JOURNAL OF GEOPHYSICAL

RESEARCH, VOL. 100, NO. B12, PAGES 24,509-24,519, DECEMBER

10, 1995

The dynamics of rapidly eraplaced terrestrial lava flows and implications for planetary volcanism StephenBaloga ProxemyResearch, Inc., Laytonsville,Maryland NASA Goddard SpaceFlight Center, Greenbelt, Maryland

Paul D. Spudis Lunar and Planetary Institute, Houston, Texas

John E. Guest Universityof London Observatory,London, England

Abstract. The Kaupulehu 1800-1801 lava flow of Hualalai volcanoand the 1823 Keaiwa flow from the Great Crack of the Kilauea southwestrift zone had certainunusualand possibly unique propertiesfor terrestrialbasalticlava flows.Both flowsapparentlyhad very low viscosities, high effusionrates,and uncommonlyrapid ratesof advance.Ultramafic xenolith nodulesin the 1801 flow form stacksof cobbleswith lava rindsof only millimeter thicknesses.

Thevelocity of thelavastream in the1801flowwasextremely high,at least10m s-• (more than40kmh-•). Observations andgeological evidence suggest similarly highvelocities forthe 1823flow. The unusualeruptionconditionsthat producedtheselava flowssuggesta floodlike mode of emplacementunlike that of most other present-dayflows.Although considerable efforthasgoneinto understanding the viscousfluid dynamicsand thermalprocesses that often occurin basalticflows,the unusualconditionsprevalentfor the Kaupulehuand Keaiwa flows necessitatedifferent modelingconsiderations.We proposean elementaryflood model for this type of lava emplacementand showthat it producesconsistentagreementwith the overall dimensionsof the flow, channel sizes,and other supportingfield evidence.The reconstructeddynamicsof theserapidly emplacedterrestrial lava flowsprovide significant insightsabout the nature of these eruptionsand their analogsin planetaryvolcanism. Introduction

The 1800-1801 eruption of the Hualalai volcano and the 1823eruptionof Kilaueavolcanoprovideunique opportunities for investigatingthe dynamicsof rapidly emplacedlava flows. These eruptions featured unusually high rates of discharge combined with relatively low lava viscosities, conditions thoughtto be presentin certainplanetarysettings.The overall dimensions,planimetricforms,and generalmorphologiccharacteristicsof these Hawaiian flows seem to suggestphysical processesof emplacementthat are usually consideredcommonplacein basalticvolcanism,for example,laminar viscous fluid flow, coolingby radiation, incipient crystallization,and the formation of an insulatingsurficialcrust. However, it has longbeenknownthat theseflowsare relativelyunusualamong terrestrial basaltic eruptions [e.g., Stearns,1926; Richter and Murata, 1961] in featuringsuchhigh volumetricrates of effusion and low magmaviscosities. More than 50 yearsago,Nichols[1939] correctlypointed out that sufficientlyslow-movinglava flowsshouldbe treated quantitatively by laminar viscousfluid dynamics,rather than a semiempiricalformalismfor turbulentwater flow. Nicholsalso noted, in effect, the definingconditionfor a sufficientlyslowmoving laminar flow, namely, that the movement of any iliaCopyright1995 by the American GeophysicalUnion. Paper number 95JB02844. 0148-0227/95/95 JB-02844505.00

ment within the flow is locally parallel to all other filaments. With noteworthyinsight,Nichols[1939,p. 294] proclaimedthat inferences about the dynamicsof lava flows "can be determined only if it is known whether the flow was turbulent or laminar." A slight overstatementof the casefor viscousfluid flow precedesthis insightful remark, that "gravity does not succeedin makingthe liquid of low viscosityflow more rapidly than the moreviscousliquid,becausethe liquid of low viscosity uses up energy in turbulence.... "When a free, unconfined upper surface is present, additional volumetric flow rate can alwaysbe usedto increasethe meanflowvelocity,althoughthis is not as efficient for turbulent flows. Inferences drawn by Nicholsfor the Alika flow of the 1919Mauna Loa eruptionand a distalsegmentof the McCartysflow in New Mexico included a laminar flow regime and flow advance rates of about 2 m/s and 0.2 m/s, respectively. Numerous works treating lava emplacementwith the conceptof laminar,viscousfluid flow haveappearedsinceNichols' publication. With the aid of more detailed observationsof basalticlava flow emplacement,we know today that the concept of laminar flow with all elements moving in parallel is certainlynot alwaysvalid. In addition,advancerates for many basaltic lava flows have been significantlygreater than the range of values inferred by Nichols. The classical view of turbulence

in fluid flow features

statis-

tically correlated,time-dependentfluctuationsin the components of flow velocity. We still cannot make the assertionof turbulencein lava flows in the classicalsense.However, pure 24,509

24,510

BALOGA

ET AL.: RAPIDLY

laminar flow is clearly somewhat of a special case due to mixing, disruptionof streamlines,irregular flowbedtopography, breaksin slope,and other factors[seeCrispand Baloga, 1994].It is unfortunatethat the term "disruptedflow" [Daughertyand Franzini, 1965] has never been widely acceptedin the scientificliterature for naturally occurringconditionsbetween pure laminar flow and classicalturbulence. Field evidence and several critical observations

indicate that

the emplacementof the 1800-1801 Hualalai and 1823 Kilauea flows resembleda rapid flooding by lava, rather than a relatively slow-moving,laminar advanceof a coolingand crystallizingviscousfluid. In thiswork we resurrectan approachthat has residedin desuetudefor more than half a centuryand proposea floodlike model for low-viscosity, high effusionrate flows. We show that a reconstructionof the flow dynamics providesquantitativeand qualitativeresultsthat are consistent with severaltypesof field evidence.The dynamicsof suchflows are scaledfor gravitydifferencesto illustratethe influenceof a planetarysettingon theseunique terrestrialanalogs.

EMPLACED

LAVA FLOWS

dentlyestimatea flowvelocityneartheventof at least10 m s-•, baseduponthe uphillflowof lavaarounda preexisting cone. Lavachannels nearthebedsof xenolithnodulescanbe quite large, up to 80 m acrossand 18 m deep (Figure 3). The drainageof the nodulebeds,both towardthe upperreachof theflowandbelowthehighway, thepresence of lavaoverspills, and the lack of a dramaticlongitudinalincreasein flow thickness,all provideevidencefor a very low viscosityhostfluid. The 1823 Kilauea

Lava Flow

The 1823 Keaiwa lava flow from the southwest rift zone of

Kilauea volcano also possesses several remarkable features

that suggest unusualeruptionconditions[Steams,1926].The flowis a basaltictholeiiticpahoehoelavaflowthat changes to a slabbyaa surfaceat its distalmargins.The flowventis a linear featurecalledtheGreatCrackandisbetween 1 and10mwide(in mostplacesaveraging about3 m) and about10 km in length (Figure4). Theflownearitssource ventisthin,ranging from2 to 30 cm. Nowhere does the flow thickness exceed about 2 m

[Steams, 1926]and1.5m isa reasonable estimate of theaverage.

Geological Observations of the 1801 and 1823 Lava

Flows

The flowtraveleddownslopeat an angleto itssourcevent.Thus flow lengthsvaryfrom 1.5 to about5 km. The total area of the

flowisabout25km2,ofwhichsome unknown butprobably small

Reconstructingthe dynamicsof flow emplacementrequires fractionis submarine[Steams,1926]. severaltypesof geologicalobservations to ascertainthe imporObservations of the eruptiontestifythat lavaissuedfrom the tant physical processesthat governed flow behavior. These Great Crackveryrapidlyin the form of a sheetthat quickly observations

include the fluid behavior

of the molten

lava at

flowedto the ocean,overwhelming a villageand overtaking

various stationsalong the path of the flow, field evidenceon fleeingnatives[Ellis,1842;Stearns,1926].Althoughthis flow flow behaviorand fluidity, and dimensionaland morphologic containsno xenoliths,accretionary ballsof lava displayvery considerations. Detailsof the knownandinferredpropertiesof thin (centimeter sizeor smaller)layersof lavaplastering, sugthe flows considered

The 1801 Hualalai

here are summarized

Lava

below.

Flow

The geologyof thisflow, alsoknownasthe Kaupulehuflow, is describedin detail by Richterand Murata [1961],McGetchin and Eichelberger[1975],Jacksonet al. [1981], and Guestet al. [1995].Briefly,it is an alkalicbasalticaa flow,havingabundant channels,erupted from a fissurezone at about 1700 m elevation alongthe northwestrift of Hualalai volcano(Figure 1). The flow trendsdownslopeto the north for a distanceof about

gestinga veryfluidmagma.Duringthe eruption,flowadvance wassorapidthat the lavaactuallyoverrodea preexisting cone, as occurred near the vent of the 1801 Hualalai

lava flow. The

lavaapparently wentabout11m up thesideof a preflowcinder cone (calledthe lava-plastered cones[Stearns, 1926]).This evidencesuggests velocitiesof flow of at least 15 m s-• comparableto the velocityinferred for the 1801 Hualalai lava flow [Guestet al., 1995].

15 km to the sea and, from there, about another 5-6 km into

Elementary Flood Model

the ocean.The remainingdepositof lava is relativelythin at the margins of the flow with thicknessesof 2-3 m near its sourcevent and thickeningonly slightlyat the marginsto about

The lowviscosity of theselavaflowsandtheir rapidratesof advancesuggests a type of geologicmassmovementmore like

a precipitous floodingof lava than a relativelyslow-moving, viscous fluidflow.The residualdeposits showthat,for practical The 1801 flow is remarkable for its beds of xenolith nodules purposes, viscosity changesalongthe flowpathswereinsignif[Richterand Murata, 1961].The xenolithsappearto havebeen icant in both cases.This contrastswith other basalticflows, emplacedas a cobblelag depositand, in someplaces,as an suchasthe 1A flowfrom the 1984Mauna Loa eruption,that overbankleveedeposit[Guestet al., 1995].Rindsof lavaon the underwenta rheologicchangeof severalordersof magnitude nodules(Figure 2) indicatethat the effectiveviscosityof the duringthe 5 daysof advance[Crispet al., 1994;Moore,1987]. host lava duringnodule emplacementwas extremelylow, of Crystallization of microlitesoccurredduringthe advanceof the 5-8

m near the ocean.

theorderof 102-103 P (10•-102Pas) [McGetchin etal., 19'76]. Mauna Drainageis similarat all locationsalongthe path of the flow where nodulebedsare exposed.There is no evidencefrom the rindsfor a significantsystematic changein viscosity,or other rheologicparameters,with distancefrom the vent.

Loa 1A flow, as well as the formation of a thick crust.

As the viscosityincreasedtoward the front of the flow, the

thickness increased dramatically to 25 m in places.Although

perhapsnot as dramatic,many of the lobes of the Puu Oo eruptionshowedsignificantthickeningwith distancefrom the The flow traversed the 15 km from the vent to the sea vent and a slowingof the advancerate that hasbeen attributed extremelyrapidly. The timescalefor the developmentof a to rheologicchanges[Finkand Zimbelman,1990].Unlike the significantcrust is typically tens of minutes to a few hours 1801and 1823eruptions,theseflowswere emplacedoverpe[Crispand Baloga,1990].The lack of any significantcruston riods,typically,of a few days. the flow suggestsan emplacementtime of that order. This For both the Mauna Loa 1A eruptionand many of the inferenceis alsosupported by the documented time of approxi- lobate Pu Oo flows,the durationof emplacement was suffimately1 hourquotedbyEllis[1842].Guestetal. [1995]indepen- cientlylongthata significant crustalcomponent formedontop

BALOGA ET AL.: RAPIDLY EMPLACED LAVA FLOWS

24,511

156 ø

20 ø

155 ø45'

I

Zone

L(km)

Slope (o)

1 2 3 4

3.5 4 5 2.5

12.5 9.5 3.8 1.2

I

W (m) 1400 1000-1300 1200 2000-3000

0

2

4

6

8

I

I

I

I

I

Contours

10 km I

in Meters

o

pACIF I 19o45 '

c flow

EXPLANATION

'Zo•

flow

120o

ß ..:.'...'. - *•:.• ..

•g•

....

..... :•.:-..::,..

•.•:::, ,:.::' ....

.....;.:...... ..... ..:.•":::•.' :-.•:..-............ •:,;•. •:;;:,: ;•;•;:•:•. ..... ......... :.?..:,;;,:;. ;•.:..::,• .......,.-*"•':'.,:-:. ......... • .. :'.......... :......,::•,;;•;•)•;.;., ............. ,•:;'•'""%" ,........ .

...........

½.' --"

.:.....;:..

•;:?•'};.--::

.:.::...

:.:... :..:.: ..:...,,.

..

. .

ß...x.

•;:;:.-•--x•-.•?-...•..,..: ::.-•.•. ;::--;::. '•:......... :..:

............... .......,.:•:-:•,: :.:,:,.:•:½ ::.::.....:.. •::.-::.;:•::.:;.•: