s

ctivity ongiant dikesin arcuate

tonics, welling. We

owards theof about

arsis trend.pdomingteristics ofrosion may

Icarus 176 (2005) 44–56www.elsevier.com/locate/icaru

Unveiling the origin of radial grabens on Alba Patera volcanoby finite element modelling

Beatrice Cailleaua,∗,1, Thomas R. Walterb, Peter Janlea, Ernst Hauberc

a Institute of Geosciences, Department of Geophysics, Kiel University, Otto-Hahn-Platz 1, D-24118 Kiel, Germanyb MGG/RSMAS, University of Miami, Rickenbacker Causeway 4600, Miami, FL 33149, USA

c Institut für Planetenforschung, Deutsches Zentrum für Luft- und Raumfahrt (DLR), Rutherfordstrasse 2, D-12489 Berlin, Germany

Received 11 July 2004; revised 22 January 2005

Available online 24 March 2005

Abstract

The Tharsis region is an 8000-km-wide structural dome that incorporates a concentration of the main volcanic and tectonic athe Planet Mars. The area of structural doming is characterised by giant radial graben-dike systems. Nested on a set of theseto the northern side of Tharsis, is Alba Patera, one of the largest volcanoes in the planetary system. The regional dikes there arearrangement and imply an E–W to NW–SE regional extension at Alba Patera. To assess the influence of regional and local tecstudied the dike orientations on the volcano with Viking mosaic data and simulated plausible stress fields with finite element modefound that the influence of a NW–SE regional extension was strong near the volcano centre but decreased rapidly in importance tnorthern pole, i.e., far from the Tharsis centre. By combining this regional stress with a broad uplift that is due to a buoyancy zone1400 km in lateral extent and centred under Alba Patera, we reproduced the radial pattern of dike swarms that diverge from the ThRegional tectonics may have dominated the early stages of dike injection. During the evolution of Alba Patera, however, local ucontrolled the dike pattern, supporting the idea of a hotspot under Alba Patera. The well-expressed dike geometry and characAlba Patera provide an ideal example for comparative study with analogue hotspots on Earth where plate tectonics and active ecomplicate the reconstruction of volcanic and tectonic history and the understanding of involved geodynamic processes. 2005 Elsevier Inc. All rights reserved.

Keywords:Mars; Alba Patera; Tharsis; Hotspot; Tectonics

tec--kmgi-rmasiseter

theon-

at-enstheum-dno-tric-3;

1. Introduction

Today, the Planet Mars is in a phase of volcanic andtonic quiescence. However, the 11,000-m high and 8000wide volcanic dome of the Tharsis region, transected byant tectonic features, bears witness to past intense defotion activity. At the northernmost boundary of the Thardome, the volcanic edifice of Alba Patera has a diam

* Corresponding author. Fax: +1-305-421-4632.E-mail addresses:bcailleau@rsmas.miami.edu(B. Cailleau),

twalter@rsmas.miami.edu(T.R. Walter),pjanle@geophysik.uni-kiel.de

(P. Janle),ernst.hauber@dlr.de(E. Hauber).

1 Now at: MGG/RSMAS, University of Miami, Rickenbacker Causeway4600, Miami, FL 33149, USA.

0019-1035/$ – see front matter 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.icarus.2005.01.017

-

up to 2700 km and a 7000-m-maximum elevation (Fig. 1).Up until now, the tectonics of Alba Patera have beenparticular object of attention. The volcano and its envirment are intensely breached by grabens (e.g.,Raitala, 1988).We distinguish the grabens of Alba Patera into two cegories including: (a) north to northeast trending grabthat converge at Alba Patera and connect to Tharsis insouth, and (b) concentric grabens that circumscribe the smit of Alba Patera (Fig. 2). The concentric grabens formeduring later stages of Alba Patera and various volcatectonic models were proposed to explain this concenity (e.g., McGovern et al., 2001; Scott and Wilson, 200

Cailleau et al., 2003).

In this study, we focus on the earliest formation ofgrabens with a radial orientation on Alba Patera. These

s. Tontes.p

Radial grabens on Alba Patera volcano 45

Fig. 1. The Tharsis region. The features extending 3000 km radial from the Tharsis plume centres are grabens related to underlying giant dike swarmhe starsrepresent tectonic centres identified byMège and Masson (1996). Tharsis is characterised by the 27-km-high Olympus Mons and the three Tharsis MAt the northern periphery of the Tharsis bulge, Alba Patera represents an important feature in this highly fractured landscape. This volcano develoed in three

main stages; a widespread volcanic unit is superposed by the centralised main shield about 1300 km across and 5 km high(Scott and Tanaka, 1986). The last

, cove rom MOLA

ra-eirex-

,ted

adat 5-um-an-

e todi-

ighe

re-)traln aingdial

weenare

n-)sion,of

ensthe

unit is the steeper summit cone about 400 km across and 2 km highscience team(Neumann et al., 2001).

grabens may be regarded radial relative to Tharsis anddial relative to Alba Patera, which poses the problem of thorigin. The grabens were generally attributed to regionaltension related to the activity of Tharsis (e.g.,Tanaka et al.1991) or defined as a Tharsis peripheral rift that deflecaround Alba Patera’s volcanic centre(Raitala, 1988). Wise(1976)combined low regional extension with a volcanic loon a thick plate, and estimated the graben orientationskm depth to be generally E–W and diverted around the smit. The method produced no radial structures and thusother process is implied.Watters and Janes (1995)acknowl-edged that radial grabens could have formed by uplift dua diapir. However, by modelling the ascent of a 224-kmameter diapir, they predicted radial stresses as being h

than hoop stresses beyond a 400-km distance from Alba Patera’s centre, which cannot explain the radial patterns of theextensive Tantalus Fossae (Fig. 2). Although they attributed

ring part of the concentric grabens. The shaded relief of Tharsis is f

r

the radial grabens to regional tectonics, they added nogional stress in their modelling.Turtle and Melosh (1997varied the direction of the regional stress around a cenpoint off to the southeast and added a volcanic load othin elastic lithosphere overlying a fluid mantle. Regardthe southern radial grabens and part of the northern ragrabens situated near the summit cone, the angles betpredicted and observed orientations of radial structuresin the range of 0 to 15◦. However, the discrepancies icrease up to 30◦ towards the pole.McGovern et al. (2001combined three stress fields: a regional constant extena growing volcano and uplift from a deep buoyant load320 km radius. The models partly fit the concentric grabnear the summit but do not match the radial grabens to

-north-western part of Alba Patera, nor the regional struc-tures with pit craters to the east (see Fig. 9 inMcGovern etal., 2001). The results were not superimposed to the grabens

(intes) are gtera

46 B. Cailleau et al. / Icarus 176 (2005) 44–56

Fig. 2. Lineaments of Alba Patera. The volcano is characterised by sets of grabens: N–S to the south (oldest Ceraunius Fossae), radial to the northermediateaged Tantalus Fossae), and concentric around the summit (young Alba Fossae to the west and Tantalus Fossae to the east). The Catenae (black linrabenswith pit craters related to giant dike swarms that initiated from the Tharsis centres (e.g.,Scott et al., 2002). The Catenae on the eastern flanks of Alba Pa

display a northeast trend that corresponds to a deviation of dike swarms observed at the northern and southern periphery of the Tharsis dome (Fig. 1). The

anks s that foOLA

-

a-bens

uc-

nu-oseses

surericaling

western Catenae are curved similar to the Alba Fossae on the upper flon the summit can be seen. The shaded relief of Alba Patera is from M

beyond 50◦ latitude, which in this study are of primary importance.

Recent studies showed the difficulty of linking Alba Ptera radial grabens purely to Tharsis tectonics. The gradiffer from Tharsis models of surface loading(Banerdt andGolombek, 2000). Based on converging patterns, the strtures were assigned to the Alba Patera centre(Mège and

Masson, 1996; Anderson et al., 2001). Moreover, gravityanomalies at Alba Patera point to a low density zone ofabout 1200 km in diameter, a size that may be explained

suggesting a local influence on regional structures. Two large calderarmedscience team(Neumann et al., 2001).

by a hotspot (e.g.,Janle and Erkul, 1991) that likely exertedan important influence on the graben geometry. Usingmerical modelling and structural observations, we propa model of interplay between regional and local procesleading to an Alba Patera hotspot interpretation. To asfull understanding of the processes involved, the numeresolution is decomposed into the following steps: dom

alone, addition of an E–W regional extension and use of amore complex regional stress field. We demonstrate that thepattern of the tectonic structures adds constraints to the lat-

on A

lly,itiesthe

tru-

og-sorenian

dy is

enss

todflas-

,rob-eep

nalrcu-;sdetheen-tre,

rth-–W

ba

re-l.,e-

lisedichp of

old-ossaeic)

the

adialnac-

allelde,dialdsns.s the—

truc-erplayollowal to

radialcon-the

ge in(greyreasesfrom

Radial grabens

eral dimension of a plausible hotspot in the mantle. Finaour results are compared with geophysical data. Similarwith hotspot tectonics and volcanism on Earth strengthenmodel of the origin of Alba Patera.

2. Regional stress fields and radial structures

2.1. Tharsis structural dome

The Tharsis dome is attributed to intense crustal insion and volcanic activity(Smith et al., 2001). Wide upliftfrom a rising plume may also contribute to the present topraphy (Harder, 2000). The tectonic activity in the Tharsiregion was predominant in the early history of Mars bef3.7 Ga ago and decreased until the middle-late Amazo1.4–0.3 Ga ago(Anderson et al., 2001). The relationshipbetween martian epochs and ages indicated in this studerived fromHartmann and Neukum (2001).

Tharsis is characterised by a giant system of grabformed by tensile stresses (Fig. 1). Radial to arcuate grabenmay be induced by isostatic or flexural stresses dueTharsis volcanic loading(Turcotte et al., 1981; Sleep anPhillips, 1985; Banerdt et al., 1992)and/or by spreading othe load in presence of a ductile layer between the etic upper crust and the strong upper mantle(Tanaka et al.1991). On the other hand, the extensive grabens are pably the surface expression of dikes controlled by a dsource (e.g.,McKenzie and Nimmo, 1999). Doming aboveplumes, or inflation of a wide deep reservoir and a regiostress dominant at the periphery, account for a radial to aate arrangement of giant grabens(Mège and Masson, 1996Wilson and Head, 2002)(Fig. 1). Thus, the Tharsis stresfield was not constant in time and location. Its magnituprobably waned with time. Decrease of magnitude fromdome centre to the surroundings is likely for such a ctralised rise. At a greater distance from the Tharsis centhe orientation of fractures is also not constant. At the noern periphery, the orientation of extension passes from Eto NW–SE, which has important implications for the AlPatera tectonics.

2.2. Structures on the Alba Patera local dome

North of the Tharsis rise, the volcano Alba Patera (Figs. 1and 2) developed in three main stages defined byScott andTanaka (1986). The edifices sizes are constrained by thecent MOLA data(Neumann et al., 2001; McGovern et a2001). The first stage consists of the formation of a widspread plateau followed in the second stage by centravolcanism that gave rise to the main volcanic shield, whis as large as 1300 km across and 5000 m high. On to

this, the construction of the 2000-m-high summit cone oc-curred in the third stage. Abundant grabens with radial andconcentric orientations dissect the broad volcanic region. In

lba Patera volcano 47

early stages, two sets of radial structures formed: theest Ceraunius Fossae to the south and the Tantalus Fto the north (Figs. 1 and 2). During later stages, concentrgrabens, which were fully described inCailleau et al. (2003and are not further considered here, developed aroundsummit region.

The Ceraunius Fossae are radial to Alba Patera and rto the Tharsis bulge(Raitala, 1988). Exact ages of grabeformation are difficult to determine owing to possible retivation at different stages.Tanaka (1990)found that a firsttectonic episode formed northeast striking grabens parto the Tharsis Montes. A later and main tectonic episoprobably related to a Syria Planum or Claritas centred rafaulting in the Noachian or in the early Hesperian periomore than 3.6 Ga ago, formed the north-striking grabeThe northern part of the Ceraunius Fossae, cutting acrosmain shield unit was reactivated—or continuously activeuntil the early Amazonian (2.9–2.1 Ga)(Tanaka, 1990). The

Fig. 3. Lineament analysis of Alba Patera. The orientations of tectonic stures are used to constrain the responsible stress fields and the intbetween local and regional stresses. Grabens or the underlying dikes fthe orientation of maximum compressive horizontal stress or are normthe minimum compressive horizontal stress. The N–S set of grabensto both the Tharsis centre and Alba Patera indicates strong regionaltrol by an E–W extension. To the north, giant dikes are strictly radial tovolcano showing that a local doming existed at Alba Patera. The chandyke trend observed on the northern upper flanks of the main shieldarea) evidences the influence of a NW–SE regional extension that decin magnitude toward the north. The grey area contours the lava flows

the main shield of Alba Patera(Scott and Tanaka, 1986). Concentric struc-tures are younger than the radial grabens and imply other processes notstudied here.

isyis un

48 B. Cailleau et al. / Icarus 176 (2005) 44–56

Fig. 4. Numerical modelling of local doming and regional extension. (a) Displacements and stresses arising from local doming are calculated in axmmetricfinite elements models. The elastic model is fulfilled in an elastic half-space; part of the mesh is presented here. The source of deep upward forcesknown

and ideally modelled by an elliptic area that has a densityρ lower than the density of surrounding rock. (b) The effect of a regional extension is also tested.

ed inx ition

atedhetic-to

niustec-0;

orthkmce

oba-,ra-onic, af-,os-

-an-

iod;nhus

byhee arsig

thecal

herefieldsa-h,reshersis

odel

reeween–W

tera,iant

-main

de-ingted

The surface axisymmetric stresses calculated previously are transferr(Twiss and Moores, 1992).

N–S oblong topography of the Ceraunius Fossae, relto linear intrusion and fissure-fed flows, differs from tAlba Patera centralised construction. Owing to this parular volcanic shape as well as their orientation relativethe regional tectonics and their older age, the CerauFossae were thought to originate from the volcanic andtonic activity of Tharsis (e.g.,Raitala, 1988; Tanaka, 199Anderson et al., 2001).

The Tantalus Fossae are deep grabens extending nward from the upper flanks of Alba Patera up to 1000far into the north polar plains. Owing to the long distanbeyond the uplifted Tharsis region, the grabens are prbly the surface expression of dikes (e.g.,Mège and Masson1996; Ernst et al., 2001). These lineaments are orienteddial to Alba Patera and arose at an intermediate tectand volcanic stage at the time of widespread volcanismter the formation of the Ceraunius Fossae(Anderson et al.2001). The dike intrusions responsible for the Tantalus Fsae probably fed part of the widespread flow field(Parfittand Head, 1993; Mège, 2001). The southern part of the Tantalus Fossae is observed on the more recent main shieldmay imply continual intrusive and tectonic activity. However, dike injection is thought to occur within a short perof a few tens of days (e.g.,McKenzie and Nimmo, 1999Wilson and Head, 2002), which contrasts with the time spabetween the formations of the plateau and main shield. Treactivation late in Alba Patera’s history, as suggestedTanaka (1990)for the Ceraunius Fossae, is more likely. Tnorthern and southern parts of the radial Tantalus Fossacontinuous and thus genetically connected but present a

nificant change of direction. The orientation of the structuresnearer to the summit is parallel to regional tectonic structures(eastern Catenae), and is compatible with a NW–SE exten-

to the (, y) coordinates system using the method of Mohr diagram before add

-

d

,

e-

sion. At farther distance, the structures fan out radial toAlba Patera centre implying a major influence of the lostress field.

2.3. Interplay between local and regional stress fields

Structural studies of Alba Patera are extensive and tis a general consensus that regional and local stresswere both involved in the graben formation on Alba Ptera (e.g.,Wise, 1976; Raitala, 1988; Turtle and Melos1997). However, as shown in the Introduction, the structuwere explained differently. In the following, we describe tmain observations that indicate interplay between Thaand Alba Patera and that allow us to develop a new mfor Alba Patera’s early history (Fig. 3).

A regional control on Alba Patera is evidenced by thmain tectonic characteristics. The Ceraunius Fossae betAlba Patera and the Tharsis centre shows an early Eregional-tensile-stress field radial to Tharsis (e.g.,Tanaka,1990; Turtle and Melosh, 1997) (Fig. 2). Tantalus Fos-sae are confined to the north-eastern side of Alba Pawhich is also the north-eastern side of Tharsis where gdike swarms propagated from the Tharsis centre(s)(Raitala,1988; Scott et al., 2002)(Fig. 1). The southern part of the radial Tantalus Fossae that cuts across the flanks of theshield is parallel to the north-east regional Catenae(Raitala,1988). For simplification, we use the term Catenae toscribe not only the craters’ chains but also their boundnormal faults and the underlying dikes that were injecfrom the Tharsis region.

However, local influences also exist. During the forma-tion of the Tantalus Fossae and the volcanic plateau, therewas no comparable activity in the area of Ceraunius Fossae,

on A

s-rsisfol-

theal.,t-ila

tresra’s

om-ress-

onaonic

era,re-ensSecev-odears-ikesur-

ectfore

di-our

dis-cal-

-e of

mheng,inedest

thecedfugh

icalon.tal

ce

e-areevi-to aon,

ikethess,

by

ts in-f the

zero.ess,dial to

Radial grabens

although this area is closer to the Tharsis centre(Andersonet al., 2001). Thus, dike injection underlying Tantalus Fosae arose from the Alba Patera centre and not from ThaTo the north, the Tantalus Fossae does not properlylow the Tharsis regional trend, but is rather radial tovolcano centre(Mège and Masson, 1996; Anderson et2001) (Fig. 1). On the western lower flanks of Alba Paera, the regional Catenae have a concentric pattern simto the local Alba Fossae on the upper flanks(Mouginis-Mark et al., 1988; Raitala, 1988)(Fig. 2). The westernCatenae likely formed after the eastern Catenae when sfields responsible for concentric grabens on Alba Patemid and upper flanks became dominant(Cailleau et al.,2003).

The complex patterns of the radial Fossae imply a ding local to Alba Patera and a decrease of regional stinto the northern regions (Fig. 3). The next sections demonstrate the close relationship between the local and registress fields on the Alba Patera area in the early tectstages.

3. Numerical models

3.1. Method

In order to reproduce the radial grabens of Alba Patwe investigate the interactions of local updoming and amote stress field. The radial dikes underlying the grabwere probably feeders to the widespread plateau (seetion 2.2). It is possible that the plateau resulted from seral phases of dike injection since the formation of flobasalt provinces extends across a period of one million y(Mège, 2001; Wilson and Head, 2002). However, the absence of crosscutting and the regular fanning of the dindicate that the stress field did not change significantly ding their injections and that the plateau growth did not affthe dike patterns. Moreover, the radial dikes formed bethe main shield and the summit cone (see Section2.2). Thus,we think that the topographic loading effect of the three efices was negligible so the topography is not included inmodels.

The modelling procedure consists of two steps. First,placements and stresses arising from local doming areculated using the finite element program TEKTON(Meloshand Raefsky, 1983). Uplift caused by lighter or hot material and/or input of magma, is simulated by a decreasdensity relative to the surrounding rocks (Fig. 4). The re-gion of density difference is modelled elliptically in forand, for simplification, is called the “buoyant zone.” Tphysical meaning of this region, i.e., plume, underplatimagma chamber or intrusive complex, can be determa posteriori from the characteristics of the model that b

correlate with the observed structures as discussed furthebelow. The results are sensitive to density contrasts be-tween the buoyant zone and the surroundings. Hence, the

lba Patera volcano 49

.

r

s

l

-

density of surrounding rocks is fixed to the value ofcrustal density. Models are axisymmetric and are produwith an elastic flat half-space.Figure 4 presents part oa mesh. The bottom and lateral sides are distant enofrom the buoyant zone to avoid boundary effects. Vertboundaries of the model are fixed in a horizontal directiThe bottom of the grid is fixed. The density of the crusrocks is 2900 kg/m3 (Zuber, 2001), the Young’s Modulusis 50 GPa and the Poisson ratio is 0.25 (e.g.,Turcotte andSchubert, 1982). Gravitational acceleration on Mars’ surfais 3.7067 m/s2.

Second, uplift is combined with an E–W or NW–SE rgional extension. With a regional stress, the modelsno longer axisymmetric. Surface stresses from the prous models of subsurface loading are transformed in(xy) plane view before addition to the regional extensifollowing the method of the Mohr diagram(Twiss andMoores, 1992). The studied grabens are related to dintrusions. We are thus interested principally in (a)orientation of maximum compressive horizontal streand (b) the magnitude of stress difference (σ1 − σ3).These two parameters condition the directions followeddikes.

Fig. 5. Example of axisymmetric surface stresses and displacemenduced by a deep buoyancy zone. The horizontal and vertical radii ocore areRx andRz. d is the depth to its top,�ρ the difference of densityrelative to the surroundings. On the surface, the vertical stresses areDikes follow the orientation of maximum compressive horizontal strwhich corresponds here to the radial stress. Hence, dikes propagate ra

rthe summit. Around the summit, however, the difference between the hoopand radial stresses is so little that no preferential orientation in absence ofregional stress is detected.

Icaru

m are-racetryaxisnghenn is

therelatheofs obe-s thh ofo obilar

;

Pa-thethe

plot-

e ofl)f therac-rn of

-east.are

aveield,

d tore-of

sssae.om-

ri-ust.nicand

lin-

50 B. Cailleau et al. /

3.2. Pure doming

Here, we describe the stresses that may form frodecrease in density within the buoyant zone only. Nogional stresses are applied. Doming of Alba Patera is chaterised by maximum tensile stresses at the axis of symm(Fig. 5). Hoop and radial stresses are equal near theof symmetry, thus no preferential orientation of fracturican be predicted on the summit part of the dome. Wthe hoop stress exceeds the radial stress, dike injectioradial to the doming centre. Varying the dimensions ofbuoyant zone does not change the radial-hoop stresstionship, but does affect the wavelength and amount ofdeformation. Widening the horizontal or the vertical sizethe source or increasing its depth enlarges the regionfaulting on the free surface. Higher density differencetween the buoyant zone and surrounding rocks augmentmagnitude of stress but does not modify the wavelengtthe deformation on the surface. Greater stresses are alstained for shallower buoyant zones. These results are simto precedent numerical experiments (e.g.,Stofan et al., 1991Banerdt et al., 1992).

Figure 6superposes the observed structures of Albatera and the predicted directions followed by dikes, i.e.,orientation of minimum compressive stress. Note thatmodels are transformed into planetary coordinates and

ted in Mercator projection to be compared with the Viking

40◦ latitude and 250◦ longitude. In order to compare the numerical results wimeters to degrees and from the aerocentric to the areographic system (theThe GMT plot is given in Mercator projection.

s 176 (2005) 44–56

-

-

f

e

-

not fit the linear Ceraunius Fossae on the southern sidAlba Patera (Fig. 2), which requires that a further (regionastress field was active there. Whatever the dimensions oAlba Patera dome, the model reproduces the radial chater of the Tantalus Fossae to the north and also the pattethe regional Catenae and Mareotis Fossae to the northHowever, some discrepancies for the radial structuresfound at a distance of 200–400 km from the summit (Fig. 6).In addition to the local source, a regional stress may hinfluenced the structures on the flanks of the main shwhich is investigated further in the following section.

3.3. E–W regional extension

A regional extension of constant magnitude is addethe mechanism of buoyancy-driven doming. Where thegional stress is oriented dominantly E–W, the orientationlikely faulting or dike propagation is N–S. Regional stremay have contributed solely to the N–S Ceraunius FosSubsequent reactivation may have occurred with local ding.

Figure 7is the combined result of two distinct numecal models. In both models, a buoyant zone uplifts the crOn the right-hand side, the influence of regional tectostress is added. For a combination of regional extensionlocal doming of Alba Patera, we predict a change from

ear N–S fractures near the summit to radial orientation at

” to-

reson with thel reprodure i

position

images. Models of doming are compared to the northern sideof Alba Patera; the fan shape of the predicted structures does

the east and west flanks, giving an “hourglass shapethe structural pattern (Fig. 7). At the periphery, the frac

Fig. 6. Predicted dike orientation induced by doming centred on Alba Patera superposed with the northern part of the volcano (Viking mosaic with alutionof (1/256◦)/pixel, about 230 m). Radial directions are predicted independent of buoyancy zone dimension and depth. There is striking correlatiofan shape of Tantalus Fossae at a distance far from the summit, indicating control by the local stress field. Near the summit, however, the modecesthe structures poorly, i.e., the observed bend on the northern upper flanks and the direction of the north-east regional structures to the east. Thes also noagreement with the concentric structures close to the summit. These are younger grabens that imply a different process. The doming centre ised at

th the observed structures, the coordinates (x, y) of the models are changed fromplanetographic system). Mars is understood as a sphere of 3386 km mean radius.

on A

sionmuminedimumc-ma

ii ats for

arbi-houtg thes th

arerea-ingdialtnes

theof

e forartrtioncan

one ofsultwe

an-theod-

de

uresthe

onalhaveflu-60flu-artly

ofandlain-s on

s oferalf thee isry,

re-60outm-ssesntlythetraloadks.w-oes

e theave-

ionsof

Wol-ver

eld.o bee-

lythe

one

Radial grabens

Fig. 7. Doming (left) and combination with an east–west regional exten(right). The dashes indicate the predicted surface orientation of maxicompressive horizontal stress followed by dikes (top). Black lines josome of these dashes to better represent the dike pattern. The maxof differential stressσ1 − σ3 indicates a preferred area of magma injetion (bottom contours). The inflation or buoyancy-driven upward magmigration is simulated by a density decrease of 200 kg/m3 relative to thesurrounding in a core of 1000 and 25 km horizontal and vertical rada 120-km depth to its top. At the centre, the radial and hoop stressedoming have a maximum magnitude of 10 MPa, compared to thetrary 3 MPa regional stress. The corresponding differential stress witregional stress is also 10 MPa, the minimum stress at the centre beinvertical stress, i.e., zero on the surface. The black dashed ellipse showcontour of a volcano representative of Alba Patera’s main shield.

tures follow the regional N–S trend again. Thus, theretwo areas influenced by the regional stresses for twosons. First, around the axis of symmetry, local upwarpinduces no preferred orientation of faulting because raand hoop stresses are nearly the same (Fig. 5). Second, athe periphery, the stress field due to local doming waand the regional stress becomes dominant. Regardingposition of maximum stress difference, the combinationregional stresses with local doming augments the chancfaulting or dike injection in the northern or southern pof the volcano compared to the eastern or western po(Fig. 7). The linear pattern predicted around the summitstill not explain the north-east graben pattern observedthe flanks of the main shield. The eastern occurrencthe Tantalus Fossae also departs from the symmetric refound with an E–W regional stress. In the next section,

test the hypothesis that the regional stress is oriented NW–SE and of decreasing magnitude with distance from Thar-sis.

lba Patera volcano 51

e

s

3.4. NW–SE regional extension

Based on the observation that the trend of the radial Ttalus Fossae on the northern upper flanks is similar topattern of the eastern regional Catenae, we investigate mels of doming with a NW–SE extension with a magnituthat decreases towards the north (seeFig. 3). The correlationbetween numerical models and the Alba Patera structdepends on the regional stress, i.e., the amplitude andstyle of magnitude decrease. A linear decrease of registress towards the north was tested and was found tofew effects on the northern upper flanks and too much inence on the lower flanks. A cosine decrease from 40 to◦latitudes was chosen to approximate partly the waning inence of regional stress radial to the Tharsis centre and pthe change of extension direction—here the orientationregional extension is kept constant. The functional formlength scale of the regional stress was successful in exping the complex geometry of younger concentric grabenAlba Patera(Cailleau et al., 2003).

As stated in the previous section, there are two areafault orientation influenced by the regional stress. The latextent of these areas depends on the horizontal size odoming source. The horizontal radius of the buoyant zonfirst tested for 300 and 500 km. At the northern periphethe faulting follows the regional trend. To avoid this, thegional stress must be zero at lower latitudes, i.e., below◦.However, a 300–500 km radius for the source is ruledfor the following reason. On the volcano summit, local doing leads to hoop stresses much higher than radial streso that the addition of a regional stress does not significaaffect the faulting orientations. On the other hand, whenuplifted area is too large, i.e., beyond 900 km, the cenarea of faulting influenced by the regional stress is too brand reaches the strictly radial pattern on the lower flanThis is true even if the regional stress magnitude is loered. Increasing the lateral size of the doming source dnot proportionally augment the stress magnitude becausdeformation or stresses are distributed over a larger wlength.

Correlation between predicted and observed orientatof grabens is better when modelling a buoyant zonehorizontal radius around 700 km (Fig. 8). The estimatedhorizontal radius is strikingly similar to the 650 km E–radius of the Alba Patera main shield, although the vcanic topography has not been included. This is howemuch larger than the 350 km N–S radius of the main shiThe corresponding upward displacements are found trealistic, i.e., from 500 to 2000 m, similar to swell elvations associated with terrestrial plumes(Crough, 1983;Sleep, 1990). In this work, constraints can be obtained onon the horizontal extent of the buoyant zone. We testedvertical extent and depth to the top of the buoyancy z

within realistic values: about 50–200 km for vertical extentand 20–100 km for depth. We found that these parametershave few effects on the geometrical results because their

00 kgto the tD. To find

ecrea

52 B. Cailleau et al. / Icarus 176 (2005) 44–56

Fig. 8. Addition of local doming and a NW–SE regional stress with south–north decrease of magnitude. To simulate uplift, a density decrease of 2/m3

relative to the surrounding is applied within an elliptic core. The core has a horizontal radius of 700 km and a vertical radius of 50 km. The depthopof the core is 50 km. The maximum tension and maximum upward displacement are found to be 28 MPa and 930 m at the axis of symmetry in 2correlation with the direction of the dikes, the local doming is added with a regional NW–SE extension. The magnitude of the regional extension dses ina cosine form from 6 MPa at 40◦ latitude to zero at 60◦ latitude. Best results are obtained when the centre of uplift is situated at 40◦ latitude and at 100 km

east of the 250◦ line of longitude. To the right, the differential stress is presented. The minimum is found to the west in agreement to the absence of radial

and s asphic

f the

is

cedd at

n of

asthern

isnaercu-antm–E

dictethe

-;

byob-

f this

ionden-ndedialionaltentn.

tionralismress.

ys a

m-omeant

thisthis

sub-we

able

thin

structures there. The maximum matches the eastern Tantalus Fossaewell as from the aerocentric to the areographic system (the planetogra

magnitudes are much lower than the horizontal extent obuoyancy zone.

An overall good correlation in the orientation of faultsobtained when the maximum regional stress at 40◦ latituderepresents about 25% of the maximum local stress induby updoming and when the centre of doming is situate40◦ latitude and 100 km east of 250◦ longitude. The cal-culated maximum stress difference predicts the positiolikely failure or preferred area of dike intrusion (Fig. 8). Thisshows that radial structures should be confined to the eern side in accordance with the observations on the nortside of Alba Patera. The maximum of differential stresssituated in the region transected by the regional CateOur axisymmetric models simulate a buoyancy zone cilar in planform. Other (non-circular) shapes of the buoyzone may result in a slightly different position of maximudifferential stress that would better correspond to the NTantalus Fossae. The match between observed and predike orientations is generally good. We have reducedmaximum angle of discrepancy down to 5◦ compared to previous models (see Introduction;Turtle and Melosh, 1997McGovern et al., 2001).

4. Discussion

A comparison between the dike orientations predictedthe numerical models and the geometry of lineamentsserved on Alba Patera leads to a better understanding o

remote volcano. Based on the crosscutting relationship ofgrabens,Cailleau et al. (2003)suggested recently that AlbaPatera was the location of local subsidence in later stages o

Catenae. The coordinates (x, y) of the models are changed from meters to degreesystem). The GMT plot is given in Mercator projection.

-

.

d

volcano-edifice evolution. The concentric fracture formatin the later stages is related to a long-term increase insity in the crust by accumulation of intrusive material acooling (Cailleau et al., 2003). In this paper we studied thearly stages of Alba Patera, and the formation of the ragraben system. Our study suggests a combination of regand local loading mechanisms, which provides a consismodel for Alba Patera’s early and later stages of evolutio

Below we discuss the use of the flat plate approximain our models followed by a comparison of our structuresults with gravity data that support a hotspot mechanfor Alba Patera. Finally, we show that Alba Patera shamultiple characteristics with hotspots on Earth and Venu

4.1. Flat plate approximation

Membrane stresses are invoked when a load displalarge size in comparison to the planet’s radius.Turcotte etal. (1981)inferred an approximate threshold of 1300-khorizontal extent beyond which membrane stresses becimportant. Although the size of the Alba Patera buoyzone determined from the graben geometry approachesthreshold, the models using a flat plate developed instudy fit with the observed structures (Fig. 8). Based on thiscorrelation, we suppose that for the case of Alba Paterasurface load, planet curvature had negligible effect. Sinceconsider the plate to be very thick, the result is comparwith the findings byJanes and Melosh (1990)(Fig. 9). Fora 1400-km-wide load, the elastic thickness must be as

f

as 20 km to be in the transition zone where both membranestresses and bending stresses support the load. Otherwise,the plate model is suitable(Janes and Melosh, 1990). For

on A

hichsitionthe lod-nd ae neg

toarsis

der-ignlting

the

-hortntlethes isob-etry

y ae-

sity;2)achncy

rceom-e-

l.,to

) an

urts onracts

el-npro-x-

almem-iusd on

st andom-sed

aiin-

0;in

risonressnta-dy,and a

Radial grabens

Fig. 9. Ranges of load width and lithosphere thickness on Mars for wthe lithosphere can be treated as a flat plate, a shell or is in the tranzone where both membrane stresses and bending stresses supports(afterJanes and Melosh, 1990). The width of Alba Patera subsurface loaing determined in our study is around 1400 km. For such a width ahalf-space as used in our models, planet curvature is expected to havligible effect. In Janes and Melosh’s study, Mars’ radiusRp is understoodto be 3398 km, andq is the support parameter: the ratio of buoyancyflexural support. Cases for surface loadings of Tharsis region and ThMontes volcanoes are indicated by the dotted lines.

such a comparison, loading by a buoyancy zone is unstood to be similar to surface unloading for which the sof stresses are inversed and the domains of different fauremain unchanged. This does not take into considerationpossible effect of loading depth.

As demonstrated byWatts (2001), the lithosphere’s elastic thickness depends on the time-scale of loading. For speriods of loading, the relaxation processes of the maare not significant and the elastic plate is thick. Butlithosphere behaves as a thin plate if a loading proceslong. In our models we used a thick plate. Therefore theserved correlation between the Alba Patera graben-geomand our elastic half-space models may imply not onlnegligible effect of planet curvature but also a short timscale of loading. From gravity data, the large low-denzone may still be visible today(Janle and Erkul, 1991McGovern et al., 2001; Arkani-Hamed and Riendler, 200.The buoyant zone evidenced in our study might rapidly rea stationary state, possibly at the level of neutral buoya(e.g.,Wilson and Head, 2002).

4.2. Comparison with gravity modelling

The lineament study and our numerical models reinfothe idea that Alba Patera was the centre of a hotspot. A cparison of our results with gravity modelling follows. ThViking Orbiter II and MGS missions show a large lowdensity zone (e.g.,Janle and Erkul, 1991; McGovern et a2001). The origin for such a gravity anomaly may be due(a) a sediment pile, (b) a thickened crust, (c) flexure, (d

altered lithosphere of reduced density, or (e) an anomalousupper mantle and/or lower crust (plume, hotspot)(Crough,1983).

lba Patera volcano 53

ad

-

Fig. 10. Comparison of the plume model proposed byJanle and Erkul(1991)based on gravity data from the Viking Orbiter II mission with omodel of buoyancy zone derived from the geometry of radial lineamenAlba Patera. The top layer represents the crust. Where the plume intewith the crust, the density was understood to be 3120 kg/m3. See Discus-sion section for further details.

Figure 10presents a plume model for Alba Patera devoped byJanle and Erkul (1991). Good correlation betweeobserved and calculated gravity anomalies along a S–Nfile was found for a plume radius of 600 km, a vertical etent of 300 km, a density contrast of−10 kg/m3 with themantle and+120 kg/m3 with the crust, and a mean crustthickness of 50 km. An uplift of 1 km was estimated frocalculating the volume expansion induced by excess tperature. This corresponds strikingly to the 700-km radand 50-km-deep buoyancy zone that we modelled basethe geometrical characteristics of radial dikes (Fig. 8). Weemphasise that two very different methodologies suggesimilar mechanism with similar geometries. The size adepth strongly supports the idea of a buoyant mantle analy underneath Alba Patera. To simulate buoyancy, we ua density contrast in the range between 100 and 200 kg/m3

similar to the mantle plume density contrast for Haw(Richards et al., 1989). Other studies assumed density cotrast from 1 to 10% (e.g.,Lachenbruch and Morgan, 199Sleep, 1990). The discrepancy of one order of magnitudethe chosen density contrast does not affect the compawith the geophysical model. Density contrast modifies stmagnitude. However, it does not change the stress orietion that gives the theoretical dike geometry. In our stua 5% density contrast, the presence of flexural stresses

smaller hotspot volume (due to an elliptic form and a smallervertical radius) lead to a 930-m uplift (model ofFig. 9). Thisis very similar to the 1 km inferred from gravity modelling

Icaru

os of

rav-

sticat-ionshick

potsand

spot

ismuter-,edions

itharthc-

nal

ti-d by

;of

nae95)etersive

theotseep

lower-thetived the

lsoed byubse-9;

lba4

54 B. Cailleau et al. /

by Janle and Erkul (1991). Similar results obtained in twdifferent methods would then constrain the magnitudethe regional stress and local doming to be, at 40◦ latitude,around 6 and 28 MPa, respectively. Based on MGS gity data and flexural models of volcanic loading,McGovernet al. (2002)suggested the estimated 40–60-km-thin elalithosphere to be indicative of a domal rise for Alba Pera. This was supported by negative density perturbatcombined with topography and elastic stresses within a tshell(Arkani-Hamed and Riendler, 2002).

4.3. Alba Patera, a classic hotspot?

Alba Patera displays signatures that resemble hotson Earth and other planets. Using temporal, geometric,geophysical constraints, we illustrate below that the hotmodel is applicable to Alba Patera.

Alba Patera has been a location of persistent volcanduring several 100 Ma time-spans(Tanaka, 1990; Cailleaet al., 2003). Similar time-spans are also discussed forrestrial hotspots and mantle plumes(Jellinek and Manga2004). The 1000–2000-km extent of Alba Patera’s upliftregion and dike length is comparable to such dimenson terrestrial hotspots(Mège, 2001). Also the dike con-figuration on Alba Patera shares strong similarities wdike swarms related to hotspots or mantle plumes on Ethat diverge from a focal centre into two or three diretions (Fig. 11) (Ernst et al., 1995). Hence, it is unlikelythat Alba Patera radial tectonics resulted only from regio

rifting as demonstrated in this study. Besides long radial

(GDS) configuration on Alba Patera shares strong similarities with dike swadirections. The origin of hotspots may include rifting along diverging plates othe volcanic edifice. The Alba Patera hotspot was, in turn, related to the Tha

s 176 (2005) 44–56

provinces on Earth (e.g.,Richards et al., 1989; Courtilloet al., 2003). A high-density intrusive core was also evdenced by the late concentric grabens on Alba Patera anshort-wavelength gravity anomalies(McGovern et al., 2002Cailleau et al., 2003)analogous to the Muskox intrusionthe Mackenzie plume in Canada(Mège, 2001). The grabensstructures of Alba Patera also show similarities with coroon Venus(Cyr and Melosh, 1993; Watters and Janes, 19.Coronae are low-relief features of 100–2600 km in diamcharacterised by radial and concentric dikes, and extenvolcanism(Stofan et al., 1992).

The origin of hotspots on Earth and Venus includesfollowing processes initiating at different depths: (i) Hotspmay be the result of plumes which originate at depths as das the core-mantle boundary(Morgan, 1972). These plumesmay themselves cause secondary plumes at the upper-mantle boundary(Courtillot et al., 2003). Stofan and Smrekar (2005)suggested the coronae on Venus to besurface expression of secondary plumes. (ii) An alternamechanism for hotspot implies subduction processes anremelting of crust trapped in the upper mantle(Sandwelland Schubert, 1992; Foulger and Natland, 2003). (iii) Thegeneration of the low density mantle material may abe a consequence of shallow crustal processes causextensional lithospheric stress, pressure release, and squent partial melting(Magee and Head, 1995; Sheth, 199Anderson, 2000).

Shallow and deep origins were both suggested for APatera (e.g.,Janle and Erkul, 1991; Anderson et al., 200).

Rather than an individual plume, the Alba Patera hotspot

ular

rm

dike swarms, Alba Patera displays initial widespread volcan-ism comparable to flood basalt provinces or large igneous

may be the result of a secondary plume from an irregTharsis mega-plume system(Janle and Erkul, 1991). The

Fig. 11. Hotspots characters. Two structures on Earth (left and centre images, afterErnst et al., 1995) and Alba Patera on Mars (right). The giant dike swa

rms related to hotspots on Earth that diverge from a focal centre into two orthreer a mantle plume. The GDS of Alba Patera are caused by a local hotspot beneathrsis mega plume.

on A

, af-ult-aveon;unt

tera

eepmsnseeldhusedterar-

thisvity

the

thetha

vol-theion

his irms. Weolledan-

herad-all-

ankefulslie

for

.C.,

.R.,ars.

rtianhys.

n ofSci.

nicss,

7.anic

of

pes308.165–

near

iant8.rth,

ence

eo-

n oflogy4.neral

Earth

an-

tism174,

withhys.

elt:ia

eu-the

769–

M.,ad,tioneo-

the

arthheir164.rsis

dip

Radial grabens

Tharsis mega-plume formed a broad uplift and extensionfecting the area of Alba Patera prior to its growth. The resing lithospheric thinning under Ceraunius Fossae may hinfluenced the location of the volcano at its northern portilithospheric thinning was shown to favour a greater amoof pressure-release melting(Magee and Head, 1993). Melt-ing of a detached volcano root was proposed for Alba Paand the concentric grabens(Scott and Wilson, 2003). Thistheory would be applicable for the late stages once a droot formed but cannot account for initial radial dike swarand widespread volcanism. Or Alba Patera may be a coquence of crustal deformation due to the Tharsis stress fipassive upwelling and upper mantle decompression. Tthe mechanism by which the Alba Patera hotspot formremains to be determined. Regardless of the Alba Pahotspot origin, it likely initiated in response to the Thasis mega-plume. However, the further development ofanomaly, as well as the local volcanic and tectonic actiat Alba Patera, was, in any case, largely uncoupled fromTharsis dome.

5. Conclusion

Combining numerical modelling and the geometry ofradial structures observed on Alba Patera, we suggestupdoming arose from a buoyancy zone centred at thecano. The regional stress inferred from the pattern ofnorthern radial structures is a NW–SE regional extenswhose magnitude decreases from the Tharsis centre. Tin accordance with the arcuate trend of giant dike swaobserved at the northern periphery of the Tharsis bulgepropose that the early stages of Alba Patera were contrby a local hotspot causing wide doming, flood basalt volcism, and radial dike swarm formation.

Acknowledgments

Support was provided to B. Cailleau by the “DeutscForschungsgemeinschaft DFG” through a grant to the guate school of Geomar, Kiel University “Dynamik globKreisläufe im System Erde” (GRD171). T. Walter acknowedges a grant from the DFG (WA 1642). The authors thEllen Stofan and the editor Peter Thomas for their carreviews. We are also grateful to Richard Ernst and LeBleamaster for their helpful comments.

References

Anderson, D.L., 2000. The thermal state of the upper mantle; no rolemantle plumes. Geophys. Res. Lett. 27, 3623–3626.

Anderson, R.C., Dohm, J.M., Golombek, M.P., Haldemann, A.F

Franklin, B.J., Tanaka, K.L., Lias, J., Peer, B., 2001. Primary centersand secondary concentrations of tectonic activity through time in thewestern hemisphere of Mars. J. Geophys. Res. 106, 20,563–20,585.

lba Patera volcano 55

-,,

t

s

Anderson, R.C., Dohm, J.M., Haldemann, A.F.C., Hare, T.M., Baker, V2004. Tectonic histories between Alba Patera and Syria Planum, MIcarus 171, 31–38.

Arkani-Hamed, J., Riendler, L., 2002. Stress differences in the malithosphere: constraints on the thermal state of Mars. J. GeopRes. 107, 5119.

Banerdt, W.B., Golombek, M.P., 2000. Tectonics of the Tharsis regioMars: insights from MGS topography and gravity. Lunar. Planet.31. Abstract 2038.

Banerdt, W.B., Golombek, M.P., Tanaka, K.L., 1992. Stress and tectoon Mars. In: Kieffer, H.H., Jakosky, B.M., Snyder, C.W., MatthewM.S. (Eds.), Mars. Univ. of Arizona Press, Tucson, AZ, pp. 249–29

Cailleau, B., Walter, T.R., Janle, P., Hauber, E., 2003. Modeling volcdeformation in a regional stress field: implications for the formationgraben structures on Alba Patera, Mars. J. Geophys. Res. 108.

Courtillot, V., Davaille, A., Besse, J., Stock, J., 2003. Three distincts tyof hotspots in the Earth’s mantle. Earth Planet. Sci. Lett. 205, 295–

Crough, S.T., 1983. Hotspot swells. Annu. Rev. Earth Planet. Sci. 11,193.

Cyr, K.E., Melosh, H.J., 1993. Tectonic patterns and regional stressesVenusian coronae. Icarus 102, 175–184.

Ernst, R.E., Head, J.W., Parfitt, E., Grosfils, E.B., Wilson, L., 1995. Gradiating dyke swarms on Earth and Venus. Earth Sci. Rev. 39, 1–5

Ernst, R.E., Grosfils, E.B., Mège, D., 2001. Giant dyke swarms: EaVenus and Mars. Annu. Rev. Earth Planet. Sci. 29, 489–534.

Foulguer, G.R., Natland, J.H., 2003. Is “hotspot” volcanism a consequof plate tectonics? Science 300, 921–922.

Harder, H., 2000. Mantle convection and the dynamic geoid of Mars. Gphys. Res. Lett. 27, 301–304.

Hartmann, W.K., Neukum, G., 2001. Cratering chronology and evolutioMars. In: Kallenbach, R., Geiss, J., Hartmann, W.K. (Eds.), Chronoand Evolution of Mars. Kluwer Academic, Norwell, MA, pp. 165–19

Janes, D.M., Melosh, H.J., 1990. Tectonics of planetary loading: a gemodel and results. J. Geophys. Res. 95, 21,345–21,355.

Janle, P., Erkul, E., 1991. Gravity studies of the Tharsis area on Mars.Moon Planets 53, 217–232.

Jellinek, A.M., Manga, M., 2004. Links between long-lived hot spots, mtle plumes, D′′, and plate tectonics. Rev. Geophys. 42. RG3002.

Lachenbruch, A.H., Morgan, P., 1990. Continental extension, magmaand elevation; formal relations and rules of thumb. Tectonophysics39–62.

Magee, R.K., Head, J.W., 1993. Large-scale volcanism associatedCoronae on Venus: implications for formation and evolution. GeopRes. Lett. 20, 1111–1114.

Magee, K.P., Head, J.W., 1995. The role of rifting in the generation of mimplications for the origin and evolution of the Lada Terra–LavinPlanitia region of Venus. J. Geophys. Res. 100, 1527–1552.

McGovern, P.J., Solomon, S.C., Head, J.W., Smith, D.E., Zuber, M.T., Nmann, G.A., 2001. Extension and uplift at Alba Patera: insights fromMOLA observations and loading models. J. Geophys. Res. 106, 23,23,809.

McGovern, P.J., Solomon, S.C., Smith, D.E., Zuber, M.T., Simons,Wieczorek, M.A., Phillips, R.J., Neumann, G.A., Aharonson, O., HeJ.W., 2002. Localized gravity/topography admittance and correlaspectra on Mars: implications for regional and global evolution. J. Gphys. Res. 107.

McKenzie, D., Nimmo, F., 1999. The generation of martian floods bymelting of ground ice above dykes. Nature 297, 231–233.

Mège, D., 2001. Uniformitarian plume tectonics: the post-Archean Eand Mars. In: Ernst, R.E., Buchan, K.L. (Eds.), Mantle Plumes: TIdentification Through Time. Spec. Pap. Geol. Soc. Am. 352, 141–

Mège, D., Masson, P., 1996. A plume tectonics model for the Thaprovince, Mars. Planet. Space Sci. 44, 1499–1546.

Melosh, H.J., Raefsky, A., 1983. Anelastic response of the Earth to a

slip earthquake. J. Geophys. Res. 88, 515–526.

Morgan, W.J., 1972. Convection plumes and plate motions. AAPG Bull. 56,203–213.

Icaru

up-

ber,ys.

mentand

ars.

hot

eric

orialale

ikes

s ofeo-

deep

logy

ter-phys.

le-.J.,eu-.B.,X.,

991.. 96,

.L.,teris-rigin,378.largees?s.),ecial

ssae

ofphys.

tin-

ofphys.

rtian

ork.tions

idge

orth-

sur-dels

56 B. Cailleau et al. /

Mouginis-Mark, P.J., Wilson, L., Zimbelman, J.R., 1988. Polygenic ertions on Alba Patera. Bull. Volcanol. 50, 361–379.

Neumann, G.A., Rowlands, D.D., Lemoine, F.G., Smith, D.F., ZuM.T., 2001. Crossover analysis of MOLA altimetric data. J. GeophRes. 106, 23,753–23,768.

Parfitt, E.A., Head, J.W., 1993. Buffered and unbuffered dike emplaceon Earth and Venus: implications for magma reservoir size, depth,rate of magma replenishment. Earth Moon Planets 61, 249–281.

Raitala, J., 1988. Composite grabens tectonics of Alba Patera on MEarth Moon Planets 42, 277–291.

Richards, M.A., Duncan, R.A., Courtillot, V.E., 1989. Flood basalt andtrack: plume heads and tails. Science 246, 103–107.

Sandwell, D.T., Schubert, G., 1992. Evidence for retrograde lithosphsubduction on Venus. Science 257, 766–770.

Scott, D.H., Tanaka, K.L., 1986. Geologic map of the western equatregion of Mars. U.S. Geol. Surv. Misc. Invest. Ser., Map I-1802-A, sc1:15,000,000.

Scott, E., Wilson, L., Head, J.W., 2002. Emplacement of giant radial din the northern Tharsis region of Mars. J. Geophys. Res. 107.

Scott, E., Wilson, L., 2003. Did the Alba Patera and Syria Planum regionMars lose their lithospheric roots in convective overturn events? J. Gphys. Res. 108.

Sheth, H.C., 1999. Flood basalts and large igneous provinces frommantle plumes: fact, fiction and fallacy. Tectonophysics 311, 1–29.

Sleep, N.H., 1990. Hotspots and mantle plumes: some phenomenoJ. Geophys. Res. 95, 6715–6736.

Sleep, N.H., Phillips, R.J., 1985. Gravity and lithospheric stress on therestrial planets with emphasis on the Tharsis region of Mars. J. GeoRes. 90, 4469–4489.

Smith, D.E., Zuber, M.T., Frey, H.V., Garvin, J.B., Head, J.W., Muhman, D.O., Pettengill, G.H., Phillips, R.J., Solomon, S.C., Zwally, HBanerdt, W.B.E., Duxbury, T.C., Golombek, M.P., Lemoine, F.G., Nmann, G.A., Rowlands, D.D., Aharonson, O., Ford, P.G., Ivanov, AJohnson, C.L., McGovern, P.J., Abshire, J.B., Afzal, R.S., Sun,

2001. Mars Orbiter Laser Altimeter: experiment summary after the firstyear of global mapping of Mars. J. Geophys. Res. 106, 23,689–23,722.

s 176 (2005) 44–56

.

Stofan, E.R., Bindschadler, D.L., Head, J.W., Parmentier, E.M., 1Corona structures on Venus: models of origin. J. Geophys. Res20,933–20,946.

Stofan, E.R., Sharpton, V.L., Schubert, G., Baer, G., Bindschadler, DJanes, D.M., Squyres, S.W., 1992. Global distribution and charactics of coronae and related features on Venus: implications for oand relation to mantle processes. J. Geophys. Res. 97, 13,347–13

Stofan, E.R., Smrekar, S.E., 2005. Large topographic rises, coronae,flow fields and large volcanoes on Venus: evidence for mantle plumIn: Foulger, G.R., Natland, J.H., Presnall, D.C., Anderson, D.L. (EdPlates, Plumes and Paradigms. Geological Society of America. SpPaper 388. In press.

Tanaka, K.L., 1990. Tectonic history of the Alba Patera–Ceraunius Foregion of Mars. Lunar. Planet. Sci. 20, 515–523.

Tanaka, K.L., Golombek, M.P., Banerdt, W.B., 1991. Reconciliationstress and structural histories of the Tharsis region of Mars. J. GeoRes. 96, 15,617–15,633.

Turcotte, D.L., Schubert, G., 1982. Geodynamics: Applications of Conuum Physics to Geological Problems. Wiley, New York.

Turcotte, D.L., Willemann, R.J., Haxby, W.F., Norberry, J., 1981. Rolemembrane stresses in the support of planetary topography. J. GeoRes. 86, 3951–3959.

Turtle, E.P., Melosh, H.J., 1997. Stress and flexural modeling of the malithopsheric response to Alba Patera. Icarus 126, 197–211.

Twiss, R.J., Moores, E.M., 1992. Structural Geology. Freeman, New YWatters, T.R., Janes, D.M., 1995. Coronae on Venus and Mars: implica

for similar structures on Earth. Geology 23, 200–204.Watts, A.B., 2001. Isostasy and Flexure of the Lithosphere. Cambr

Univ. Press. 472 pp.Wise, D.U., 1976. Faulting and stress trajectories near Alba volcano, N

ern Tharsis ridge of Mars. Geol. Romana 15, 430–433.Wilson, L., Head III, J.W., 2002. Tharsis-radial graben systems as the

face manifestation of plume-related dike intrusion complexes: mo

and implications. J. Geophys. Res. 107.

Zuber, M.T., 2001. The crust and mantle of Mars. Nature 412, 220–227.