JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 103, NO. E5, PAGES 11,135-11,158, MAY 25, 1998
Characteristics of lunar mare deposits in Smythii and
Marginis basins: Implications for magma transport
mechanisms
R. Aileen Yingst and JamesW. Head III
Departmentof GeologicalSciences,Brown University,Providence,RhodeIsland
Abstract. An analysisof 34 lunarlava flows andpondsin the easternlimb Smythiiand
Marginisbasinswasundertaken
to examineandmodelthefirststagesof secondary
crustal
formationandassess
processes
involvedin magmatransportanderuption.In orderto isolatethe
characteristics
of singleeruptiveepisodes,
wefocused
ondiscrete
marepondsadjacent
to themajor
mafia.Meanvaluesfor areasandvolumesof deposits
estimated
to begoodcandidates
for single
eruptive
phases
arelarge,approximately
950-1000
km2and-200km3respectively.
These
eruptive
volumesarecommensurate
with the largestknownterrestrialeruptions(floodbasalts).The lack of
geomorphological
structures
indicativeof shallowmagmareservoirs
indicates
thatdeep,probably
subcrustal
sourceregionsareprevalent.
With respectto crustalthickness
relationships,
the
magnitudeandfrequencyof eruptiveeventsareobservedto be greatestin areasof thinnestcrust.
Specifically,regionsof majormaria(Mare Smythii,Mare Marginis)occurin areasof thinnest
crust,while isolatedpondsoccurwherethe crustis relativelythicker.This is consistent
with the
correlation
observed
globallybetweencrustalthickness
andthemagnitude
andfrequency
of
eruptiveevents.Agesof volcanicflowsandpyroclastic
eventsrangefromEarly Imbrianto
Imbfian-Eratosthenian
butareconcentrated
mostheavilybetween3.80-3.60Ga, suggesting
a
periodof peakvolcanism
beginning
around3.85Ga andlastingapproximately
200 Ma. However,
theexistence
of dark-halocraterclusters
in non-mare
unitswithintheregionsuggests
thepresence
of cryptomaha,which wouldindicatean earlieronsetof volcanismanda volumeof marematehal
potentiallygreaterthanthatcurrentlyexposedonthesurface.Typicalnearest-neighbor
distances
suggestdepositsderivefrom reservoirs<-100 km in diameter.The observations
madehereare
consistent
with a magmatransportmodelin whichplumesrisingdiapiricallystallat a density
boundaryunderthelunarcrustandpropagate
dikesto thesurfacethroughoverpressufization.
1. Introduction
Lunar volcanic deposits (maria)represent only a small
fraction (17%) of the surfacearea of the Moon [Head, 1976]. In
terms of lunar evolution, however, these mare deposits are a
cruciallink betweenour understandingof the initial stagesof
primary crustal formation from the lunar magma ocean, and
subsequent
thermalevolutionthat produceda partial secondary
crust [Taylor, 1989]. The characteristics of these mare
deposits, their size and thickness, age and composition,
morphology and setting, associatedfeatures and distribution
across the lunar surface, provide important information for
decipheringthe processesresponsiblefor their generationand
formation. For example, the distributionof depositsacrossthe
Moon is highly heterogeneous.As seen in Figure 1, the
nearside of the Moon has a significantly higher density of
mare depositsthan the farside. The reasonfor this dichotomy
is unclear but must depend in part on the underlying
mechanismsdriving magma throughthe crust to the surface.It
has been suggested,for instance, that crustalthickness plays
an important role in the efficacy of magma transport [e.g.,
Solomon, 1975; Head and Wilson, 1992; Robinson et al.,
Copyright1998 by the AmericanGeophysicalUnion.
Papernumber98JE00736.
0148 -0227/98/98 JE-00736509.00
1992], due to the densitycontrastbetween mare basalt magma
and the highlandscrust.Thus the mare distribution dichotomy
may be due to the nearside-farside
crustalthickness asymmetry
observedby the Apollo and Clementine missions [e.g., Zuber
et al., 1994; Neumann et al., 1996].
In order to resolve suchquestions,it is vital to reconstruct
the conditions for ascent and eruption of magma so that
subsurface processes may be constrained and modeled.
Classification and analysis of common mare deposit
characteristics provides a diagnostic tool in this regard. In
previous efforts we have compiled and analyzeda databaseof
characteristics for isolated mare deposits and individual
eruptive eventsoccurring in basins on the lunar westernlimb
and farside[Yingst and Head, 1994, 1997a]. Here we expand
this studyto includeisolatedmaredeposits
in the easternlimb
basinsof Smythii and Marginis. We comparethe resultsfrom
these basins to those of previous average volume and
distributionestimatesin order to build a statisticalpicture of a
typical lunar eruptive phase.We then interpret these results in
terms of source region geometry and crustal thickness
relationships, in order to understandthem within the larger
framework of magma transportand eruptionmechanisms.
2. Method
The first stepin our approachis to modelthe morphology
of a typicallunarvolcanicphase,definedhere as a single dike
11,135
11,136
YINGSTAND HEAD:SMYTHII AND MARGINISBASINLUNARMAREDEPOSITS
emplacementevent that may have rangedfrom a small, short
durationeruption up to a high-flux, longer duration eruption
lasting severalyears. Accordingly, it is necessaryto isolate
those characteristicswhich are common to lunar eruptive
events. This task is hamperedby the complex nature of the
largemaria. Althoughflow frontsare locally observedin some
regions [Gifford and El-Baz, 1981; Schaber, 1973], the
morphology of eruptive episodes is difficult to identify
againstthe backgroundof other depositsdisplaying a variety
of compositions,ages and stratigraphicpositions.
In light of these constraints,we have adoptedan approach
similar to the one used to categorize and analyze pond
characteristics in other farside basins [Yingst and Head,
1997a], where analysisis limited to significantpopulationsof
discrete,isolatedmarepatches,or ponds [Beals and Tanner,
1975; Whitford-Stark, 1982], regionsmore likely to represent
individual volcanic phases.Such groupings are most common
on the lunar limbs and farside,as seenin Figure 1.
For the purposeof comparing possible models of magma
transport,regionsof studywere limited to thosebasinswhose
age, diameter, depth, morphology and associated crustal
thicknessesdifferedfrom regions previously examined(e.g.,
Orientale and SouthPole-Aitkenbasins[Gaddis,1981; Yingst
and Head, 1997a], Australebasin [Hiesinger et al., 1996]). On
the basis of these criteria, lava ponds in the Smythii and
Marginis basins on the easternlimb were mappedand their
areasdetermined.Ages, modesof occurrence,topography, and
associated
featureswere identifiedusing LunarOrbiter, Apollo
and Clementine data. Evidencewas sought in each discrete
pondfor multipleflowsin orderto isolatethosecharacteristics
commonto individualflows or eruptiveepisodes.To this end,
pondsshowinga homogeneous
albedo,color, and craterfrequency
distribution,as well as a lack of characteristics
indicative of multiple flows (overlapping flows, several
potentialsourcevents,etc.), wereestimated
to be individual
eruptive phases. Thus, although it is not possible
unambiguously
to determinethe numberof flows represented
by a marepond,the depositsdescribed
abovewereconsidered
best candidatesfor estimates of individual eruptive phases.
Clementinemultispectral
datahasbeenusedin the past to test
the efficacyof estimatingindividualeruptivephasesin this
manner [Yingst and Head, 1997b]. However, Clementine
imagesof the Smythii/Marginisregion of the easternlimb
tend to have low phase angles. Because the photometric
propertiesof Clementineimagesat low phaseangles change
very rapidly [Nozetteet al., 1994] in a mannerthat is currently
not precisely modeled[e.g. McEwen, 1996; Pieters et al.,
1997], multispectral data for the Smythii/Marginis region
cannot
at this time be used as an accurate indicator
of subtle
variationsin soil mineralogy.For this reason,we have chosen
not to include the Clementine multispectral dataset in this
work.
Pond thicknesseswere calculatedusinga variety of methods
describedin previousstudies[Yingst andHead, 1997a; Gillis
et al., 1997]. These thicknesseswere used to estimate the total
volume of each deposit or flow. Finally, associatedcrustal
6t
iø
225 ø
45 ø
270 ø
45 ø
225 ø
90 ø
0ø
90ø
180ø
225ø
40 ø
O
¸
--,SPA
MR
225ø
•o
270 ø
225ø
•
~
0o
-75 ø
_75ø -60 ø
45 ø
45ø
,,
it
90 ø
45ø1
_60ø _75ø
225 ø
180ø
'
-75 ø -60 ø
225ø
Figure 1. Map showing the location of mare depositson the lunar surface.Maria are shown in black.
Distributionis concentrated
on the nearside,while the farsideand limbs have very few deposits.Regions
mentionedin the text are notedas follows:SPA, SouthPole-Aitken;Or, Orientale;SM, Smythii/Marginis;and
MR, Mendel-Rydberg.Map after Schultzand Spudis[1983].
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
11,137
thicknessesbased on Clementine altimetry models were noted
[Zuber et al., 1994; Neumann et al., 1996]. Results from these
analysesare presentedin Tables 1 and 2.
3. Resultsand Interpretation
3.1.
Setting
The Smythii and Marginis basins, shown in Figures 1 and
2, lie on the eastern limb of the Moon. While they are both
datedas pre-Nectarian,Smythii has a more well-definedring
structure,one that appearsto overlap Marginis basin. It is thus
consideredyounger [Wilhelms, 1987]. Marginis basin is
approximately580 km diameter[Wilhelms, 1987], although
exactmeasurements
are difficult becauseonly partial segments
of the ring structureremain. We assumefor the purposesof this
study that Marginis basin is circular, extrapolating from the
semicircularring structurenoted by Wilhelms and El-Baz
[1977].Thissuggests
a basinareaof-2.6 x 105km:. Smythii,
by comparison,
is larger(840km diameter
and5.54 x 105km:
in area) and better preserved.It displays an extensive central
maredeposit(Mare Smythii) and a subcircularcentral furrowed
and pitted plains-type deposit (notedINfp) upon which many
floor-fracturedcraters lie [Schultz, 1976]. Most mare ponds
within central Smythii basin lie within these craters.
3.2.
Areas
and
Volumes
Thirty-nine volcanic deposits (shown in Figure 2) were
mapped in the Smythii and Marginis basins; their
characteristicsare describedin Tables 1 and 2. Two deposits
(Mare Smythii and Mare Marginis) have areal extents and
estimated volumes at least 10 times greater than any other
ooo•o•o•o•o••
pond in the region. These regions display a complex
morphology and have evidence of multiple flows [Wilhelms
and El-Baz, 1977], someof which are suggestedto be younger
than Apollo 12 basalts (3.20 Ga [Spudisand Hood, 1992]).
Thesetwo depositsare more similar to the major mafia than
small, discrete lava ponds. For the purposes of creating a
statisticalpicture of individual eruptive phases,we thus focus
on analysisof the small individual deposits.We then compare
the resultsto characteristicsof the larger Mafia Smythii and
Marginis.
Of the remaining37 deposits
in bothbasins,five show
evidencethat they represent
multipleeruptiveepisodes
(differing crater-frequencydata [Wilhelms and El-Baz, 1977]).
Onemultiphase
deposit
(Joliot;
pond1 in Marginis
basin)
shows morphological evidence of being comprisedof more
thanoneeruptivephasein that flow boundaries
are evident.
However,the preciseboundaries
of theseflows have not yet
been determinedand the flows have not been dated. The
remaining four of the five multiphase ponds (Cam0ens,
Haldane,Kiess, and TassoS in Smythii basin) have each been
divided into two distinct deposits of differing ages and
eruption
histories
EI-Baz,
1977],
yielding
41
separable
volcanic[Wilhelms
deposits.and
The
deposits
noted
as Late
Imbrian/Eratostheniandark material (EId) in these four craters,
aswellasSmythiiW, Kastner
NEandWidmannstatten,
are
interpretedto be dark mantle deposits rather than effusive
eruptions
whose
thicknesses
canbeestimated
bythemethods
statedabove and are thereforenot consideredas mare ponds.
Although
it ispossible
thatthese
EIddeposits
contain
effusive
as well as pyroclastically emplacedelements, becauseof the
ZZ•
:•
11,138
YINGSTAND HEAD:SMYTHIIAND MARGINISBASINLUNARMAREDEPOSITS
m
n.
o
0
••0•0•000••••000•0•0000•
o oo
<<
<
<
<
•
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
l
•
rgtni
Ma sBasin
/
/
'
.• •
%
\
t '..'.:
•
.:"'
/
50 km
I
....
-,.............
IHubble
MareMargmls
•
• •oliot
•3
20øN
• ..•
•:.:.'
4 / %Al-Biruni
........
.:
10
..... :::
::::.;:"'•
qodlard
I
:......::.'.
11 "':'•
::.•.:"
:..•,
....
':::::':
••6
'":'::"?
............
•:::....::"
• •n Yunus
12
:;".•:•.•.:.•.:•.::•...•..
'"'"'
'"":":'
......
' 5'"•
,..x....
;•7
...
..'
10øN
/
I
/
/
7•
.:..:...,..:
),)
.......
?'"': 1 •x
8
-
t
"•:•'"::
/
2
Nepe•
•
•
•
•
/ • .........
.:.:::.:..::.
..............
•::
........
•................
.::..• /
f
/
.......
?:
••: ::'
.....
...•.•..
ß
.................
...xt I
...
':"• •
••
:'.
.....
•:.
M•e Smythii
I 1112v:;;:•,
'd•½::..•:.
•"•<.:"
:-:::
............
:::•:.:..:..
18
':'-•-........:'
..............
•.......
.?:'..•:.'/
.....
ß......
1•3
4x)
X•
•_
/
o
"...........
7", "?
....
W• •__•
22k •2• •
Hira
YamaI
Kfistner
10øS
!
!
Smyii Basin
80øE
Im 1
I'%! Im2
I
I
90OE
100øE
Eld
Floor-fractured
crater
Figure 2a. Sketchmap of Smythii/Marginis basin region showing the location of mare deposits.Ponds
studiedare indicatedandcorrelateto numbersin Tables 1 and 2. Older Late Imbrian mare material (Im0 is
indicatedin black. YoungerLate Imbrian marematerial (Im2) is shown in grey, while striped areasdenote
Eratosthenian/Imbrian
dark mantlematerial.Floor-fractured
cratersare indicatedby a dot pattern. Ages noted
for maredepositsare baseduponHiesingeret al. [1997] and Wilhelmsand El-Baz [ 1977]. Regionswithin Mare
Marginis dated through crater size-frequencydata by Hiesinger et al. [1997] that are discussedin the text are
noted by Roman numerals.
11,139
11,140
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
Smythii
Marginis
30 ÷ -
15
10
-5
-10
-15
-75
-80
-85
Longitude
-90
-95
-100
(Degrees)
:
-4000
-2000
0
2000
Figure 2b. Airbrushmapof the Smythii/Marginis
basinssuperposed
ontothe Clementine
altimetrydataof
theregion(basedon datafromZuberet al. [1994]),showingthetopography
andsettingof thebasin.
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
mantling nature of pyroclastic deposits, it is difficult to
'distinguish betweenmaterial emplacedby effusive flow and
material emplacedby pyroclastic means. Consequently,if
sufficient pyroclastics exist in a region to create any
ambiguitywe do not treat the deposit further as an effusive
eruptiveepisode.The natureof thesedark mantle depositswill
be discussedin more detail below. Thus, on the basis of their
characteristics(Tables 1 and 2) 34 out of 41 separablevolcanic
deposits and flows in Smythii and Marginis basins are
classified here as mare ponds and of these, 33 ponds are
interpretedto representindividualeffusiveeruptivephases.
Areas for the 34 mare pondsin the Smythii/Marginisregion
are shown as an area-frequencydistribution in Figure 3.
Estimates of individual eruptive phases are shown in Figure
3b. Joliot, which is assumedto be a multiphase pond whose
individual flows are not currently measuredindividually, is
excludedfrom this figure.
11,141
cratersbasedon crater diameters[Pike, 1977• 1980]. The depth
of excavationwas then determinedfrom this value [StOffler et
al., 1975] and the pond thickness was estimated, where dark
ejecta indicates a minimum depth and bright, presumably
highland ejecta indicates a maximum depth for the mare
deposit. Crater depth/diameterrelationships derived by Pike
[1977] were also usedto estimate the geometry and depth of
fill for flooded or partially embayed craters, where such
existed, so that the thickness of the embaying unit could be
determined.For ponds lying on relatively fresh crater floors,
the measureddiameter of the floor was compared to the
calculated floor diameter of the original crater based on
depth/diameter measurements[Pike, 1977]. The difference
between
these two values was then used to derive a thickness
estimate[Whitford-Stark, 1979]. Other indicatorsof deposit
thickness
used were
shadow
measurements
on
flow
fronts
where available, and partially buried topography other than
craters,
where the elevation of the pre-existing topography
Pondareasrangefrom170 km2 to 6575 km2 (meanvalue
1120 km•) but areconcentrated
in the lowerportionof this was known from topographic maps or other sources.For a
range. For those pondsjudged to be single eruptive episodes, more comprehensive treatment of these pond thickness
areasalsorangefrom170km• to 6575km•, with a meanvalue estimate methods the readeris directed to Gaddis [1981] and
of 965 km•. The total area of all mare ponds in the
Smythii/Marginisregion is approximately38,015
Lava pond areas, Smythii/Marginis basins
representingless than 5% of the total area of these basins.
Coverage by lava ponds is higher within Marginis basin.
About21,950 km•, or 8% of the Marginisbasinarea,is
covered
with lavaponds,compared
to 16,065km•, or 3% of
the area of Smythii basin. Mare coverage for this region
(ponds
andmajormaria)isabout102,845km•, or 13%of the
•Medi•n!
i
i
i
157--_•---Mean
....i.............
i............................
ii.............
-i...........
total area in the two basins. For Marginis this equates to
approximately
56,895 km•, or 22% of the basin,while for
Smythiitotalmarecoverage
is 45,950km2,whichis 8% of the
basin area. Thus, effusive mare deposits in these two basins
make up approximately 13% of the total area of the basins,
with ponds representing about 5% and the maria about 8%
(Figure 2a).
Pondsin the Marginis region tend to be concentratedin the
southernhalf of the basin,the only exceptionsbeing the large
craterfloor ponds Joliot and Hubble (ponds 1 and 2). Pond
lO
-
0 i•
....
0
1000
depositper 171,000 km•, the depositdensitywithin the
centralbasinring(area-120,000 km•) is oneper6,600 km2.
In general terms, the mean pond density for the entire
3000
I iII
4000
I I ii II ', i,,,',
5000
6000
I
7000
Area (kmx)
densityfor Marginisbasinis 1 pondper20,000km•, whilefor
thesouthportionof the basinit is 1 per 12,000 km•. For
Smythii, deposits that occur within the central region are
relatively evenly spacedin the confines of the central ring,
while those ponds occurring outside this region are
concentratedin the northwest portion of the basin. Thus,
while the pond density for the basin as a whole is only one
i ,, I, i I,,
2000
Individual eruptive areas,Smythii/Marginis basins
20 I,,,,I,,,,I,,,,
:
edi
I ....
I ....
I ....
I,,,•
b
n
15'Miear,
.......
:.............
:.............
...............
r.............
."
............
Smythii/Marginis
regionis aboutonepondper 121,000
but the concentrationof mare depositsappearsto be relatedto
the state of preservation of the basin. In the better preserved
Smythii basin, deposits are highly concentratedin the center
of the basin, while in Marginis basin the distribution of
depositsis muchmore diffuse.Deposits in Marginis are not
containedby the highly degradedtopography of the central
2000
3000
4000
5000
6000
7000
0
1000
depressionor basin rings.
Area
(km
2)
Volumesof pondsand individualflows were estimatedusing
the variousmethodsdescribedin Gaddis[ 1981] and Yingst and Figure 3. Area-frequencydistribution (a) for all mare ponds
Head [1997a]. Specifically, in those caseswheresuch craters and (b)for those ponds and flows estimatedto be individual
are available, pond thicknesseswere estimatedby calculating eruptive phases. Mean and median values are indicated by
the depth of relatively young, optically mature post-mare labeled arrows.
.....
_
11,142
YINGSTAND HEAD:SMYTHII AND MARGINISBASINLUNARMAREDEPOSITS
Yingst and Head [1997a]. The techniquesusedfor each lava
pondare indicatedin Tables 1 and 2, ar•.dthe compileddata are
shownin a frequencydistributionplot in Figure 4.
The total calculated volume for mare ponds in the
Smythii/Marginis
regionis 6755 km3, where3445 km3 is in
Smythiibasin,and3310km3 is in Marginisbasin.As shown
15
:
-- ..•
km3 to 1045 km3, with an averagevalueof 200 km3. The
:
:
i
!
!
::
.
i
:
:
:
'.
.
:
:
ß
.....
i.............
',.............
i..............
.............
i..............
i...........
:-
largest volumes occur within the large craters (e.g., Neper,
Kiess, and Helmert-Kao). In the caseof those ponds or flows
estimatedto be individual eruptive events, volumes also range
from15 km3 to 1045 km3, asshownin Figure4b. The total
I
volume of ponds in Smythii basin is estimated to be -3450
km3,or about19% of the total volumeof marematerialin the
0
basin. Similarly, the total volume of ponds in Marginis basin
is estimatedas -3305 km3, or 12% of the total volumeof basin
The mean value for volumes of individual
floor areas
areas
: ': :"--"::•
I Crater
h•ter-crater
:
-- ; ........
•.............
•.............
-i..............
i.............
i..............
::.............
g
in Figure 4a, the rangefor pond volumes in both basins is 15
mare material.
'
i I
:i
-•
,
1000
', i
2000
3000
4000
5000
6000
7000
Area (km2)
Figure
5. Area-frequencydistribution plot for mare ponds
eruptive
phases
for theregionasa wholeis 195 km3 (190 km3 occurring within crater floors and inter-crater highland
averagefor Smythiibasinand270 km3 for Marginisbasin). regions.
Suchaveragevalues bracket the mean volume for individual
eruptive phases in the Orientale/Mendel-Rydberg basins
(about240 km3[YingstandHead,1997a]).
3.3.
Lava pond volumes, Smythii/Marginis basins
!
•
!
!
:
:
:
:
:
Modes
of
Occurrence
We noted two modes of occurrencefor mare ponds: (1)
within crater floors, and (2) in inter-crater highland regions.
These are shown in terms of area- and volume-frequency
distribution plots in Figures 5 and 6. For the
Smythii/Marginis region as a whole, the majority of ponds
and flows occuron craterfloors (24, representing71% of all
occurrences).In the caseof single eruptive episodes,most of
the discernible flows (23 out of 34) lie within craters. In
6
- '
300
600
900
1200
1500
Volume (km3)
Individual eruptive volumes,Smythii/Marginis basins
15 • , , • , , • , , I , , • , , • , , • , , •
-•Median
i i
10
i
i
i
I , , • , ,.
i
i
ib
.............................................................................
,
:
''1''1''
300
600
900
1200
1500
Volume(km3)
Figure 4. Volume-frequencydistribution (a) for all mare
ponds and (b) for those ponds and flows estimatedto be
individual eruptive phases. Mean and median values are
indicatedby labeled arrows.
addition,crater-flooroccurrences
tend to representthe highest
volume mare ponds.Thus a regionalpreferencefor craterfloors
is observedboth for mare pond occurrenceand total volume
extruded.The majority of these crater floor ponds lie in the
relatively young (Late Imbrian-aged) cratersthat dominatethe
floor of Mare Smythii, suggesting that these ponds may
representa later stage of volcanism, an observation that we
will returnto later. Of the craterfloor pondsthat occurin older
craters, eight (33% of crater floor occurrences)lie in preNectariancratersoutsidethe central basin, and only .4 ponds
(17% of crater floor occurrences)occur in Nectarian-aged
craters. In contrast, 12 ponds, or almost 50% of crater
occurrences,lie within Imbrian-agedcraters.Two such ponds
(ponds 10 and 11 in Smythii basin; Figure 2a, Table 2) are
shownin Figure 7. The cratersin which these ponds occurare
characterizedby updomedfloors, moats, and a system of
subconcentric fractures, previously classified by Schultz
[1976] as type III floor-fracturedcraters.It has been suggested
that this type of upliftedand fracturedcratermorphologyis due
to endogenic modification dueto viscous relaxation of crater
topography over time [Danes, 1965; Baldwin, 1968; Hall et
al., 1981] or through surfacefailure in responseto igneous
intrusion [Schultz, 1976; Brennan, 1975; Wichman and
Schultz, 1995]. Eruption of magma has resulted in partial
burial of severalof the floor fracturesin Cam6ens.Similarly,
the pond in Doyle appearsto have partially buried sections of
the fracture system in the southwesterncrater floor. Mare
emplacement subsequent to fracture formation is thus
suggestedfor both ponds.
Two separatevolcanic episodesare discerniblein Cam6ens,
distinguished by differences in albedo, texture, and crater
density.The northerndeposithas fewer cratersthanthe flow in
YINGSTAND HEAD:SMYTHII AND MARGINISBASINLUNARMAREDEPOSITS
11,143
The diffusedark material associatedwith the northern deposit
Inter-crater volumes
Crater floor volumes
8
appearsto disregardboundariesformedby the preexisting
topography. Thus, the above morphology suggests a
pyroclasticemplacement
mechanism
for this depositrather
than an effusiveone [Lucchitta, 1972; McGetchin and Head,
1973; Pieters et al., 1973, 1974; Wilhelrns, 1987]. A 15 km
long linear rille extendsfrom the northwestern
edgeof the
pyroclastic
materialin Cam6enscraterto the southern
rim of
0
200
400
600
800
1000
t
Doyle
1200
Volume (km3)
N
i•
Figure 6. Volume-frequencydistributionplot for mare ponds
occurring within crater floors and inter-crater highland
regions.
":'"'."
the centralportion of the crater,a fact that led Wilhelmsand
El-Baz [1977] to interp.ret
thisdepositas younger.In addition,
the northerndepositdisplays a very low albedo and a more
diffuseboundarythan the centraldeposit,suchthat while the
centraldepositrepresentsan effusivevolcanic eruption, the
northerndepositis denotedas a darkmantledeposit[Wilhelms
linear
rill•i
second
rim
andfracture
system
•l
and EI-Baz, 1977]. The northern deposit extends beyond the
crater rim into the surroundinghighlands, forming finetexturedlobesof a moreruggedrelief than the centraldeposit.
moat
•
/
I
.D ø
10 km
,
,I
,
85øE
Map Units
YoungerLate Imbrian mare material
Late
Imbrian/Eratosthenian
dark
material
Crater
rim
Fracture
Linear
rille
Figure 7b. Geologic map of ponds 10 and 11 in Smythii
basin, noted in Figure 2a. The upper Late Imbrian mare
material of the pondsis shownin grey while stippled areasare
noted
Figure 7a. Lunar OrbiterphotoLO IV 18 of ponds10 and 1 1
in Smythii basin, noted in Figure 2a.
as Late
Imbrian/Eratosthenian
dark
mantle
material
[Wilhelms and EI-Baz, 1977]. Crater rims are outlined by
dashed lines. The Imbrian-aged mare plains have flooded
significant portions of the floors of the Imbrian-agedcraters
Doyle and Cam6ens, while the dark mantle deposit has
blanketed the northern portion of Cam6ens crater. A 15 km
linear rille cuts across the dark mantle deposit in Cam6ens
crater. Several subcircularfractures within eastern Doyle and
along the rim of Cam6ens are evidence of tectonic evolution
before or concurrentwith mare emplacement.
11,144
YINGST AND HEAD: SMYTHII AND MARGINISBASINLUNAR MARE DEPOSITS
Doyle, cuttingthroughthe youngerCamOensdarkdeposit.In
contrastto this pair of deposits,the pond in Doyle crater is
10 km
smaller in both area and volume, has a higher albedo and
showsa roughertexturethanth.ecentralflow in CamOens.The
mare material approximately follows the concentriccontours
N
of the crater rim, but is confined to the crater's western
portion. No other featuresaside from crater floor fractures
appearto be associatedwith this pond.
On the basis of the precedingobservations,the history of
this region may be reconstructed.Some time during the
Imbrian period impa.c.
ts createdthe initial cratersDoyle and
CamOens. Fracturing of the crater floors followed. Mare
material was then emplaced,embaying some fracturesin both
cratersand filling the floor of CamOens.A later episodeof
pyroclastic activity occurredin the northern moat region of
CamOens, manfling a portion of the crater rim and the
surroundinghighlands. Either concurren,twith, or after this
last volcanic episode, a linear rille formed from the
northwesternrim region of CamOensto the southernrim of
Doyle crater.
The remainderof ponds in Smythii/Marginis (10 deposits,
or 29% of all occurrences) lie in inter-crater regions,
apparently unrelated to any other formation or feature. An
example of such a pond is located north of A1-Biruni crater
(pond 3 in Marginis Basin; Figure 2a), shown in Figure 8. In
the caseof this pond, younger Late Irabrian mare material has
flooded a region of lower-elevation furrowedhighlands north
of A1-Biruni crater [Wilhelrns and EI-Baz, 1977], embaying
several small cratersand burying the hummockypre-existing
topography.There are no vents or other structuresvisible that
would suggestthe sourceof the mare material.Instead,effusive
eruption of mare material appears to have buried any
responsiblesourcevents. Smaller craters on the surfaceof the
deposit are evidenceof younger (post-Late Irabrian) impacts.
No other outstandingfeaturesare presentin this mare deposit.
Thus the morphology and placementof this pond in a region
of low elevationmake this deposit typical of pondsoccurring
in inter-crater areasin Smythii and Marginis basins.
20øN
92øE
Map Units
YoungerLateImbrian marematerial
Crater
rim
Figure 8b. Geologic map of pond 3 in Marginis basin,
shownin Figure 2a. Mare materialis shownin grey; cratersare
notedby dashedlines. Late Imbrianmaterialhas coveredthis
inter-craterregion, embayingthe local highlandsand small
craters.
On the basis of these observations,the geologic history of
this pond may be determined.Formation of the many large
surroundingcratersoccurredfirst, including the formation of
the Nectarian-agedA1-Birunicrater to the south. Formation of
the furrowedinter-craterplains occurredduring or immediately
after this period, both in this region and in central Smythii
basin.This was followed by emplacementof the mare pond by
effusive volcanism. Finally, subsequentto emplacement of
mare material, continued impact activity formed the small
younger craters in the mare pond and the surrounding
highlands.
For the Marginis basin, the majority of ponds (seven
ponds,or 58% of occurrences)lie in inter-crater regions, with
areasrangingfrom 175km:to 3595km:,andvolumes
ranging
from 20 km 3 to 715 km 3. The remainder occur in crater floors
(five ponds,or 42% of occurrences),displaying areasranging
from290 km• to 6155 km•, andvolumesin the rangeof 60
km3 to 710km3.In contrast,
pondswithinSmythiibasintend
.
.....i..•.':..:.
'....., "•::'"•
..::•.--...•....::;..
,- • % -:-.
,-.- .•:..•
.....
.
.-;i:
:--".....-'
...
...•;:
% -.:'•:•,
.
.
to occur most frequently in crater floors (19 ponds,
representing86% of occurrences),
with areasrangingfrom 170
km• to 6575 km•. Volumesfor craterfloor occurrences
range
from15 km3 to 1045 km3. Threeponds(14%of occurrences)
Figure8a.Lunar
Orbiter
photo
IV 165ofpond
3 inMarginislie in
basin, shown in Figure 2a.
inter-crater regions, showing areas ranging from 255
km• to 1675km•, andvolumes
in therangeof 25 km3 to 250
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
11,145
km3.All multiphase
pondsoccurwithincraterfloors.The pond pond. Mare ridges are seen in Mare Smythii and Mare
associatedwith crater Haldane (pond 12 in Table 2) breaches
the crater rim and displaysmaterialboth on the craterfloor and
in the inter-crater region to the north. However, because
topographic evidence suggeststhat the sourcevent for this
pond lies within the interior of crater Haldane,this pond is
classifiedas a crater floor occurrencein Figures5 and 6, and in
Table
2.
3.4.
Associated
Features
Linear rilles, mare ridges, dark-halo craters, and floorfractured craters are among the variety of features found
associatedwith Smythii/Marginis region lava ponds, as noted
in Tables 1 and 2. Examples of these features are labeled in
Figure 9. For example, linear rilles are seen in Neper crater
Marginis, as well as in the crater Joliot. They tend to be
associatedwith large, contiguousmaria. Clusters of dark-halo
impact craters lie within the non-mare units north and east of
Mare Marginis, and circumferentialto Mare Smythii [Schultz
and Spudis,1979, 1983]. As was noted before, several ponds
are also associatedwith floor-fracturedcraters(e.g., Figure 7
[Schultz, 1976]).
Linear rilles (simple graben) occurredin four ponds (12%);
all of these occurrenceswere in craters(e.g., Cam/3ens,pond
11 in Figures 2a and 7). Volumes for these ponds range from
40 km3to 780 km3, with an averagevolumeof 300 km•. The
nonarcuate nature of these linear rilles, along with their
association with mare deposits, favors a volcanic mode of
origin, suchas the near-surfaceemplacementof a dike [Pollard
et al., 1983; Head and Wilson, 1993].
Mare ridges were observedin the crater Joliot, as well as in
Mare Smythii and Mare Marginis, the largestvolume deposits.
Dynamic models for the formation of these ridges [e.g.,
Melosh,1978;Solomonand Head, 1980;Pullanand Lambeck,
1981; Golombek, 1985] tend to attributeridge origin to the
stressesplaced on the lithospherefrom large volumesof mare
material.
Such a conclusion
is consistent
with
the observed
association of mare ridges in Smythii and Marginis basins
with the largest and thickest mare deposits.
Dark-halo impact cratershave been mappedby Schultz and
Spudis [1979, 1983] in highland regions surroundingMare
Smythii and Mare Marginis. Specifically, these cratersappear
to be concentratedwithin central Smythii basin and in the A1Khwarizmi/King and Lom.onosov-Flemingbasins east and
north of Smythii respectively. Such cratershave been usedas
indicatorsof hidden volcanicmare material,termedcryptomare
[Headand Wilson, 1992], that has been buried by basin or
craterejectadeposits[Schultzand Spudis, 1979, 1983; Hawke
and Bell, 1981; Bell and Hawke, 1984; Antonenko et al.,
1995]. The presenceof clustersof dark-halo craters suggests
that there was early volcanic activity in this region, the nature
of which is not fully representedby the surfacemare deposits.
In addition, elevated concentrationsof Mg, Fe and Ti, as well
as low A1 concentration in some portions of A1Khwarizmi/King [Clark and Hawke, 1991] suggest the
presence of a partially obscured basaltic component
[Schonfeld and Bielefeld, 1978; Hawke et al., 1985]. Thus
volumes indicated by mare ponds should be considered a
minimum value for the total mare volume in this region.
ß
:•-•.:•.
'::::•--•.i•;;•
.... ..*...--,..-:?
..%./.-;..'
There
fractures
Figure 9. Examples of featuresfound associatedwith the
Smythii/Marginisregion lava ponds. The Lunar Orbiter photo
LO IV 18 (Figure 9a) shows a linear rille within Neper crater,
Smythii basin (pond 1 in Figure 2a and Table 2), while the
Apollo 15 photoAS15-95-12991(Figure9b) showsexamples
of a mare ridge (MR) in Mare Smythii and a dark-halo crater
(DHC) nearMare Smythii. See Tables 1 and 2 for a summaryof
feature
locations.
are
several
floor-fractured
craters
in
the
Smythii/Marginis region [Schultz, 1976; Wolfe and EI-Baz,
1976; Wichman and Schultz, 1995]. Many of these craters
have lava ponds associatedwith them. Eight ponds (21% of
occurrences)occur in floor-fractured craters (e.g., Haldane
[Wolfe and EI-Baz, 1976], Doyle and Cam6ens, Figure 7). The
associated
with
these craters
are concentric
to the
crater rim, implying tectonic modification of the original
crater such as uplift of the local topography by some
mechanism [e.g., Schultz, 1976], rather than a strictly
endogenicorigin suchas dike.emplacement[e.g., Pollard et
al., 1983; Head and Wilson, 1993]. The nature of floorfractured crater formation
3.5.
will be discussed in more detail later.
Stratigraphy
Initially, the volcanic deposits of the Smythii/Marginis
regionwere dividedby Wilhelmsand EI-Baz [1977] into older
YINGST
ANDHEAD:
SMYTHII
ANDMARGINIS
BASIN'LUNAR
MARE
DEPOSITS
11,146
(lower) and younger (upper) Late Imbrian mare materials, and
Late Imbrian/Eratostheniandark material, displaying albedo of
mareor lower (Figure 2). As noted previously, Wilhelrnsand
E1-Baz [1977] have identified these latter deposits as dark
mantle deposits.The very low albedo, fine textureand rugged
relief of these deposits, along with the diffuse boundaries
displayedby many of them (e.g., Cam6ensin Smythii basin,
shownin Figure 7), are all consistentwith a pyroclasticorigin
[Lucchitta, 1972; McGetchin and Head, 1973; Pieters et al.,
noted that age determination for the earliest reference
[Wilhelrns and E1-Baz, 1977] was basedpartly on differences
in albedo.The designationof sevenpondsby these authorsas
Late Imbrian/Eratosthenian (EId) was due to their relative
albedo, lower than other Late Imbrian-aged mare deposits.
However, the low albedoof thesedepositsis most likely due to
their pyroclastic origin rather than youth. Thus for these
depositsalbedowouldbe suspect
asan indicatorof age.
Craterage-frequency
dataaccumulated
recently[Hiesingeret
1973, 1974; Wilhelrns, 1987]. We thus interpret the seven al., 1997] suggestthat a great dealof volcanismin Marginis
depositsandflows notedasEId by Wilhelrnsand E1-Baz[ 1977] basin occurredin a relatively limited timespan, ranging from
to be pyroclastic deposits of upper Late Imbrian or possibly
Eratosthenianage. On the basis of albedo characteristicsand
the density of small superposedcraters,the depositshave been
furtherdividedby Wilhelrnsand E1-Baz[ 1977] into 41 separate
and datablelava ponds, dark mantle depositsor flows. These
41 volcanic deposits are shown in an age-frequency
distribution plot in Figure 10. Hiesinger et al. [1997] have
datedseveralpondsand mare regions in Marginis basin using
cratersize-frequencydistributionmeasurements.
Theseagesare
displayedin Figure 10 as well. For thosepondswhoseagesdo
not agree between the two studies (three ponds in Marginis
basin), datesderived from the crater size-frequencydistribution
study [Hiesinger et al., 1997] have been usedrather than the
agessuggestedby Wilhelrnsand E1-Baz[1977]. It shouldbe
'•EasternMar•;inis
(III;3.59
Ga)]
3.65
"--
Southern
Marginis
(II;
3.65
Ga),
]
Goddard(3.66Ga)
•Central
Mar•inis
(I;3.70
-3.62
3.70•oliot
u•
(3.76
- 3.65Ga)]
25
3.75
ß•
•Ibn
2c
Yunus(3.78Ga)]
3.80"-- .•Hubble (3.80
Ga)l
•
15
z
Eratosthenian-Imbrian
darkdeposits
[ Wilhelmsand EI-Baz,1977]
ß
ImI
..
Im2
Imbrian
-3.80-3.60
Ga. Based on the available
Time before present(Ga)
Figure 10. Distribution of mare pond ages within the
Smythii/Marginis basin area.The inset stratigraphiccolumn
represents stratigraphic ages assigned by Hiesinger et al.
[1997] on the basis of crater size-frequencydistributions for
someMarginis pondsand portionsof Mare Marginis indicated
by Romannumeralsin Figure2a. Note that the clusterof ages
around~3.80-3.60 Ga representsregions widely separatedin
spaceand that in the caseof the pondin Ibn Yunus and areasof
Mare Marginis directly south of this crater, neighboring
regions display very disparateages.
it
is
Hiesingeret al. [1997] may be typical of the entire periodof
volcanic activity in Marginis. In addition, becausethose
pondsin Smythii basin datedas EId on the basisof albedo
differences[Wilhelrnsand E1-Baz,1977] are likely dark dueto
compositionratherthan age, it is possiblethat thesedeposits
mayhave beenformedexclusivelyin the upperLateImbrian.
Thusthe weightof evidencesuggeststhat, for Marginis basin
specifically,andby extrapolationthe region in general,the
majorityof volcanicactivitywas in the latter part of the Late
Imbrian period.
Out of 41 individually mappedvolcanic depositsand mare
flows within the Smythii and Marginis basins, most (31
deposits,or 76%) are datedas Im2 (youngerLate Imbrian)
[Wilhelrnsand El-Baz, 1977; Hiesinger et al., 1997]. Of the
remaining10 deposits,two (5% of occurrences)
areIm• (older
Late Imbrian), seven (17%) are Late Imbrian/Eratosthenian
dark mantle depositsand one (in Hubble crater, pond 2 in
Marginis basin, Figure 2a) is dated as 3.80 Ga,
contemporaneouswith the Early/Late Imbrian division. The
oldest ponds (Imp) are confined to the eastern portion of
Marginis basin, while the dark mantle deposits (EId) rim the
central interior of Smythii basin, occurring within, or
mantling the rims of, several of the prevalent Imbrian-aged
craters there. For the Marginis basin area, two of the 12
depositsare Im• (17%) andnine (75%) are Im2, while the pond
occurringin Hubble crater (pond 2 in Marginis basin, Figure
2a) straddlesthe Early/Late Imbrian periods. Likewise, all but
one of the depositsin Smythii basin are datedeither Im2 (21,
accountingfor 76% of deposits)or Late Imbrian/Eratosthenian
(seven deposits, or 24% of occurrences).Only one (Erro NW,
pond2 in Figure 2a, Table 2) is datedImp. This distribution of
observedsurfacedeposits implies that volcanism was active
over a similar time period for both basins,with volumetrically
small activity in the early part of the Late Imbrian and more
volumetrically significant activity during the latter part of the
Late Imbrian. Such activity is consistent with the volcanic
flux measured for the Moon
Eratosthenian
evidence,
reasonableto assumethat the stratigraphic ages indicated by
as a whole. Volcanic
flux estimates
basedon lunar eruption rates [Hartmannet al., 1981] suggest
that > 90% of the volume of known volcanic deposits were
eraplacedin the Late Imbrian period (3.80-3.20 Ga), peaking
at-200 Ma into the Late Imbrian period(around3.60 Ga [Head
and Wilson, 1992]), while < 5% was emplaced in the
Eratosthenian period [Head and Wilson, 1997]. The
distributionof volcanism during peak activity was widespread,
involving most of the large nearsidebasins, including nearby
Crisium [Wilhelrns, 1987]. Thus, basedupon the ages of the
surfacemare deposits,we concludethat the most volcanically
active period in Smythii/Marginis basins (the Late Imbrian)
coincides with the most widespreadand volcanically active
time in lunar history.
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
3.6.
Distribution
Nearest-neighbordistanceswere measuredfor the 34 mare
pondsin Smythii/Marginisby takingthe averagevaluefor the
measured
center-to-center
distanceof the five closestdeposits
to eachpond, after Yingst andHead [1997a]. A frequency
distributionplot for the averagenearest-neighbor
separation
distances
in the Smythii/Marginisregionis shownin Figure
11. Center-to-center
separationdistances,shownin Figure
1la, rangefrom50 to 250 km andaretypicallyin the rangeof
60-150 km, while edge-to-edge
separationdistances,
shownin
Figure 1lb, rangefrom 15 to 160 km, with a mean value of 6 5
km. Theserangesaresimilarto thoseobservedfor pondsin
11,147
thickness may be analyzed. Mean values for center-to-center
distances, shown in Figure 1l c, are estimated to be 85 km
within the inner basin ring and 160 km between the inner and
outer rings. Mean values for edge-to-edgedistances,shown in
Figure 1ld, are calculatedto be 40 km in the basin center and
110 km outside the basin center. The fact that ponds are not
only more frequentbut are also more closelypacked within the
basin center suggests that pond distribution and spacing is
dependentto someextent upon topography and the associated
crustal
thickness.
4. Comparison and Global Context
the South Pole-Aitken and Orientale basins on the western
The significant numberof lava pondsanalyzedin this and
other studies[e.g., Gaddis, 1981; Yingst and Head, 1997a;
Smythii basinmay be dividedinto two groups;those within Hiesinger et al., 1996] permit comparisons to be made
the confinesof the centralring (14 deposits)and those outside between the characteristicsof ponds in disparateregions,
the centralbasin ring (eight ponds),so that comparisonof allowing improved modeling of lunar eruptive commonalties
distributionwith respectto varying topographyand crustal in a global sense. Values for these characteristics are
limb and farside [Yingst and Head, 1997a]. Ponds within
Center-to-centerdistancesfor lava ponds
in Smythii/Marginis basins
Center-to-center distances,
Sm•thii
basin
lava
>onds
,•,,l•,,,I,,,,I,,,,I,,,,I,,,,l,,,,I,,,,
12
.... ' .... M;li'h'
10
' ' ' .... ' .... ' .... ' ....a
ean
_
0
50
100
150
200
250
300
350
400
Distance (km)
Distance (km)
Edge-to-edgedistancesfor lava ponds
in Smythii/Marginis
ba•sins
12
Mediah
, , , , I
Edge-to-edgedistances,
Smythii
basin lava ponds
, , , , I , , , , I i , i , I , • , • I , , , , I , ,
: : ::::.:..:..:..L.L.'L."..'L.:..'LL.L."..L.::
10-
_
, , I , , , ,
b
: :::
: :::
::
:::
..........
..........
t...........
!............
'i...........
i'...........
i......... ]--•-i--i-i--i-•--•-•--i-i--i-i--i-•-i-• Interior
,
,
,
_
0
0
:-':..:..L_
50
100
150
200
250
Distance (km)
300
350
400
Distance (km)
Figure 11. Frequencydistributionplot of averagenearest-neighbordistancesmeasuredfor ponds_in the
Smythii/Marginisbasin region. Center-to-centerdistances(Figure 1l a) are larger than edge-to-edgedistances
(Figure l lb) becauseof the irregular shape of many of the ponds examined. Also shown are frequency
distribution plots of center-to-center(Figure 1l c) and edge-to-edge(Figure 11d) distancesfor ponds within
Smythii basin alone, demonstratingthe differencebetweendepositsin the interior of the basin and those in
the exterior rings.
11,148
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
summarizedin Table 3. As has been previously stated, we
interpretthe 33 flows and pondsnoted as individual phasesin
Tables 1 and 2 to be our current best estimate
of individual
eruptive events.It is this subsetof depositswith which we will
now be concerned.
4.1.
Areas
and
Volumes
As statedpreviously, the mean areal extent of those lava
ponds in Smythii/Marginis basins which are estimated to be
thebestcandidates
for individualeruptiveepisodes
is 965 km:,
while the averagevolumeis 195 km3. This comparesto
approximatemeanareasof 1115 and 2080 km2 and mean
volumesof approximately
240 and860 km3 for Orientaleand
SouthPole-Aitken basins, respectively. A volume-frequency
distribution plot for these deposits is shown in Figure 12.
From these data, it is
clear that
South Pole-Aitken
is
characterized
by a widerrangeof values, with typically higher
volumes for individual ponds, while both Orientale and
Smythii/Marginis have a narrower range of values, with
occurrences
peakingat lowervolumes(lessthan150kmS).
In all basins,however,mare pondsdisplayvolumesthat are
high by terrestrialstandards.For example, typical flows for a
single eruption from a shallow sourceregion such as Hawaii
averagelessthan 1 kms per eruption[Peterson
andMoore,
1987]. On the otherhand,the Laki eruptionin Iceland, one of
the largesthistoricalterrestrialeruptions,was measuredat -12
kms [J6nsson,1983], whichis comparable
to the lowest
volumes observedin this study. A more striking terrestrial
comparison can be made using flood basalt provinces,
volcanic regions believed to be associatedwith deep-seated
sourceregions [Campbelland Griffiths, 1990]. For example,
the Roza Member of the Columbia River Basalt province has
an estimated
volumeof 1200 kms [Tolanet al., 1989]. Such
large volumessuggestby comparisonthat deep, rather than
shallowsourceregionsare indicatedin the basinsstudied.
4.2.
Morphology
and Associated Features
The morphology and structures associated with mare
depositsare indicators of the conditions that existed during
magmaextrusion. The typical morphology for lava ponds in
this region is similar to that observed in South Pole-Aitken
and Orientale basins.Pondstend to be relatively smooth, with
no domes, calderas, or other similar structures evident.
Features are limited to linear rilles, dark-halo craters, and
tectonic
structures associated
with
floor-fractured
craters.
We
will examine each of these features in detail below.
4.2.1.
Linear
rilles.
Linear rilles
were found to be
associatedwith four depositsin the Smythii/Marginis region
(12% of pond occurrences).
The meanvolumefor theseponds
is 300 km3,whichis higherthanthemeanvaluefor deposit
volumesin this region. The numberof pondsassociatedwith
linear rilles in Smythii/Marginis is higher than that for lava
pond occurrencesin South Pole-Aitken basin (two occurrences,
or 4% [Yingst and Head, 1997a]), an observation attributable
to the poor resolution and viewing angle of the images
available for the previous study. In contrast, linear rilles
occurredin -- 35% of deposits examined in Orientale basin
[Gaddis,1981; Yingst and Head, 1997a], generally being
found in depositswith the highestvolumes.Linear and arcuate
rilles have been interpretedto be linked to impact basin
structure[Mason et al., 1976] and the emplacementof mare
deposits in the basins via flexural deformation related to the
•
o
•
o
YINGSTAND HEAD:SMYTHII AND MARGINISBASINLUNARMARE DEPOSITS
11,149
Individual eruptive episodes,lunar limbs and farside
50
i
Smythii/Marginis ponds
South Pole-Aitken ponds
Orientale ponds
40
•
ß
3o
•
¸
20
i
i
ß
ß
ß
.
10
ß
,
i
ß
0
500
1000
1500
2000
2500
.
3000
3500
Volume(km3)
Figure 12. Volume-frequencydistributionplot for lava pondsinterpretedto representthe best candidatesfor
individualeruptivephases(seeYingstand Head [1997a]andTables1 and2). Thereare 39 depositsnotedin the
South Pole-Aitken basin region, 31 in the Orientale/Mendel-Rydberg basin area, and 33 in the
Smythii/Marginis region.
maredepositload [Solomon and Head, 1980]. However, some
linear rilles have been interpreted on the basis of their
nonarcuateshapeand their association with volcanic deposits
to be the surfacemanifestation of a dike injected to nearsurfacedepths [Headand Wilson, 1993]. None of the five
linear rilles noted in this study follow the contours of an
impact structurerelatedto a mare deposit that would suggesta
flexural origin [e.g., Solomon and Head, 1980]. For these
linear rilles, then, the currentbest interpretationappearsto be
one in which graben form in response to local stresses
producedby the near-surfaceemplacement
of a dike propagated
from depth.
4.2.2.
Dark-halo craters. Dark-halo impact craters
have beenpreviouslymappedin variousnon-mareunits in the
northeast portion of Marginis basin and in southernSmythii
basin[Schultzand Spudis, 1979]. In general, thesecraterslie
within the central portion of Smythii basin, as well as within
the highly degradedpre-Nectarian[Wilhelrns, 1987] basins
Lomonosov-Flemingand A1-Khwarizmi/Kingnorth and east of
Smythii. The presenceof these dark-halo crater clusters is
indicative of the existence of cryptomaria in this region
[Schultzand Spudis, 1979; Hawke and Bell, 1981; Bell and
Hawke, 1984], suggesting that the onset of mare volcanism
might have been earlier than is currently indicated by the
inferredagesof the known surfacemaredeposits[Schultzand
Spudis,1979, 1983; Hawke et al., 1985]. If cryptomariai s
presentin the Smythii/Marginisregion as suggested,the total
mare volume is greater than that suggestedby the surface
deposits. We may thus use the size and distribution of darkhalo cratersboth to identify cryptomare and to estimate the
volumetricsignificanceof cryptornatedeposits[Schultzand
Spudis, 1979, 1983; Hawke and Bell, 1981; Bell and Hawke,
1984; Antonenkoet al., 1995]. For example,if we assumethat
the centralportion of Smythii basin was originally filled with
cryptornatematerial to a depth of-500 m (a depth similar to
that of the present Mare Smythii), this yields a cryptornate
volumeof all the mare depositsin Smythii basin analyzedin
thisstudy(18,390km3)represents
only27%of this potential
cryptomarevolume estimate.
As another example of the potential volumetric
significanceof cryptornatematerial, let us examinethe preNectarian
basins
Lomonosov-Fleming
and
A1Khwarizmi/King. We note the areal extent of cryptornatein
Lomonosov-Fleming
to be roughly65,000 km2, andin A1Khwarizmi/King
to be about80,000km2baseduponestimates
by Schultzand Spudis[ 1983]. If we then assumethat the
estimatedareaof cryptomarefor theseolderbasinsrepresents
a
fill of-500 m depth, we calculatea cryptomarevolume of
about32,500 km3 in Lomonosov-Fleming
basin and 40,000
km3 in A1-Khwarizmi/King
basin. Thesevaluesbracket the
cryptornatevolumeestimatedfor Smythii basin but are more
than twice the total volume of mare material currently visible
on the surface of Smythii. These first-order calculations
demonstratethe possibilitythat cryptomarematerial may be a
significant component in the eastern limb Smythii and
Marginis basins.
What could have obscuredthese potential cryptornate
deposits? One possibility is that cryptomare material was
obscured through local mixing by subsequent impacts.
Impactors that form craters deep enough to excavate the
underlying highlands may have contributed to mare soil
contamination through deposition of ejecta containing
highland material. This process has undoubtedly occurred.
Because the areal extent of proposed cryptomare regions
associatedwith southern Smythii, Lomonosov-Fleming and
A1-Khwarizmi/King are commensuratewith regions such as
Maria Smythii and Orientale [Schultzand Spudis, 1983], it is
likely that local mixing processes from small (<10 km
diameter) craterswould not be solely sufficientto obscurelocal
cryptomaria. However, several young (Imbrian-aged) craters
with diameters
> 25 km lie in and around these older basins.
These may have depositedsignificantamountsof ejecta which
volumeof approximately
70,000km3, significantly
morethan obscuredearlier (pre-Imbrian) mare deposits. Craters such as
either Mare Smythii or Mare Marginis. In fact, the total
Langemanand Lobachevsky in A1-Khwarizmi/King, Chang
11,150
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
Heng and Lomonosovin Lomonosov-Fleming,
and Haldanein
Smythii are good candidatesfor this process.
Anotherpossibility is that cryptomariawereblanketedby
emplacement of extensive ejecta deposits through basin
formation.
It is believed that the effects associated with the
Orientale basin-forming event may have reached--1500 km
away into SouthPole-Aitken basin [Wilhelms, 1987; Headet
al., 1993]. This suggeststhat basin-forming events have
potentially extensive effects. Thus nearby basins that formed
subsequent to Smythii, Lomonosov-Fleming, and A1Khwarizmi/Kingmay havebeenin close enoughproximity to
havecontributedto the obscuringof any previouslyemplaced
cryptomaria. For the caseof Smythii basin, there are several
Nectarian-aged
craterswhichmight have contributedmantling
ejecta; among these are A1-Biruni and Hubble in Marginis
basin, Lomonosov in Lomonosov-Flemingbasin, and most
significantly, Neper and Jansky craters within the outer
Smythii basin ring. In addition, the minimal differencein
laccoliths would serve as reservoirs for magma, which then
moveto the surfacethroughthe resultingfractures.Currently,
however,no featuressuchas large shields have been observed
that would suggestderivation of the associatedlava ponds
through this type of low-pressure, low-effusion rate
mechanism. Thus, if floor-fractured craters are formed from the
shallow igneous intrusion of a laccolith, that laccolith does
not appearto be the subsequent
reservoirfor the pond. Rather,
both pond andlaccolith (if suchexists) musthave originated
from a deeper sourcethat would provide the high driving
pressuresconsistent with morphologies like those in this
region, as suggestedby Wichman and Schultz [ 1995]. Indeed,
sucha scenarioof laccolithemplacementmanifesting itself as
a systemof fractureswould be consistent with the range of
features predicted by Wilson and Head [1996], for surface
manifestationsof magmapropagatedthrough dikes fed from
sub-crustal source regions. However, the hydrostatic
arguments[e.g., Solomon, 1975] invoked by Wichman and
A1/Si ratios between the western furrowedplains within
Smythii and the adjacent highlands [Andre et al., 1977],
Schultz [1995] in order to emplace the relatively dense
laccolith high into the lower density crustmay not be fully
particularly in the region southeast of Crisium [Clark and applicable.Magma transportsolelyby hydrostatic
rise may be
Hawke, 1987], indicates a chemically homogeneous overly simplified in terms of the ability of a dike to remain
stratigraphiclayer, the most likely sourcefor which is Crisium
ejecta [Andre et al., 1977]. Thus the formation of Crisium
open through 20-50 km of crust [Headand Wilson, 1992]. In
addition, for formation of the associatedpond to occur, a
basinis alsoa candidateevent for obscurationof cryptomare.
For Lomonosov-Fleming and A1-Khwarizmi/King, the
formation of Smythii and Crisium basins subsequent
to the
•ormation of thesetwo basins [Wilhelms, 1987] wouldhave
contributedejecta material to mantle cryptomaria in both
basins.Due to the extremeageof Lomonosov-Fleming
and A1Khwarizmi/King [Wilhelms, 1987], several Nectarian basins
are also potential candidates
for contributingto the mantling
mechanism
otherthanhydrostatic
rise mustthen be employed
which overcomesboth the density barrier presentedby the
of cryptomare material, such as Mendeleev and Moscoviense.
On the basis of the above evidence, we conclude that the
existenceof dark-haloimpactcratersassociatedwith Smythii
and Marginisbasinsrepresentcryptomariaburiedprimarilyby
the mantling ejecta of younger impact events. Thus the
significance of early volcanic activity may have been
underestimated
[e.g., Schultzand Spudis, 1979, 1983; Hawke
et al., 1985; Head and Wilson, 1992], and estimates of the
onsetof volcanismandthe total volumeof maredepositsfor
this region should both be regarded as minimum values
pendingconstraints
on cryptomarevolumeand stratigraphy.
4.2.3. Floor-fractured craters. Eight ponds(21% of
occurrences)on the floor of Smythii basin occur in floorfracturedcraters(e.g., Doyle and Cam6ens, Figure 7). Ponds
occurring in floor-fractured craters are similar in texture,
albedo, and volume to other lava ponds in the region,
suggestinga similar emplacementmechanism. However, it is
also observedthat ponds within floor-fracturedcraterstend to
occurin the young Late Imbrian-aged craters that ring Mare
Smythii. Craters of this age, which are abundanton the floor
of Smythii basin, formed concurrently with the most
voluminousperiod of mare emplacement.
We previously noted two general modelsfor the formation
of floor-fracturedcraters: (1) igneousintrusionand (2) viscous
relaxation. Both models dependstrongly on the history of
thermal activity for a given region. The igneous intrusion
model [Schultz, 1976; Brennan, 1975] involves the shallow (a
few hundredsto a few thousands of meters from the surface)
injection of a laccolith or sill beneath a crater, which drives
crater modification through floor uplift [Wichman, 1993;
Wichmanand Schultz, 1995]. Accordingto the model, such
lunarcrust,andthe decreasein driving pressureresulting from
laccolith emplacement.Finally, it must be noted that a model
of laccolithic intrusion necessarilyrequiresa large numberof
dikesin the lunar crust.Sincethis modelassumes
the magma
column supporting these laccoliths extends into the mantle,
the propagation of conduits into the overlying crust is
probably not affected by the formation of a crater on the
surface,at least20 km above.Thus,for everydike that actually
emplacesa laccolith in proximity to a craterto producefloor
uplift, there must be several that either do not reach nearsurfacelevels, or are not emplacedundera crater. Unless the
densityof dikesin the crust(a numberwhich is currentlyvery
poorly constrained)is very high, the sheer number of floorfracturedcratersin this regionseemsto favor a regional,rather
than a local origin.
In contrast to this model, the relaxation model employs
local crustal heating during mare emplacementto lower local
viscosity. This allows viscous relaxation to occur at a faster
rate than in other regions [Danes, 1965; Baldwin, 1968; Hall
et al., 1981]. This model has the benefit of not requiring
shallow emplacement of high-density material into a lowdensity (brecciated) crust, making it more consistent with
morphological and geophysical constraints. In addition, the
proximity of floor-fracturedcratersto Mare Smythii suggests
that the volcanism associatedwith Mare Smythii could have
been the source of heat requiredto produce lower crustal
viscosity in the surrounding area. The concentration of
volcanicdepositsin the Smythii basin centersuggeststhat the
heat flux might have been strongest here, so that viscous
relaxation might have been more prominent. However, this
doesnot explain why Mare Orientale has a larger volume than
Mare Smythii [Head, 1982] andwould presumablyhave had an
even greatereffect on the local crust,yet has no floor-fractured
craters.
Wichman and Schultz [1995] note in their observations a
positive correlation between the diameterof a floor-fractured
craterand the extent of crateruplift. This suggeststhe process
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
11,151
occurrencein impact structureswas observedfor depositsin
the Orientale,Mendel-Rydberg,and SouthPole-Aitkenbasins.
derivative
of the above models is one in which thermal
In theseregions, approximately70% of all pondswerefound
conditions
create a favorable environment
for both viscous
in craters, superimposedbasins, or within the low-lying
relaxation and volcanism.Becausesignificant mantle uplift is
Orientalebasinring [Yingstand Head, 1997a].For all regions,
indicatedin Smythii basin [Neumannet al., 1996], this may ponds found in craters and superimposed
basins generally
have provided a mechanismfor relatively near-surfacecrustal displayedhigher averagevolumesthan those lying in interheating and thus regional viscous relaxation. Neumannet al.
crater highlands. Many of the largest lava ponds (e.g., the
[1996] state that there is a decreasein the relief of the lunar
Apollo depositsin South Pole-Aitken, Kiess andHelmert-Kao
Moho (uplift of the mantle) with increasing basin age, such cratersin Smythii) lie within the deepestbasinsand the largest
that the oldest basins are the most isostatically compensated. craters.Together, these observationssuggestboth a higher
total volume of extrusion, and a higher frequencyof eruptive
Specifically, older pre-Nectarian basins, such as South PoleAitken, Fecunditatis, Australe, and Tranquillitatis, are eventsfor topographicdepressions
and lows than for highland
isostatically compensated but the younger Smythii is not
regions.These observationsare consistentwith the modelof
[Neumannet al., 1996], suggestingthat it was only in the later Head and Wilson [1992], where local-scale variations in
part of the pre-Nectarian period that the lunar lithosphere was crustal thicknessaffect mare extrusion. Further implications of
strong enough to maintain a high state of stress. Thus older this model are discussedin more detail later.
involved must be a regionalphenomenonor must explain why
craters alone are affected. A model which we suggest as a
basinswould have no evidence of localized (crater) relaxation
because at that time lateral movement
of crustal material
would
have made long wavelength relaxation of the entire basin
possible. By contrast, during the later stages of the preNectarian, the crust would have been sufficiently strong to
resist isostatic compensation. Consequently, conduction
throughthe crustof the heat providedby the uplifted mantle in
Smythii would have yielded a relatively thermally mobile
basin center. This would have presumably been a conducive
environment
for local
relaxation.
Such local
relaxation
of
relatively young craters would not be evident in younger
basins such as Crisium [Wilhelms, 1987], becausebasin floor
craterswouldhavebeenburiedby subsequent
voluminousmare
emplacementepisodes.Thus, the lack of floor-fracturedcraters
in Crisium would be a consequenceof the volcanic activity that
emplacedMare Crisium[Head et al., 1978]. It shouldbe noted
that recent analyses of Orientale basin using Clementine
images[Headet al., 1997] suggestthat Kopff cratermay very
likely be a floor-fracturedcrater.If suchis the case,it would be
a significantfinding in terms of determining the nature of the
thermal environment in which floor-fractured craters form,
since the young Orientale basin would have a very different
thermal
structure
than the older basins.
Finally, this model does not dependon hydrostaticrise as a
mechanismfor fracture formation or mare extrusion.Magma in
the above model passes directly from reservoir to surface
through dikes (the mechanism known to transport magma
through brittle crust), held open by a state of
overpi'•½•SSurization
in the source region instead of
hydrostatically[Head and Wilson, 1992]. Suchreasoningalso
removesthe obstacleof finding a mechanismto decipherwhy
ponds are associatedwith floor-fractured craters. Instead, this
model decouplesthe mechanismresponsiblefor the formation
of floor-fracturedcraters(local relaxation dueto heating from
below) with that for the emplacement of mare deposits
(overpressurizedreservoirsgeneratedby the sameheating).We
therefore postulate that uplift of the lunar mantle provided
sufficient heat to viscously relax the crust in a local sense,
forming floor-fractured craters, and yielded an accessible
sourceof magma for lava pond emplacement.
4.3.
Modes
of
Occurrence
Ponds in Smythii/Marginis show a preference for
deposition in impact craters, as shown in Figures 5 and 6.
Occurrencesof ponds in craters accounted for 70% of all
individual eruptive deposits. A similar preferencefor pond
4.4
Areal
Distribution
Average volume distribution,frequency,and spacingof lava
ponds in various regions yield important constraints on the
different characteristics and geometry of magma reservoirs
associatedwith these mare deposits. Typical volumes for lava
ponds in the limb and farsideregions are within the range of
195 to 860 km3 [Yingst and Head, 1997a; this study].
Adopting a geometryof 100 x 100 x 0.25 km for the dikes
feedingthese flows [Head and Wilson, 1992], and assuming
that the total volume (dike plus pond) typically represents
about 1% of the total volume of the reservoir [Blake, 1981],
then each pond could potentially representan eruption from a
reservoir with a volume of -270,000-340,000 km3. If we
further assumean ideal spherical magma reservoir, such a
volume yields a diameter of approximately 80-90 km. This
diameter range is similar to the average range of nearestneighbor distances for all basins examined. Note that a
differenceof-600 km• in mean volume (suchas that found
between
the mean volumes
of the limb
areas and South
Pole-
Aitken) translates into only a 10 km difference in the
calculatedreservoir radius. The presence and spacing of lava
ponds may thus provide an indication of the frequency and
geometryof magma reservoirsat depth.
While average nearest-neighbor distance values are an
important factor in determining reservoir geometry, nearestneighbordistancesfor individual pondsmay yield information
regarding pond clustering associatedwith the parent source
regions. For example, if we accept the above estimates for
reservoirdiametersas reasonable,the fact that many ponds are
separatedfrom one anotherby distancesof less than 100 km
(such as the Doyle-CamOens Haldane region; Figures 2a, 7)
suggeststhat such clusters may represent a population of
ponds derived from one reservoir. In other cases, ponds are
separatedby much larger distances. For example, the pond
northwestof Erro crater(pond 2 in Figure 2a) is separatedfrom
the next nearest pond by -150 km. Individual ponds such as
theseare candidatesfor separatereservoirs. In order to assess
the relevanceof nearest-neighbordistancesin understanding
the sequenceof emplacementfor these deposits, spectral and
age characterizations,as well as constraintson the extent of
possiblecryptomaria,are required as a next step.
The general distributionof ponds in Smythii and Marginis
basins appearsto be related to basin degradationstate. In
Smythii, pondsare concentratedwithin the central portion of
the basin, while in Marginis pond distributionis more diffuse.
11,152
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
This correlation suggests a connection between pond
occurrenceand basin age. One plausible explanation is that
differencesin the thermal regime of each basin influencedthe
concentration of magma reservoirs. Thus, becauseSmythii is
younger (and thus uncompensated[Neumannet al., 1996]),
heat was focusedwhere mantle uplift occurred,namely, in the
central basin. This is where the greatestnumberof reservoirs
formed, or alternatively, where diapirs were able to penetrate
to a shallower depth. Conversely, Marginis, an older
compensatedbasin [Neumannet al., 1996], would have had no
such concentration of a heat source in the basin center, so that
reservoirdistribution,and thus associatedpond concentration,
would be more diffuse. It is also possible that the observed
pond distributionis due to mantle heterogeneities
that relate to
mechanismsnot currently well constrained.
4.5.
Crustal
Thickness
Relationships
It is clear that the areal and volumetric
mean
values
for
pondsin Smythii and Marginis basinsare more comparableto
the youngerOftentale basin than to SouthPole-Aitken basin.
In fact, the mean value for volume of magma extrudedin an
individual eruptive episode in the three smaller basins is
nearlyequal(-195-250km3),whilethatfor SouthPole-Aitken
is more than 3 times larger. In addition, as previously
mentioned,there are many more relatively low-volume ponds
in the limb basins than in South Pole-Aitken
basin.
There are
several hypotheses which might explain the greater mean
volumesof individual eruptiveeventson the farside.
SomeSouth Pole-Aitken lava ponds might be the products
of multiple flows, and thusyield a larger mean volumeestimate
for individual eruptive episodesthan reality dictates.Although
this possibility cannot be discounted,there is currently no
evidence (e.g., variations in albedo, different crater densities,
other characteristicsdiscussedabove)to suggestthat this is
the case in a general sense. High resolution multispectral
imaging data is currently being utilized to further test this
possibility(e.g., Clementinemultispectralimage data [Yingst
and Head, 1997b]), but preliminary data support the
interpretation that the vast majority of the basins represent
distinct episodes.
It is possible that, given a relatively homogeneous
distributionof dikes within the crustbelow these regions, the
older SouthPole-Aitken basin was volcanically active longer
and thus sampled more large volume eruptions over time.
However, it has been suggestedthat a lower thermal gradient
existed for this region at the time of basin formation
volume of lava ponds and regions of topographic lows
associatedwith thin crust (compareFigures2a and 2b). This
relationship is shown for the farside in Figure 13, which
displaysthe volume of mare material observedin SouthPoleAitken
basin as a function
of the estimated
crustal thickness
(shown in increments of 5 km). Volume has been normalized
to the total
with
area within
each 5 km
South Pole-Aitken
increment
of crustal
basin
thickness.
associated
Thus
the
volume at each increment of Tc represents the effective
thickness
of mare fill
that would exist if the total
volume
of
mare material lying at that thickness value were spreadin a
uniformlayer acrossthe correspondingarea. For the example
of South Pole-Aitken basin, the bulk of mare material and the
majority of ponds occurwherethe crust is thinnest (Tc < 50
km; only 40% of the basin area). Areas of thicker crust (50-70
km) containa smaller number of ponds and a very small total
volume [Yingst and Head, 1997a]. Smythii and Marginis
basinsalso follow this general trend. In generalterms,regions
of major mafia (Mare Smythii and Mare Marginis) occur in
areasof thinnest crust, while ponds tend to occur in regions
that have thicker crust. Overall, it can be seenin Figure 14 that
most ponds occur in regions where crustal thickness is
estimated to be < 50 km. About 61% of the total
volume
of
pondsanalyzedoccurswithin just 31% of the region, the area
corresponding
to Tc < 50 km. Only 39% occurin the remaining
areasof thicker crust. In terms of eruption frequencyfor mare
ponds,21 eruptiveoccurrences
(62%) lie at Tc< 50 km, while
13 (38%) occur at values of 50 km or above. In addition, we
can analyze the contribution of the major maria (associated
with crust 40 km thick or less) to the number of eruptive
events,by estimating the numberof eruptionsthese deposits
represent. We assumea typical volume range for individual
eruptivephasesin the Smythii/Marginis region of 190-270
km3, as suggested
by the volumedataenumerated
earlier.
Dividing these averagesby the combined volume of Mare
Smythii and Mare Marginis yields -200-300 potential
0.16
0.140.12-
0.080.06-
0.04-
[Solomon et al., 1982; Neumann et al., 1996], which would
have resultedin fewer magma sources,with shorter-cooling
times and possibly lower degreesof partial melting. This
would have
served to decrease the number
and volume
of
eruptions within the farsidebasin comparedto those on the
lunar limbs. Another possibility is that factors such as the
numberof tappaNemagma reservoirs,or the degreeof partial
melting associated with these reservoirs, are variable on a
global scale. These latter factors cannot currently be
established or ruled out.
Finally, the averageeruptedvolume may be a function of
variationsin crustalthicknessof different lunar regions [e.g.,
Head and Wilson, 1992; Robinson et al., 1992]. Estimates of
lunar crustalthickness(Tc) derivedfrom altimetrydataobtained
by the Clementine laser altimeter [Zuber et al., 1994] show
that there is a very close correlation between the average
0.02-
o ::::I::::I::::I::
10
20
............
30
40
50
60
70
80
Crustal Thickness(km)
Figure 13. Total volume of lava extrudedcomparedto crustal
thickness
for
South
Pole-Aitken
basin.
In order
to
avoid
sampling bias, the total volume of mare material has been
normalized to the amount of surfacearea occupiedby each
estimated crustal thickness
value. Thus the numerical value at
each incrementof crustal thicknessrepresentsthe thicknessof
mare fill that would exist if the total volume of mare material
lying at each crustalthickness were spreadin a uniform layer
acrossthe areacorrespondingto that crustalthicknessvalue.
This figure showsthe inverse correlationbetweenthe volume
of mare material
and the thickness of the lunar crust.
YINGST AND HEAD: SMYTHII AND MARGINIS BAS1NLUNAR MARE DEPOSITS
11,153
We conclude, on the basis of the very strong local and
0.07
global correlations observedbetween the magnitudeand
frequencyof magmaeruptionandcrustalthicknessdifferences,
0.06
that it is likely that the ability of magmato reachthe surface
from subcrustalreservoirsis directly related to the thicknessof
the intervening crustal column. This connection has been
observedin otherregionsof the Moon [e.g., Robinsonet al.,
1992]), and hasbeensuggested
by Headand Wilson [1992] to
be the resultof diapirsstalledundercrustalcolumnsof varying
0.05
0.04
0.03
height.A schematicrepresentation
of this modelfor the limb
and farside basins is shown in Figure 16. Assuming a
0.02
o.o
o
0
20
40
60
80
100
120
Crustal Thickness (5 km bins)
Figure 14. Total volume of lava extrudedcomparedto crustal
thicknessfor Smythii/Marginisbasins, normalizedto amount
of surfacearea as in Figure 13.
eruptive phasesthat these regions may represent. Thus, the
highest frequency of eruptive episodes appears to have
occurredin regionsof the thinnestcrust.
In terms of maria in specific basin regions, we have
observed that the total extruded mare volume is a function
of
the thicknessof the intervening crustalcolumn. Determining
the role of crustalthicknessin mechanismsof mare transport,
however,is ultimately a global issue.Thus a usefulexerciseis
to consider
estimates
of
the
total
lunar
mare
volume
as a
functionof Tc,where the large maria whosevolumesare known
are included. Although the number of individual eruptive
episodescannot be estimatedbecauseof the reasonsstatedin
our approach,an estimateof the total volume of mare material
depositedat each crustalthickness value acrossthe Moon is
possible.This is shownin Figure 15, where the estimatedtotal
volume of mare material for depositson the lunar limbs and
farside, as well as the nearsidecontiguousmare regions for
which volume estimatesare available, is plotted against the
estimatedcrustalthickness(displayedin incrementsof 5 km).
These regionsinclude South Pole-Aitken and Orientale basins
[Yingst and Head, 1997a], Crisium, Humorum, Nectaris,
Imbrium, and Serenitatis nearside basins [Solomon and Head,
1980], and Smythii and Marginis basins (this study). As
relatively homogeneousdistribution of magma sources
throughout
the lunar mantle,mareemplacement
in this model
is dependent
on the level of overpressurization
which these
sourceregionsreachuponstallingat a boundarydefinedby the
low density lunar highland crust. This overpressurization
would,in turn, drive the propagationof dikes from depth. For
diapirs at equal levels of overpressurization,those dikes
emplacedinto regionsof thinner crustreachthe surfacemore
readilythan thosepropagatinginto thicker crust.This model
implies that the height of the overlying crustalcolumnis
pivotalin maretransportand distribution.The amountof this
intervening crust determineswhetherdikes propagatedfrom
overpressurized
sourceregions are able to extrudeonto the
surface(at areasof thinnest crust)or must stall and freeze (at
regions where the crust is thicker). A high degree of
correlation
between
crustal
thickness
and the
number
and
magnitude of individual eruptive events is required. The
observations discussed here for Smythii/Marginis are
consistentwith both the analyses of other lava ponds on the
lunar limbs and farside, and with the predictions implied by
this model. This model also predictsthat, becausedikes must
be driven from depthsof at least the base of the crust, each
episodeof mareemplacement
is likely to be associated
with a
high effusion rate and thus large volumes of basalt.
Emplacement episodes would be evidenced by relatively
smooth deposits showing a lack of featuresassociatedwith
shallow sourceregions(e.g., calderas,large shieldvolcanoes).
Again, these predictionsare consistentwith what we have
observedfor lava pondsin Smythii/Marginis andon the lunar
limbs and farside[Yingstand Head, 1994, 1996, 1997a].
A Lunar
surface
area
represented
by
Tc<40
km
=12%
< 40 km = 66%
before, volume has been normalized to the total area associated
with each 5 km bin of crustal thickness
so that effective
mare
thickness for each Tc value is displayed.As shown in this
figure, the total volume of mare materialis inversely relatedto
crustalthicknessin that areasof thinnest correspondingcrust
show the greatest amount of mare material on the surface.
Specifically,66% of the total mare volume measuredoccurso n
only 12% of the lunar surface,the arearepresentedby T• < 40
km. In addition, it is seen that regions on the nearsidewhich
are characterizedby the large volumesof the contiguousmaria
also have the thinnest crust. As stated above, the number of
•.•
0.5
20
40
60
80
100
120
Crustal Thickness(5 km bins)
Figure 15. Crustal thicknessversus the total volume of mare
material for the Moon, normalized to amount of surface area as
eruptiveoccurrencescorrespondingto the total mare volume
in Figure13.Regions
werechosen
onthebasisof availability
cannot currently be determined. This is to be expected,
of volumeestimates.TheseregionsincludeSouthPole-Aitken
however, if crustal thickness is directly related to eruption and Orientale basins [Yingst and Head, 1997a], Crisium,
volume. The number of occurrences becomes more difficult
to
Humorum, Nectaris, Imbrium and Serenitatis nearside basins
judge as Tcdecreasesbecausethe increasedvolume of deposits [Solomon
and Head, 1980],andSmythii/Marginis
basins(this
effectively obliteratesour ability to discernindividual flows.
study).
11,154
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
•
Crisium
Basin
Basin
Crust
mare depositdistribution,in that mare depositswould havehad
a longer time to accumulatein an older basin. Specifically,
ponds in South Pole-Aitken and Australe were emplaced
throughoutthe Late Imbrian period [Wilhelms and El-Baz,
1977; Wilhelms et al., 1979; Hiesinger et al., 1996], while
pondsin Smythii/Marginisand Orientalewere emplacedduring
the upper Late Imbrian and Eratosthenian periods [Wilhelms
and EI-Baz, 1977; Scott et al., 1977; Hiesinger et al., 1997].
However, the volcanic activity in all regions have the same
apparentstartingpoint. Only the numberof depositsemplaced
during each period are different. The formation of the farside
Figure 16. Model of lunar crust/mantle history [after Head and eastern limb basins predates the apparent onset of mare
and Wilson, 1992]. A range of overpressurizationconditions volcanism. In addition, ejecta deposits from the Orientale
within the diapir-like reservoirs is predicted, due to stalling
basin have emplaceda stratigraphicdatum at - 3.8 Ga, burying
undercrustalcolumnsof varying height. This excesspressure any pre-Orientale mare material at least in South Pole-Aitken
instigates the production of individual dikes that propagate and areasin the proximity of Orientale [Head et al., 1993].
towardthe surface.On the easternlimb pictured here (not to Thus post-Orientale mare emplacement in South Pole-Aitken
scale), dikes are more likely to reach the surfacein areas of
and Orientale appearsto have occurredover similar periods in
relatively thinner crust (such as the basins, representedby A
the case of these basins.
and B), while in areas of thicker crust dikes reach the surface
An alternate theory, which we propose here, is that, due to
thin crust, some areasof the Moon were filled preferentially
duringthe early stagesof lunar volcanism and remainedactive
for several periods,while other nearby regionswere filled only
4.6.
Stratigraphy
duringthe later stages.Considerfor example, Crisium basin,
The local stratigraphic profile for the Smythii/Marginis which lies so close to the Smythii/Marginis region that their
region (shownin Figure 10) suggeststhat, since the deposits basin rings overlap [Wilhelms and EI-Baz, 1977]. Crisium
in this region are all relatively close in age, there may have maria representsthree different and comparatively voluminous
beena periodof active volcanism in this region around3.85stagesof volcanism [Solomon and Head, 1980], spanning a
3.60 Ga. However,almost 200 myr appearsto have elapsed time period from possibly the earliest stages of volcanic
betweenthe emplacementof the pond in Ibn Yunuscrater and history to less than 3.5 Ga [Boyce et al., 1977; Head et al.,
the portion of Mare Marginis duesouthof that crater(region 1978]. Crisium basin has been estimated to have a
III in Figure 2a). Similar observations were made for mare correspondingcrustal thickness which varies from 12-25 km,
depositsin Grimaldi basin northeastof Orientale [Yingst and while crust in the Smythii/Marginis region ranges from 25-60
Head, 1994]. Grimaldi basin is 170 km in diameter. In this
km [Zuber et al., 1994]. If the thicknessof the crust determines
relatively small area, three flows of varying compositionsare the likelihood of a magma conduitreaching the surface,as has
represented,datedat 3.12, 2.79, and 2.49 Ga [Greeleyet al.,
been suggestedabove, regions with a relatively thinner crust
1993; Williamset al., 1995]. The large periodsof quiescence would have preferentiallybeen filled first, and would have been
between emplacementperiods for these flows occurring in
subjectto voluminousmare emplacementfor longer periods of
relatively small areassuggeststhat (1)parent sourceregions time than nearby areas. Basins such as Crisium, Australe and
have a very long cooling time (200-750 Ma) or (2) magma SouthPole-Aitken would thus display a wider range of deposit
reservoirsare being emptiedand then replenishedfrom depth. agesbecausethey have been active for longer periodsof time
On the basis of our present understanding,we favor the due to the thinner crust in that particular region. Thicker crust
interpretation that these deposits represent the sequential in surroundingregions (such as Smythii/Marginis in the case
emplacement
of materialderivedfrom differentsourceregions of Crisium) would thus cause local mare flooding to be
at depth, whose formation was separatedin time by several constrainedto a later, possibly more active stage. Regions of
hundredmillion years.
very thick crust (such as that surroundingAustrale) would show
The age-frequency distribution of lava ponds in
no volcanic activity. Wilhelms [1987] dates the deposits
Smythii/Marginis displayed in Figure 10 suggeststhat the within Mare Crisium as relatively younger than those in
mare deposits in this region are somewhat younger in Marginis
basin,andcontemporaneous
withthosei• Smythii
comparison to ponds on the farside [Wilhelms et al., 1979;
basin. However, if the volcanic flux on the eastern limb was
Hiesingereta!., 1996]. In Smythii/Marginis,only five ponds, similar to that calculatedfor the Moon during the Late Imbrian
all of which are within the older Marginis basin, are datedas [Hartmann et al., 1981; Head and Wilson,1992], it is possible
older Late Imbrian. (Because the formation of Orientale defines
that many Early Imbrian flows emplacedin the lowest lying
the baseof the Late Imbrian period,no pondscanbe older than areasare now obscuredbecausethey were coveredup by later,
Late Imbrian (3.80 - 3.20 Ga) in Orientalebasin.) Smythii and more voluminousflows. More preciseage determinationof the
Marginishave a large populationof depositsclassifiedeither deposits within Crisium and other regions, as well as the
as Late Imbrian, or undivided Late Imbrian/Eratosthenian dark
testing of the models presentedhere, awaits crater-frequency
mantle deposits. In South Pole-Aitken and Australe, by dating derived using high-resolutionClementine images.
comparison, a large percentageof the subdividedponds are
Finally, it is possible that the mare material observed on
datedas older Late Imbrian in age. This observationmay be the surfaceis not indicative of the full stratigraphic range of
explainedby several hypotheses.
mare emplacement episodes. As has been noted previously,
As has been suggestedabove, the extreme age of South Nectarian-Imbrianfurrowed and pitted material is extensive in
Pole-Aitken and Australebasins may have played a role in
the region north of Mare Marginis and in a subcircularregion
rarely (C).
YINGST AND HEAD: SMYTHII AND MARGINIS BAS1NLUNAR MARE DEPOSITS
around Mare Smythii [Wilhelms and El-Baz, 1977]. The
geomorphology of this material in Smythii basin is
particularly striking in that its boundary is very closely
definedby the inner ring of the basin, similar to the boundary
of volcanicmaterial that fills the central portion of Orientale
basin. It has thus been suggestedthat this region represents
basaltthat hasbeenreworkedby subsequent
crateringor other
11,155
fewer Imbrian craters exist within Marginis than within
Smythii, implying a connectionbetweenthe age of the
associated
craterand the emplacement
mechanism.Headand
Wilson [1992] suggestthat local thinning of the crustin the
vicinityof impactcraterscreatesa favorableenvironmentfor
mareemplacement.
Formationof a craterthinsthe localcrustal
regime, but doesnot affect the overall pressureconditions at
subcrustaldepths. All things being equal, a dike propagating
beneath a crater would thus have a greater chance of
in A1/Si ratiosto surrounding
highlandssoilsanddissimilarto
nearbymaredeposits,a soil composedsolely of highly intersecting the surfacebecausethe local crust was thinner.
This implies that young craters (Imbrian-aged), which would
reworked basalt is not indicated [Andre et al., 1977]. However,
there is only a minimal differencein A1/Si ratios between be relatively fresh during the period of local volcanism, would
furrowedplainssoilsandhighlandsoilsto thewestof Smythii be somewhat more likely candidates for mare emplacement
than older craters(pre-Nectarianor Nectarian) which wouldbe
basin,an observationthat has been interpretedas indicating a
chemicallyhomogeneous
ejectalayer emplacedprior to the more degradedand thusmore shallowat the time of active mare
LateImbrianperiodof marevolcanism[Andreet al., 1977; volcanism. The greater frequencyof Imbrian-aged craters in
Smythii basin couldthus provide more favorable regions for
Wilhelms and El-Baz, 1977]. This plains-type material may
processes
[Stewartet al., 1975].Because
thisregionis similar
thus cover older cryptomaria,which might have filled vast
mare
extrusion.
As indicatedpreviously, it is also possible that there are
differencesin the thermal history of the region that favor
cratersis a possible indicator of a scenariosuch as this eraplacementof mare within cratersformed aroundthe time of
[SchultzandSpudis,1979, 1983; HawkeandBell, 1981; Bell mare extrusion. Superposition relationships suggest that
andHawke, 1984; Antonenko et al., 1995]. The geochemical Smythii basin is younger than basins such as Fecunditatis,
data provided by the Apollo X-ray and gamma-ray Australe, and Tranquillitatis [Wilhelms, 1987]. These older
spectrometers
did not directlyindicatethe presenceof any basins display a compensationsignaturewhich is lacking in
the youngerSmythii basin [Neumannet al., 1996], indicating
mare materialin the light plains region [e.g., Adler et al.,
portionsof the basin in the Early Imbrianor earlier.As
previously
noted,the presence
of features
suchas darkhalo
1972; Metzgeret al., 1973; Wilhelms and El-Baz, 1977].
However,multispectral
datafrom Mariner 10 hasbeenusedto
identify a region northeastof Mare Marginis that has a
relativelystrongspectralsignaturein the 0.40/0.56 gm ratio
and thus is spectrally bluer than the surroundingsoils
[Robinsonet al., 1992]. This area,identifiedas intermediate
that the thermal
structure of the Moon
evolved
so that there
was a point in the pre-Nectarianafter which cooling of the
outer layers of the Moon preventedbasins from reaching a
fully compensated state. Thus minor crustal thickness
differences and increases in mantle uplift causedby the
formationof impact craterswouldhave been smoothedout in
in composition
betweenmareandhighlands
units[Robinson the older basins. In Smythii basin, where differencesin local
et al., 1992], correspondsto the furrowedplains unit of compensationwould still have existed, craters might have
Wilhelms and El-Baz [1977]. Suchspectrallyblue units have provideda more favorableenvironmentfor extrusion.
beenpreviouslyassociated
with cryptomare
deposits[Metzger
and Parker, 1979; Davis, 1980]. Thus the geologic setting of
this furrowedplains unit, combinedwith its multispectral
signature,suggeststhat soils in this region contain a
cryptomare
component.
Examination
usinghigherresolution
Clementinemultispectraldatain orderto searchfor further
mineralogicindicationsof marematerialwithin this plains
unit wouldbe a next stepin resolvingthe stratigraphicprofile
of volcanic activity in Smythii/Marginis.
5.
Local-Scale
Variations
Thus far, we have presented observations of the
characteristicsof discretemare ponds on the Moon's eastern
limb and the general trendswhich these observationsfollow
with respectto other similar localized depositson the lunar
western limb and farside. While the deposits within the
Smythii/Marginis study area are consistent with the trends
noted, there is one local-scale variation of mare pond
characteristicsfor which there is currently insufficient data to
make any firm conclusions,but is notedhere in the context of
future studies. This variation
relates to a difference
between
the
mode of occurrence of ponds in the Smythii basin and
• Marginis basin.
In Smythii basin, more than 86% of the mare deposits
studiedoccurin relatively young (Imbrian-aged) craters, while
in Marginis, less than 50% lie within craters. Thus the
differencein frequencyof occurrence
may be dueto the fact that
6.
Conclusions
Analysis of 41 mare ponds and dark mantle deposits has
provided information on associated features, modes of
occurrence, and the range and frequency distribution of
eruption areasand volumesin Smythii and Marginis basins.
On the basisof this analysis,we find the following.
1. The majority of deposits likely to represent single
eruptiveepisodeshave areaswhich lie in the range of 170 to
6575 km2, with an averagevalueof 965 km2. This translates
intoa meanareaof 720 km: for Smythiibasinand1830 km:
for the Marginis basin region.
2. Volumes
for thesedeposits
rangefrom15 to 1045km3,
with a meanvalueof 195 km3. For Smythiibasinthe mean
valuefor deposit
volumeis 190km3,whilefor Marginisbasin
it is 270 km3.
3. Deposits tend to occur preferentially within areas of
locally thinned crust (craters). In terms of pond density,
concentration of mare deposits appears to be related to the
state of preservation of the basin, such that deposits are
highly concentratedin the center of the better preserved
Smythii basin, while in Marginis basin the distribution of
depositsis more diffuse.
4. No definitive morphological evidence, such as large
shield volcanoes or collapse calderas,was found for shallow
crustal magma reservoirs in association with these ponds.
11,156
YINGST AND HEAD: SMYTHII AND MARGINIS BASIN LUNAR MARE DEPOSITS
Features associated with isolated lava ponds (such as linear
rilles) are consistentwith emplacementand extrusion onto the
surfacethroughdikes from deep,perhapssubcrustalreservoirs.
In terms of floor-fracturedcraters, some models suggestthat
the
existence
of
such
craters
indicates
shallow
who providedphotographic
support,and fundingfrom NASA Grant
NAGW-713 to J.H. from the National Aeronautics and Space
AdministrationPlanetaryGeologyand GeophysicsProgram.
laccolithic
intrusion. However, such a model also requiresthat both the References
laccolith and the pond stem from the same deep-seated
reservoir.
5.
The
presence of
such
indicators
as
geochemical/multispectralanomalies and abundantdark-halo
impact cratersin the Smythii/Marginis region suggeststhe
presenceof previously deposited cryptomare material. Based
upon the distributionof these indicators, suchcryptomaria, if
it exists, may be a significant volumetric component of the
total mare material in the area (e.g., potentially twice the
volume of presentlyexposedmare deposits).
6. Peak volcanismin this region as representedby exposed
mare depositsappearsto have occurredat around3.80-3.60 Ga.
Becauseof the possible existence of cryptomare material, as
indicatedby dark-halo craters,this shouldbe consideredas the
latest period at which the onset and peak of volcanic activity
could have occurred.In a preliminary analysis, Spudis and
Hood [1992] have reportedevidencefor the presenceof young
mare depositsin Mare Smythii, possiblyyounger than Apollo
12 mare material (-3.20 Ga), although the location of their
crater counts within Smythii were not specified. Further crater
size frequency distribution data for Smythii may reveal
evidence for young deposits.
We have compared the characteristics of discrete mare
depositson the easternlimb with those of similar deposits on
the western limb and farside in order to put Smythii and
Marginis ponds into a global context. On the basis of the
observationsmade, we infer the following.
1. Volumes of eruptive events on the Moon are very large
comparedto most terrestrialeruptions, and seem to have their
best
terrestrial
analog,
geomorphologically
and
volumetrically, in flood basalts rather than small volume
eruptionsderived from shallow reservoirs.
2. Nearest-neighbor distancesaverage about 60-100 km,
suggestinga constraint for the diameterof sourceregions to
within or below this range. Typical lava pond volumes and
nearest-neighbordistancesfor ponds on the limbs and farside
of the Moon suggest that magma is derived from subcrustal
reservoirs<-100 km in diameter.Thus many ponds,especially
those lying in the furthest rings of their respective basins, are
likely to be solitary representatives of their source regions.
By the same reasoning, other ponds that lie more tightly
spaced,notably thosewithin the central portions of the basin
regions, may be membersof a pond cluster which originates
from a single sourceregion.
3. For each individual basin, the volume and frequencyof
eruptions is related to the amountof crustthrough which the
magmamust pass.Those areaswith thinnercrusthave a greater
volume and numberof mare occurrenceson average relative to
regions of thicker crust. Although the number of eruptive
events cannot currently be determinedin the large contiguous
maria for the Moon as a whole, regions of thinner crust
correspondto areasof higher total mare volume.
Acknowledgments.We gratefully acknowledgethe assistanceof
Irene Antonenko, B. Ray Hawke and an anonymousreviewer who
providedreviews of earlier versionsof this manuscript,Peter Neivert,
Adler, I., et al., The Apollo 15 X-ray fluorescenceexperiment, Proc.
Lunar Sci. Confi,3rd, 2157-2178, 1972.
Andre, C.G., R.W. Wolfe, I. Adler, P.E. Clark, J.R. Weidner, and J.A.
Philpots,Chemical character of the partially flooded SmythiiBasin
basedon AI/Si orbitalX-ray data,Proc. Lunar Planet.Sci. Conf.,8th,
925-931, 1977.
Antonenko, I., J.W. Head, J.F. Mustard, and B.R. Hawke, Criteria for
the detectionof lunar cryptomaria,Earth Moon Planets,69, 141-172,
1995.
Baldwin,R.B., Rille patternin the lunar crater Humboldt,J. Geophys.
Res., 73, 3227-3229, 1968.
Beals,C.S., and R.W. Tanner,Crater frequencieson lava-coveredareas
relativeto the Moon's thermalhistory,Moon, 12, 63-90, 1975.
Bell,J.F.,andB.R. Hawke, Lunar dark-haloedimpactcraters: Origins
and implicationsfor early mare volcanism,J. Geophys.Res., 89,
6899-6910, 1984.
Blake, S., Volcanism and the dynamicsof open magma chambers,
Nature, 289, 783-785, 1981.
Boyce, J.M., G. Schaber,and A. Dial, Age of Luna 24 mare basalts
based on crater studies, Nature, 265, 38-39, 1977.
Brennan,W.J., Modification of pre-mare impact craters by volcanism
and tectonism, Moon, 12,449-461,
1975.
Campbell, I.H., and R.W. Griffiths, Implicationsof mantle plume
structurefor the evolution of flood basalts,Earth Planet. Sci. Lett. 99,
79-93, 1990.
Clark, P.E., and B.R. Hawke, The relationshipbetween geologyand
geochemistry
in the Undarum/Spumans/Balmer
region of the Moon,
Earth Moon Planets, 38, 97-112, 1987.
Clark, P.E., andB.R. Hawke, The lunarfarside: The natureof highlands
eastof Mare Smythii, Earth Moon Planets,53, 93-107, 1991.
Danes, Z.F., Reboundprocessesin large craters,Astrogeol.Stud.Ann.
Progr. Reg. A (1964-1965), pp. 81-100, U.S. Geol. Surv., Reston,
Va., 1965.
Davis, P.A., Iron and titanium distribution on the Moon from orbital
gamma spectrometry and implicationsfor crustal evolutionary
models,J. Geophys.Res., 85, 3209-3224, 1980.
Gaddis,L.R., Photogeologic
andremotesensingstudiesof regionallunar
volcanism,M.Sc. thesis,46 pp., Brown Univ., Providence,R.I., 1981.
Gifford, A.W., and F. EI-Baz, Thicknessesof lunar mare flow fronts,
Moon Planets, 24, 391-398, 1981.
Gillis, J.J., P.D. Spudis,and B.J. Bussey,The geology of Smythii and
Marginisbasinsusingintegratedremote sensingtechniques:A look
at what's aroundthe corner,Lunar Planet. Sci. Conf.,28th, 419-420,
1997.
Golombek, M.P., Fault type predictionsfrom stress distributionson
planetarysurfaces:Importanceof fault initiation depth,J. Geophys.
Res., 90, 3065-3074, 1985.
Greeley, R., et al., Galileo imaging observationsof lunar maria and
relateddeposits,J. Geophys.Res.,98, 17,183-17,205, 1993.
Hall, J. L., S.C. Solomon, and J.W. Head, Lunar floor-fractured craters:
Evidence for viscousrelaxationof crater topography,J. Geophys.
Res., 86, 9537-9552, 1981.
Hartmann, W.K., et al., Chronology of planetary volcanism by
comparativestudiesof planetarycratering,in BasalticVolcanismon
the Terrestrial Planets,editedby membersof the Basaltic Volcanism
StudyProject,pp. 1049-1127,Pergamon,Tarrytown,N.Y., 1981.
Hawke, B.R., and J.F. Bell, Remotesensingstudiesof lunar dark-halo
impact craters: Preliminaryresultsand implicationsfor early
volcanism,Proc. Lunar Planet.Sci. Conf.,12th,665-678, 1981.
Hawke,B.R., P.D. Spudis,and P.E. Clark, The origin of selectedlunar
geochemicalanomalies: Implicationsfor early volcanismand the
formationof light plains,Earth MoonPlanets,32, 257-273, 1985.
Head,J.W., Lunar volcanismin spaceand time, Rev. Geophys.Space
Phys., 14, 265-300, 1976.
Head, J.W., Lava flooding of ancient planetary crusts: Geometry,
thickness,andvolumesof floodedlunarimpactbasins,Moon Planets,
26, 61-88, 1982.
YINGST AND HEAD: SMYTHII AND MARGINISBASINLUNAR MARE DEPOSITS
Head, J.W., and L. Wilson, Lunar mare volcanism: Stratigraphy,
eruptionconditions,andthe evolutionof secondarycrusts,Geochim.
Cosmochirn.Acta, 56, 2155-2174, 1992.
Head, J.W., and L. Wilson, Lunar grabenformationdue to near-surface
deformationaccompanyingdike emplacement,Planet. Space Sci.,
41,719-727,
1993.
11,1•7
multispectralimagesof the easternlimb and farside of the Moon, J.
Geophys.Res.,97, 18,265-18,274,1992.
Schaber,G., Lava flows in Mare Imbrium: Geologic evaluation from
Apollo orbital photography,Proc. Lunar Sci. Conf, 4th, 73-92, 1973.
Schonfeld,E., and M. Bielefeld,Correlationof dark mantledepositswith
high Mg/A1 ratios, Proc. Lunar Planet. Sci. Conf., 9th, 3037-3048,
1978.
Head,J.W., andL. Wilson,Lunar mare basaltvolcanism:Early stages
of secondarycrustalformationand implicationsfor petrogenetic Schultz, P.H., Floor-fractured lunar craters, Moon, 15, 241-273, 1976.
Schultz,P. H. and P. Spudis,Evidencefor ancientmarevolcanism,Proc.
evolution and magma emplacementprocesses,Lunar Planet. Sci.
Lunar Planet. Sci. Conf, loth, 2899-2918, 1979.
Conf., 28th, 545-546, 1997.
Head,J.W.,J.B.Adams,T.B. McCord,C. Pieters,and S. Zisk, Regional Schultz,P. H. and P. Spudis,The beginningand end of lunar mare
volcanism, Nature, 302, 233-236, 1983.
stratigraphy and geologic history of Mare Crisium, Geochim.
Cosmochim.
Acta, 9, suppl.,43-74, 1978.
Head, J.W., S. Murchie, J.F. Mustard, C.M. Pieters,G. Neukum, A.
Scott,D.H., J.F. McCauley, and M.N. West, Geologic map of the west
side of the Moon, U.S. Geol. Surv. Map, 1-1034, 1977.
McEwen,R. Greeley,E. Nagel, and M.J.S.Belton,Lunar impact
Solomon, S.C., Mare volcanism and lunar crustal structure, Proc. Lunar
basins: New data for the western limb and far side (Orientale and
Sci. Conf.,6th, 1021-1042,1975.
SouthPole-Aitken
basins)fromthefirstGalileoflyby,J. Geophys. Solomon,S.C. and J.W. Head, Lunar mascon basins: Lava filling,
Res., 98, 17,149-17,181, 1993.
tectonics,
and e.volution
of the lithosphere,
Rev. Geophys.,18, 107Head, J.W., C.M. Weitz, C.M. Pieters, G. Neukum, J. Oberst, H.
Hiesinger,A. Cook,andM. Wahlisch,The Maunder- Kopff crater
paradox: AnalysisusingClementinedata,Ann. Geophys.,Space
Planet.Sci., 15, suppl.3, C 790, 1997.
HiesingerH., R. Jaumann,G. Neukum, and J.W. Head, Mare Australe:
141, 1980.
Solomon,S.C., R.P. Comer, and J.W. Head, The evolution of impact
basins: Viscous relaxation of topographicrelief, J. Geophys.Res.,
87, 3975-3992, 1982.
Spudis, P.D., and L.L. Hood, Geological and geophysical field
New resultsfromLunarOrbiterandClementine
UV/VIS imagery,
investigations
from a lunarbase at Mare Smythii,in Lunar BasesH,
Lunar Planet. Sci. Conf, 27th, 545-546, 1996.
NASA Conf Publ., 3166, 163-174, 1992.
Hiesinger,H., U. Wolf, R.A. Yingst,andJ.W. Head,A new view of the Stewart,H.E., J.D. Waskom, and R.A. De Hon, Photogeologyand basin
stratigraphyof mare depositsin MarginisBasin,Lunar Planet. Sci.
configurationof Mare Smythii,Proc. Lunar Planet. Sci. Conf., 6th,
Conf, 28th, 571-572, 1997.
2541-2551, 1975.
J6nnson,J., Eldgos • s0gulegumtfma • Reykjanesskaga
(Volcanic St/Sffler,D., D.E. Gault, J. Wedekind, and G. Polkowski,Experimental
eruptionsin historicaltime on the ReykjanesPeninsula,southwest
hypervelocityimpact into quartz sand: Distributionand shock
Iceland)(in IcelandicwithEnglishsummary),Ndmirufraedingurlun, metamorphism
of ejecta,J. Geophys.Res.,80, 4062-4077,1975.
52, 127-139, 1983.
Taylor,S.R., Growth of planetarycrust, Tectonophysics,
161, 147-156,
Lucchitta,B.K., The Apollo17 landingsite,Nature,240, 259-260,1972.
1989.
Mason,R., J.E. Guest,andG.N. Cooke,An Imbriumpatternof graben Tolan, T. L., S. Reidel, J.L. Anderson, M.H. Beeson, K.R. Fecht, and
on the Moon, Geol.Assoc.LondonProc., 87, 161-168,1976.
McEwen, A.$., A preciselunar photometricfunction,Lunar Planet $ci.
Conf, 27th, 841-842, 1996.
McGetchin, T.R. and J.W. Head, Lunar cindercones,Science, 180, 6871, 1973.
Melosh,H.J., The tectonicsof masconloading,Proc. LunarPlanet.$ci.
Conf, 9th, 3513-3525, 1978.
Metzger, A.E., and R.E. Parker,The distributionof titanium on the lunar
surface,Earth Planet. Sci. Lett., 45, 155-171, 1979.
D.A. Swanson, Revisions to the estimates of the areal extent and
volumeof the ColumbiaRiver Basalt Group, Spec. Pap. Geol. Soc.
Am., 239, 1-20, 1989.
Whitford-Stark, J.L., Charting the southernseas: The evolutionof the
lunar Mare Australe,Proc. Lunar Planet. Sci. Conf, loth, 2975-2994,
1979.
Whitford-Stark,J.L., A preliminaryanalysisof lunarextra-marebasalts:
Distribution,compositions,
ages,volumes,and eruptionstyles,Moon
Planets, 26, 323-338, 1982.
Metzger, A.E., J.I. Trombka, L.E. Peterson,R.C. Reedy, and J.R.
Arnold, Lunar surface radioactivity: Preliminaryresultsof the Wichman, R. W., Post-impactmodificationof craters and multi-ring
basinson theEarthandMoon by volcanismand lithospheric
failure,
Apollo 15 and Apollo 16 gamma-rayspectrometerexperiments,
Science, 179, 800-803, !973.
Neumann,G.A., M.T. Zuber, D.E. Smith,and F.G. Lemoine, The lunar
crust: Globalstructureand signature
of majorbasins,J. Geophys.
Res., 101, 16,841-16,863, 1996.
Nozette, $. et al., The Clementine mission to the Moon:
overview, Science, 266, 1835-1839, 1994.
Scientific
Peterson,D.W. and R.B. Moore, Geologichistory and evolutionof
geologicconcepts,
Islandof Hawaii, U.S. Geol. Surv.Prof Pap.,
1350, 149-189, 1987.
Pieters,C., T.B. McCord,$. Zisk,andJ.B.Adams,Lunarblackspotsand
natureof the Apollo 17 landingarea, J. Geophys.Res., 78, 58675875, 1973.
Pieters,C., T.B. McCord, M.P. Charette,andJ.B. Adams,Lunar surface:
Identification
of the dark mantlingmaterialin the Apollo 17 soil
samples,Science,183, 1191-1194, 1974.
Pieters, C.M., $. Tompkins, G. He, J.W. Head, and P.C. Hess,
Mineralogyof the mafic anomalyin the SouthPole-AitkenBasin
(SPA): Implications
for excavation
of the lunarmantle,Geophys.
Ph.D. thesis,Brown Univ., Providence,R. I., 1993.
Wichman, R. W., and P. H. Schultz, Floor-fractured craters in Mare
Smythiiand west of OceanusProcellarum: Implicationsof crater
modificationby viscousrelaxationand igneousintrusionmodels,J.
Geophys.Res., 100, 21,201-21,218, 1995.
Wilhelms,D. E., The geologichistoryof the Moon, U.S. Geol. Surv•
Prof Pap., 1348, 302 pp., 1987.
Wilhelms, D. E., and F. E1-Baz, Geologic map of the east side of the
Moon, U.S. Geol. Surv. Map, 1-948, 1977.
Wilhelms,D.E., K.A. Howard, andH.G. Wilshire, Geologic map of the
southside of the Moon, U.S. Geol. Surv.Map, I-1162, 1979.
Williams, D.A., R. Greeley, G. Neukum, R. Wagner, and S. Kadel,
Multi-spectralstudiesof westernlimb andfarsidemariafrom Galileo
Earth-Moon Encounter 1, J. Geophys. Res., 100, 23,291-23,299,
ß
1995.
Wilson, L., and J.W. Head, Lunar linear rilles as surface manifestations
of dikes: Theoreticalconsiderations,
Lunar Planet. Sci. Conf.,27th,
1445-1446, 1996.
Pike,R.J.,Apparentdepth/apparent
diameterrelationfor lunarcraters,
Wolfe, R.W., and F. E1-Baz, Photogeologyof the multi-ringedcrater
Haldanein Mare Smythii, Proc. Lunar Planet. Sci. Conf, 7th, 2903-
Proc. Lunar Planet.Sci. Conf.,8th, 3427-3436, 1977.
Pike,R.J.,Geometricinterpretation
of lunar craters, U.S. Geol. Surv.
Yingst,R.A., andJ.W. Head,Lunarmaredepositvolumes,composition,
Res. Lett., 24, 1903-1906, 1997.
Prof Pap., 1046-C, C1-C177, 1980.
Pollard,D.D., P.T. Delaney, W.A. Duffield, E.T. Endo, and A.T.
Okamura, Surface deformation in volcanic rift zones,
Tectonophysics,
94, 541-584, 1983.
Pullan, $., and K. Lambeck, Mascons and loading of the lunar
lithosphere,Proc. Lunar Planet.Sci., 12B, 853-865, 1981.
Robinson,
M.$., B.R. Hawke, P.G. Lucey, and G.A. Smith,Mariner 10
2912, 1976.
age, and location: Implicationsfor source areas and modes of
emplacement,Lunar Planet.Sci. Conf, 25th, 1531-1532,1994.
Yingst,R.A., and J.W. Head, Characteristics
of mare depositson the
eastern limb of the Moon: Implicationsfor magma transport
mechanisms,
Lunar Planet.Sci. Conf, 27th, 1479-1480,1996.
Yingst,R.A., andJ.W. Head,Volumesof lunarlavapondsin SouthPoleAitken Basin and Orientale Basin: Implicationsfor eruption
11,158
YINGSTAND HEAD:SMYTHII AND MARGINISBASINLUNARMAREDEPOSITS
conditions,transportmechanisms,
and magma source regions,J.
Geophys.Res.,102, 10,909-10,931,1997a.
Yingst,R.A., andJ.W. Head, Multispectralanalysisof mare deposits
in
SouthPole-Aitken basin,Lunar Planet. Sci. Conf.,28th, 1609-1610,
R. A. YingstandJ. W. HeadIII, Department
of Geological
Sciences,
Brown University,Box 1846, Providence,RI 02912 (e-mail:
[email protected])
1997b.
Zuber,M.T., D.E. Smith,F.G. Lemoine,and G.A. Neumann,The shape
and internal structure of the Moon from the Clementine mission,
Science, 266, 1839-1843, 1994.
(ReceivedJuly29, 1997;revisedFebruary18, 1998;
acceptedMarch 3, 1998.)