JOURNAL

OF GEOPHYSICAL

RESEARCH, VOL. 89, NO. B10, PAGES 8391-8406, SEPTEMBER

30, 1984

Calderas'A Planetary Perspective C. A. Wood

NASA JohnsonSpaceCenter

Caldera-bearing volcanoes arethelargestvolcanicedificesona planetandareexpressions of long-lived, near-surface magmachambers.On earthonlythreemorphological classes of calderasare common: shield calderas(e.g., Mauna Loa, Hawaii), stratoconecalderas(e.g., Crater Lake, Oregon),and ashflow calderas (e.g., VailesCaldera,New Mexico). Eachof theseclasseshasa characteristiccalderasize,magmatype,and tectoniclocation.No calderashavebeenrecognizedon the moon or Mercury, but Mars haslarge shield calderasas well asthe uniquehighlandpateracalderas.The latter may representan explosive,ultramarie styleof volcanism,unlikeany volcanicactivityin modemearthhistory.Radarimagessuggest that Venus hasat leastone largecaldera-bearingshieldvolcanoand possiblecalderasof othertypes.Jupiter'ssatellite Io hasabundant, large shieldlikecalderasbut no shieldmassifs.The numberand variety of calderason a planetmay be proportionalto planetmass,consistentwith a previoussuggestion that volcaniccomplexity increases with planetsize.Io isan exception,butheatingin Io isgenerated tidallyratherthanby thedecayof radioactive elements,so that its calderasare not related to its size. The occurrenceof shieldvolcanoeson four solar systembodiessuggests that thermal and tectonicmechanisms(i.e. hot spots)giving rise to constructionof shield mountainsand associatedcollapsedsummitsare basic processesin planetary geology.Hot spotvolcanismisexpectedto be mostimportant on planetslackingplatetectonics.On earth, largeshieldshavelifetimesof onlya few millionyears,but on Mars, largeshieldsmayhavebeenactivefor

1.5x 109years. Shield calderas oneartharesmaller than20kmindiameter, butknownorsuspected shield calderason Mars, Venus,and Io haveconsiderablylargerdiameters.Assumingthat calderadiametersare relatedto magmachamberdiametersanddepths,magmachamberson otherplanetsare significantlylarger and/ or deeperthan their counterpartson earth. Terrestrialmagmachambersare smallbecauseterrestrial shieldvolcanoesare smallin comparisonto shieldson Mars and Venus,and terrestrialshieldsare small becausethe earth'smobileplatesmovegrowingshieldsaway from their deepsources,thusdistributing magmaalonga narrowchainratherthan pilingit up arounda centralvent. Finally, the longlava flowsof Mars are also explained if the volume of erupted magma has any simple relation to magma chamber volume.

INTRODUCTION

Much newintereststemmedfrom studiesof ignimbritesand the recognition of resurgentcalderas as a distinct caldera type Volcanoesare morphologicalexpressions of magmacompo- [Smith and Bailey, 1968].Following Williams[ 1941],calderas sition,eruptionprocesses, andregionaltectonics.For example, are commonlydefinedas large, more or lesscircular,volcanic observationof the low-angleshieldshapeand longlavaflowsof collapsedepressionswith diametersconsiderablylarger than OlympusMons led to interpretationsof basalticmagmas,effu- any includedvent. Somewhatarbitrarily, calderashave been sive eruptions, and possiblehot spots on Mars [Greeley, definedashavingdiameters> 1km [e.g.,Francis,1976]or > 1mi 1973; Cart, 1973]. Similarly,the lack of unambiguoussilicic [e.g., Macdonald, 1972],with smallerdepressions beingcalled volcanoesand pyroclasticflowson Mars impliesthat compared craters.In practice,thesediametersdo separatenearlyall crato earth, itscrustisdeficientin volatilesand/ or the planetlacks ters (i.e., scoria cones,maars, tuff rings, and summit pits in a thick graniticcrust[Francisand Wood, 1982].Suchdeduc- stratovolcanoes), mostof whichare simpleeruptionvents,from tions are examplesof comparativeplanetology,usingknowl- depressions resultingfrom collapseinto partially drainednearedgeof landformsand processes on one planet (typically, the surfacemagmachambers.The only real ambiguityin usingthis earth) to interpret featuresand historieson others. In this diameterdependentdefinition occursfor maars,which are as reviewI considercalderasin a comparativeplanetologycontext large as 3 km. However, the mediandiameter of maarsis only to provideperspective for studentswho investigateonlyterres- 0.8 km [ Wood, 1984a],whereasthe mediandiameterof the 220 trial calderasandto searchfor broadscalerelationships between calderaslistedby Pike and Clow [1981] is 7 km. calderasand planetarycharacteristics. Caldera

Classes

TERRESTRIAL CALDERAS

As a guideto interpretingpossiblevolcanicdepressions on other planetsand satellites,I first summarizeinformation on the types, geometries,numbers,distribution,and formation processes of terrestrialcalderas.Investigationsof calderaorigins were widespreadduring the first third of this century,but little newwork appearedfor 2:5yearsfollowingthe publication of Williams'[1941]classicmemoir"Calderasandtheirorigins."

Thispaperisnot subjectto U.S. copyright.Publishedin 1984 by the AmericanGeophysicalUnion. Paper number 4B0330.

Althoughvariousclassificationschemesdistinguishup to 12 caldera typesbasedon morphologyand eruption mechanism [e.g., Oilier, 1969; Williams and McBirney, 1979; Pike and Clow, 1981],only threegeneralmorphologicalclassesof calderas are important on the earth. Theseare calderasin basaltic shield volcanoes, calderas that behead stratovolcanoes, and

large calderas associatedwith broad ash flow sheets.These classeswill be referred to here as shield calderas, stratocone calderas, and ash flow calderas to avoid conflicts with traditional, more restrictivetypes named after specificexamples (e.g., Hawaiian, Krakatoan, Valles,etc.) and to providea morphologicalclassificationnecessary for planetarycomparisons.

8391

8392

WOOD: CALDERAS--APLANETARY PERSPECTIVE

.

,.

i'}-;:::.; .:..

:,,.>./'.

%....:

-;.-.:.•:,...'...:.;.. ,/......:.(-:'•;'• .... .:.... .......

• ".•:,%:::..,.•.....:.........• .... .•>;....:. ß ...:... -=•f ;....'•.

...?..

..... •,,(:.... :: •.:'-;•--:'½. .,..

....

.

.

.:..x.,.......:....., ..,.. :.h...::...,:.,7;/: '

'.;',-.-.-•., ,•-. L' "•{•2 .;•-; """ ........

: ,-;i:•:-;" .., ............ ½7:•.'"":-.' "--" •"*"""; "'::' "7•':"/•'•':;" •½..--: ............................ :-.:• '-*"•'i ,.:t.• •'"t:"•T-•?• -..•- "? 1 km and hence, by zones/ rifts, and 20%at both oceanicand continentalhot spots. general usagebasedon terrestrial experience,would be called Accordingto the presentdata, a.shflow calderasare equally calderas.However,if pitsform simplyby collapseinto underlyabundant in rifts/extensiona,l zones and above subduction ing voids(e.g., by collapseof a frozen lava lake roof after the zones,but manyolderfeaturesuncataloged by Pike[ 1978]exist lake has drained), then lunar voidscould be 6 times larger than in theBasinandRangeofthewestern UnitedStates.A.dd!tion- terrestrial voids before a solidified roof would collapse.This ally, ongoingmapping suggeststhat a few hundredash flow number,basedonly on gravitydifferences,is reducedto 4.6 if (3.3 g/cm3 versus2.6 for calderasexistin Mexico[Swansonand McDowell, 1983].Thus, the higherdensityof lunarbasalts although ash flow calderasoccur elsewhere,the Basin and earth basalts)is considered.Thus,exceptfor planetarydifferRange extensionalpr0yince of westernNorth America may encesin gravity, craterson lunar domeswould probablybe < 1 have the highestconcentrationon earth. Note also that the km. Therefore, despitetheir size, lunar dome cratersare no diameters of ash flow calderas in the East African rift are more calderasthan are the pits on small Icelandicshieldvolcaconsiderablysmaller(averageD = 11kin) thanthosein the rest noes; hencethe standardcalderasizedefinition is not adequate of the world (averageD = 24 kin), suggesting that theremay be to distinguishdifferentcraterstructuresfound on otherplanets. Although some scientistshad earlier argued that nearly all tectonicallydistinctsubclasses of ash flow calderas. This summaryof terrestrialcalderasdemonstratesthat cal- lunar craterswere of volcanicorigin [e.g., Simpson,1966],an deraclasses, originallydefinedonthebasisof morph•ology, also overwhelmingvariety of evidencefrom crater statistics,field have distinct ranges indiameter, petrolo•y, andtectonic setting.mapping, shock studies,and experimentalcratering clearly demonstratesthat almostall craterson the moon (and Mercury, LUNAR CALDERAS? Mars, and varioussatellites)wereformedby impactsof cosmic debris(see Taylor [1982, p. 62] for a reviewof the volcanic/ Volcanismhasbeenpervasiveon the moon, but nearlyall is impact controversy).Thereare, however,a few cratersthat do

and near spreadingcenters(Tortuga Island, Gulf of California; Erta Ale, Ethiopia). Fifty-five percentof catalogedstrato-

represented by thevastlavaflowsthatmakeuptheflat mare plains[Head, 1976].Conspicuousvolcanicconessuchas large

TECTONIC ENVIRONMENTS TERRESTRIAL CALDERAS

shield volcanoes or stratovolcanoes do not exist on the moon.

Identifiable volcanic edifices include small cones, dark halo

craters, anddomes [Head,1976], withthel•ossibility thatcertain impactlikecratersmay also be volcanic. Lunar domesare low, roughly circular structureswith con-

vexslopes andsummit craters (Figure 4).Basal diameters range from 3 to 17 km, flank slopesare lessthan 5ø, and summit cratersare 1-3 km across[Head and Gifford, 1980]. Most researchers[e.g. Head and Gifford, 1980; Strain and EI-Baz, 1980]concludethat lunardomesare small,monogenetic shield

o

7O

so

5O

Z

O 3O

30

lO

10

SR R HSSZ SHIELD CALDERAS

of Idaho. The majordiff•erence is that •ummitpitsin terrestrial shield'sbasal diameters),whereaslunar domeshave kilometer

-

7O

volcanoes similar to those in Iceland and the Snake River Plain

smallshields areonlyatfewhundred meters w•de(6-10%of the

OF

SR:SPREADING

Fig. 3.

SR R HSSZ STRATOCONE CALDERAS

RIDGE, R=RIFT,

HS:HOT

SR R HSSZ ASHFLOW CALDERAS

SPOT, SZ:SUBDUCTION

ZONE

The distribution of tectonicenvironmentsfor terrestrial cal-

sizecratersthat are 19-26%of theirbasaldia,meter[Headand dera classes,basedon data of Pike [1978], as combinedin Figure 2.

Gifford, 1980].This observationintroducesthefirst complica- These statisticsinclude 35 shieldcalderas, 131 stratoconecalderas,and tion (i.e. scale)when comparingcalderason earth and other 33 ash flow calderas.

WOOD:

CALDERAS--A PLANETARY PERSPECTIVE

8395

Fig. 4. Lunardomes,suchasthesenorth of the craterHortensius,are low moundswith centralpits. Domesare apparently small, monogeneticshieldvolcanoessimilar to thosethat occurin Icelandand the SnakeRiver Plainsof Idaho. Theselunar domeshavebasaldiametersof about 7 km and summitpitsabout 1.5km wide accordingto Head and Gifford [ 1980].Lunar orbiter frame IV-133HI.

not have the morphologyof normal impact cratersand may be volcanic in origin. The crater Kopff, for example, has equal angle inner and outer rim slopes,is relatively shallowwith a smooth floor, and has no central peak (Figure 5). The feature that most clearly suggests that Kopff is unusualis its secondary craters,whichare smallerand closerto Kopff than is the case for impact craters (R. Strom, personal commmunication,

1983). Despite theseanomalies,Kopff might simply be an impactcraterformedin someunusualcircumstances, and in fact it doesnot look like any terrestrialvolcano.

The discussionof possiblelunar volcanic features could be extendedto includeother inconclusiveor unlikely possibilities [e.g., Green, 1971]. A fundamental point about lunar volcanism, however, is that there is no compelling evidence for calderassimilar to any of the terrestrial classes,nor for any uniquely lunar types. As Head [1976] pointed out, shallow, crustalmagma chambers,as requiredfor caldera collapse,do not appear to occur on the moon. This may be due to the prevalenceof deep fractures associatedwith the formation of large impact basins,which allowedmagmasto risedirectlyto

8396

WOOD: CALDERAS--A PLANETARYPERSPECTIVE

,.

ß..,,:. ;?

...... c.' .......

; ?'".'?'

....

..:•..::..•':}.'."-¾ '4"'.

:::.. •.,: • •:-:: ........ :,•:-.:. ..... •

..........

.... ;:: ,,:...;:;•:• -,:: ..... .::.-:-,-.;...:,-::; .... ,...,..,:..... ..:.... :½•*? ..;X;•; ....... ..;•,:::; :;:>.::.; ....... ,..: ....... ,.:, --.:: ..., ...... .•:'.,.,:.: ':;,;"::.:,. '":":L•:::':',';;;;":¾;:-**'.': •.::.: ...... .':'"-;:' ';;; t./:'..',½*?:,.' :...•½.,/•:;,,,.½ ....

......:.......... :.-:: ',:'".::::: •,.........

•.'

ß ....:.:,::.::•:.**s ........................... .:,:-,:,';:--•:--•;."',;.::,.."... '?'-• :!:/":"::'*".. ," ....... '.*:":**• ......... '"">..,;;•:,"'. ....?;" ',.:.'•?xa•?'f•:?g•7.•:;;..'*"-:/:? ........ ... -'.: .........

?. ':. '- ß:?' ""-,::**-,,'L..**::.' *'--•'" *'.....,-.4':-•;*./:-:-:.;,; '

..*•". ......... :'.• ........... ":;•....... -: .'

ß "/'•:"'. ?":%'.?** %-'.: ;-'-•: ';?•?' **:'""*;•:•-**:.½:;:-.. ::?•-::;::.:;*'

'....:".:":."'. ".. ,, ....-;:;' ..7-.'....-.':.-' ...:"

,;,........... ----;: . :•::.::.;..:-;•:"...:...,.•" 20

km, whereas theearth,with3 •/2timesmorelandarea,may have an order of magnitudefewer large calderas. The large sizeand low relief of calderason Io suggested to Carr et al. [1979] and Schaber[ 1980, 1982]that ash flow calderassuchas that at Yellowstone,Wyoming, might be the best terrestrialanalog.Thiscomparison,basedsolelyon thegeometriccharacteristics of thecalderas,overlookssignificantmorphologicaldifferences.The scallopedrimsand the steepwall scarps of the Ionian calderas are also characteristic

of terrestrial shield

calderasbut are rare in ashflow calderas.Similarly,resurgent centers and rim domes are common

in ashflow

calderas on

earth, but havenot beenidentifiedin any calderaon Io. Additionally, the long, narrow flows radiatingfrom many calderas terrestrial shield volcanoes, However, the Ionian craters are not on Io (Figure 13)are similar to the lava flows whichconstruct formedon mountainsbut appearto lie on flat plains[ Carret al., terrestrial shieldvolcanoes[Carr et al., 1979; Schaber,1980, 1979].The diametersof 93 calderasin the 12.5%of Io with the 1982],but whichare rare on ashflow calderavolcanoes.On the highest-resolutionrange from lessthan 14 km (the effective basisof the radiatingflowsand calderamorphology,the Ionian resolution)to 200 km and average54 km [Schaber,1982].Carr cratersseemmost like the calderason tops of terrestrialshield

WOOD:

CALDERAS--A PLANETARY PERSPECTIVE

..:•..i!

......

..:.:" ..

•

''

-'....... .',.;

•,...... .•-:- ,:; ....

4'

--,...

.•

...;;7'"' ,.;,

: --".

"•'•",,: .i•:.:"•:'' ';

.......

.

,:;:• : ..:....

....... ::,--. ::•.....

.:** ...................

.';•; ....... :::-•

....,:.;--;.,.-'-

..... ..:

"'.:..L.}

•'---.. .....

.,

•

.:.-. .:..

.•-.: .... '::'. :.

.

..... •:. 4

: ..:......

'"?:-...

'""'"• .... .::. ::.,

..;• •;,.:..'-

,,,;?.; -½:::::....:.-

......... ...-?:,. ..

-:::::---"-:./:•-;-, ,'.....:'-; .......... ; .....--"•,,..,::',::. -:-:: "-;%..;%.-.; *"'

.

.

-..

I':•.

'................ 8...................... •:':-:??....... ....;'.... ;' ................................. .... ' ;;'--•,,;• ....... %;::(:::'??' ........... ½;•%::"•': ...... :•;, ./' ...- ..

..... ;•,

%

..

;*'"'""% %..;.,;-:-*".' -'½

'%* ....:•"::.:.? ......

. ;•....

::.•..;:,.,•. .:..:. --

...,

•.--? ".;.;::,... ,:,.•..,,.;,:., .... .:• ::,.-? . ..;.,• -½•; .... -• •

........

.•

.:.:.•;--•; ...•:.% .,........

....

"?•

....... .;......,• '"'2-' •-,,• •:•-;-

,...• '"'.

'•'%'""'".-m•' ---, :.;;;"-:'.--.--.0.., '""7"•;'4.

•,::,'....

.:

•,

,::.

'":•.•

. :

,.:;:.,;.::.-.....;

:.....