q 2003 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or editing@geosociety.org.Geology; February 2003; v. 31; no. 2; p. 175–178; 4 figures. 175

Olympus Mons, Mars: Detection of extensive preaureolevolcanism and implications for initial mantle plume behaviorElizabeth R. FullerJames W. Head III

Department of Geological Sciences, Brown University, Providence, Rhode Island 02906, USA

ABSTRACTThe early volcanic history and structural development of Olympus Mons, Mars, has

been obscured by the formation of the large circum-Olympus aureole deposits in the EarlyAmazonian. Mars Orbiter Laser Altimeter data reveal an enormous preaureole extrusiveflow unit in Amazonis Planitia that is more than 300 km wide and extends ;1800 kmradially away from the Olympus Mons aureole. We interpret this volcanic unit to repre-sent proto–Olympus Mons lava flows emplaced during the latest Hesperian or earliestAmazonian. The orientation and localization of the deposits appear to be due to channelingof the flows into a broad depression between the generally north dipping slope of AmazonisPlanitia and the southwestward-dipping slope of the Alba Patera flanks. Formation of theaureole blocked this depression and caused subsequent summit flows to pond in thecircum-Olympus trough and flow around the aureole lobes. Emplacement of this proto-Olympus flow unit formed a barrier along the northern margin of Amazonis Planitia,causing ponding of later lava flows and outflow channel events debouching there. Thelarge volume and great lateral extent of the unit imply abundant magma supply and higheffusion rates during the initial stages of Olympus Mons construction; this unit may bethe result of the initial impingement of a melt-rich mantle plume head.

Keywords: volcanism, Olympus Mons, Mars, plumes, lava flows.

INTRODUCTIONOlympus Mons, the tallest known volcano

in the solar system, sits on the northeasternflank of the Tharsis rise of Mars. It is sur-rounded by a several hundred kilometer widering of material known as the aureole (out-lined in Fig. 1), which is thought to have beencaused by massive deformation and landslid-ing of the edifice flanks (e.g., Morris and Ta-naka, 1994; Francis and Wadge, 1983). Theearly history of Olympus Mons is poorly un-derstood because the aureole deposits have de-stroyed and obscured evidence of early andflanking eruptions (Scott and Tanaka, 1986).In their analysis of the Olympus Mons region,Morris and Tanaka (1994) showed that thestructure of the aureole deposits is consistentwith gravitationally induced collapse, butpointed out that such a mechanism would notbe expected to bury the entire precollapsestructure. They therefore looked for evidenceof preaureole lava flows extending beyond theaureole, but were unable to find such struc-tures in the Viking images (Morris and Tana-ka, 1994). Previous maps, generated from Vi-king images, show Amazonis and ArcadiaPlanitiae (Fig. 1) to be blanketed by a seriesof region-wide volcanic units: the five-member Arcadia Formation, a volcanic for-mation without an apparent source that per-sisted throughout the Amazonian (Scott andTanaka, 1986).

This study uses data from the Mars OrbiterLaser Altimeter (MOLA) and images from theMars Orbiter Camera (MOC) to examine ev-

idence of preaureole volcanism. MOLA data,and the derived data products, have permittedthe analysis and detection of a significantnumber of geologic features that were onlyhinted at in Viking image data. For example,gradient maps, detrended topography maps,and slope maps generated from MOLA data(e.g., Fig. 2A) have revealed details of the ge-ology and structure of Amazonis Planitia (Ful-ler and Head, 2002) and shown that it differsfrom that originally mapped by Scott and Ta-naka (1986). Among the new units detectedwas a very large flow unit extending north-west across Amazonis Planitia and into Ar-cadia Planitia from the margins of the Olym-pus Mons aureole (Fig. 1). Here we documentthe nature and detailed stratigraphic relation-ships of this unit and examine evidence for itsprovenance and age.

OBSERVATIONSGeneral Characteristics

Stratigraphically, the lobate flow unit (Figs.1 and 2) overlies the Upper Hesperian VastitasBorealis Formation (Scott and Tanaka, 1986),a unit laterally correlative with the outflowchannels and thought to represent a potentiallyice-rich sublimation residue remaining fromthe outflow events (e.g., Kreslavsky and Head,2002). The Vastitas Borealis Formation is;100 m thick and overlies Lower Hesperianridged plains of volcanic origin (Head et al.,2002); the ridges on the surface of this unitcan be seen to protrude through the VastitasBorealis Formation, but are largely buried bythe lobate flow unit (Fig. 2). The eastern con-

tact of the flow unit is the sharp, scarp-likewestern boundary of the topographically high-er Early Amazonian aureole deposits, sug-gesting that the aureole unit was emplaced ontop of the flow unit, and is thus younger.These general relationships place the age ofthe flow unit as latest Hesperian or earliestAmazonian, older than any mapped depositsfrom Olympus Mons.

MorphologyThe preaureole Olympus Mons flow unit is

extremely long (;1780 km) and wide (310–420 km), with a maximum thickness of ;100m; as exposed, it is the longest flow unit yetdetected on Mars, longer than any identifiedin the flow catalog of Peitersen et al. (2002).The flow unit is on a very shallow slope(;0.00388); the substrate decreases in eleva-tion only 120 m over the unit’s 1800 km ex-tent. It is approximately constant in width,suggesting preferential forward advance in-stead of lateral spreading. The regional topog-raphy appears to be channeling the flow, in-hibiting lateral movement in favor of forwardflow. The nearly constant width along the flowunit is common for long (.50 km) flows, ac-cording to a study of 145 martian and terres-trial lava flows (Peitersen et al., 2002).

The surface of the flow unit shows a rough,rubbly texture (Fig. 3A), occasionally inter-rupted by round depressions meters to tens ofmeters across (e.g., Fig. 3B). These depres-sions are not seen in MOC images to the northand south, suggesting that they are unique tothis flow unit. Possible origins include impactcraters, collapse depressions, or deflation pits.Global MOLA roughness data reveal the pres-ence of a very young, latitudinally dependent,fine-grained mantling deposit overlying muchof the flow unit (Kreslavsky and Head, 2000);MOC images show extensive eolian modifi-cation of the surface. These data indicate thatthe current surface has been altered in the ;3b.y. since its emplacement, explaining why theflow unit was not recognized in Viking im-ages, which show only the extensively modi-fied surface.

Flow units on Mars can be formed by lavaflows, lahars (e.g., Christiansen, 1989), or acombination of both (e.g., Tanaka et al., 1992;Russell and Head, 2001). Christiansen (1989),using Viking images, mapped lahars whoseoverall morphology is broadly similar to thepreaureole flow unit. Russell and Head (2001),

176 GEOLOGY, February 2003

Figure 1. Location of newly mapped flow unit and nearby volcanic provinces,overlaid on topography and gradient maps generated from Mars Orbiter LaserAltimeter. Scarp around upper portion of Olympus Mons edifice and OlympusMons aureole deposits are outlined. Note that scales above and below 0 m(martian datum) are different. Lambert conformable conic projection, 90–2008W,0–708N.

Figure 2. A: Regional topography overlaid on gradient map, generated from Mars Orbiter Laser Altimeter data. B: Geologic sketch mapoverlaid on (A). Note smooth, (low-relief) plains surrounding flow unit, sharp northern contact of flow unit and very subtle southern contact,very smooth surface of Amazonis Planitia, and pervasive wrinkle ridges texture. AF—Acheron Fossae; OMA—Olympus Mons aureole.Arrow points to crater shown in Figure 3D.

using MOLA data and MOC images, have ex-amined the region and interpreted the broad,lobate structures as lava flows and the narrow-er, higher-relief units as lahars. The broad, lo-bate morphology of the preaureole flow unittherefore appears more consistent with a lava-based than a lahar-based origin. MOLA datafurther suggest that the preaureole flow unit iscomposed of a series of discrete lava flows.The data reveal a few discernable flow mar-gins on the surface of the unit (mapped in Fig.2). Figure 3C, MOC images taken across acrater rim, reveals laterally extensive layers inthe crater walls. These could be a result ofejecta layering, but their cohesiveness and dif-ferential weathering patterns are more sugges-

tive of solid rock units. Their overall mor-phology is similar to layers seen in MOCimages elsewhere and interpreted to be lavabeds (e.g., McEwen et al., 1999). They alsoresemble exposed cliffs within the ColumbiaRiver Basalt Group. It should be noted, how-ever, that apparent layers within the ColumbiaRiver Basalt Group are often entablature tiers,i.e., laterally persistent patterns of coolingjoints (e.g., McMillan et al., 1989); the layersvisible in Figure 3C may similarly representcooling effects instead of a series of discreteflow events.

The southern portion of the flow unit formstwo rectilinear depressions in northeasternAmazonis Planitia (centered at 328N, 1508W).

The western margin of these basins is approx-imately parallel to the wrinkle ridges thatsweep across this region (mapped in Fig. 2;see also discussion in Head et al., 2002), sug-gesting that the southernmost flows abutted awrinkle ridge and were deflected to form theenclosed basins. These southernmost flowsmay be the oldest in the flow unit; the upper-most flows (mapped individually in the south-west corner of Fig. 2) appear to curve aroundthese basins.

The flow unit is surrounded by closely as-sociated smooth plains (Fig. 2) that also over-lie the Vastitas Borealis Formation. MOC im-ages show several features interpreted to berelated to ice or water. At the margins of theflow unit, subparallel meandering channelscarve through the substrate (e.g., Fig. 3D), andare interpreted to represent melting and mo-bilization of ground ice. The meandering plan-form of the channels and the extensive eolianreworking suggest that the smooth plains areeasily eroded. We interpret these deposits tobe related to the emplacement of the flow unit,most likely resulting from melting of the un-derlying ice-rich sediments of the Vastitas Bo-realis Formation during emplacement of thesuperposed flow unit, causing melting anddrainage out along the flow margins.

ProvenanceThe source and proximal extent of the flow

unit are buried by the Olympus Mons aureole.There are two nearby major volcanic edifices,Alba Patera and Olympus Mons (Fig. 1). Theflow unit extends approximately radially toOlympus Mons. If it originated at the Olym-pus Mons summit, it would be on the orderof 2500 km long. If it originated from theAlba Patera summit, it would be nearly 3500km long.

Prior to the emplacement of this unit, the

GEOLOGY, February 2003 177

Figure 3. Mars OrbiterCamera images. A: Sur-face of preaureole flowunit. M07-03360, 148.948W,36.778N. B: Surface of flowunit; black arrows indicatecircular depressions. M15-01063, 163.608W, 39.468N.C: Crater rim showing lay-ers that may represent in-dividual lava-flow events.M18-01140, 164.068W,42.858N. D: Channels flow-ing away from flow unit.M09-03601, 172.678W,51.808N. Images areavailable courtesy ofMalin Space ScienceSystems.

regional topography was dominated by the re-gional slope, grading downward to the northfrom the martian southern highlands towardthe northern lowlands (Fig. 1). The flanks ofthe Alba Patera edifice formed a slope upwardto the northeast, and the flanks of the pre-aureole proto–Olympus Mons volcanic edifice(Morris and Tanaka, 1994) would have formeda similarly steep slope to the southeast. Thistopography would have preferentially chan-neled flows extending north from OlympusMons or southwest from Alba Patera along arelatively narrow trough to the northwest. Aflow extending south or southwest from AlbaPatera would first have banked against thepreaureole Olympus Mons volcanic edifice orthe Noachian (i.e., pre-Hesperian) AcheronFossae (Fig. 1). Reconstructing the likely ex-tent of this preaureole edifice indicates thatlava flowing from Alba Patera would have hadto climb over the flanks of Olympus Mons toflow into Amazonis Planitia. In addition, thekilometer-scale roughness of the flow unit

(Kreslavsky and Head, 2000) supports an or-igin at Olympus Mons (Fuller and Head,2002): as discussed in previous papers in moredetail, lava flows from a single source typi-cally have internally consistent surface rough-ness. There are no other preaureole OlympusMons lava flows to which this flow unit canbe compared, but comparison with flows fromAlba Patera and the Tharsis Montes showsthat those flow surfaces are significantlyrougher (at 2.4 km base-length roughness)than this preaureole flow unit. On the basis ofthese observations, we conclude that the mostlikely provenance of this flow unit is thepreaureole Olympus Mons.

Stratigraphic ImplicationsThe flow unit and its surrounding smooth

plains overlie the Lower Hesperian ridgedplains and the Upper Hesperian Vastitas Bo-realis Formation, but are truncated by theLower Amazonian Olympus Mons aureole(Fig. 4A). MOLA data show that the northern

margin of the flow unit stands out sharplyfrom surrounding topography, while thesouthern margin is less distinct (Figs. 2A and4B). This is consistent with the interpretationthat the southern margin was repeatedly em-bayed and mantled by younger volcanic andoutflow events in Amazonis Planitia (Fullerand Head, 2002). MOLA data reveal that theflow unit mapped here crosses the contacts ofall five members of the Arcadia Formation,mapped on the basis of Viking data by Scottand Tanaka (1986) and previously thought tohave been emplaced throughout the Amazo-nian. Thus, the Arcadia Formation is in needof redefinition in future studies.

DISCUSSIONSeveral researchers have proposed a

gravity-spreading origin for the OlympusMons aureoles (e.g., Francis and Wadge,1983; Morris and Tanaka, 1994). Morris andTanaka (1994) argued that ice-lubricated grav-itational collapse implies the presence of an

178 GEOLOGY, February 2003

Figure 4. A: Stratigraphic column (afterScott and Tanaka, 1986); approximate timingof preaureole proto–Olympus Mons flow isindicated by arrow. Amazonian is youngestmartian era; time of Hesperian-Amazonianboundary is ca. 3.5 Ga. B: Interpretive crosssection. Surface is generated from Mars Or-biter Laser Altimeter gridded topography at1558W; other units are shown with approxi-mate thicknesses (to scale). Elevation val-ues are referenced to martian datum. Hr—Lower Hesperian ridged plains. VE—Verticalexaggeration.

existing large volcanic edifice, evidence ofwhich would be lava flows still visible beyondthe margins of the aureole; they considered thefact that they had not found any in Vikingimages to be a weakness of the model. Thisstudy therefore strengthens their interpretationand supports their primary conclusion, thecollapse of preexisting edifice margins as anorigin for the aureole.

The very large volume (50,000–75,000km3) of the exposed flow unit is approximate-ly half the volume of the Columbia River Ba-salt Group (Tolan et al., 1989), and this is aminimum estimate, because the flow unit like-ly originated closer to the Olympus Monssummit region. The lack of such extensiveflow units later in the history of OlympusMons suggests that the early stages of eruptiveactivity were characterized by a greater mag-ma supply rate and potentially higher effusionrates than later (Wilson et al., 2001). This be-havior could be a function of source mantleplume geometry, where early massive flowswere fed by a plume head and subsequentlower flow volumes were due to the diminish-ing volume of the plume tail. This dichotomyin eruptive styles between plume head andplume tail is well documented for terrestrialplumes (e.g., Ernst and Buchan, 2001). Of ad-ditional significance is the relationship be-tween plume head size and initial eruptionvolume. The size of the plume head is pri-marily a function of the depth at which theplume originated (Campbell, 2001); the enor-mous volumes of lava erupted here suggestthat a source plume may have originated deepin the interior, perhaps at the martian core-mantle boundary.

ACKNOWLEDGMENTSWe thank Patrick McGovern, Karl Mitchell, Devon Burr,

and James Zimbelman for helpful discussions. We grate-fully acknowledge the use of Mars Orbiter Laser Altimeterdata, available at http://ltpwww.gsfc.nasa.gov/tharsis/mola.html, and Mars Orbiter Camera images, processed byMalin Space Science Systems, available at http://www.msss.com/mocpgallery/. We thank Patrick McGovernand Richard Ernst for their thoughtful reviews and helpfulcomments.

REFERENCES CITEDCampbell, I.H., 2001, Identification of ancient mantle

plumes, in Ernst, R.E., and Buchan, K.L., eds.,Mantle plumes: Their identification through time:Geological Society of America Special Paper 352,p. 5–21.

Christiansen, E.H., 1989, Lahars in the Elysium regionof Mars: Geology, v. 17, p. 203–206.

Ernst, R.E., and Buchan, K.L., 2001, The use of maficdike swarms in identifying and locating mantleplumes, in Ernst, R.E., and Buchan, K.L., eds.,Mantle plumes: Their identification through time:Geological Society of America Special Paper 352,p. 247–265.

Francis, P.W., and Wadge, G., 1983, The OlympusMons aureole: Formation by gravitational spread-ing: Journal of Geophysical Research, v. 88,p. 8333–8344.

Fuller, E.R., and Head, J.W., III, 2002, Amazonis Plan-itia: The role of geologically recent volcanismand sedimentation in the formation of the smooth-est plains on Mars: Journal of Geophysical Re-search, v. 107, 2002JE001842 (in press).

Head, J.W., III, Kreslavsky, M.A., and Pratt, S., 2002,Northern lowlands of Mars: Evidence for wide-spread volcanic flooding and tectonic deformationin the Hesperian Period: Journal of GeophysicalResearch, v. 107, 10.1029/2000JE001445.

Kreslavsky, M.A., and Head, J.W., III, 2000, Kilometer-scale roughness of Mars: Results from MOLAdata analysis: Journal of Geophysical Research,v. 105, p. 26,695–26,711.

Kreslavsky, M.A., and Head, J.W., III, 2002, The fateof outflow channel effluents in the northern low-lands of Mars: The Vastitas Borealis Formationas a sublimation residue from frozen ponded bod-ies of water: Journal of Geophysical Research, v.107, 2001JE001831 (in press).

McEwen, A.S., Malin, M.C., Carr, M.H., and Hart-mann, W.K., 1999, Voluminous volcanism on ear-ly Mars revealed in Valles Marineris: Nature,v. 397, p. 584–586.

McMillan, K., Long, P.E., and Cross, R.W., 1989, Ve-siculation in Columbia River basalts, in Reidel,S.P., and Hooper, P.R., eds., Volcanism and tec-tonism in the Columbia River flood-basalt prov-ince: Geological Society of America Special Pa-per 239, p. 157–167.

Morris, E.C., and Tanaka, K.L., 1994, Geologic mapsof the Olympus Mons region of Mars: U.S. Geo-logical Survey Miscellaneous Investigations MapI-2327, scale 1:2,000,000.

Peitersen, M.N., Zimbelman, J.R., and Bare, C., 2002,Analysis of geomorphometric properties of mar-tian and terrestrial long lava flows: Lunar andPlanetary Science Conference XXXIII (CD-ROM), abstract 1026.

Russell, P.S., and Head, J.W., III, 2001, The Elysium/Utopia flows: Characteristics from topographyand a model of emplacement: Lunar and Plane-tary Science Conference XXXII (CD-ROM), ab-stract 1040.

Scott, D.H., and Tanaka, K.L., 1986, Geologic map ofthe western equatorial region of Mars: U.S. Geo-logical Survey Miscellaneous Investigations Se-ries Map I-1802-A, scale 1:15,000,000.

Tanaka, K.L., Chapman, M.G., and Scott, D.H., 1992,Geologic map of the Elysium region of Mars:U.S. Geological Survey Atlas of Mars Series MapI-2147, scale 1:5,000,000.

Tolan, T.L., Reidel, S.P., Beeson, M.H., Anderson, J.L.,Fecht, K.R., and Swanson, D.A., 1989, Revisionsto the estimates of the areal extent and volume ofthe Columbia River Basalt Group, in Reidel, S.P.,and Hooper, P.R., eds., Volcanism and tectonismin the Columbia River flood-basalt province:Geological Society of America Special Paper 239,p. 1–120.

Wilson, L., Scott, E.D., and Head, J.W., 2001, Evidencefor episodicity in the magma supply to largeTharsis volcanoes: Journal of Geophysical Re-search, v. 105, p. 1423–1433.

Manuscript received 12 June 2002Revised manuscript received 10 October 2002Manuscript accepted 12 October 2002

Printed in USA