VOL. 84, NO. BI4
JOURNAL OF GEOPHYSICAL RESEARCH
DECEMBER 30, 1979
The SubglacialBirth of Olympus Mons and Its Aureoles
CARROLL
ANN
HODGES
AND
HENRY
J. MOORE
U.S. GeologicalSurvey,Menlo Park, California 94025
The vast volcanicplains and shieldsof Mars, togetherwith the thermal, spectrographic,and morphological evidencefor water ice at the poles,for severalpercentwater 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.Relatively small mesasand buttes, with and without summit craters, are remarkably similar to the table
mountainsof Iceland that formed by subglacialeruption during the late Quaternary period. Table mountainstypicallycomprisefoundationsof pillow lava and palagonitizedtuff breccias(m6berg),overlainby
subaeriallava flows that commonlyculminated in a typical shield volcano;sometable mountains,however, failed to reachthe subaerialstageand thuslack the cap rock. Subglacialfissureeruptionsproduced
ridgescomposedof pillow lava and m6berg. Conical knobson steep-sidedMartian plateausare reminiscent of the small Icelandic shield volcanoesatop m6berg pedestals.Candidate table mountainson Mars
are especiallynumerousin 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 possibility that Olympus Mons and the broad lobate aureole depositsaround its basemay have had similar
subglacialbeginnings.This hypothesisrequiresan ice cap severalkilometersthick in the vicinity of this
enormousshield during its initial stagesof eruption. The amounts of water ice required do not appear
excessive,given the limits to presentknowledgeof the water budget throughoutthe planet's history. A
mechanismfor localizing suchice, however,is required.
INTRODUCTION
The morphologiesof numeroussmall landforms on Mars
are not readily interpretablein terms of conventionalimpact,
volcanic,or erosionalprocesses.Such featuresinclude circular
to rectangularmesas and buttes, with and without summit
craters(Figures1 and 2), and 'two-story'structures
consisting
of conicalor irregularknobsatop broad, flat, steep-sidedpedestals(Figure2a). Somelandformsexhibitmorphologies
consistentwith either volcanic or impact origins,so their genetic
classificationis not obvious.These types of featuresare especially abundantin the 'northern plains,' betweenabout latitude 40øN and the margin of the polar ice cap. Features that
are clearly volcanicalso occurin parts of this region, specifi-
cally, north of the Tharsisand Elysium volcanicprovinces.
Well-defined
lava flows on the flanks of Alba Patera are inter-
mingled with conical cratered domes a few kilometersacross
that resemble small volcanic cones; clusters of such features
have beennoted elsewhereas well (Figure 2b).
If terrestrialanalogsare to be found for theseMartian landforms,they are mostprofitablysoughtwheregeologicand climatic conditions are similar to past or present environments
on Mars; Iceland, with its combination of active volcanism
and glaciation,is a promisinglocale to explore [Allen, 1977,
1978, 1979;Hodges, 1977; Hodgesand Moore, 1978a, b; Williams, 1978].The table mountainsof Iceland (Figures3a and
[1960, 1966a,b, 1967],Sigvaldason[1968], and Preusser[1976].
These distinctivelandforms may be roughly circular but more
commonly are subrectangular,with abrupt edges and steep
flanks that slopeas much as 35ø or more (Figures 5a and 5b).
The top is flat or gently convex, with or without a crater at the
summit, and is composedof aa and pahoehoe flows which in
most cases developed into a small subaerial shield volcano
[Kjartansson,1960].Underlying the subaerialflowsis a pedestal-like foundation of 'm6berg,' the Icelandic term for palagonitized tuffs and breccias[Kjartansson,1960]. The m6berg
mass in turn overlies and is partly intercalated with a basal
core of pillow lava. Some table mountains lack the subaerial
basalt cap (Figure 6).
The pillow lava and m6berg pedestalsindicate that water
was abundant at the site of extrusion. The prevailing genetic
hypothesisis diagramedLnFigure 7 [Einarsson,1968].Basaltic
magma was extruded subglacially into water-filled caverns
melted into or through a late Quaternary ice sheet.The initial
extrusionsof pillow lava were emplaced beneath meltwater
and ice. As the pile of pillow lava grew upward, confining
pressuresof the overlying meltwater and ice decreased,and
the consequentincreasedvesiculationof the magma, augmentedby influx of water into the vent, initiated phreatomagmatic explosions [Jones, 1970]. Sideromelane tuff breccias,
subsequentlyaltered to palagonite,were then depositedatop
pile reachedthe surfaceof
3b),whichformedasa resultof subglacial
volcanicextrusion, the pillow lava. If the accumulating
are our major topic of interesthere, augmentedby reference the meltwater lake, subaerialflowswere emplaced,commonly
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 disstagesof the Icelandic table mountains[Kjartansson,1966a; ruption occurred,and tuff brecciaswere depositedas steeply
dipping sediments,much like the foreset beds of a delta, but
Moore and Fiske, 1969].
in this caseoverriddenby topsetbeds of lava [Jones,1969].
ICELAND'S TABLE MOUNTAINS
With subsequentwithdrawal of the ice, the subaerial flows
The relatively small, isolatedplateaustermed table moun- formed a resistantcap overlying the less resistantm6berg
tains (Figure 4) have beendescribedin 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 Pleistocenestageof table mountain development.
Crateredhills or mesasof m6bergthat are not cappedby lava
Copyright¸ 1979 by the American GeophysicalUnion.
to oceanic shield volcanoes such as Mauna Loa, which have
Paper number 9B 1125.
0148-0227/79/009B-
1125501.00
8061
8062
HODGES AND MOORE: SECOND MARS COLLOQUIUM
a
b
c
d
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
Iceland which are formed by subglacial volcanic eruption (compare
with Figure 3). Terracesand ridgesoccur near the base.Smaller, similar featuresare nearby. Horizontal and approximate vertical scalesof
profile are same scaleas photograph.(Height estimatedfrom shadow
length.) Sun at left. (Latitude 49øN, longitude 225ø; Viking photograph 009B 15.)
--- 6 km
Fig. 2a. Rectangular to subcircular Martian mesasthat resemble
Icelandic table mountainswith (arrow a) and without (arrow b) obvious 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 aperuptions) presumably representextrusionsthat did not reach proximately twice the scaleof 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 1000m of total relief
[Van Bemmelenand Rutten, 1955]. The contrast in morphology between strictly subaerial shieldsand table mountains in
Iceland [Jones,1969] is readily apparent on topographicmaps
of the two types of structures(Figure 8).
The tuyas of British Columbia also have been interpreted as
productsof subglacialPleistoceneeruptions [Mathews, 1948],
as have similar volcanic rock successionsin Antarctica [Le
Masurier, 1972] and Alaska [Hoare and Coanrad, 1978]. Analyses of subglacial shield-type eruptions and resulting landforms are applicable to the submarine development of larger
oceanic shield volcanoes [Kjartansson, 1966a, 1967; Jones,
1966; Moore and Fiske, 1969].
OCEANIC
SHIELD
.
..i
•'
:'*:i.'*"
ß
...
...:.
.
;•;•i;..;:•::
:::::::::::::::::::::::::::::::::::::
"'"'::•
"':.•i
""•
*' *<":'
n
.....":'-};:.
VOLCANOES
Moore and Fiske [1969] documented a threefold structure
analogousto that of the table mountains in the shield volcanoes 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 superstructure 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 susceptibility to masswastingand erosionof the brecciasand pillow
Fig. 2b. Small,possiblyvolcanic,cratereddomes(arrow).Dialava), but the lower flank is not as steepas in the subglacial grammaticprofiledrawnat scalelargerthan that of photograph;
no
case. Kjartansson [1966a] inferred that the same stages and verticalexaggeration.
(Heightestimatedfrom shadowlength.)Sunat
035A62.)
productsof eruption characterizeddevelopmentof the Surtsey left. (Latitude40øN,longitude15ø;Vikingphotograph
HODGESAND MOORE:SECONDMARS COLLOQUIUM
8063
Fig. 3a. Landsatview of three table mountainsin northern Iceland:(a) Baejarfjall;(b) Kviholafj611;
(c) Gaesafj611.
GaesaO611
(c) has
a capof subaeriallava;the othertwo tablemountains
apparentlyconsist entirely of m6berg. All three have summit craters. Compare a
with probableMartian counterpartin Figure 1. North at top; sun at
right.(Landsatimage2094-11403,band7.)
Island shield,off the southcoastof Iceland, but recent drilling
showedno evidenceof pillow lava. The subaqueous
sectionat
Fig. 4. Aerial photographof GaesaQ611
with summitcrater (arSurtseyconsistsentirely of tuffs,down to the preexistingsea
row a) and cap of subaerialflows(b) that are superposed
on m6berg
floor at about 180 m (J. G. Moore, personalcommunication, exposedon steepslopes(c). Postglacialflowsindicatedby d. North at
1979).
LANDFORMS
ON THE NORTHERN
PLAINS OF MARS
top; sunat lowerright. Profileat approximatelysamescaleas photograph; no verticalexaggeration[Van Bemmelenand Rutten, 1955,
Plate XIX].
A preliminarysurveyof Viking photographshas yielded the heightsmentionedhere and shownin the illustrationpro-
numerousexamplesof landforms that resemblerather strikingly the Icelandictable mountainsand m6berg hills, or possiblevariantsthereof(Figure 10). Thesestructures,like those
in Iceland, are a few kilometers acrossand of the order of a
few hundredmetershigh. (We mustemphasize,however,that
files were estimated on the basis of shadow lengths and are
thus subjectto revision;both sun angleand image-processing
techniquesaffect suchestimates.)The featuresare commonly
associated with other landforms more convincingly inter-
preted as volcanic. For example, the isolated flat-topped
mesasof Figure 1la are adjacentto two crateredconesthat
resembleterrestrial subaerialvolcanoes(Figures 11b and 11c).
In someplains areas (notably, westernElysium Planitia and
southeasternAcidalium Planitia), domesand coneswith summit craters are numerous;their morphology is different from
that of typical impact craters of similar size that appear
equallyfresh(Figure2b). The configuration
of the plainssurface in the Elysium area is irregular and pitted (Figure 12),
much like the surfaces of terrestrial
lava flows.
The geomorphicevidencethat volcanismoccurredlocally
in the northern plains is persuasive;we proposethat some of
the unusualmorphologiesmay have resultedfrom interaction
-,,,,-2,5'km
of magmawithicein a process
analogous
to thatof thesubglacialeruptionsthat formedtable mountainsin Iceland.This
explanationwould require surfaceice severalhundred meters
thick to have existed beyond the present north polar cap of
Mars, a condition that could perhapshave occurredwith rela-
tively slightmodificationsin temperatureor pressureand/or
Fig. 3b. Verticalaerialphotograph
of tablemountainsa and b water content of the Martian atmosphere.A comprehensive
shown
in Figure
3a,withthirdventandm6berg
lobevisible
south-analysis
(in progress)
of thespatial
distribution
of thesekinds
westof juncture.
Northat top;sunat lowerleft.E-W profileacross
circular
table
mountain
atapproximately
same
scale
asphotograph;
offeatures
isnecessary
inorder
todetermine
therelation,
if
novertical
exaggeration.
(U.S.AirForce
photograph
AF-55-AM-3,
any,between
theiroccurrence
andlatitude
ontheplanet.
Our
roll 121,frame13394.)
preliminaryresultsare consistent
with thoseof Allen [1979].
8064
HODGES AND MOORE: SECOND MAR•-•COLLOQUIUM
Fig. 5a. Herdubreid, highestof the Icelandic table mountainswith over 1000m of relief, as viewed from the northeast.
Snow-coveredshield volcano is superposedon subaerial flows at top of steepcliff; intercalated palagonite tuff breccia and
columnar basalt form the pedestalfoundation. (Photographby C. A. Wood.)
OLYMPUS
MONS
AND
ASSOCIATED
FEATURES
The enormousshield volcano Olympus Mons, 26 km high
and 600 km acrossat its base, may also have had subglacial
beginnings. The shield is at an elevation of 2-3 km on the
northwest flank of the 'Tharsis bulge,' a northeast trending
topographicridge 10 km above Mars datum [U.S. Geological
Survey, 1976]. The bulge is surmounted by a chain of three
shield volcanoes, each of which is somewhat smaller than
Olympus Mens, but like the larger shield, all attain elevations
of 26 km. OlympusMens is distinctivenot only for its size but
also for the prominent escarpment,up to 6 km high [Blasius,
1976], that rings its base. In addition, it is nearly surrounded
by peculiar grooved terrain informally termed aureole deposits [Carr, 1973], which appear to consistof a sequentialseries
O'
•3o0
/-/e rdo
of superposedarcuate lobes,differing in degreeof degradation
and burial. The distal margin of one of the oldest lobes is
about 600-800 km north and west of Olympus Mens; younger
lobes are closerto the shield (Figures 13a and 13b). A gravity
anomaly requiring excessmass at depth has been resolved
over the northwest lobe of the aureole by line-of-sight tracking of a Viking orbiter [Phillipsand Bills, 1979;Sjogren, 1979].
Smooth plains materials [Scott and ½arr, 1978] embay and
partly bury thesedepositsin an annular belt at the baseof the
shield. The exposedarea of the aureoles,with some allowance
for overlap,is about 1.3x 106km2.
The prominent escarpmentencircling the base of Olympus
Mens is perhaps its most perplexing feature, for there is no
apparent counterpart around subaerial terrestrial volcanoes.
The scarp does, however, suggestanalogy with the subglacial
re t d
I$OO
m
,•00rn
!
i
i
illill
Subaerial and
postglacial
volcanic
i
i
i
!
i
i
!
ll!lllllH
llll
I1
Ill
llll
l!ll]ll/ll
i
!
i
llll
!
I
i
i
i
il
I
ß
i
I
l!
I
I
i
!
Its
Subaerial
top basalts
1
II
!
Palagonite
tuff breccias
Columnar
basalts
rocks
Fig. 5b. Diagrammatic crosssectionthrough Herdubreid (Figure 5a); intercalated columnar basaltsand palagonite
tuff brecciasform the base,overlain by subaerialtop basaltsand surmountedby shield volcano [ Van Bernrnelenand Rutten, 1955].
HODGESAND MOORE: SECONDMARS COLLOQUIUM
8065
IV
Fig. 6. Aerial photograph of BaejarOall (arrow a) and Kviholafj/S11(b) (also shown in Figure 3), which are composedalmost entirely
of m6berg. These two table mountains either formed subglacially
without subaerial venting and thus were not capped by subaerial
flows or were eroded after emplacement so that basaltic cap rocks
were removed. Both have summit depressions.Low sun angle en-
chances
topographic
irregularities;
compare
withFigure
3b.Northat
top; sunat lower right [ Van Bemmelenand Rutten, 1955,Plate XXV].
iliiiiiiiiiii:::
.... : ......
':::ii::::'::t'
,i;!ii!iill
table mountain pedestalsand with the subaqueousflanks of
oceanicshields.Such an analogy implies that (1) ice existed in
the Olympus Mons vicinity when eruptions began, (2) the
vent or vents eventually surfaced above the ice, and (3) the
enormoussubaerial shield was emplacedatop the subglacially
confined platform. Lava flows cascadingover the scarp indicate that Olympus Mons continued to grow long after the
postulatedice receded;such flows are evidencethat the scarp
existed during late growth of the shield and is not simply a
postvolcanicerosionalproduct. Similar relations occur in Iceland.
The grooved, eroded appearance of the aureole lobes and
their specificassociationwith Olympus Mons lead us to suggest further that these depositsalso may be products of volcanic eruptionsbeneath or into an ice cap, but like the m6berg ridges in Iceland these extrusionsfailed to surface above
the ice or meltwater lakes. The age of the lobes is uncertain
becauseof the difficulty in recognizingcraterswithin this rugged terrain. In a preliminary effort, however, we classifieddepressionsand circular features (>2.5 km across)as definite
craters,probable craters,and possiblecraters,and our analysesshowed that the aureoles appear to,have preceded development of Olympus Mons; additionally, they are probably as
old as or older than the Chryse area [Dial, 1978] and cratered
plains [Scottand Carr, 1978].
The aureole
materials
bear some resemblance
to the Icelan-
dic m6berg hills that are not capped by lava, one of which is
portrayed in the low-sun photographof Figure 6. The m6berg
depositsof Iceland are highly susceptibleto erosion by wind
Bedrock
Glacier
Meltwater
Pillow lava
and subsurfacewater and commonly exhibit irregular groand dikes
tesque shapesand closeddepressionswithout surficial drainage [Kjartansson,1960, Preusser,1976]. Several ragged mesas
that appearto be remnants,or outliers,of the aureoledeposits
are perhaps the best Martian analogs for table mountains;
Scree
Lava
Tuff and
Foreset bedthey are about 250 m high (as inferred from shadow length
breccia
ded pillowmeasurements),with craterlike depressions,and occur 200breccia
300 km southof the baseof OlympusMons (Figure 14).
If the aureole lobes representsequentialvolcanic deposits,
Fig. 7. Development of an Icelandic table mountain from subthe eruption centers appear to have migrated within a rela- glacialbirth (I, bottom)to subaerialerosion(IV, top). (Modified after
tively confined area before eventually being sustainedat the Einarsson[ 1968].)
8066
HODGES AND MOORE: SECOND MARS COLLOQUIUM
Kilometers
19o30 '
0
•.-J--!64o50
'
A
10
20
30
4O
50
VERTICAL
EXAGGERATION
X2
SEA
LEVEL
•
6,000'
! 2,000'
SEA LEVEL
B
.....
6,000'
• "•
12,000'
c
C'
SEA
LEVEL
•
6,000'
D
6,000'
D'
SEA LEVEL
'
6,000'
! 2,000'
! 4,000'
Pillowlava
Clasticrocks,
includingphreatic
Subaeriallava flows
(M•in shield)
Intrusive rocks
explosion ash
t
Fig. 9. Geologic sectionsthrough Kilauea and Mauna Kea volcanoes, Hawaii, showing sequentialstagesof developmentfrom initial
emplacementof pillow lavas (A-A') through constructionof subaerial
shield (D-D') [Moore and Fiske, 1969].
SHIELD
TABLEMOUNTAIN
Fig. 8. Contrastingtopographyof subglaciallyerupted shieldvolcano that formed table mountain (bottom) and subaerialshield (top)
in Iceland; grid is 1 km, contour interval is 20 m, north at top. Profiles
at approximatelysamescaleas map; no vertical exaggeration.(From
AMS SeriesC762, sheet 5721 III, Hrutafell, Iceland, 1:50,000, May
1951.)
site of Olympus Mons. According to this model the peculiar
aureole terrain representserosional remnants of thick (of the
order of a kilometer in places)m6berg, which was not capped
by subaerial lava. The cratered mesasof Figure 14 appear to
be vestigesof the initial m6berg surface,vulnerable to erosion
by wind and meltwater becauseof the poorly consolidated,
brecciated, and tuffaceousnature of the deposits.In addition
to explaining the peculiar dissectedfabric of these deposits,
which are apparently unique on Mars, this hypothesisalso is
compatible with the gravity data [Phillipsand Bills, 1979];the
aureoles represent volcanic centersand, as is characteristicof
the table mountains, there should be numerous dikes and sills
at depth.
Alternative Modelsfor Origin of the
Aureoles and Basal Scarp
Alternative explanations thus far suggestedfor either the
basal scarp or the aureole depositsof Olympus Mons pose
various difficulties. Several hypotheses involve pyroclastic
flows, landslides,gravity sliding, gravity thrust faults, debris
flows, or lava flows originating from the vicinity of Olympus
Mons; some of these addressthe scarp problem but do not ac-
Fig. 10. Martian cratered mesa amid irregular, hummocky patterned ground;narrow terraceoccursnear baseof flanks.Ratio of crater diameter to mesa diameter is small in comparisonto ratios of crater diameter and surrounding ejecta diameters of impact craters (see
Figure 2d). Diagrammatic profile drawn approximately at scale of
photograph, no vertical exaggeration.(Height estimatedfrom shadow
length.) Sun at left. (Latitude 47øN, longitude 230ø; Viking photograph 009B10.)
HODGES AND MOORE: SECOND MARS COLLOQUIUM
8067
c
snovv
Fig. 11. (continued)
count for the gravity anomaly. Those involving preexisting
volcanoes or domical plutons account for the gravity but are
improbable in other respects and have no relation to the
scarp. None adequately accounts for both the scarp and the
gravity results.
.•}i:•.•.•'•:;'' "•:'•:'-•
";..•"•'-:':-:
'-'-"',
"----:•*
....
'.........
"'
......
....
:}:':....•:-..:.-....:
;:?;•-•-,....
'...
.-.....
:•:;•
..:.--:::.•.:;-"
...•--'c
xt..:•'
....--.:..:.•%.
.• '-'-•', L..;:"
.....
..5:.:
........
snow
Fig. 11. Martian mesas(a) that resemble Icelandic table mountains with subaerial lava caps; these features are closely associated
with (b) a crateredconeand (c) a crateredpeak that stronglyresemble
terrestrial volcanic cones.Unlike impact craters,length of slope from
center to rim of cratered cone (Figure I lb) is equal to the length of
slope from rim to edge of flank, and crater flanks appear smooth. If
cratered features of Figures I lb and 1l c are volcanic, then mesas
(Figure 1la) also may be volcanic, possiblyoriginating as subglacial
eruptions subsequentlycapped by subaerial flows. White patcheson
flanks of all three landforms may be 'snow' localized on north facing
slopes.North is approximately at top in all photographs.Diagrammatic profiles not to scale (no shadows from which to estimate
heights).(a) Latitude 77øN, longitude 68ø, sun at lower fight, Viking
photograph070B27. (b) Latitude 78øN, longitude66ø, sun at bottom,
Viking photograph 070B29. (c) Latitude 79øN, longitude 75ø, sun at
lower right, Viking photograph070B09.
.?:..
6 km
Fig. 12. Complex terrain west of Viking 2 landing site that appearsto be at leastpartly volcanic.Dark irregularpatch(a) is probably a lava flow, and the irregularsurroundingsurface(b) resembles
that of manyterrestrialflows(seeFigure4, unit d). Crateredmesa(c)
may be a table mountain.Diagrammaticprofiledrawn approximately
at scaleof photograph.
(Heightestimatedfrom shadowlength.)(Latitude 44øN, longitude238ø;Viking photograph009B22.)
8068
HODGES
ANDMOORE:
SECOND
MARSCOLLOQUIUM
1.50o-w
•i;•40•
N
0
I•_:::.
.......
t5•N
lOO
L:.... ...:..... J
SCALE
INKILOMETERS
250N
-
Fig.13a. Olympus
Monsandsurrounding
lobate
aureole
deposits.
Lobes
appear
tohavebeenemplaced
sequentially
(asnumbered
in figure13b),andthedeposit
atthefarthest
periphery
onthenorthwest
(a)appears
tobeamong
theoldest.
Aureole
b apparently
issuperposed
onaureoles
at a andc.Onlyarcuate
remnants
areexposed
onthesouthandeast.Crateredplateaus
ofFigure14areatd.Base
ofOlympus
Monsatthescarp
isabout
600kmacross.
North-south
profile
along
longitude
133.25øW
isapproximately
toscale
asshown
(larger
thanthatofphotograph
mosaic);
novertical
exaggeration.
Compositeof Mariner 9 photographs.
HODGES AND MOORE: SECOND MARS COLLOQUIUM
8069
145øW
35øN
Fig. 13b. Outlinesof OlympusMensshieldandsurrounding
aureolelobes,indicating
a possible
nineseparate
eruptive
centers.Numberedfrom presumedoldest(1) to youngest.Outline of OlympusMens is about600 k_rnacross.
King and Riehle [1974] proposedthat the scarp resulted
Carr [1975] suggestedthat great massesof the volcano
from extensivewind erosionand slumpingof a pyroclasticash could have been dislodgedand emplacedas landslidedeposits
flow or ignimbrite foundation beneath the shield. The founda- hundreds of kilometers away; but had the shield originally
tions of terrestrial shield volcanoes, however, are not known continued its gentle slope outward, merging with the surto include substantialpyroclasticphasesexcept during late rounding plain, there is no obvious reason for the accumu-
submarine
growth;
-subaerial
pyroclastic
activityusuallyoccurs as essentiallythe last gaspin shielddevelopment,as exemplified at Mauna Kea and Haleakala in Hawaii. Extensive
ash flows are commonly associatedwith more silicic stratovolcanoes,not basalticshields.Also the dependenceon wind
to erodethroughlava flowsat the baseof a shieldthat initially
slopedgraduallyto the surroundingplainsappearsto require
an intense 'sandblasting'symmetrically on all sides of the
shieldin order to removethe resistantbasaltand exposethe
ash flows; such a processmust have occurred within a relatively restrictedtime before later lava flows cascadedover the
scarp.King and Riehle [1974] do not addressthe origin of the
aureoles,but E. C. Morris (oral communication,1979) has
proposedthat thesedepositsalso are a seriesof ash flows.
lated flows to have becomeuniformly unstableat the low, outermost margin of the flanks. Terrestrial landslides occur on
preexistingscarpsor steepslopes.Furthermore,presenttopography would have required uphill sliding on the southeast
[U.S. GeologicalSurvey, 1976].
An alternative, somewhat similar origin for the aureoles,
suggestedby J. G. Moore (oral communication, 1978), involves subaqueousdebris flows. If the vent were surrounded
by a body of meltwater, then subaerial flows from the shield
would tend to decrepitateupon entering the water, and, augmented by the aqueous medium, substantial flow breccias
could perhapshave moved over the great distancesrequired,
more or lessas density currents.The resulting depositsmight
be similar to m6berg without the pillow lava. A recentinvesti-
8070
HODGESAND MOORE:SECONDMARS COLLOQUIUM
pose instability on the crust by loading at the volcanic pile
and consequentlifting and sliding of the outermost flanks of
the volcano.
Harris [1977] proposed that the aureoles were a series of
gravity thrust faults in the plains materials caused by em-
placementof and crustalloadingby the hugevolcanicshield.
Thrusting of surrounding ground may indeed have occurred;
in fact, the offsetin the outer margin of the northwestsegment
of lobe b (Figure 13a) resemblesrather strikingly the map pattern of the Appalachian Pine Mountain overthrust [Thornberry, 1965]. However, the distinctive outlines of the lobes that
define discretebodiesof material and can be traced nearly to
the volcano'sbase(Figure 13b), the elevatedbut virtually accordant surfaceof the deposits,and, particularly, the development of the shield'sbasal scarpseeminadequatelyexplained
by this means.Neither of the gravity-induced mechanismsdescribedabovewould accountfor the gravity anomaly.
If the depositswere the eroded base of a larger Olympus
Fig. 14. Table mountains (a) southwestof Olympus Mons (Fig- Mons or of similar but separateshields[Carr, 1973], the eroure 13a), probably remnant outliers of ridged-and-groovedaureole
several
terrain (b), here interpreted as m6berg deposits(palagonitized tuffs sionalagent or processrequired to remove successively
and breccias)and pillow lavas. Texture of aureole terrain (b) resem- structureson the scale of Olympus Mons, with destructionof
bles that of m6berg hills in Figure 6 (featuresa and b). Profile drawn each before the next is built, is difficult to envision. A similar
at scale larger than that of photograph; no vertical exaggeration. problem is raisedby the hypothesisof Head et al. [ 1976],who
(Height estimated from shadow length.) (Latitude 10øN, longitude
attributed the scarp to erosionby unknown agent or process
142ø;Viking photograph044B13.)
of the prevolcanocratered terrain in virtually the entire northern hemisphereof Mars. The absenceof scarpsaround other
gation (J. G. Moore, personal communication, 1978) of a older shieldsis not explained,however,nor is the origin of the
landslide off the island of Hawaii, however, revealed that in a aureoles addressed.It would appear that Olympus Mons is
subaqueousenvironment the coefficientof friction as inferred considerablyyoungerthan developmentof the hemispheredifrom the ratio of length of runout to height of fall [Howard, chotomy on Mars, inasmuch as the plains at the base of
1973]is about 17, essentiallyequivalent to that of comparable Olympus Mons, now surfacedby young lava flowspartly from
subaerialslides,whereassimilar computationfor the aureoles the shield, are topographically continuouswith the older motsuggests
that ratiosas large as 117would be required.
tled plains to the north [Scott and Carr, 1978]. The age of the
All landslidehypothesespresentsomedifficulty in explain- older plains constrainsthe time of the northern hemisphere
ing the location of the most extensive volume of aureole de- plains development. The lava flows cascadingover the basal
posits on the northwest side, where the volcano's flank is scarpof the volcano indicate that the scarpexistedwell before
broadest.The scarp appears to be a nearly constantheight volcanismceased.Erosion of domical plutons (K. R. Blasius,
around
theentireshield,
irrespective
of variations
in apparent unpublished manuscript, 1974), although consistentwith the
volume and location of aureole lobes. None of the above slide
gravity data [Phillips and Bills, 1979; Sjogren, 1979], does not
or debris flow models accountsfor the gravity anomaly re- explain apparent age differencesand superpositionof the auquiring excessmass beneath the aureole lobes northwest of reole lobesbecausean extensiveerosionperiod would plane
Olympus Mons.
such coalescingplutons to a relatively continuoussurface, irSteep fault scarpsof limited extent occur on the subaerial respectiveof ages of intrusion. If the aureolesß
are simply the
flanks of Hawaiian volcanoes,exemplified by the Hilina Pali, deeply eroded remnantsof old subaeriallava flows [McCauley
more than 300 m high, on the southeast flank of Kilauea. et al., 1972; Morris and Dwornik, 1978], such lavas must have
Analogy with the continuousscarp at the base of Olympus been entirely different physically and/or chemically from the
Mons, however, seems inappropriate because the Hawaiian long, narrow flows that now cover the surrounding plain;
scarpsform in the subaerialshieldand possiblyare dependent these were clearly derived from the shield volcanoes and reon the existenceof the incompetent material in the submarine semble morphologically typical terrestrial basalt flows from
baseof the shieldas well as on the break in slopeat sea level. central vents [Greeley, 1973; Carr et al., 1977].
Further, the continuity of the Martian scarpis not duplicated
The volcanic explanation for the aureoles that at present
by the fault scarps in Hawaii, which have a discontinuous seemsto us most compatiblewith their configurationand disscallopedpattern (J. G. Moore, oral communication,1978). tribution as well as with the gravity data is subglacialerupSimilar kinds of slump scars can be identified inboard from tion. The pillow lavas and tuffaceous breccias thus accumuthe main scarpat Olympus Mons (Figure 13b), but it is post- lated would be analogous in origin, though clearly not in
ulatedhere that their formationwasdependenton the preexistingand unrelated basal scarp.
Gravity sliding on a subsurfaceplane has also been suggested as an explanation for the scarp and aureole (M. H.
Carr, oral communication, 1978). In this model the plane of
lubrication is the interface between permafrost and a presumed underlying zone with liquid pore water. Although this
model has not been examined in detail, it evidently would im-
scale, to those Icelandic table mountains on which subaerial
basaltseither were removedby erosionor were not emplaced.
The cratered mesas of Figure 14,, and the massive aureole
lobesas well, probablyare not subaeriallavas,nor were they
likely capped by such lava, but they may represent instead
thick subglacially developed m6berg that has been modified
by erosion.A complex network of intrusionsand dikes would
thus be present beneath the aureoles to produce the excess
HODGESAND MOORE: SECOND MARS COLLOQUIUM
8071
massat depth requiredby the gravitydata. The gravity high the disc-widespectrumof Mars. Palagoniteis also a reason-
northwest
of'OlympusMonsis currentlyinterpreted
by Phil- able petrologicinterpretationof samplesanalyzed by both Vilipsand Bills [1979]as evidencefor magmaticsourcerocksbeneath the aureoles, and this analysisis consistentwith the
model proposedhere.
An obviousdifficulty with this explanationfor the aureoles
is the mechanismthat permittedsuchextensiveeruptionsbeneath ice and meltwaterwith little or no subaerialventing.
Subaqueous
pillow lavasgenerallytend to pile up becauseof
rapid chilling [Sigvaldason,1968]. As in subaerial flows on
earth, however, a high rate of effusion [Shaw and Swanson,
1969;Walker, 1973]apparentlypermitsextensivespreadingof
submarineextrusionsas sheet flows [Holcomb, 1978; Ballard
et aL, 1979].The total volumeof aureolematerialmay be of
king landers;major constituentsof the Martian fines,as of the
Icelandic palagonites,are iron-rich clays. Toulminet al. [1977]
postulated that explosive interaction of iron-rich basaltic
magma and subterranean ice could readily account for the
large quantitiesof clay-rich material apparentlypresentat the
Martian
surface.
The interaction of magma with water and ground ice or
permafrostappearsinescapableif the climatic regime of Mars
during volcanic epochswas similar to its current state. Landforms similar in size and morphology to the Icelandic table
mountains,however, suggestthat large expansesof surfaceice
may have existedover someparts of the planet during various
the orderof 1-4 x 106km3, and someindividuallobesprob- times in the past. On the basisof shadow lengths the heights
ably contain as much as 5 x 105km3. An ice sheet 1-2 km of the pedestalsdescribedhere are generallyof the order of a
thick might not have been breached if such volumes were few hundred meters, and therfore the requisite thicknessof
erupted from different vents at different times and at very ice, at least in the northern latitudes, could probably have
rapid rates.Walkerand Blake [1966]describeda m6bergridge existedin the pastwith relatively slightmodificationsin atmoin Iceland that formed as lava flowed 22-35 km beneath a valspheric temperature of pressureand/or an increase in the
ley glacier at the relatively low gradient of about 0.018, and amount of water.
they speculatedthat an enormous'j6kulhlaup'(catastrophic A more demanding challenge to current models of volatile
flood),causedby the eruption,musthaveformeda subglacial budgetand planetary history, however,is engenderedby our
conduit down which the lava subsequentlyflowed. Subaerial interpretationof Olympus Mons and adjacentfeatures,which
Martian lavas have flowed on gradients as low as 0.002 callsfor extensiveice in the Tharsis regionof Mars. Assuming
[Moore et aL, 1978],so that subglacialflowson similar gradi- that the top of the talus on the Olympus Mons scarp is apentsare possible.Olympus Mons is on the flank of the Tharsis proximatelyat the baseof the subaerialflows,the thicknessof
bulge,and a low westwardgradientof about0.002 likely ex- ice requiredto form the m6berg foundationwould be perhaps
isted when eruptions began; subglacialconfinementand the 3-4 km.
low gradient unrestrictedby valley walls may have induced
The m6berg pedestal on which Olympus Mons rests, acbroadlateralspreadingor pondingfrom eachof the 8-10 pos- cording to this hypothesis,would, like the scarpsof terrestrial
tulatedsubglacialvents(Figure 13b).
table mountains, readily retreat by mass wasting and wind
Alternatively,if the magmareservoirs
wereinitially shallow erosion,but the initial relief of the scarpcould be explained as
within the crustand becameprogressively
deeperas the crust a primary feature, an inherent consequenceof the volcano's
thickened,then low pressureheadsmay have existedduring subglacial beginnings. Furthermore, like the Pacific guyots
the initial aureoleeruptions,permittinglittle vertical growth. and shields[Moore and Fiske, 1969], the huge Olympus Mons
As the planet cooled and magma reservoirsdeepened,suf- edifice probably subsidedsomewhat,causingthe annular deficientpressurehead must have been attainedto permit the pressionaround its base. Subsequentdepositionof lava, eovertical growth of Olympus Mons [Carr, 1976a, b] and the lian debris, and mass-wasted materials in this circumferential
three large shieldsnear the crestof the Tharsisridge,each of depressionobscuredaureoledepositsthat probably existedat
which also reaches 26 km in elevation.
the base of the scarp, as suggestedby the outcrop pattern. If
As magma contactedice, vast meltwater lakes probably the broad aureoles do consistof m6berglike materials, they
formed, although surfacescould have remained frozen. The would supply an extensivesourcefor the palagonitic debris
existenceof liquid water at the surface would of course re- that appearsto be widely distributedover the planet.
quire atmosphericconditionsdifferentfrom thoseof the present, a difficulty encounteredas well in attemptsto explain Ice Volume and the Volatile Budget
Martin channel configurations.
The morphology of the Olympus Mons complex seems
rather well explained by the table mountain analogy, but acThe Role of Ice
countingfor an ice accumulationof perhaps7-10 x 106 km3
If the Icelandicanalogyis to be drawn,the formerpresence posessome difficulty. Some current estimatesof the volatile
of more extensive surface ice on Mars must be established.
The pervasiveevidencefor a substantialthicknessof ground
ice has been described elsewhere [Carr and Schaber, 1977;
Masurskyet eL, 1977;Soderblomand Wenner,1978].A largely
water-icecompositionof the north polar cap appearsconclu-
budget, based on argon isotope ratios, permit a maximum
amount of water-outgassedequivalent to a planet-wide thicknessof only about 9.5 m [Andersand Owen, 1977]. However,
the amount of H20 intrinsic to Mars is a matter of some dis-
pute, and it appears that the total volatile inventory of the
sive, and its present thicknesshas been estimated at as much planet is not yet well defined. Bogard and Gibson[1978] enuas 1 or 2 km [Farmeret eL, 1976;Kiefferet eL, 1976].Water in merated the difficulties of using noble gas ratios and comViking samplesmeasuredabout0.3-2% [Biemannet eL, 1977]. parisonswith chondritesand statedthat presentdata are inGeochemicaland spectraldata from the Martian surfacefur- adequatefor determining the relative abundancesof volatile
ther indicate that magma may have interacted with H20 to elementsoriginally supplied to Mars. Lewis [1974] proposed
form sideromelane,the precursorof palagonite. Soderblom that Mars had an initial water content6 times (per unit mass)
and Wenner[1978]observed
that the 0.3- to 2.5-/.tinspectrum that of earth, basedon an equilibrium condensationmodel of
of antarcticpalagonitesamplesis the closestapproximationto solarsystemformation. McElroy et al. [1977],usinga nitrogen
8072
HODGES AND MOORE: SECOND MARS COLLOQUIUM
isotope analysis,suggestedthat H20 could have been abun- at lower elevations.The maximum thicknessof the ice cap on
dant enough to form a planet-wide surfaceice layer 200 m Greenland is currently 3.4 km [Flint, 1971]; even thicker
thick. If this last amount were concentratedin a 30ø spherical sheetsmay have existedon earth during the Pleistocene.
A possiblealternative to the extensivecoverageby glacial
cap of ice centeredon Olympus Mons, a 6-km-thick ice sheet
could result. Assuming adequate amounts of water, the tem- ice is the interactionof magma with more readily explicable
peratureof the planetary surfacewould dictatethat ice be the ground ice. If suchinteraction occurred,tuffaceousflows and
dominant phase, unless atmospheric pressure increased, as brecciasmay have erupted until the conduit was insulated
from further entry by meltwater. But under approximately
may have occurredduring timesof extensivevolcanism.
The ice coveragecalled for here is probably bestderived as these conditions on earth, small maars and tuff cones are the
a localized cap. The possibility that the rotational axis may predictablelandforms.On the SewardPeninsula(Alaska),anhave wandered during the courseof the planet's history has gular blocksof loessare incorporatedin the ejectaaround sevbeen proposedby Murray and Malin [1973]and by Ward et al. eral maars, indicating (along with other evidence)that the
[1979],who cite as evidencethe positionof the Tharsisbulge sedimentswere frozen at time of eruption (D. M. Hopkins,
with its excessmassessentiallyat the equator.They consider oral communication, 1979). Only the subglacial and subit more probablethat the bulgeand volcanicshieldsgrew else- marine eruption style appearsto producemorphologiescomwhere (i.e., when the rotational axis was differently located) parable to those of the table mountains and the Olympus
and caused a shift of the poles, such that the excessmass Mons shield and associated features on Mars.
would be repositionednearer the equator.This postulateis
SUMMARY
further supportedby the large basin (~800 km across)and reThe morphologiesof small Martian buttesand mesas,espesultingmassdeficiencynear the presentsouthpole. Conceivably, the hypothesizedglacial stageoccurredif and when the cially concentrated in latitudes north of 40 ø, are similar to
Tharsis region was nearer the pole before adjustmentof the thoseof Icelandic table mountainsthat are presumedto have
formed by subglacialeruption. The geologicand climatic replanet to the excessmass.
An alternative and perhapsmore appealing causeof a lo- gimesof Mars may be or have been analogousto the environcalized ice cap is the relief of the Tharsis bulge itself, regard- ment of Iceland, where volcanism and glaciation have interless of its latitudinal position. Under favorable atmospheric acted throughout much of recent geologic history. In the
conditions,ice may have accumulatedat high altitude and northern latitudes of Mars the candidate table mountain
been sustainedbelow 10 kin, even at equatorial latitudes. structuresare associatedwith landformsand terrainsthat apThus both ice and volcanism would have followed the deforpear more obviouslyvolcanic,and thus it seemsa reasonable
mation that producedthe bulge and, presumably,the exten- postulatethat a relatively slightmodificationin positionor exsive fracturing of the old cratered plains. Stratigraphic diffi- tent of the polar cap could have allowed volcanic extrusion
culties(M. H. Carr, oral communication,1978)that arisewith into or throughice.
More problematical is the localization of a sufficientthickthe polar wander hypothesisare therebyeliminated.
The means by which such ice could be preferentially con- ness of ice in the Thatsis region to account for our intercentrated as a result of altitude changesdependson atmo- pretation of Olympus Mons and its surroundingaureolesby
spheric conditions in addition to water concentration.The initially subglacial eruptions. The basal scarp around the
present atmospheredecreasesin temperaturewith altitude enormousshield is readily explained if the shield were built
atop a pedestalof pillow lava and palagonitizedtuff breccias,
from about 240øK at the surface to about 100øK at 200 kin,
with about a 50ø decreasein the first 20 km. Higher temper- or m6berg. The position, apparent sequence,and grooved,
atures must have occurred at the surface when and if fluvial
serrate texturesof the aureole lobes are compatiblewith their
action occurred to form or modify channels. According to interpretation as separate subglacial eruptives that did not
Masurskyet al. [1977] fluvial channelagesvary from 3.5 to 0.5 reach the surface and remained as expansesof m6berg unb.y. During periodswhen liquid water existedat low altitude, capped by subaerial lava flows. Massive palagonite would
ice may have accumulatedcontemporaneously
at higher alti- thushave beensubjectedto wind erosionand transport,partly
accounting for the widespread distribution of such material
tudeson the flanks of the Thatsis ridge.
An ice cap could have been sustained throughout inter- implied by spectraland Viking geochemicaldata. Initial lomittent volcanism, as is the case in Iceland today. Williams calization of the ice could perhapsbe attributed to the 10-kin
[1978]has suggested
that the blocky,conspicuously
ridged ter- elevation of the Tharsis bulge.
The total water budget may indeed be an obstacleto this
rain at the northwestbaseof Arsia Mons is bestexplainedas a
sequenceof recessionalmoraines from a glacier that devel- hypothesis,but there appearsto be at presenta broadrangeof
oped on the volcano or existed prior to eruption and was volume estimates,dependent on analytical technique, and
maintained as the volcano grew. Similar depositsoccur in an little knowledgeof the possiblevariationsof near-surfacevolapronlike lobe at the northwestbaseof PavonisMons. How- atile concentrationsthroughout Martian history. Photogeoever, there are no such deposits around Ascraeus Mons, logic interpretationspresentedhere suggestthe former exisyoungest of the large Tharsis shields [½arr, 1976a; Wood, tenceof more extensiveglacial ice than is currentlymeasured
1976], so that the 'Ice Age' may have vanishedby the time the or inferred; the feasibility of such an ice age dependsultiAscraeusMons eruptionsbegan.Although the presentsurface mately on developmentof a compatible and defensiblemodel
of the Olympus Mons shield appearsto be youngerthan As- for atmosphericevolution, and further geologicinvestigations
craeusMons [Wise et al., 1979],eruptionscould have begun at of Mars may require sucha model.
the larger volcano during an earlier glacial stage.If a thick ice
Acknowledgments.We are indebtedto a number of reviewerswho
cap existedon the Tharsisbulge, it probably was in a dynamic offered critical comments. Especially helpful were Robert B. Harstate, with ice accumulatingrapidly at highest elevations as graves,who brought to our attention the possibleanalogybetween
the massflowed outward and downward into regionsof thaw Icelandic landforms and those on Mars, and Richard S. Williams,
HODGES AND MOORE: SECOND MARS COLLOQUIUM
8073
who kindly sharedhis extensiveknowledgeof Icelandicgeologyand
lithospheric spreading centers (abstract), Eos Trans. AGU, 59,
1104-1105, 1978.
whosethinking about Martian featureshad earlier progressedalong
lines similar to ours. We benefitedgreatly from severaldiscussions Howard, K. A., Avalanche mode of motion: Implications from lunar
with JamesG. Moore, who providedconsiderablebackgroundfor our
examples,Science,180, 1052-1055, 1973.
comparisons
with Hawaiianvolcanicshields.In additionto detailed Jones, J. G., Intraglacial volcanoesof south-west Iceland and their
significancein the interpretation of the form of the marine basaltic
reviewsby thesethree geologists
we receivedextremelyhelpful comvolcanoes,Nature, 212(5062), 586-588, 1966.
mentaryfrom Michael H. Cart, PeterH. Schultz,GeorgeE. McGill,
and Lisa A. Rossbacher.Carlton C. Allen, Dennis L. Orphal, and Jones,J. G., Intraglacial volcanoesof the Laugarvatn region, southRobin T. Holcomb also read the manuscriptand offered useful sugwest Iceland, I (with discussion),Quart. J. Geol. Soc. London, 124,
part 3(495), 197-211, 1969.
gestions.
Jones,J. G., Intraglacialvolcanoesof the Laugarvatnregion,southwestIceland, II, J. Geol., 78(21), 127-140, 1970.
REFERENCES
Allen, C. C., Interactions of volcanismand ice--Earth, Mars, and icy
satellites, Bull. Amer. Astron. Soc., 9, 541, 1977.
Allen, C. C., Subglacialvolcanism-terrestriallandforms and implications for planetary geology,Report of Planetary Geology Program,
1977-1978, NASA Tech. Memo., 79729, 194-195, 1978.
Allen, C. C., Volcano/ice interaction on Mars, Lunar Planet. Sci.
X(I), 18-20, 1979.
Anders, E., and T. Owen, Mars and Earth, Origin and abundanceof
volatiles, Science, 198, 453-465, 1977.
Ballard, R. D., R. T. Holcomb, and T. H. Van Andel, The Galapagos
rift at 86øW, 3, Sheet flows, collapsepits and lava lakes of the rift
valley, J. Geophys.Res., 84, 5390-5406, 1979.
Biemann, K., J. Oro, P. Toulmin III, L. E. Orgel, A. O. Nier, D. M.
Anderson, P. G. Simmonds, D. Flory, A. V. Diaz, D. R. Rushneck,
J. E. Biller, and A. L. Lafieur, The search for organic substances
and inorganic volatile compounds in the surface of Mars, J.
Geophys.Res.,82(28), 4641-4658, 1977.
Blasius,K. R., Topical studiesof the geologyof the Tharsis region of
Mars, Ph.D. thesis,85 pp., Calif. Inst. of Technol., 1976.
Bogard, D. D., and E. K. Gibson, Jr., The origin and relative abundancesof C, N, and the noble gaseson the terrestrialplanetsand in
meteorites, Nature, 271, 150-153, 1978.
Carr, M. H., Volcanism on Mars, J. Geophys.Res., 78, 4049-4062,
1973.
Carr, M. H., Geologic map of the Tharsis quadrangleof Mars (MC9), Map 1-893,U.S. Geol. Surv.,Reston,Va., 1975.
Cart, M. H., The volcanoes of Mars, Sci. Amer., 234, 32-43, 1976a.
Cart, M. H., Elevation of martian volcanoesas a function of age, Reports of accomplishmentsof Planetology Programs, 1975-1976,
NASA Tech. Memo., TMX-3364, 152-153, 1976b.
Carr,M: H., and G. G. Schaber,
Martianpermafrost
features,
J.
Geophys.Res.,82(28), 4039-4054, 1977.
Cart, M. H., R. Greeley, K. R. Blasius,J. E. Guest,and J. B. Murray,
Some Martian volcanicfeaturesas viewed from the Viking orbiters,
J. Geophys.Res.,82(28), 3985-4015, 1977.
Dial, A. L., The Viking I landing site crater diameter-frequencydistribution, Reports of the Planetary Geology Program, 1977-1978,
NASA Tech. Memo., 79729, 179-181, 1978.
Einarsson,T., Jardfraedi,Saga bergsog lands(Geology, the history of
the rocks and landforms), 335 pp., Nal og Menning, Reykjavik,
1968.
Farmer, C. B., D. W. Davies, and D.C.
Kieffer, H. H., S.C. Chase, Jr., T. A. Martin, E. D. Miner, and F. D.
Palluconi, Martian north pole summer temperatures:Dirty water
ice, Science, 194, 1341-1344, 1976.
King, J. S., and J. R. Riehle, A proposedorigin of the Olympus Mons
escarpment,Icarus, 23, 300-317, 1974.
Kjartansson, G., The M6berg formation, On the geology and geo-
physicsof Iceland, in Proceedingsof the XXI Sessionof the International GeologicalCongress,Guide to ExcursionNo. A2, edited by
S. Thorarinsson,pp. 21-28, Reykjavik, 1960.
Kjartansson,G., Stapakenninginog Surtsey(A comparisonof tablemountains in Iceland and the volcanic island of Surtsey off' the
southcoastof Iceland), Ndttt•rufraedingurinn,
36(1-2), 1-34, 1966a.
Kjartansson,G., Sur la r6cessionglaciaire et les types volcaniques
dans la r6gion du Kj/31ursur le plateau central de l'Islande, Rev.
Geomorphol.Dyn., 16(1), 23-39, 1966b.
Kjartansson, G., Volcanic forms at the sea bottom, Iceland and Mid-
OceanRidges,Reykjavik,Visindaf61ag
islendinga,
Rit. 38, pp. 5366, Soc.Sci. Island., Reykjavik, 1967.
Le Masurier, W. E., Volcanic record of Antarctic glacial history: Implicationswith regardto Cenozoicsealevels,Spec.Publ. 4, pp. 5974, Polar Geomorphol. Inst. of Brit. Geogr., London, 1972.
Lewis, J. S., The chemistryof the solar system,Sci. Amer., 230(3), 5065, 1974.
Masursky, H., J. M. Boyce, A. L. Dial, G. G. Schaber,and M. E. Strobell, Classification and time of formation of Martian channels
basedon Viking data, J. Geophys.Res., 82, 4016-4038, 1977.
Mathews, W. H., 'Tuyas,' fiat-topped volcanoesin northern British
Columbia, Amer. J. Sci., 245, 560-570, 1948.
McCauley, J. F., M. H. Cart, J. A. Cutts, W. K. Hartmann, H. Ma-
sursky,D. J. Milton, R. P. Sharp,and D. E. Wilhelms, Preliminary
Mariner 9 reporton the geologyof Mars, Icarus, 17, 289-327, 1972.
McElroy, M. B., T. Y. Kong, and Y. L. Yung, Photochemistry
and
evolution of Mars' atmosphere:A Viking perspective,J. Geophys.
Res., 82, 4379-4388, 1977.
Moore, H. J., D. W. G. Arthur, and G. G. Schaber,Yield strengthsof
flows on the Earth, Mars, and Moon, Proc. Lunar Planet. Sci. Conf
9th, 3351-3378, 1978.
Moore, J. G., and R. S. Fiske, Volcanic substructure inferred from
dredgesamplesand ocean-bottomphotographs,Hawaii, Geol.Soc.
Amer. Bull., 80, 1191-1202, 1969.
Morris, E. C., and S. E. Dwornik, Geologic map of the Amazonis
quadrangle
of Mars(MC-8), Map 1-1040,U.S. Geol.Surv.,Reston,
LaPorte, Mars: Northern
summerice cap--Water vapor observationsfrom Viking 2, Science,
194, 1339-1341, 1976.
Flint, R. F., Glacial and Quaternary Geology,892 pp., John Wiley,
New York, 1971.
Greeley, R., Mariner 9 photographsof small volcanic structureson
Mars, Geology,1, 175-180, 1973.
Harris, S. A., The aureole of Olympus Mons, Mars, J. Geophys.Res.,
82(20), 3099-3107, 1977.
Head, J. R., M. Settle, and C. A. Wood, Origin of Olympus Mons escarpment by erosion of pre-volcano substrate,Nature, 263, 667668, 1976.
Hoare, J. M., and W. L. Coontad, A tuya of Togiak Valley, southwest
Alaska, J. Res. U.S. Geol. Surv., 6, 193-201, 1978.
Hodges,C. A., Basalticring structuresof the Columbia Plateau and
possibleextraterrestrialanalogs,in Lunar ScienceVIII, pp. 449451, Lunar ScienceInstitute, Houston, Tex., 1977.
Hodges,C. A., and H. J. Moore, Tablemountainsof Mars (abstract),
in Lunar and Planetary ScienceIX, pp. 523-525, Lunar and Planetary Institute, Houston,Tex., 1978a.
Hodges,C. A., and H. J. Moore, The subglacialbirth of Olympus
Mons, Geol. Soc.Amer. Abstr. Programs,10(7), 422, 1978b.
Holcomb, R. T., Interpretation of submarinepahoehoeflows along
Va., 1978.
Murray,B.C., and M. C. Malin, Polarwanderingon Mars, Science,
179, 997-1000, 1973.
Phillips,R. J., and B. G. Bills,Mars:Crustand uppermantlestructure, Second International Colloquium on Mars, NASA Conf
PubL, 2072, 65-67, 1979.
Preusser,H., The Landscapes
of Iceland.'Typesand Regions,363 pp.,
W. Junk, The Hague, 1976.
Scott,D. H., and M. H. Cart, Geologicmap of Mars, Atlas of Mars,
scale1'25,000,000,Misc.Invest.Map 1-1083,U.S. Geol. Surv.,Reston, Va., 1978.
Shaw,H. R., and D. A. Swanson,Eruptionand flow ratesof floodbasalts, in Proceedingsof the SecondColumbiaRiver Basalt Sym-
posium,editedby E. H. Gilmour and D. Stradling,pp. 271-299,
EasternWashingtonStateCollegePress,Cheney,Wash.,1969.
Sigvaldason,
G. E., Structure
andproducts
of subaquatic
volcanoes
in
Iceland, Contrib.Mineral PetroL,18(1), 1-16, 1968.
Sjogren,
W. L., Marsgravity:High-resolution
results
fromVikingOrbiter 2, Science,203, 1006-1010, 1979.
Soderblom,L. A., and D. B. Wennet, Possiblefossil H:O liquid-ice
interfaces in the martian crust, Icarus, 34, 622-637, 1978.
Thornberry,W. D., RegionalGeomorphology
of the UnitedStates,609
pp.,JohnWiley, New York, 1965.
8074
HODGESANDMOORE:SECONDMARSCOLLOQUIUM
Toulmin, P., A. K. Baird, B.C. Clark, K. Keil, H. J. Rose, Jr., R. P.
Christian, P. H. Evans, and W. D. Kelliher, Geochemical and mineralogicalinterpretationof the Viking inorganicchemicalresults,J.
Geophys.Res., 82, 4625-4634, 1977.
U.S. GeologicalSurvey,Topographicmap of Mars, Misc.Invest.Map
1-961, scale 1/25,000,000, Reston, Va., 1976.
Van Bemmelen,R. W., and M. G. Rutten, Tablemountains
of Northern Iceland (and Related GeologicalNotes), 217 pp., E. J. Brill, Leiden, Netherlands, 1955.
Walker, G. P. L., Lengths of lava flows, Phil. Trans. Roy. Soc. Lon-
of Mars: The role of the Tharsis uplift, J. Geoœhys.Res., 84, 243259, 1979.
Williams, R. S., Geomorphic processesin Iceland and on Mars: A
comparativeappraisalfrom orbital images,Geol. Soc.Amer. Abstr.
Programs, 10(7), 517, 1978.
Wise, D. U., M.P. Golombek, and G. E. McGill, Tharsisprovinceof
Mars: Geologic sequence,geometry and a deformation mechanism,
Icarus, 38, 456-472, 1979.
Wood, C. A., Morphological evolution of shield volcanoeson Mars
and Earth (abstract),Eos Trans.AGU, 57, 344, 1976.
don, Set. A, 274, 107-118, 1973.
Walker, G. P. L., and D. H. Blake, The formationof a palagonite
brecciamassbeneatha valley glacierin Iceland(with discussion),
Quart. J. Geol.Soc.London,122, part 1(485),45-61, 1966.
Ward, W. R., J. A. Burns,and O. B. Toon, Pastobliquity oscillations
(Received April 2, 1979;
revisedJuly 20, 1979;
acceptedJuly 27, 1979.)