VOL. 84, NO. BI4 JOURNAL OF GEOPHYSICAL RESEARCH DECEMBER 30, 1979 The Subglacial Birth of Olympus Mons and Its Aureoles CARROLL ANN HODGES AND HENRY J. MOORE U.S. GeologicalSurvey, Menlo Park, California 94025 The vast volcanic plains and shieldsof Mars, togetherwith the thermal, spectrographic,and morpho- logical evidencefor water ice at the poles,for severalpercent water in Viking soil samples,for ground ice or permafrostover much of the planet, and for the existenceof surfacewater at some time in the past, suggestthat magma and water or ice may have interactedduring evolutionof the planet'slandscape. Rel- atively small mesas and buttes, with and without summit craters, are remarkably similar to the table mountainsof Iceland that formed by subglacialeruption during the late Quaternary period. Table moun- tains typicallycomprise foundations of pillow lava and palagonitizedtuff breccias(m6berg), overlain by subaeriallava flows that commonly culminated in a typical shield volcano;some table mountains,how- ever, failed to reachthe subaerialstage and thus lack the cap rock. Subglacialfissure eruptions produced ridgescomposed of pillow lava and m6berg. Conical knobs on steep-sidedMartian plateausare reminis- cent of the small Icelandic shield volcanoesatop m6berg pedestals.Candidate table mountains on Mars are especiallynumerous in the region betweenlatitude 40øN and the margin of the north polar cap, and interaction of lava with a formerly more extensiveice cap may have occurred.More significantis the pos- sibility that Olympus Mons and the broad lobate aureole depositsaround its base may have had similar subglacialbeginnings. This hypothesisrequires an ice cap severalkilometers thick in the vicinity of this enormous shield during its initial stagesof eruption. The amounts of water ice required do not appear excessive,given the limits to presentknowledge of the water budget throughout the planet's history. A mechanismfor localizing suchice, however, is required. INTRODUCTION [1960, 1966a,b, 1967], Sigvaldason[1968], and Preusser[1976]. The morphologiesof numerous small landforms on Mars These distinctivelandforms may be roughly circular but more are not readily interpretable in terms of conventionalimpact, commonly are subrectangular,with abrupt edges and steep flanks that slopeas much as 35ø or more (Figures 5a and 5b). volcanic, or erosionalprocesses. Such featuresinclude circular to rectangularmesas and buttes, with and without summit The top is flat or gently convex, with or without a crater at the craters(Figures 1 and 2), and 'two-story'structures consisting summit, and is composedof aa and pahoehoe flows which in of conicalor irregularknobs atop broad, flat, steep-sidedped- most cases developed into a small subaerial shield volcano estals(Figure 2a). Somelandforms exhibit morphologies con- [Kjartansson,1960]. Underlying the subaerialflows is a pedes- sistentwith either volcanic or impact origins, so their genetic tal-like foundation of 'm6berg,' the Icelandic term for pala- classificationis not obvious.These types of features are espe- gonitized tuffs and breccias[Kjartansson, 1960]. The m6berg cially abundant in the 'northern plains,' betweenabout lati- mass in turn overlies and is partly intercalated with a basal tude 40øN and the margin of the polar ice cap. Features that core of pillow lava. Some table mountains lack the subaerial are clearly volcanicalso occurin parts of this region, specifi- basalt cap (Figure 6). cally, north of the Tharsis and Elysium volcanicprovinces. The pillow lava and m6berg pedestalsindicate that water Well-defined lava flows on the flanks of Alba Patera are inter- was abundant at the site of extrusion. The prevailing genetic mingled with conical cratered domes a few kilometers across hypothesisis diagramedLn Figure 7 [Einarsson,1968]. Basaltic that resemble small volcanic cones; clusters of such features magma was extruded subglacially into water-filled caverns have been noted elsewhereas well (Figure 2b). melted into or through a late Quaternary ice sheet.The initial If terrestrialanalogs are to be found for theseMartian land- extrusions of pillow lava were emplaced beneath meltwater forms,they are mostprofitably sought where geologic and cli- and ice. As the pile of pillow lava grew upward, confining matic conditions are similar to past or present environments pressuresof the overlying meltwater and ice decreased,and on Mars; Iceland, with its combination of active volcanism the consequentincreased vesiculation of the magma, aug- and glaciation,is a promisinglocale to explore [Allen, 1977, mented by influx of water into the vent, initiated phreatomag- 1978, 1979;Hodges, 1977; Hodges and Moore, 1978a, b; Wil- matic explosions [Jones, 1970]. Sideromelane tuff breccias, liams, 1978].The table mountainsof Iceland (Figures3a and subsequentlyaltered to palagonite,were then depositedatop 3b), whichformed as a resultof subglacialvolcanic extrusion, the pillow lava. If the accumulatingpile reachedthe surfaceof are our major topic of interest here, augmentedby reference the meltwater lake, subaerialflows were emplaced,commonly to oceanic shield volcanoes such as Mauna Loa, which have capped by a shield volcano. As the subaerial flows advanced subaqueouseruptive historiesparallel with the subglacial into the meltwater at the margin of the pile, explosive dis- stagesof the Icelandic table mountains[Kjartansson, 1966a; ruption occurred,and tuff brecciaswere depositedas steeply Moore and Fiske, 1969]. dipping sediments,much like the foreset beds of a delta, but in this caseoverridden by topset beds of lava [Jones,1969]. ICELAND'S TABLE MOUNTAINS With subsequentwithdrawal of the ice, the subaerial flows The relatively small, isolated plateaustermed table moun- formed a resistantcap overlying the less resistantm6berg tains (Figure 4) have been describedin detail by Van Bernrne- foundation. The thicknessof the ice sheet, indicated by the len and Rutten [1955], Jones [1966, 1969, 1970], Kjartansson height of the m6berg pedestals,was apparently 500-1000 m during the Pleistocenestage of table mountain development. Copyright ¸ 1979 by the American GeophysicalUnion. Crateredhills or mesasof m6bergthat are not cappedby lava Paper number 9B 1125. 8061 0148-0227/79/009B- 1125501.00 8062 HODGES AND MOORE: SECOND MARS COLLOQUIUM a b Fig. 1. The small circular plateau with summit crater (arrow) is one of the best candidates for a Martian table mountain. It has a mor- phology and dimensions that are similar to the table mountains in c d Iceland which are formed by subglacial volcanic eruption (compare with Figure 3). Terraces and ridgesoccur near the base.Smaller, simi- lar featuresare nearby. Horizontal and approximate vertical scalesof --- 6 km profile are same scaleas photograph.(Height estimatedfrom shadow length.) Sun at left. (Latitude 49øN, longitude 225ø; Viking photo- graph 009B 15.) Fig. 2a. Rectangular to subcircular Martian mesas that resemble Icelandic table mountains with (arrow a) and without (arrow b) obvi- ous summit craters.Two-story structureand typical impact craters in- (Figure 6) and m6berg ridges (formed by subglacialfissure dicated by arrows c and d. Diagrammatic profiles drawn at ap- eruptions) presumably representextrusions that did not reach proximately twice the scale of photograph;no vertical exaggeration. (Heights estimatedfrom shadowlengths.) Sun at left. (Latitude 45øN, the surfaceof the ice or meltwater [Kjartansson,1960]. longitude 20ø; Viking photograph026A30.) These Icelandic features are relatively small, generally a few kilometers across and a few hundred meters high; the highest,Herdubreid (Figure 5), has over 1000 m of total relief [Van Bemmelenand Rutten, 1955]. The contrast in morphol- ogy between strictly subaerial shieldsand table mountains in . Iceland [Jones, 1969] is readily apparent on topographic maps of the two types of structures(Figure 8). The tuyas of British Columbia also have been interpreted as ..i•' ;•;•i;..;:•::::::::::::::::::::::::::::::::::::::: productsof subglacialPleistocene eruptions [Mathews, 1948], ... ß:'*:i.'*" "'"'::• as have similar volcanic rock successionsin Antarctica [Le Masurier, 1972] and Alaska [Hoare and Coanrad, 1978]. Anal- ...:. yses of subglacial shield-type eruptions and resulting land- "':.•i""•*' *<":' n..... ":'-};:. forms are applicable to the submarine development of larger oceanic shield volcanoes [Kjartansson, 1966a, 1967; Jones, 1966; Moore and Fiske, 1969]. OCEANIC SHIELD VOLCANOES Moore and Fiske [1969] documented a threefold structure analogousto that of the table mountains in the shield volca- noes on the island of Hawaii (Figure 9): (1) pillow lava and fragmentserupted from deepwatervents, overlain by (2) vitric explosiondebris, littoral cone ash and breccia erupted from shallow water vents or at the interface of subaerial lava and water, and (3) the subaerial shield that forms the super- structure of the volcano. As in the table mountains, the con- tact between the'subaerial shield and the submarine pedestal is marked by a distinct break in slope (becauseof the suscepti- bility to masswasting and erosionof the brecciasand pillow Fig. 2b. Small,possibly volcanic, cratered domes (arrow). Dia- lava), but the lower flank is not as steepas in the subglacial grammaticprofile drawn at scalelarger than that of photograph;no case. Kjartansson [1966a] inferred that the same stages and verticalexaggeration. (Height estimated from shadowlength.) Sun at productsof eruption characterizeddevelopment of the Surtsey left. (Latitude40øN,