2.

S . TESFAYE ET AL.544

The increase in fault block width away from fault D is interpreted

to be produced by the increasing strength of the hanging-wall

block as it increases in thickness in the down-dip direction of the

graben-bounding fault. Assuming a wedge-shaped hanging wall,

thinnest near the escarpment and thicker toward the graben centre,

and given the same material property (basalt in this case), fault

spacing will be wider where layer thickness is greater during

extensional deformation (e.g. Mandl 1987).

Discussion

Displacement transfer in the Dobe–Hanle graben area manifests

itself as relay zones between synthetic fault segments and as

accommodation zones between interacting rift segments. The

geometries within the study area provide clues to the evolution

of displacement transfer as regional extension gradually in-

creases. Initially, en echelon fault or rift segments are separated

by dipping but otherwise undeformed ramps, as shown by the

incision of old stream drainage patterns (Figs 12 and 15a). As

displacement increases, the ramps presumably steepen, but more

importantly, become broken up to form elongate fault blocks,

usually rotated along antithetic faults (Figs 5, 6, 12 and 15a, b).

Crosscutting faults develop near the hinge zones of the dipping

ramps as extension persists (Figs 5, 9 and 10). These faults

appear to originate, at least in some cases, as open (mode I)

fractures (Fig. 11). With continued extension, some of the

crosscutting faults link together, forming a continuous fault

joining originally separate fault or rift segments and breaching

the relay or accommodation zone (Figs 5, 9 and 11).

Faults in both the relay and accommodation zones are

dominantly antithetic. The associated fault blocks are rotated (up

to 308) toward the grabens. In contrast, the grabens are generally

symmetric, bounded on both sides by major faults and separated

by horizontal horsts. The underlying reason for the observed

predominance of antithetic faulting geometries in the study area

Fig. 15. (a) Perspective image of the South

Dobe composite zone generated from the

ASTER image and digital terrain elevation

data showing the rotated fault blocks within

the relay ramp and the undisturbed footwall

block in the background. A trace of an old

stream channel that runs across four fault

blocks is shown at the centre of the image.

(b) Field photograph of the South Dobe

composite zone showing the horizontal

basaltic layers in the footwall (horst) and

rotated fault blocks within the relay ramp.

The picture was taken from fault D looking

towards fault A (refer to (a) for

orientation).

Fig. 16. Topographic profile across the South Dobe composite zone

constructed from field survey data. The relay ramp is cut by antithetic

faults bounding rotated fault blocks (numbered). The fault blocks tend to

become narrower and steeper closer to fault D. Vertical exaggeration is

3:1. Location of the profile is shown in Figure 12.

RELAY AND ACCOMMODATION ZONES, DOBE – HANLE GRABENS 545

is not clear. In the Dobe–Hanle accommodation zone, however,

the presence of both rotated and unrotated fault blocks appears

to be influenced by the accommodation zone breaching faults. In

the Dobe–Hanle accommodation zone the southern extent of the

basinward rotated fault blocks coincides with the termination of

the accommodation zone breaching faults (Figs 9, 10, 11 and

13). South of this boundary, faults form a series of narrow,

elongated horst and graben structures that display ramp dip but

no rotation. A possible explanation for such geometry is that for

the fault blocks in the ramp (hanging wall) to rotate they need to

be physically separated from the footwall by the accommodation

zone-breaching faults. This implies that the process of accom-

modation zone breaching precedes the rotation of fault blocks in

the ramp. It cannot be ascertained, from the available data, if this

is the case elsewhere in the Middle Dobe relay zone or South

Dobe composite zone.

The occurrence of open fractures at the leading edge of

Dobe–Hanle accommodation zone breaching faults signifies the

importance of precursory structures in the initiation and develop-

ment of extensional faults (Crider & Peacock 2004). The

identification of accommodation zone breaching faults, open

fractures, and lineaments close to the west Hanle graben-bound-

ing fault (Figs 11 and 13) indicates the advanced stage of the

accommodation zone breaching process. In addition, the spatial

alignment of these structural elements suggests that the accom-

modation zone breaching is progressing from the Dobe graben

towards the Hanle graben. Moreover, the presence of accommo-

dation zone breaching faults, open fractures and lineaments

between the interacting Dobe–Hanle graben-bounding faults

suggests that the west Hanle graben-bounding fault and the

southernmost Dobe graben-bounding fault (fault D) could be

connected at depth (e.g. Peacock & Parfitt 2002). Whether the

breaching process will be completed or not depends on the

persistence of the extensional deformation that produced it in

the first place. The 1989, moderate magnitude earthquake

sequence (mb 5–6) in the middle of Dobe graben (Fig. 10)

indicates that extensional deformation in the area is still continu-

ing (Sigmundsson 1992).

Conclusions

Relay and accommodation zones in the central Afar depression

result from regional, crustal-scale extension. They serve to

transfer displacement between fault and rift segments, respec-

tively. We have analysed displacement transfer over a range of

scales from fault segments that are several kilometres apart to

rift segments that are tens of kilometres apart. Characteristic

structures in the Dobe–Hanle region include dipping relay ramps

between overlapping faults, elongate fault blocks bounded by

faults that dip antithetically to the graben-bounding faults, and

crosscutting faults that link the major boundary faults. Observed

geometries suggest that these structures develop in this order as

net extension increases through time. The presence of accommo-

dation zone breaching faults, open fractures and lineaments

between the interacting Dobe–Hanle graben-bounding faults

indicates the evolving nature of the accommodation zone breach-

ing process. The spatial alignment from north to south of the

crosscutting faults, open fractures and lineaments indicates that

the breaching process is progressing from the Dobe graben

towards the Hanle graben.

We thank R. Bilham, whose support and encouragement made this work

possible. C. Ebinger, S. Agar and C. Childs are thanked for commenting

on an earlier version of the manuscript. We are grateful for the

Table 1. Fault block parameters along profile K–K9 (Figs 12 and 16)

Block number Distance fromfault D (m)

Block dip(degrees)

Block width(m)

18 44 25 8817 117 15 5716 202 16 11415 341 24 16414 434 24 2213 494 23 9812 669 24 25211 1053 18 51510 1325 10 309 1544 17 4088 1909 15 3217 2241 16 3436 2610 16 3954 3016 11 4183 3517 12 5832 4294 4 9721 5367 4 1173

Block distance is measured from fault D to the centre of each block.

Fig. 17. Patterns of fault block width and dip along profile K–K9 shown

in Figure 12. (a) The plot shows the width of the fault blocks increases

as the distance from the graben-bounding fault D increases. (b) Fault

blocks with smaller width tend to rotate more than those with greater

width.

S . TESFAYE ET AL.546

constructive comments made by D. Peacock and an anonymous reviewer,

whose input improved the content of the manuscript. The research work

was supported by NASA grants NAG5-2584 and NAG5-3072. During the

course of the investigation S.T. also received support from a CIRES

Graduate Fellowship at the University of Colorado, Boulder. The

assistance of the Geology and Geophysical Department and Geophysical

Observatory of Addis Ababa University and the Ethiopian Institute of

Geological Surveys in facilitating fieldwork is greatly appreciated.

References

Abbate, E., Passerini, P. & Zan, L. 1995. Strike-slip faults in a rift area:

a transect in the Afar Triangle, East Africa. Tectonophysics, 241,

67–97.

Acton, G.D., Stein, S. & Englen, J.F. 1991. Block rotation and continental

extension in Afar: a comparison to oceanic micro-plate systems. Tectonics,

10, 501–526.

Barberi, F. & Santacroce, R. 1980. The Afar Stratoid Series and the magmatic

evolution of East African rift system. Bulletin de la Societe Geologique de

France, 22, 891–899.

Barberi, F. & Varet, J. 1977. Volcanism of Afar: small-scale plate tectonics

implications. Geological Society of America Bulletin, 88, 1251–1266.

Barberi, F., Ferrara, G., Santacroce, R. & Varet, J. 1975. Structural evolution

of the Afar triple junction. In: Pilger, A. & Rosler, A. (eds) Afar

Depression of Ethiopia. IUCG Scientific Reports, 14, 38–54.

Barberi, F., Civetta, L. & Varet, J. 1980. Sr isotopic composition of Afar

volcanics and its implication for mantle evolution. Earth and Planetary

Science Letters, 50, 247–259.

Beyene, A. & Abdelsalam, M.G. 2005. Tectonics of the Afar Depression: a

review and synthesis. Journal of African Earth Sciences, 41, 41–59.

Bosworth, W. 1985. Geometry of propagating continental rifts. Nature, 316, 625–

627.

Cartwright, J., Trudgill, B. & Mansfield, C. 1995. Fault growth by segment

linkage: an explanation for scatter in maximum displacement and trace length

data from the Canyonlands Grabens of SE Utah. Journal of Structural

Geology, 17, 1319–1326.

Childs, C., Watterson, J. & Walsh, J. 1995. Fault overlap zones within

developing normal fault systems. Journal of the Geological Society, London,

152, 535–549.

CNR–CNRS (Afar team) 1973. Geology of Northern Afar (Ethiopia). Revue de

Geographie Physique et de Geologie Dynamique, XV, 443–490.

Crider, J.G. & Peacock, D.C.P. 2004. Initiation of brittle faults in the upper crust:

a review of field observations. Journal of Structural Geology, 26, 691–707.

Dakin, F., Gouin, P. & Searle, R. 1971. The 1969 earthquakes in Serdo

(Ethiopia). Bulletin of the Geophysical Observatory, Haile Selassie I

University, 13, 19–56.

Dawers, N.H. & Anders, M.H. 1995. Displacement-length scaling and fault

linkage. Journal of Structural Geology, 5, 607–614.

Densmore, A.L, Dawers, N.H., Gupta, S., Allen, P.A. & Gilpin, R. 2003.

Landscape evolution at extensional relay zones. Journal of Geophysical

Research, 108, 2273, doi:10.1029/2001JB001741.

Ebinger, C.J. 1989a. Geometric and kinematic development of border faults and

accommodation zones, Kivu–Rusizi rift, Africa. Tectonics, 8, 117–133.

Ebinger, C.J. 1989b. Tectonic development of the western branch of the East

African rift system. Geological Society of America Bulletin, 101, 885–903.

Ferrill, D.A., Stamatakos, J.A. & Sims, D. 1999. Normal fault corrugation:

implications for growth and seismicity of active normal faults. Journal of

Structural Geology, 21, 1027–1038.

Gouin, P. 1979. Earthquake History of Ethiopia and the Horn of Africa.

International Development Center, Ottawa, Ontario, Publications, 118E.

Griffiths, P.S. 1980. Box-fault systems and ramps: atypical associations of

structures from the eastern shoulder of the Kenya Rift. Geological Magazine,

117, 579–586.

Hayward, N. & Ebinger, C. 1996. Variations in the along-axis segmentation of

the Afar Rift System. Tectonics, 15, 244–257.

Hoffmann, C., Courtillot, V., Feraud, G., Rochette, P., Yirgu, G., Ketefo,

E. & Pik, R. 1997. Timing of the Ethiopian flood basalt event and

implications for plume birth and global change. Nature, 389, 838–841.

Huggins, P., Watterson, J., Walsh, J.J. & Childs, C. 1995. Relay zone

geometry and displacement transfer between normal faults recorded in coal-

mine plan. Journal of Structural Geology, 17, 1741–1755.

Jackson, J. & Blenkinsop, T. 1997. The Bilila–Mtakataka fault in Malawi: an

active, 100-km long, normal fault segment in thick seismogenic crust.

Tectonics, 16, 137–150.

Lahitte, P., Gillot, P.-Y., Kidane, T. & Courtillot, V. 2003. New age

constraints on the timing of volcanism in central Afar, in the presence of

propagating rifts. Journal of Geophysical Research, 108, 2102, doi:10.1029/

2001JB001689.

Larsen, P.H. 1988. Relay structures in a Lower Permian basement-involved

extension system, East Greenland. Journal of Structural Geology, 10, 3–8.

Lepine, J.C. & Hirn, A. 1992. Seismotectonics in the Republic of Djibouti, linking

the Afar Depression and the Gulf of Aden. Tectonophysics, 209, 65–86.

Mandl, G. 1987. Tectonic deformation by rotating parallel faults: the ‘bookshelf’

mechanism. Tectonophysics, 141, 277–316.

Manighetti, I., Tapponnier, P. & Gillot, P.Y. et al. 1998. Propagation of

rifting along the Arabia–Somalia plate boundary: into Afar. Journal of

Geophysical Research, 103, 4947–4974.

Manighetti, I., Tapponnier, P., Courtillot, V., Gallet, Y., Jacques, E. &

Gillot, P.Y. 2001. Strain transfer between disconnected, propagating rifts in

Afar. Journal of Geophysical Research, 106, 13613–13665.

Mohr, P. 1971. Tectonics of the Dobi graben region, central Afar, Ethiopia.

Bulletin of the Geophysical Observatory, Haile Selassie I University, 13,

73–89.

Morley, C.K., Nelson, R.A., Patton, T.L. & Munn, S.G. 1990. Transfer zones

in the East African Rift system and their relevance to hydrocarbon

exploration in rifts. AAPG Bulletin, 74, 1234–1253.

Morton, W. & Black, R. 1975. Crustal attenuation in Afar. In: Pilger, A. &

Rosler, A. (eds) Afar Depression of Ethiopia. IUCG Scientific Reports, 14,

55–65.

Nelson, R.A., Patton, T.L. & Morley, C.K. 1992. Rift segment interaction and

its relation to hydrocarbon exploration in continental rift systems. AAPG

Bulletin, 76, 1153–1169.

Peacock, D.C.P. 2002. Propagation, interaction and linkage in normal fault

systems. Earth-Science Reviews, 58, 2002.

Peacock, D.C.P. & Parfitt, E.A. 2002. Active relay ramps and normal fault

propagation on Kilauea Volcano, Hawaii. Journal of Structural Geology, 24,

729–742.

Peacock, D.C.P. & Sanderson, D.J. 1991. Displacements, segment linkage and

relay ramps in normal fault zones. Journal of Structural Geology, 13, 721–

733.

Peacock, D.C.P. & Sanderson, D.J. 1994. Geometry and development of relay

ramps in normal fault systems. AAPG Bulletin, 78, 147–165.

Peacock, D.C.P., Knipe, R.J. & Sanderson, D.J. 2000. Glossary of normal faults.

Journal of Structural Geology, 22, 291–305.

Rosendahl, B.R., Reynolds, D.J. & Lorber, P.M. et al. 1986. Structural

expression of rifting: lessons from Lake Tanganyika, Africa. In: Frostick,

L.E., Renaut, R.W., Reid, I. & Tiercelkin, J.J.et al. (eds) Sedimentation

in the African Rifts. Geological Society, London, Special Publications, 25,

29–43.

Sigmundsson, F. 1992. Tectonic implications of the 1989 Afar earthquake

sequence. Geophysical Research Letters, 19, 877–880.

Souriot, T. & Brun, J. 1992. Faulting and block rotation in the Afar triangle, East

Africa: The Danakil ‘crank-arm’ model. Geology, 20, 911–914.

Tapponnier, P., Armijo, R., Manighetti, I. & Courtillot, V. 1990. Bookshelf

faulting and horizontal block rotations between overlapping rifts in southern

Afar. Geophysical Research Letters, 17, 1–4.

Tesfaye, S. 2005. Fault population investigation in the Dobe and Guma grabens,

Central Afar, Ethiopia. Journal of African Earth Sciences, 41, 437–444.

Trudgill, B. & Cartwright, J. 1994. Relay-ramp forms and normal-fault

linkages, Canyonlands National Park, Utah. Geological Society of America

Bulletin, 106, 1143–1157.

Ukstins, I., Renne, P., Wolfenden, E., Baker, J., Ayalew, D. & Menzies,

M.A. 2002. Matching conjugate volcanic rifted margins: 40Ar/39Ar chronos-

tratigraphy of pre- and syn-rift bimodal flood volcanism in Ethiopia and

Yemen. Earth and Planetary Science Letters, 198, 289–306.

Varet, J. & Gasse, F. 1978. Geology of Central and Southern Afar (Ethiopia and

Djibouti Republic). CNRS, Paris.

Vellutini, P. 1990. The Manda–Inakir Rift, Republic of Djibouti: a comparison

with the Asal Rift and its geodynamic interpretation. Tectonophysics, 172,

141–153.

Walsh, J. & Watterson, J. 1991. Geometric and kinematic coherence and scale

effects in normal fault systems. In: Roberts, A., Yielding, G. & Freeman,

B. (eds) The Geometry of Normal Faults. Geological Society, London, Special

Publications, 56, 193–203.

Received 11 January 2007; revised typescript accepted 3 July 2007.

Scientific editing by Tim Needham

RELAY AND ACCOMMODATION ZONES, DOBE – HANLE GRABENS 547