surface morphology of active normal faults in hard rock: implications for the mechanics of the asal...

Earth and Planetary Science Letters 299 (2010) 169ndash179

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage wwwe lsev ie rcom locate eps l

Surface morphology of active normal faults in hard rock Implications for themechanics of the Asal Rift Djibouti

Paul Pinzuti a Arnaud Mignan b Geoffrey CP King a

a Institut de physique du globe de Paris 1 rue Jussieu - 75238 Paris cedex 05b Institute of Geophysics ETH-Zurich Switzerland

Corresponding author Institut de physique du globeParis cedex 05 France

E-mail address pinzutiipgpfr (P Pinzuti)

0012-821X$ ndash see front matter copy 2010 Elsevier BV Adoi101016jepsl201008032

a b s t r a c t

a r t i c l e i n f o

Article historyReceived 19 February 2010Received in revised form 2 July 2010Accepted 29 August 2010Available online 26 September 2010

Editor RD van der Hilst

KeywordsAsal Riftnormal faultsdikesslip partitioning

Tectonic-stretching models have been previously proposed to explain the process of continental break-upthrough the example of the Asal Rift Djibouti one of the few places where the early stages of seafloorspreading can be observed In these models deformation is distributed starting at the base of a shallowseismogenic zone in which sub-vertical normal faults are responsible for subsidence whereas cracksaccommodate extension Alternative models suggest that extension results from localised magma intrusionwith normal faults accommodating extension and subsidence only above the maximum reach of the magmacolumn In these magmatic rifting models or so-called magmatic intrusion models normal faults have dips of45ndash55deg and root into dikes Vertical profiles of normal fault scarps from levelling campaign in the Asal Riftwhere normal faults seem sub-vertical at surface level have been analysed to discuss the creation andevolution of normal faults in massive fractured rocks (basalt lava flows) using mechanical and kinematicsconcepts We show that the studied normal fault planes actually have an average dip ranging between 45deg and65deg and are characterised by an irregular stepped form We suggest that these normal fault scarps correspondto sub-vertical en echelon structures and that at greater depth these scarps combine and give birth to dippingnormal faults The results of our analysis are compatible with the magmatic intrusion models instead oftectonic-stretching models The geometry of faulting between the Fieale volcano and Lake Asal in the Asal Riftcan be simply related to the depth of diking which in turn can be related to magma supply This new viewsupports the magmatic intrusion model of early stages of continental breaking

de Paris 1 rue Jussieu - 75238

ll rights reserved

copy 2010 Elsevier BV All rights reserved

1 Introduction

Tectonic-stretching models have been previously proposed toexplain the process of continental break-up through the example ofthe Asal Rift (Djibouti) In this kind of models (eg Dunbar andSawyer 1989 Jackson and McKenzie 1983 Kusznir et al 1991 Linand Parmentier 1990) the normal faults are considered a secondaryconsequence of stretching which controls the deformation associatedwith upwelling of hot viscous material As extension proceeds thebrittle zone of the crust is chopped up by multiple generations ofnormal faults that form at high angles rotate to lower angles and aresubsequently cut by new high angle faults (eg Jackson andMcKenzie 1983 Tapponnier and Francheteau 1978) Because oftheir low angle fault plane the normal faults accommodate thesubsidence and the extension of the crust

In the Asal Rift this mechanism is in agreement with the presenceof sub-vertical normal faults at the surface and with the results of 3Dspatial distribution of seismicity beneath the Asal Rift (Doubre et al

2007ab) These results propose lower dips (50ndash60deg) for the faultslocated far from the rift central axis and vertical dips for the normalfaults bounding the rift inner floor However the seismicity beneaththe Asal Rift does not seem related to vertical fault planes butassociated with nucleationopening of tensional fractures (Aki 1984Doubre and Peltzer 2007 Shimizu et al 1987) around the volume ofhot rocks (see the discussion in this paper) Moreover thismechanism implies normal fault blocks tilted away from the riftaxis Restoration of the Asal Rift topography (De Chabalier andAvouac 1994) paleomagnetic evidence (Manighetti 1993) andbasalt flows vectors (Stieltjes 1980) rule out this assumption thecurrent slopes of the topographic surface being inherited from theinitial shape of a central volcano the Fieale (De Chabalier and Avouac1994)

Models of ground deformations (eg Stein et al 1991 Tarantolaet al 1979 1980) associated with the 1978 seismo-volcanic event(eg Abdallah et al 1979 Leacutepine et al 1980) require the presence ofdikes beneath the Asal Rift Down to the brittleductile zone thevertical component of displacement (subsidence) which cannot beaccommodated by dikes would be accommodated by sub-verticalnormal faults The horizontal component of displacement (extension)which is accommodated at depth by cracks filled by magma (dikes) is

170 P Pinzuti et al Earth and Planetary Science Letters 299 (2010) 169ndash179

accommodated near the surface by open cracks This geometry seemsin agreement with the current surface observations that suggest sub-vertical normal faults and open fissures and it is also used to explainthe long-term evolution of the Asal Rift (Cattin et al 2005 DeChabalier and Avouac 1994 Stein et al 1991) In these tectonic-stretching models applied to the Asal Rift the slip on normal faults isnot correlated with a dike intrusion but controlled by the founderingof blocks into the lithosphere

Alternative models (Fig 1B) suggest that most of the deepextension results from localised magma intrusion with faultsaccommodating extension and subsidence above the maximumreach of the magma column (eg Abelson and Agnon 1997 Agnonand Lyakhovsky 1995 Buck 2006 Hubert-Ferrari et al 2003) whichis proposed for Afar in a model combining geophysical andpetrological data (Fig 1B Pinzuti et al 2007a) In magmatic riftingmodels or so-called magmatic intrusion models the yield stress isreduced by an order of magnitude (eg Buck 2004 2006) comparedto the tectonic-stretching models (eg Dunbar and Sawyer 1989Jackson and McKenzie 1983 Kusznir et al 1991 Lin and Parmentier1990) In these models the normal faults result from dike intrusionwhich impose a dipping plane of 45degndash55deg to the normal faults (egAgnon and Lyakhovsky 1995) as is also suspected beneath the AsalRift where the current horizontal extension (measured across theopening faults from InSAR data) suggests that the sub-vertical faults atthe surface have shallower dipping planes at depth (Peltzer andDoubre 2006) To link the surface observations of the Asal Rift withdipping normal fault at depth this fault geometry implies some slippartitioning (Bowman et al 2003) with opening cracks accommo-dating part of the extension and normal faults accommodatingsubsidence and the remaining fraction of extension

While these two models can lead to the same observations of sub-vertical faults and opening cracks at the surface the magmaticintrusion model contributes to sub-vertical scarps which are due toslip partitioning that turn to dipping normal faults at depth and thatroot into dike intrusions This study aims at 1) proposing frommorphology of exhumed normal fault in hard rock a new conceptualmodel of normal fault evolution due to magmatic intrusion and2) exploring the consequence of this model for the long-termevolution of the Asal Rift by linking the suspected geometry ofnormal faults at depth with magma intrusion

2 Geological setting

The Asal rift is the first emergent segment of the Aden ridge whichpropagates westward on land into the Afar depression (Fig 2A) (egManighetti et al 1998) With a ~40 km length whose 15 km areemerged it currently opens at 16plusmn1 mm yrminus1 in the N40degplusmn5degE

Stretching model

THINNEDLITHOSPHERE(or upper crust)

Subvertical faults

Dikes

ASTENOSPHERE(or lower crust)

10 km

0

-10

Cracks

km

A

Fig 1 A) Schematic representation of models of rift evolution dominated by ductile stretaccommodate subsidence controlled by distributed ductile deformation at depth Ductile deextending through the crust and mantle The model shown is based on Pinzuti et al (2007a)Knox et al 1998 Makris and Ginzburg 1987 Nyblade et al 2000 Ruegg 1975) and petroloare not at the same scale

direction (Fig 2A and B) This rift is structured by a dense network offissures and sub-vertical normal faults with throws up to 200 mpropagating northwestward (N130degplusmn10deg) from the Ghoubbet Bay tothe northwest shore of the Lake Asal (Fig 2B Manighetti et al 1998)

Asal Rift opening is characterised by effusive events associatedwith incipient rifting (Richard 1979) from 853plusmn35 ky to 315plusmn53 ky(Manighetti et al 1998) While immerged context allows theformation of hyaloclastites (326plusmn15 ky) in the south part of therift magmatic activity remains effusive in the north (315plusmn53 ky and334plusmn43 ky Manighetti et al 1998) From the end of this period until~100 ky the evolution of the rift is characterised by the activity of acentral volcano the Fieale (Fig 2) which fills the inner floor andconceals previous faults with large volumes of basalt lava flows(Pinzuti 2006 Pinzuti et al 2007a) Around 50plusmn20 ky magmaticactivity decreases and the whole successive basalt lava flows thatstructure the Fieale volcano become gradually offset by normal faults(Manighetti et al 1998) The structure of the modern rift starts 40ndash30 ky ago with the development of the border faults H andα1 (Fig 2Stein et al 1991 Manighetti et al 1998 Pinzuti 2006 Pinzuti et al2007b) From this period the Fieale volcano progressively collapsesand the magmatic activity locates within the inner floor along smallvolcanic edifices and eruptive fissures (Fig 2 Stein et al 1991 DeChabalier and Avouac 1994 Manighetti et al 1998 Doubre et al2007b)

The latest magmatic episode recorded in the rift corresponds to aseismic-volcanic sequence (Mb 5 and 53) that occurred in 1978(Abdallah et al 1979 Leacutepine et al 1980) This event produced ~2 mof extension in a N40˚ direction and up to 70 cm of subsidence in theinner floor (Ruegg et al 1979) A one-week basaltic fissural eruptiongenerated at the northwestern tip of the volcanic chain gave birth tothe Ardoukoba volcano (Fig 2B Allard et al 1979) Currently themost important deformations are observed around the Fieale edificeand the northeast part of the rift (Doubre and Peltzer 2007 Doubreet al 2007ab Manighetti et al 1998 Peltzer and Doubre 2006)

3 Data acquisition

To examine competent rocks which can sustain open fissures to asubstantial depth we studied massive rock types corresponding tobasalt We measured 14 vertical topographic profiles along 4 of majornormal faults of the Asal Rift which offset successive basalt lava flowsof 1 to 5 m thick along their traces (Fig 3A) Profiles were obtainedusing a handheld laser distancemetre and anglemeasuring binocularsto determine baseline distance horizontal distance and height foreach point (Fig 3B) The instrumental error associated with themeasurements is smaller than 50 cm which is less than the typicalldquoraggednessrdquo of fault scarps It should be noted that this technology

Magmatic intrusion model

3

-20

0

-80

-60

-40 LITHOSPHERICMANTLE

CONTINENTALCRUST

NESW

Dippingfaults

Lithosphericdike

40 kmAsalLake

AbheacuteObockkm

B

10 km

ching at the base of the lithosphere or within the lower crust Steeply dipping faultsformation is assumed to occur at depths of a few kilometres B) Localised deformationwhich combined seismic refraction seismological gravimetry (Berckhemer et al 1975gical data (Pinzuti 2006 Pinzuti et al submitted for publication) Note the two figures

11deg39

11deg37

11deg35

11deg33

42deg3342deg3142deg2942deg2742deg25

km0 1 2

FiealeFieale

Ghoubbet Ghoubbet BayBay

Lake AsalLake Asal

I

H

D

G

ε1

γ1

ε2δ

α3α3

α2α2

ππ

J

I

F

Gα1 β

Ardoukoba lava flows (1978)Recent lava flows (lt~6 ky)

Major faultsMinor faultsFissures

E

γ2

42˚24 42˚30 42˚36

11˚30

11˚36

11˚42

Lake AsalLake Asal(-150 m)(-150 m)

GhoubbetGhoubbetBayBay

FiealeFieale

A B

ADENAG

MI

ALAL

SOMALIASOMALIA

Red Sea

Red Sea

NUBIANUBIA(Ethiopian(Ethiopian

highs)highs)

EastEast AfricanAfricanRiftRift

EAEA

TA

MHMH

MH-GMH-G

ARABIAARABIA

Aden RidgeRed Sea

Fig 2 A) Satellite image of the Asal Rift combined to IGN ASTER and SRTM DEM Inset regional tectonic settings where the arrows represent the rift segments and indicate thedirection of propagation MI Manda Inakir AG Asal-Ghoubbet T Tadjoura EA Erta Ale TA TatAli AL Alayta MH Manda Hararo MH-G Manda Hararo-Gobaad Modified fromManighetti et al (2001a) B) Tectonic map of sub-aerial section of Asal Rift Modified from Manighetti et al (2001ab)

171P Pinzuti et al Earth and Planetary Science Letters 299 (2010) 169ndash179

has made the study possible Other methods of collecting numerousprofiles on nearly vertical cliffs would have been very difficult toemploy because the fault scarps are high (50ndash200 m) steep and hardto climb

4 Morphology of normal faults

Each stage of normal fault evolution is clearly observable acrossthe Asal Rift The extending region as around the inner floor givesbirth to fissures spaced at intervals of metres to tens of metres(Figs 2B and 4A) The trace of the fissure often continues as a narrowdown-flexure which flanks the lengths of the nascent vertical faultscarp on the hanging wall side (Fig 4B) While the down-flexure givesway to vertical motion on this single fissure of several metres wideother fissures accommodate a substantial part of the horizontal

Ardoukoba

Fieale

Lake Asal

Ardoukoba

Fieale

1162deg

1158deg

1154deg

4245deg 4249deg

Lake Asal

Ghoubbetbay

Ghoubbetbay

4253deg4241deg

1166deg

A3A3

A1A1A2A2

A6A6A7A7

A8A8A9A9

A10A10A11A11 A5A5

A4A4

A12A12

A13A13A14A14

H

α1

γ1

γ2

Topographicprofile location

H α1 γ1 γ2

2 km2 km

Major faults

A

Fig 3 A) Satellite SPOT image of the Asal Rift combined to IGN ASTER and SRTM DEM shorepresentation describing the acquisition of topographic profile from handheld laser distanceach point (dashed line) are determined from base line distances and vertical angles meas

motion (Fig 4C) With slip accumulation the open fissure located atthe base of the free face is gradually filled by debris from the earlyevolution (Fig 4D) As the growth of the normal fault continuesdebris progressively forms a talus at the base of the fault scarp (Fig 4Eand F) When the normal faults become mature Holocene scarp canoften be seen near the top of the talus slope and fault scarps especiallyshow an irregular stepped form (Fig 4E and F)

Topographic profiles realised on major faults on the Asal Rift allowus to constrain the origin of these stepped forms The fault scarppresented in Figure 5A shows a sub-vertical face with dips between75deg and 90deg and a platform between 30deg and 45deg (Fig 5A to C) The dipof the talus is generally less than 35deg corresponding to a stable slope(Carson and Kirkby 1972) The stepped form of this profile can beexplained in several ways At a larger scale normal faults often stepforward as younger normal faults cut the hanging wall of earlier

α1

α2

Fault scarp

etc

d2

d1

0Talus

B

wing major tectonic features and the location of the topographic profiles B) Schematice metre and angle measuring binoculars Horizontal distances (d1 d2hellip) and height forurements

A B C

D E F

Fig 4 Photographs of each stage of normal fault evolution A) Fissures spaced at intervals of metres to tens of metres appear in and extending region B) Down-flexure appears withsome fissures becoming very wide (scale of several metres) C) Down-flexure gives way to vertical motion on a single fissure Other fissures also accommodate a substantial part ofthe horizontal motion D) Open fissure at the base of a vertical free face filled by debris from the early evolution E) and F)Mature fault surfaces showing stepping Note the irregularnature of the stepping A Holocene scarp can be seen near the top of the talus slope


faulting (Kusznir et al 1991) Field observations (eg consistency inthe succession of basalt lava flows Fig 4F) and previous structuralstudies (De Chabalier and Avouac 1994 Manighetti et al 1998 Steinet al 1991) exclude this possibility Differential erosion betweenbasalt flows could also explain dip variation of normal faults but therelative homogeneity of the chemical composition of the successivebasalt lava flows comprising the studied normal faults (Pinzuti 2006Pinzuti et al 2007a) rules out this assumption Moreover if erosion isresponsible for the stepped form negative correlation between thedip and the age of the normal fault should be expected which is notthe case (Fig 5D)

Since the degree of fracturing or jointing prior to fault formationdetermines whether a vertical free face can form one could proposean alternative model In Figure 6A the rock apparently had fewexisting fissures and consequently could support a nearly verticalface In Figure 6B the rock is more fractured preventing theformation of a free face At the surface level partitioning betweenextension on fissures and vertical slip on a single fissure (Fig 4Band C) provide planes of weakness leading to stepped formdevelopment It has been suggested that horizontal slip surfacescould also offset the sub-vertical faces (Bigi and Costa Pisani 2005)or that the location of the steps could simply be linked to horizontalzones of weakness (limit between basalt lava flows) These zonespossibly help the linkage between the vertical cracks Therefore inboth cases (Fig 6A and B) the form of the face is likely to havedeveloped as the fault formed

Seven other topographic profiles show clear evidence for originaltensile fissures while another three are less clear and three othershave relatively smooth slopes that do not exhibit steps (Table DR1Figs DR1 to DR3) The average dips of the studied normal faults fallbetween 45deg and 65deg (Fig 5D) These averages do not necessarilyrepresent the dip of the fault at greater depth if open fissures con-tribute to extension Taken together the surface morphology of the

normal faults is consistent with their merging at depth into morelocalised shear zones that could have dips as low as 45deg

Surface morphology of several Asal Rift normal faults and fieldobservations suggest that the stepped forms of fault scarps havedeveloped as the fault formed For validating purposes we propose aconceptual model of evolution of normal fault zones in basalt takingin consideration the tectonic context of the Asal Rift

5 Mechanical concepts

Small-scale triaxial laboratory experiments show that initialfailure in rocks always occurs in extension under different stressconditions Only once sufficient damage has accumulated do morecomplex processes of crack interaction and rotation allow shearing tooccur For low confining pressures as indicated by the Mohr circles(Fig DR4A) rock samples spall and fragment with no shearingsurface developing (Fig DR4B) At mid-level a ragged shear zoneevolves from earlier tension cracks leading to fragmentation (FigDR4C) For still greater stresses a narrow shear zone can appear (FigDR4D) and en echelon structures generally form (Fig DR5)

When applied to geological scales microscopic scale observations(Scholz 2002) show that strength S scales with rock sample size D asS~Dndash12 (Fig DR6) Thus 100 metres characteristic size rock bodieswould be an order of magnitude weaker than a 10 cm characteristicsize sample For several authors (Ashby and Sammis 1990 King andSammis 1992) the scale and distribution of the largest defectsdetermine how a rock fails If the defects are larger than a fewmetresthe behaviour of the rock cannot be directly determined by studyingsmaller samples Nonetheless as similar mechanical processes canoccur at different scales (King 1983) studying small rock samples canprovide insights into the behaviour of larger samples

As proposed elsewhere extension beneath the Asal Rift resultsfrom localised magma intrusion (Tarantola et al 1979) which

80 100 120 140 160Distance (m)

35deg

75deg

Talus Fault scarp

90deg

NW SE

0

20

40

60

80

100

80 100 120 140 160Distance (m)

A13

Hei

gh

t (m

)

Average

Dip 59degAverage

Dip 28deg

0

20

40

60

80

100

Dip

(deg)

80deg80deg

75deg75deg35deg35deg

85deg85deg

40deg40deg

0 50 100 1500

20

40

60

80

65deg

45deg

15-25 ky

Ave

rag

e d

ip (

deg)

Height (m)

15-20ky ~40 ky ~25 ky

AB

C

D

Fig 5 A) Photographs of a studied fault scarp (profile A13 Fig 3A Table DR1) The vertical line corresponds to the position of the profile measurements (perpendicular to the scarpstrike) B) Corresponding topographic profile Dark grey shading indicates rock and light grey shading talus Two steps can be seen C) Slope of profiles in B The inclination of thetalus is usually less than 35deg and hence the slope is generally stable The steps on the exposed fault scarp are highlighted by a succession of high dips (75degndash90deg sub-vertical) and lowdips (~35deg) B) Dip as a function of the height of the exposed fault Black diamonds correspond to the dip where clear evidence for earlier tension cracks is visible and white diamondscorrespond to the dip of less clear or ambiguous examples Stars indicate the slope of the talus surface There is no correlation between the dip and the mean age (Manighetti et al1998 Pinzuti 2006 Pinzuti et al 2007b) of the normal faults

A B

Fig 6 Stability of the free face A)Where the rock is relatively un-fissured a free face of 10 s of metres is stable (normal fault α2 Fig 3A) B) In more fractured rock collapse occurs asthe vertical offset develops (normal fault γ2 Fig 3A) The size of the talus wedge indicates material lost from the free face and is small for A) and substantial for B)



induced normal faults that could intercept the dike near its top (Rubinand Pollard 1988) Normal faults resulting from dike intrusion aresub-vertical at surface level since they are generated from nucleationof large-scale tension fractures andor columnar joints (Gudmundsson1992 Mastin and Pollard 1988) which gradually link up with otherfractures and create distributed shear zones between dike top and thesurface (Mastin and Pollard 1988) The opening of a vertical dike in thecrust (Fig 7A) induced two lobes of concentrated tensile stress above andon both sides of the crack tip (Pollard et al 1983) The predictedtrajectories normal to the most tensile stress are steep because the stressfield above the fracture is dominatedby thenear-tip stressfield associatedwith the opening-mode fracture and the influence of the stress-freeground surface (Okubo andMartel 1998) Thus opening cracks above thedike are sub-vertical and oriented perpendicularly to the trajectories ofthemost tensile stress (Fig 7A) This result is compatiblewith sub-verticalfractures in opening-mode at the surface as observed in the Asal Rift(Fig 4B to D)

As shown in Figure 7B Coulomb stress change caused by an openingvertical dike characterises the conditions under which failure occurs inrocks The crosses indicate the direction of shear failure (mode II) whichgive birth to normal fault plane In this example the normal fault dips at55deg but this angle can changedue tovariationsof friction coefficient porepressure or regional stress (Rubin an Pollard 1988) Numericalsimulations of magma propagation proposed that dip angle rangesbetween 45deg and 55deg (Agnon and Lyakhovsky 1995) which is inagreementwithmost of the fault plane solutions for faulting earthquakesin continental regions (Jackson 1987) and especially in Afar (Braunmil-ler and Nabelek 1990 Jacques et al submitted for publication)

6 Conceptual model

Based on the mechanical and kinematic concepts previouslydescribed we propose a conceptual model of normal fault zonesevolution due to dike intrusion (Fig 8) The principal feature of thismodel is that fractured rock medium has a characteristic size ofmeters rather than microns At the beginning of dike intrusionmodest extension results in extensive tension cracking (Figs 7Dand 8A) but cracks are more developed near the surface while athigher depth confining pressure reduces their length (Fig 8A) Thedeep cracks must be commensurately more numerous to accommo-date the same extension as the ones found at the surface Asdemonstrated elsewhere tension fractures can open at crustal depthsof 05-15 km (Gudmundsson 1992) Tension crack opening alsooccurs at greater depths due to pore pressure effects and presence of

3

2

1

123

Distance fro

0

A

Dep

th (

m)

Change in HorizontalStress (MPa)

5

67

8

Fig 7 A) Contours of the tensile stress near the top of an infinite vertical opening-moderepresent the orientations of opening cracks B) Coulomb stress change induced by the openSee text for more details

fluid-filled voids at seismogenic depth as shown for the Asal Rift orcentral Afar (Noir et al 1997 Doubre and Peltzer 2007)

As extension continues the damaged zone grows and fissuresappear at the surface (Figs 8B and 4A) Rotations from interactions ofcracks start the formation of a deep shear zone (Fig 8C) which resultsin the development of a downward flexure at the surface (Fig 4BGrant and Kattenhorn 2004) As the localised shear at depth becomesmore pronounced a surface fissure eventually opens enough to allowvertical motion (Fig 4C) Near the surface (Fig 8D) motion becomespartitioned between extension on fissures and vertical slip on a singlefissure (Bowman et al 2003)

With the accumulation of slip along the fault at depth verticaldisplacement jumps from the original fissure to an adjacent one in thehanging wall (Fig 8E) This is the same as the mechanical process thatcreates the en echelon features shown at the sub-millimetric scale (FigDR5) When crush zones dip at shallow angles vertical cracksaccommodate vertical motion and a stepped fault is created(Fig 8E) At depth the normal fault has an average dip of 45ndash55degdue to the dike environment but at surface level the normal faultplane corresponds to one or more near vertical irregular surfacesgiving a step-like appearance (Fig 8F)

Crush zones that linked thefissures promote scarpweakness and thusthe collapse of rocks Following the morphology analyses of the normalfault scarps the higher dips revealed from topographic profiles wouldcorrespond to an early open fissure and the lower dips to crushshearzones that linked the fissures The collapses of rock will progressively fillthefissures at thebaseof the scarp (Fig 4D) and then forma talus (Fig 8F)above which a small scarp is often present generally corresponding to arecent earthquake or Holocene motion (Figs 8F 4E and F Pinzuti et al2007b) The number of the observed steps and their degree of regularitydepend on the original fissure spacing and the total throw of the fault

Our model explains why the step-like appearance of the normalfault at surface level is the consequence of the fault plane evolutionbelow the surface due to dike intrusion This model links surfacemorphology with normal fault planes dipping between 45deg and 55deg atdepth which impacts the Asal Rift opening In the next section weshow that this result is compatible with normal fault spatialdistribution and long-term mechanism evolution of the Asal Rift

7 Relation between fissuring faulting and diking

At the Fieale location the trace of the Asal Rift bounding faultscurves towards the rift axis (Figs 2B and 9A) This fault patterncould be related to the interaction of the regional stress field with

0 1 2 3

2

m the dike (m)

B

Coulomb Stress ChangeDCFF (MPa)

4

5

3

6

fracture The short ticks are trajectories perpendicular to the most tensile stress anding of a vertical dike (maximum driving stressPmax=40 MPa Poissons ratio=025)

E Stepped block surface offsets

A Initial extension

F Advanced Morphology

Large verticalcracks near

to the surface

Smallercracks at

depth

Smallercracks at

depth

Holocene scarp

Bed rock

Talus

B Damage zone development

Distributednear thesurface

Localisedat depth

C Shear Zone development

Tension cracksan flexure nearto the surface

Shear at depth(45-55deg)

Vertical

D Slip partitionning

Horizontal

Shear zone at depth

Ragged shear zone at

intermediatedepth

Fig 8 Sketch showing the evolution and erosion of a normal fault in hard rock based on field observations (see Fig 5 for typical example) and assumed that the same processes thatcan operate at small scales can also operate at larger scales See text for details about each stage of the evolution


the one set up by the Fieale volcano mass (Van Wyk de Vries andMerle 1996) However this shape and spatial distribution of normalfaults can also be correlated to the long-term propagation of dikesinto the crust in agreement with the Aden ridge model (Hubert-Ferrari et al 2003) and with the overall north westwardpropagation of the rift (Manighetti et al 1998) Indeed the openingrate of an active rift which is related to plate motion cannot alwaysbe fully accommodated by dike intrusions because insufficientdriving pressure or the presence of a level of neutral buoyancy atdepth (Buck 2006 Lister and Kerr 1991) prevent dikes fromreaching the surface Thus extension must be accommodated bynormal faults that root into dikes (Agnon and Lyakhovsky 1995Rowland et al 2007 Rubin and Pollard 1988)

Consequently when magma rises close to the surface the fault riftzone is narrow It can widen when the dike intrusion does not reach ashigh (Mastin and Pollard 1988 Pollard et al 1983) A similar relationhad been previously proposed by Okubo and Martel (1998) to explainat smaller scale the ldquohourglassrdquo fault pattern of thepit craters of the EastRift Zone of Kilauea volcano (Hawaii) The principle is easily noticed onthe IGN Digital Elevation Model (DEM) of the Asal Rift (Fig 9A) wherethe bounding faults near Lake Asal are further apart than in theGhoubbet Bay region Note that some faults near the rift axis and LakeAsal may be concealed by young basalt lava flows (Fig 2B) From thetectonic feature of the Asal Rift (Fig 9A) and assuming that the majornormal faults have a dip of 55deg the long-term relation between dikeopening and depth can be estimated along the rift axis (Fig 9B to E)

A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H

J H F D α1 βγ δ ε1

α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE

Fig 9 A) IGN Digital Elevation Modelling (resolution X Y=10 m Z=1 m) showing the region between Lake Asal and the Ghoubbet Bay BndashE) Relation between dike opening anddepth along the rift axis based on surface offset (black arrows) estimated from IGN DEM topographic profiles and assuming that normal faults dip at 55deg


Near Lake Asal dikes rise to 3ndash4 km under sea level while for Fiealediking rises to 1 kmor less (Figs 9B to E and 10A) Since Fieale volcano isthe most active centre dikes can likely rise higher in its vicinity

To validate this result we compare our depth estimates with thoseobtained from elastic modelling Mechanical models show that whena dike has not yet reached the surface its internal pressurecompresses the surrounding rock and produces horizontal displace-

A Lake Asal

Fieale1 km

2 km

3 km

4 km

Fig 10 Depths reached by magma column beneath the Asal Rift A) Schema showing the deFig 9 B) Horizontal extensional strain (exx) parallel to the surface per metre average dikewidths estimated from the Figure 9 Curves are for depth dike of d=125 km (blue curve)pressure along the dike length is assumed in all models Youngs modulus and Poissons ra

ments directed away from the dike (Pollard et al 1983) Conse-quently two zones of maximum horizontal tension and extensionstrain are created at the surface on the either side of the dike (Mastinand Pollard 1988 Pollard et al 1983 Rubin 1992) These two zonesof maximum horizontal tension and extension strain will produce twoparallel zones of fissures and normal faults which lead to grabenformation above a dike The model predicts that the distance between

0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km

Distance from Asal Rift axis (km)

crease of magma level beneath the axis Depths correspond to those estimated from thethickness versus distance from the Asal Rift axis DG values correspond to the graben2 km (green curve) 27 km (yellow curve) and 33 km (red curve) A uniform drivingtio used in this model are respectively equal to 45 GPa and 025 (Grandin et al 2009)


the two highly strained zones and thus the graben width is function ofthe dike depth (Mastin and Pollard 1988 Pollard et al 1983 Rubin1992) These models usually use surface displacements produced fromdiking events to infer dike depth and opening magma chamberdeflation and slip on border faults (Grandin et al 2009 Rubin 1992Rubin and Pollard 1988 Stein et al 1991 Wright et al 2006)

Figure 10B shows the results of our numerical model predictions ofhorizontal surface strain above an infinite vertical dike localised at fourdifferent depths Using the graben widths (DG) measured fromtopographic profiles (Fig 9) we estimated the depth of dike intrusionbeneath the Asal Rift axis If the graben width really corresponds to thedistance between themaximaof thehorizontal extensional strain abovethe vertical dike the dike top would approximately range from 1 kmbeneath the Fieale volcano to 3ndash4 km near the Lake Asal These resultssuggest that the dike top depth estimated from 55deg dipping normalfaults are consistentwith those obtained fromtheelasticmodels (Figs 910A and B) Although the theoretical models show that the role of dikeintrusion in triggering faulting is clear its role in contributing to long-term rift topography is generally less acknowledged Here similaritieswith of Holocene and Quaternary deformations across and along the riftaxis (Pinzuti 2006 Pinzuti et al 2007b Rubin 1992 Stein et al 1991)show that the process of riftinghas been steady state and largely devoidof volcanism during the past (~30ndash40 ky) This suggests that faultinggenerated by repeated dike intrusion could significantly contribute tothe topography of the Asal Rift during the past 40 ky

While the intrusion of dikes does not contribute to subsidence thefault dip θ the opening of the rift u and the amount of subsidence vare related by

tan θeth THORN = 2v = ueth THORN

For the last 100 ky the reconstructed topographyof the Fieale volcanosuggests a spreading rate of 17ndash29 mm yrminus1 and a subsidence of 1ndash35 mm yrminus1 (De Chabalier and Avouac 1994) This would suggest veryshallowangle faulting (~20deg)which isnot likely Evenwith the spreadingrate suggested fromplatemotion (8ndash11 mmyrminus1 Vigny et al 2007) andthe upper limit of the subsidence rate (35 mm yrminus1) the dip is still lessthan an unrealistic 35deg The most straightforward explanation is that thesubsidencehas been suppressedby thefillingof amagmachamberor sillsaround the edifice (Cattin et al 2005 De Chabalier andAvouac 1994) Ata shorter time scale (~10 ky) the Lake Asal high stand shorelines locatedto the northwest of the edifice give a reliable subsidence rate of about8 mm yrminus1 (Stein et al 1991) Together with the plate rate this gives alikely dip of about 55deg as expected for normal faults initiating at the tip ofan opening dike in the magmatic intrusion model

8 Discussion

Our conceptual model explains why step-like appearance of theAsal Rift normal faults is the consequence of the fault plane evolutionbelow the surface due to dike intrusion This model complementsthose proposed by Mastin and Pollard (1988) and Gudmundsson(1992) Indeed these authors also suggest for different depth scalesthat normal faults induced from dike intrusion evolve from sets of enechelon tension fractures orand columnar joints but do not link thenormal fault evolution at the surface with depth deformation Ourconceptual model links surface morphology with normal fault planesdipping between 45deg and 55deg with spacing linked to depth of dikeintrusion This result is compatible with the spatial distribution ofnormal faults and the long-termmechanism evolution of the Asal Rift

The presence of a dike beneath the rift axis is suspected fromseismic reflection profiles (Ruegg 1975) which show an anomalouslow-velocity mantle under the Asal Rift Major elements compositionand trace ratio of basalt lava flows across the rift (Pinzuti 2006Pinzuti et al 2007a Pinzuti et al submitted for publication) revealthat the extension below the Asal Rift results from magma intrusion

localised at depths between 60 and 20 km (Pinzuti et al 2007aPinzuti et al submitted for publication) At shallower depthgeochemistry (Pinzuti 2006 Pinzuti et al submitted for publicationVigier et al 1999) andmagnetotelluric (Van Ngoc et al 1981) studiessuggest the presence of a magma material reservoir (2ndash4 km) andormolten material beneath the rift axis

Seismic studies (Doubre et al 2007ab) propose that the thickness ofthe seismogenic crust is about 3ndash4 km below the Fieale volcanoTomography inversion (Doubre et al 2007ab) reveals crustal structurebeneath the rift but the results are only robust for the first 3 km wherethe greater part of the seismic events is localised (Doubre et al 2007a)However Doubre et al (2007b) propose that the Asal central magmaticsystem between 5ndash7 km and 3 km corresponds to a volume of hotrocks or crystal mush deforming aseismically and possibly containingsmall pockets of partial melt These authors also suggest that the depthof the brittle-ductile transition (~600 degC isotherm) which can beconsidered as the depth of dike intrusion decreases north westwardalong the rift axis from 3 to 4ndash5 km It should be noted that this depth isnot directly deduced from seismic data but following Pollard et al(1983) and using an inner floor width of the of 85 km Despite theuncertainties seismic studies and tomography inversion reveal that therift axis is under run by a volumeof heated rocks that extends from3 kmbeneath the Fieale volcano to 4ndash5 km toward the Lake Asal which isconsistent with our magma level estimates

3D spatial distribution of micro-seismicity (Mdle28) from 23-yearrecordings beneath the Asal Rift reveals the aseismic nature of theobserved slip on the Asal faults and shows that the northern borderfaults ε1 and ε2 (Figs 2 and 9) have a 50ndash60deg dipping plane (Doubre etal 2007b) which is in agreement with our dip estimates On thecontrary the younger normal faults which bound the inner floor rifthave steepest fault planes dipping by 85degplusmn5deg These sub-vertical dipswere estimated considering vertical fault plane at the surface and smallclusters of events at depth located below the fault surface traceWhile apart of these events has normal mechanisms that indicate pure ormainly dip-slip motion on steep plane another part attests of normalfaulting on steep planes reverse faulting and strike slip The slip patchessurfaces (le1ndash2 km2) and total dissipated moments (le1012 Nm)associated with the cluster events are equivalent to 1 mm of slip on100 m-long rupture zones (Doubre et al 2007b) Thus it suggests thatthis micro-seismicity results from microscale damage fracturing of therocks embedding the fault plane preferentially at their base (Doubre etal 2007b) Consequently this micro-seismicity is not related to verticalfault planes but associated with nucleationopening of tensionalfractures (Aki 1984 Shimizu et al 1987) around the volume of hotrocksMoreover observations indicate that no causal relationship existsbetween the activation of the faults and the earthquakes between 1997and 2005 (Doubre and Peltzer 2007) During this period faulting iscontrolled by pressure changes in fluid-filledfissures connecting fault atdepth (Doubre and Peltzer 2007) and imply sub-vertical faults at thesurface with shallower dipping planes at depth to accommodate thecurrent horizontal extension (Peltzer andDoubre 2006) A recent studyabout the Manda Hararo-Dabbahu rift (Afar) also shows that a dippingnormal fault (30ndash40deg) is required to accommodate the amount ofsubsidence and extension above the dike during the 2005 rifting event(Barisin et al 2009) Grandin et al 2009 preferably suggest that thisasymmetrical surface deformation results from thedilatancy involved inthe formation of incipient normal faults above the dike with a geometryequivalent to a 60deg dipping fault and kinematics consistent withshallower dipping fault This fault is divided into a series of linked sub-vertical opening and shallow dipping dislocations as proposed in ourconceptual model in the early stage of the shear zone formation

9 Conclusion

The surface expression of active normal faults in the Asal Rift maysuggest that the faults have very high dip angles However our


conceptual model based on simple mechanical and kinematic conceptstogether with field observations suggest that near the surfacedeformation in hard rock is not localised but distributed During theformation of the normal fault many sub-vertical fissures open and oneormoreof thesemay accommodate all the vertical component of slip Asdeformation proceeds slip becomes partitioned between opening onmany fissures and vertical slip on a few Since surface features couldindicate faults with dips of between 45deg and 55deg at greater depth thesurface observations can be made consistent with a model wheredeformation is localised at depth on narrow dikes and where dikes andfaults can be related by simple kinematics Normal faults accommodateextensionwhen driving pressure or buoyancy is insufficient for the diketo reach the surface and the distribution of surface faults can be used toestimate the depth reached by the magma column

Acknowledgements

The authors would like to thank Rolando Armijo Paul TapponnierEric Jacques Gilles Peltzer Yann Klinger Raphaeumll Grandin and PhilippeLopez for useful discussions This is IPGP contribution number 3083

Appendix A Supplementary data

Supplementary data to this article can be found online atdoi101016jepsl201008032

References

Abdallah A Courtillot V Kasser M Le Dain AY Leacutepine J-C Robineau B Ruegg J-CTapponnier P Tarantola A 1979 After seismicity and volcanism relevance to themechanics of accreting plate boundaries Nature 282 17ndash23

Abelson SG Agnon A 1997 Mechanics of oblique spreading and ridge segmentationEarth Planet Sci Lett 148 405ndash421

Agnon A Lyakhovsky V 1995 Damage Distribution and Localization During DykeIntrusion In Baer Heimann (Eds) Physics and Chemistry of Dykes BalkemaRotterdam pp 65ndash78

Aki K 1984 Evidence for magma intrusion during the Mammoth Lakes earthquakes ofMay 1980 and implications of the absence of volcanic (harmonic) tremorJ Geophys Res 89 7689ndash7696

Allard P Tazieff H Dajlevic D 1979 Observations of seafloor spreading in Afarduring the November 1978 fissure eruption Nature 279 30ndash33

Ashby MF Sammis CG 1990 The damage mechanics of brittle solids in compressionPageoph 133 489ndash521

Barisin I Leprince S Parsons B Wright T 2009 Surface displacements in theSeptember 2005 Afar rifting event from satellite image matching asymmetricuplift and faulting Geophys Res Lett 36 L07301 doi1010292008GL036431

Berckhemer H Baier B Bartelsen H Behle A Burckhardt H Gebrande H MenzelJ Miller H Vees R 1975 Deep Seismic Soundings in the Afar Region and on theHighlands of Ethiopia In Pilger A Rosler A (Eds) Afar Depression of EthiopiaStuttgart GermanySchweizerbart Scientific Report 14 pp 89ndash107

Bigi S Costa Pisani P 2005 From a deformed Peri-Tethyan carbonate platform to afold-and-thrust-belt an example from the Central Appenines (Italy) J Struct Geol27 523ndash539

Bowman D King GCP Tapponnier P 2003 Slip partitioning by elastoplasticpropagation of oblique slip at depth Science 300 1121ndash1123

Braunmiller J Nabelek J 1990 The 1989 Ethiopia earthquake sequence EOS TransAm Geophys Union 71 1480

Buck WR 2004 Consequences of Asthenospheric Variability on Continental RiftingIn Karner GD Taylor B Droscoll NW Kohlstedt DL (Eds) Rheology andDeformation of the Lithosphere at Continental Margins Columbia Univ Press NewYork pp 1ndash30

Buck WR 2006 The Role of Magma in the Development of the Afro-Arabian RiftSystem In Yirgu G Ebinger CJ Maguire PKH (Eds) The Afar Volcanic Provincewithin the East African Rift System Special Publications 259 Geological SocietyLondon pp 43ndash54

Carson MA Kirkby MJ 1972 Hillslope form and process Cambridge University PressCambridge 475 pp

Cattin R Doubre C de Chabalier J-B King G Vigny C Avouac J-P Ruegg J-C2005 Numerical modelling of quaternary deformation and post-rifting displace-ment in the Asal-Ghoubbet rift (Djibouti Africa) Earth Planet Sci Lett 239352ndash367

De Chabalier J-B Avouac J-P 1994 Kinematics of the Asal Rift (Djibouti) determinedfrom the deformation of Fieale Volcano Science 265 1677ndash1681

Doubre C Manighetti I Dorbath C Dorbath L Jacques E Delmond J-C 2007aCrustal structure and magmato-tectonic processes in an active rift (Asal-GhoubbetAfar East Africa) 1 Insights from a 5-month seismological experiment J GeophysRes 112 B05405 doi1010292005JB003940

Doubre C Manighetti I Dorbath L Dorbath C Bertil D Delmond J-C 2007bCrustal structure and magmato-tectonic processes in an active rift (Asal-GhoubbetAfar East Africa) 2 Insights from the 23-year recording of seismicity since the lastrifting event J Geophys Res 112 B05406 doi1010292006JB004333

Doubre C Peltzer G 2007 Fluid-controlled faulting process in the Asal Rift Djiboutifrom 8-year radar interferometry observations Geology 35 (1) 69ndash72

Dunbar JA Sawyer DS 1989 How preexisting weaknesses control the style ofcontinental breakup J Geophys Res 94 7278ndash7292

Grandin R Socquet A Binet R Klinger Y Jacques E de Chabalier J-B King GCPLasserre C Tait S Tapponnier P Delorme A Pinzuti P 2009 September 2005Manda Hararo-Dabbahu rifting event Afar (Ethiopia) Constraints provided bygeodetic data J Geophys Res 114 B08404

Grant VG Kattenhorn SA 2004 Evolution of vertical faults at an extensional plateboundary southwest Iceland J Struct Geol 26 537ndash557

Gudmundsson A 1992 Formation and growth of normal faults at the divergent plateboundary in Iceland Terra Nova 4 464ndash471

Hubert-Ferrari A King G Manighetti I Armijo R Meyer B Tapponnier P 2003Long-term elasticity in the continental Lithosphere modelling the Aden Ridgepropagation and the Anatolian extrusion process Geophys J Int 153 111ndash132

Jackson JA 1987 Active normal faulting and crustal extension In Coward M Dewey JHancock P (Eds) Continental Extensional Tectonics Blackwell London pp 3ndash18

Jackson JA McKenzie DP 1983 The geometrical evolution of normal fault systemsJ Struct Geol 5 471ndash482

Jacques E Kidane T Tapponnier P Manighetti I Gaudemer Y Meyer B Ruegg JCAudin L Armijo R Normal Faulting During the August 1989 Earthquakes in CentralAfar Sequential Triggering and Propagation of Rupture Along the Docircbi GrabenBSSA Submitted for publication

King GCP 1983 The accommodation of strain in the upper lithosphere of the earth byself-similar fault systems the geometrical origin of b-value Pageoph 121 761ndash815

King GCP Sammis CG 1992 The mechanisms of finite brittle strain Pageoph 138611ndash640

Knox RP Nyblade AA Langston CA 1998 Upper mantle S velocities beneath Afarand western Saudi Arabia from Rayleigh wave dispersion Geophys Res Lett 254233ndash4236 doi1010291998GL900130

Kusznir NJ Marsden G Egan SS 1991 A flexural-cantilever simple-shearpure-shearmodel of continental lithosphere extension applications to the Jeanne dArc BasinGrand Banks and Viking Graben North Sea Geological Society London SpecialPublications 56 41ndash60 doi101144GSLSP19910560104

Leacutepine J-C Ruegg J-C Anis AM 1980 Sismiciteacute du rift dAsal-Ghoubbet pendant lacrise sismo-volcanique de Novembre 1978 Bull Soc Geol Fr 7 809ndash816

Lin J Parmentier EM 1990 A finite amplitude necking model of rifting in brittlelithosphere J Geophys Res 95 4909ndash4924

Lister JR Kerr RC 1991 Fluid-mechanical models of crack propagation and theirapplication to magma transport in dykes J Geophys Res 96 10049ndash10077

Makris J Ginzburg A 1987 The afar depression transition between continentalrifting and sea floor spreading Tectonophysics 141 199ndash214

Manighetti I 1993 Dynamique des systegravemes extensifs en Afar thegravese de Doctorat Univde Pierre and Marie Curie Paris 240 pp

Manighetti I Tapponnier P Gillot P-Y Jacques E Courtillot V Armijo R RueggJ-C King G 1998 Propagation of rifting along the Arabia-Somalia plateboundary into Afar J Geophys Res 103 (B3) 4947ndash4974

Manighetti I Tapponnier P Courtillot V Gallet Y Jacques E Gillot P-Y 2001aStrain transfer between disconnected propagating rifts in Afar J Geophys Res 106(B7) 13613ndash13665

Manighetti I King GCP Gaudemer Y Scholz CH Doubre C 2001b Slipaccumulation and lateral propagation of active normal faults in Afar J GeophysRes 106 13667ndash13696

Mastin LG Pollard DD 1988 Surface deformation and shallow dike intrusionprocesses at Inyo craters Long Valley California J Geophys Res 9313221ndash13235

Noir J Jacques E Beacutekri S Adler PM Tapponnier P King GCP 1997 Fluid flowtriggered migration of events in the 1989 Dobi earthquake sequence of CentralAfar Geophys Res Lett 24 2335ndash2338 doi10102997GL02182

Nyblade AA Knox RP Gurrola H 2000 Mantle transition zone thickness beneathAfar implications for the origin of the Afar hotspot Geophys J Int 142 615ndash619

Okubo CH Martel SJ 1998 Pit crater formation on Kīlauea volcano HawaiiJ Volcanol Geoth Res 86 1ndash18 doi101016S0377-0273(98)00070-5

Peltzer G Doubre C 2006 Asymmetric opening and episodic faulting in the Asal RiftDjibouti Alaska Satellite Facility Newsletter 3 (2) 2ndash3

Pinzuti P 2006 Croissance et propagation des failles normales du rift dAsal-Ghoubbetpar datations cosmogeacuteniques 36Cl-Liens avec le magmatisme PhD thesis IPGPFrance

Pinzuti P Humler H Manighetti I Gaudemer Y 2007a Spatial and temporalevolution of the magmatism in the Asal-Ghoubbet rift Afar depression GeophysRes Abstr 9 07500 EGU2007

Pinzuti P Manighetti I Gaudemer Y Finkel RC Ryerson FJ 2007b Growth andpropagation of normal faults in the Asal-Ghoubbet rift from 36Cl cosmogenic datingand offset measurements Geophys Res Abstr 9 05015 EGU2007

Pinzuti P Humler E Manighetti I Gaudemer Y submitted for publication PetrologicalConstraints on Melt Generation Beneath the Asal Rift (Djibouti) Earth Planet SciLett

Pollard DD Delaney PT Duffield WA Endo ET Okamura TA 1983 Surfacedeformation in volcanic rift zones Tectonophysics 94 541ndash584 doi1010160040-1951(83)90034-3

Richard O 1979 Etude de la transition dorsale oceacuteanique-rift eacutemergeacute Le Golfe deTadjoura (Reacutepublique de Djibouti) PhD thesis Univ of Paris sud Orsay France


Rowland JV Baker E Ebinger CJ Keir D Kidane T Biggs J Hayward N Wright TJ2007 Fault growth at a nascent slow-spreading ridge 2005 Dabbahu rifting episodeAfar Geophys J Int 171 (10) 1226ndash1246 doi101111j1365-246X200703584x

Rubin AM Pollard DD 1988 Dike-induced faulting in rift zones of Iceland and AfarGeology 16 (5) 413

Rubin AM 1992 Dike-induced faulting and graben subsidence in volcanic riftzones J Geophys Res 97 1839ndash1858

Ruegg J-C 1975 Structure profonde de la croucirce et du manteau supeacuterieur du Sud-Estde lAfar dapregraves les donneacutees sismiques Ann Geacuteophys 31 329ndash360

Ruegg J-C Leacutepine J-C Tarantola A 1979 Geodetic measurements of riftingassociated with seismo-volcanic crisis in Afar Geophys Res Lett 6 817ndash820

Scholz CH 2002 TheMechanics of Earthquakes and Faulting 2nd ed Cambridge UnivPress New York 471 pp

Shimizu H Ueki S Koyama J 1987 A tensile-shear crackmodel for themechanism ofvolcanic earthquakes Tectonophysics 144 287ndash300

Stein RS Briole P Ruegg J-C Tapponnier P Gasse F 1991 Contemporaryholocene and quaternary deformation of the Asal Rift Djibouti implications forthe mechanics of slow spreading ridges J Geophys Res 96 (B13) 21789ndash21806

Stieltjes 1980 Geological Map of Asal Rift Republic of Djibouti scale 150000 CentreNational de la Recherche Scientifique Paris

Tapponnier P Francheteau J 1978 Necking of the lithosphere and the mechanics ofslowly accreting plate boundaries J Geophys Res 83 3955ndash3970

Tarantola A Ruegg J-C Leacutepine J-C 1979 Geodetic evidence for rifting in Afar a brittleelastic model of the behaviour of the lithosphere Earth Planet Sci Lett 45 435ndash444

Tarantola Ruegg JC Leacutepine JP 1980 Geodetic evidence for rifting in Afar 2 1980vertical displacements Earth Planet Sci Lett 48 363ndash370

Van Ngoc P Boyer D Le Moueumll JL Courtillot V 1981 Identification of a magmachamber in the Ghoubbet-Asal Rift (Djibouti) from a magnetotelluric experimentEarth Planet Sci Lett 52 372ndash380

Van Wyk de Vries B Merle O 1996 The effect of volcanic constructs on rift faultpatterns Geology 24 643ndash646

Vigier N Bourdon B Joron J-L Allegravegre CJ 1999 U-decay series and trace elementsystematics in the 1978 eruption of Ardoukoba Asal rift timescale of magmacrystallization Earth Planet Sci Lett 174 81ndash97

Vigny C de Chabalier J-B Ruegg J-C Huchon P Feigl KL Cattin R Asfaw LKynbari K 2007 Twenty-five years of geodetic measurements along the Tadjoura-Asal rift system Djibouti East Africa J Geophys Res 112 (B11) 6410 doi1010292004JB003230

Wright TJ Ebinger C Biggs J Ayele A Yirgu GJ Keir D Stork A 2006 Magma-maintained rift segmentation at continental rupture in the 2005 Afar dykingepisode Nature 442 291ndash294 doi101038nature04978


accommodated near the surface by open cracks This geometry seemsin agreement with the current surface observations that suggest sub-vertical normal faults and open fissures and it is also used to explainthe long-term evolution of the Asal Rift (Cattin et al 2005 DeChabalier and Avouac 1994 Stein et al 1991) In these tectonic-stretching models applied to the Asal Rift the slip on normal faults isnot correlated with a dike intrusion but controlled by the founderingof blocks into the lithosphere

Alternative models (Fig 1B) suggest that most of the deepextension results from localised magma intrusion with faultsaccommodating extension and subsidence above the maximumreach of the magma column (eg Abelson and Agnon 1997 Agnonand Lyakhovsky 1995 Buck 2006 Hubert-Ferrari et al 2003) whichis proposed for Afar in a model combining geophysical andpetrological data (Fig 1B Pinzuti et al 2007a) In magmatic riftingmodels or so-called magmatic intrusion models the yield stress isreduced by an order of magnitude (eg Buck 2004 2006) comparedto the tectonic-stretching models (eg Dunbar and Sawyer 1989Jackson and McKenzie 1983 Kusznir et al 1991 Lin and Parmentier1990) In these models the normal faults result from dike intrusionwhich impose a dipping plane of 45degndash55deg to the normal faults (egAgnon and Lyakhovsky 1995) as is also suspected beneath the AsalRift where the current horizontal extension (measured across theopening faults from InSAR data) suggests that the sub-vertical faults atthe surface have shallower dipping planes at depth (Peltzer andDoubre 2006) To link the surface observations of the Asal Rift withdipping normal fault at depth this fault geometry implies some slippartitioning (Bowman et al 2003) with opening cracks accommo-dating part of the extension and normal faults accommodatingsubsidence and the remaining fraction of extension

While these two models can lead to the same observations of sub-vertical faults and opening cracks at the surface the magmaticintrusion model contributes to sub-vertical scarps which are due toslip partitioning that turn to dipping normal faults at depth and thatroot into dike intrusions This study aims at 1) proposing frommorphology of exhumed normal fault in hard rock a new conceptualmodel of normal fault evolution due to magmatic intrusion and2) exploring the consequence of this model for the long-termevolution of the Asal Rift by linking the suspected geometry ofnormal faults at depth with magma intrusion

2 Geological setting

The Asal rift is the first emergent segment of the Aden ridge whichpropagates westward on land into the Afar depression (Fig 2A) (egManighetti et al 1998) With a ~40 km length whose 15 km areemerged it currently opens at 16plusmn1 mm yrminus1 in the N40degplusmn5degE

Stretching model

THINNEDLITHOSPHERE(or upper crust)

Subvertical faults

Dikes

ASTENOSPHERE(or lower crust)

10 km

0

-10

Cracks

km

A

Fig 1 A) Schematic representation of models of rift evolution dominated by ductile stretaccommodate subsidence controlled by distributed ductile deformation at depth Ductile deextending through the crust and mantle The model shown is based on Pinzuti et al (2007a)Knox et al 1998 Makris and Ginzburg 1987 Nyblade et al 2000 Ruegg 1975) and petroloare not at the same scale

direction (Fig 2A and B) This rift is structured by a dense network offissures and sub-vertical normal faults with throws up to 200 mpropagating northwestward (N130degplusmn10deg) from the Ghoubbet Bay tothe northwest shore of the Lake Asal (Fig 2B Manighetti et al 1998)

Asal Rift opening is characterised by effusive events associatedwith incipient rifting (Richard 1979) from 853plusmn35 ky to 315plusmn53 ky(Manighetti et al 1998) While immerged context allows theformation of hyaloclastites (326plusmn15 ky) in the south part of therift magmatic activity remains effusive in the north (315plusmn53 ky and334plusmn43 ky Manighetti et al 1998) From the end of this period until~100 ky the evolution of the rift is characterised by the activity of acentral volcano the Fieale (Fig 2) which fills the inner floor andconceals previous faults with large volumes of basalt lava flows(Pinzuti 2006 Pinzuti et al 2007a) Around 50plusmn20 ky magmaticactivity decreases and the whole successive basalt lava flows thatstructure the Fieale volcano become gradually offset by normal faults(Manighetti et al 1998) The structure of the modern rift starts 40ndash30 ky ago with the development of the border faults H andα1 (Fig 2Stein et al 1991 Manighetti et al 1998 Pinzuti 2006 Pinzuti et al2007b) From this period the Fieale volcano progressively collapsesand the magmatic activity locates within the inner floor along smallvolcanic edifices and eruptive fissures (Fig 2 Stein et al 1991 DeChabalier and Avouac 1994 Manighetti et al 1998 Doubre et al2007b)

The latest magmatic episode recorded in the rift corresponds to aseismic-volcanic sequence (Mb 5 and 53) that occurred in 1978(Abdallah et al 1979 Leacutepine et al 1980) This event produced ~2 mof extension in a N40˚ direction and up to 70 cm of subsidence in theinner floor (Ruegg et al 1979) A one-week basaltic fissural eruptiongenerated at the northwestern tip of the volcanic chain gave birth tothe Ardoukoba volcano (Fig 2B Allard et al 1979) Currently themost important deformations are observed around the Fieale edificeand the northeast part of the rift (Doubre and Peltzer 2007 Doubreet al 2007ab Manighetti et al 1998 Peltzer and Doubre 2006)

3 Data acquisition

To examine competent rocks which can sustain open fissures to asubstantial depth we studied massive rock types corresponding tobasalt We measured 14 vertical topographic profiles along 4 of majornormal faults of the Asal Rift which offset successive basalt lava flowsof 1 to 5 m thick along their traces (Fig 3A) Profiles were obtainedusing a handheld laser distancemetre and anglemeasuring binocularsto determine baseline distance horizontal distance and height foreach point (Fig 3B) The instrumental error associated with themeasurements is smaller than 50 cm which is less than the typicalldquoraggednessrdquo of fault scarps It should be noted that this technology

Magmatic intrusion model

3

-20

0

-80

-60

-40 LITHOSPHERICMANTLE

CONTINENTALCRUST

NESW

Dippingfaults

Lithosphericdike

40 kmAsalLake

AbheacuteObockkm

B

10 km

ching at the base of the lithosphere or within the lower crust Steeply dipping faultsformation is assumed to occur at depths of a few kilometres B) Localised deformationwhich combined seismic refraction seismological gravimetry (Berckhemer et al 1975gical data (Pinzuti 2006 Pinzuti et al submitted for publication) Note the two figures

11deg39

11deg37

11deg35

11deg33


km0 1 2

FiealeFieale


Lake AsalLake Asal

I

H

D

G

ε1

γ1

ε2δ

α3α3

α2α2

ππ

J

I

F

Gα1 β



E

γ2

42˚24 42˚30 42˚36

11˚30

11˚36

11˚42



FiealeFieale

A B

ADENAG

MI

ALAL

SOMALIASOMALIA

Red Sea

Red Sea


highs)highs)


EAEA

TA

MHMH

MH-GMH-G

ARABIAARABIA

Aden RidgeRed Sea






Ardoukoba

Fieale

Lake Asal

Ardoukoba

Fieale

1162deg

1158deg

1154deg

4245deg 4249deg

Lake Asal

Ghoubbetbay

Ghoubbetbay

4253deg4241deg

1166deg

A3A3

A1A1A2A2

A6A6A7A7

A8A8A9A9

A10A10A11A11 A5A5

A4A4

A12A12

A13A13A14A14

H

α1

γ1

γ2


H α1 γ1 γ2

2 km2 km

Major faults

A




α1

α2

Fault scarp

etc

d2

d1

0Talus

B


A B C

D E F












80 100 120 140 160Distance (m)

35deg

75deg

Talus Fault scarp

90deg

NW SE

0

20

40

60

80

100

80 100 120 140 160Distance (m)

A13

Hei

gh

t (m

)

Average

Dip 59degAverage

Dip 28deg

0

20

40

60

80

100

Dip

(deg)

80deg80deg


85deg85deg

40deg40deg

0 50 100 1500

20

40

60

80

65deg

45deg

15-25 ky

Ave

rag

e d

ip (

deg)

Height (m)

15-20ky ~40 ky ~25 ky

AB

C

D


A B






6 Conceptual model


3

2

1

123

Distance fro

0

A

Dep

th (

m)


5

67

8









0 1 2 3

2

m the dike (m)

B


4

5

3

6



A Initial extension



to the surface

Smallercracks at

depth

Smallercracks at

depth

Holocene scarp

Bed rock

Talus



Localisedat depth




Vertical


Horizontal

Shear zone at depth


intermediatedepth





A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H


α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE





A Lake Asal

Fieale1 km

2 km

3 km

4 km



0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km









8 Discussion






9 Conclusion




Acknowledgements




References





































































11deg39

11deg37

11deg35

11deg33


km0 1 2

FiealeFieale


Lake AsalLake Asal

I

H

D

G

ε1

γ1

ε2δ

α3α3

α2α2

ππ

J

I

F

Gα1 β



E

γ2

42˚24 42˚30 42˚36

11˚30

11˚36

11˚42



FiealeFieale

A B

ADENAG

MI

ALAL

SOMALIASOMALIA

Red Sea

Red Sea


highs)highs)


EAEA

TA

MHMH

MH-GMH-G

ARABIAARABIA

Aden RidgeRed Sea






Ardoukoba

Fieale

Lake Asal

Ardoukoba

Fieale

1162deg

1158deg

1154deg

4245deg 4249deg

Lake Asal

Ghoubbetbay

Ghoubbetbay

4253deg4241deg

1166deg

A3A3

A1A1A2A2

A6A6A7A7

A8A8A9A9

A10A10A11A11 A5A5

A4A4

A12A12

A13A13A14A14

H

α1

γ1

γ2


H α1 γ1 γ2

2 km2 km

Major faults

A




α1

α2

Fault scarp

etc

d2

d1

0Talus

B


A B C

D E F












80 100 120 140 160Distance (m)

35deg

75deg

Talus Fault scarp

90deg

NW SE

0

20

40

60

80

100

80 100 120 140 160Distance (m)

A13

Hei

gh

t (m

)

Average

Dip 59degAverage

Dip 28deg

0

20

40

60

80

100

Dip

(deg)

80deg80deg


85deg85deg

40deg40deg

0 50 100 1500

20

40

60

80

65deg

45deg

15-25 ky

Ave

rag

e d

ip (

deg)

Height (m)

15-20ky ~40 ky ~25 ky

AB

C

D


A B






6 Conceptual model


3

2

1

123

Distance fro

0

A

Dep

th (

m)


5

67

8









0 1 2 3

2

m the dike (m)

B


4

5

3

6



A Initial extension



to the surface

Smallercracks at

depth

Smallercracks at

depth

Holocene scarp

Bed rock

Talus



Localisedat depth




Vertical


Horizontal

Shear zone at depth


intermediatedepth





A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H


α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE





A Lake Asal

Fieale1 km

2 km

3 km

4 km



0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km









8 Discussion






9 Conclusion




Acknowledgements




References





































































A B C

D E F












80 100 120 140 160Distance (m)

35deg

75deg

Talus Fault scarp

90deg

NW SE

0

20

40

60

80

100

80 100 120 140 160Distance (m)

A13

Hei

gh

t (m

)

Average

Dip 59degAverage

Dip 28deg

0

20

40

60

80

100

Dip

(deg)

80deg80deg


85deg85deg

40deg40deg

0 50 100 1500

20

40

60

80

65deg

45deg

15-25 ky

Ave

rag

e d

ip (

deg)

Height (m)

15-20ky ~40 ky ~25 ky

AB

C

D


A B






6 Conceptual model


3

2

1

123

Distance fro

0

A

Dep

th (

m)


5

67

8









0 1 2 3

2

m the dike (m)

B


4

5

3

6



A Initial extension



to the surface

Smallercracks at

depth

Smallercracks at

depth

Holocene scarp

Bed rock

Talus



Localisedat depth




Vertical


Horizontal

Shear zone at depth


intermediatedepth





A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H


α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE





A Lake Asal

Fieale1 km

2 km

3 km

4 km



0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km









8 Discussion






9 Conclusion




Acknowledgements




References





































































80 100 120 140 160Distance (m)

35deg

75deg

Talus Fault scarp

90deg

NW SE

0

20

40

60

80

100

80 100 120 140 160Distance (m)

A13

Hei

gh

t (m

)

Average

Dip 59degAverage

Dip 28deg

0

20

40

60

80

100

Dip

(deg)

80deg80deg


85deg85deg

40deg40deg

0 50 100 1500

20

40

60

80

65deg

45deg

15-25 ky

Ave

rag

e d

ip (

deg)

Height (m)

15-20ky ~40 ky ~25 ky

AB

C

D


A B






6 Conceptual model


3

2

1

123

Distance fro

0

A

Dep

th (

m)


5

67

8









0 1 2 3

2

m the dike (m)

B


4

5

3

6



A Initial extension



to the surface

Smallercracks at

depth

Smallercracks at

depth

Holocene scarp

Bed rock

Talus



Localisedat depth




Vertical


Horizontal

Shear zone at depth


intermediatedepth





A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H


α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE





A Lake Asal

Fieale1 km

2 km

3 km

4 km



0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km









8 Discussion






9 Conclusion




Acknowledgements




References








































































6 Conceptual model


3

2

1

123

Distance fro

0

A

Dep

th (

m)


5

67

8









0 1 2 3

2

m the dike (m)

B


4

5

3

6



A Initial extension



to the surface

Smallercracks at

depth

Smallercracks at

depth

Holocene scarp

Bed rock

Talus



Localisedat depth




Vertical


Horizontal

Shear zone at depth


intermediatedepth





A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H


α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE





A Lake Asal

Fieale1 km

2 km

3 km

4 km



0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km









8 Discussion






9 Conclusion




Acknowledgements




References






































































A Initial extension



to the surface

Smallercracks at

depth

Smallercracks at

depth

Holocene scarp

Bed rock

Talus



Localisedat depth




Vertical


Horizontal

Shear zone at depth


intermediatedepth





A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H


α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE





A Lake Asal

Fieale1 km

2 km

3 km

4 km



0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km









8 Discussion






9 Conclusion




Acknowledgements




References





































































A B

C D

E F

G H

J

H

F

α1

α2

α3

D β

δ

ε1

I

G

K

ε2

-100

0

100

200

300

400

500

Ele

vati

on

(m

)

NGhoubbet

Bay

Fieale

Lake Asal

10 2 km

0200400

810987654321

6 4 2 0 2 4 6 8

Ele

vati

on

(m

)D

epth

(km

)

0200400

10987654321

Dep

th (

km)

Ele

vati

on

(m

) Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

8 6 4 2 0 2 4 6 8Distance (km)

0200400

Ele

vati

on

(m

)

10987654321

Dep

th (

km)

A B

C D

G H


α1δ ε1

GHJ

IJ

π

α2α3 π ε1dm ~ 6500 m

dm = ~4000 m

dm = ~2500 m ε2

ε2

ε2

0200400

10987654321

8 6 4 2 0 2 4 6 8

Distance (km)

Ele

vati

on

(m

)D

epth

(km

)

E FI HJ

α2δ ε1

dm = ~5200 m ε2

A

B

C

DE





A Lake Asal

Fieale1 km

2 km

3 km

4 km



0

02

06

1

14

-20 -10 0 10 20-4

-2

Dep

th (

km)

Str

ain

(ex

x) p

er m

eter

dik

e th

ickn

ess

(10-4

)

B DG= ~25 km

DG= ~4 kmDG= ~52 km

DG= ~65 km









8 Discussion






9 Conclusion




Acknowledgements




References











































































8 Discussion






9 Conclusion




Acknowledgements




References







































































Acknowledgements




References





































































surface morphology of active normal faults in hard rock: implications for the mechanics of the asal...

Documents