View metadata, citation and similar papers at core.ac.uk brought to you by CORE
provided by Lancaster E-Prints
JOURNAL OF GEOPHYSICALRESEARCH, VOL. 100,NO. El2, PAGES26,327-26,338, DECEMBER 25, 1995
Explosivevolcanism on Venus:Transient volcanic explosions as a mechanismfor localizedpyroclast dispersal
SarahA. Fagentsland Lionel Wilson EnvironmentalScience Division, Institute of Environmentaland Biological Sciences, Lancaster University, Lancaster, England
Abstract.It is proposedthat transient volcanic explosions of the vulcanian type may provide a mechanismfor thegeneration and dispersal of pyroclasticmaterial on Venus. The influence of the Venusianhigh atmospheric pressure environment implies that continuous discharge plinian erup- tiveactivity is relativelyuncommon: the tendency for suppression of exsolution and expansion of magmaticgases favors effusive eruptions. However, it maybe possible for explosive activity to occur,in a fashionanalogous to vulcanianeruptions on Earth, as a resultof theaccumulation of hot,pressurized gas under a coherentrock "lid". The explosion may be initiated by the failure of thisretaining caprock, causing the catastrophic release of thehigh-pressure gas, which expands outof thevent driving the fragmented caprock material ahead of it anddisplacing the surrounding atmosphere.On Earth the driving gas may originate either from vaporization of groundwater or fromdegassing of a stalledmagma body in thenear-surface crust, whereas on Venus, where the presenceof crustalstores of volatilecompounds isuncertain, the latter option only is favored: prolongeddegassing may lead to an accumulation of gas sufficient toinitiate an explosion. This paperpresents the results of a numericalmodel describing the explosion process under boundary conditionsrepresenting the Venusian physical environment. This involves treatments of the accel- erationof thedriving gas, caprock and displaced atmospheric gas out of thevent and the subse- quentmotions and aerodynamic interactions between the atmosphere and the ejected blocks of fragmentedcaprock. In thisway, predictions of the eruption velocities and of theresulting distri- butionof (large)solid ejecta can be obtained for likely conditions on Venus. Deposits of large blockydebris are predicted to rangeup to a maximumdistance of theorder of 1 kmfrom the vent onVenus, compared with distances of severalkilometers commonly attained by ejecta from tran- sientexplosions on Earth. More typical blocky deposits may extend for only a fewhundred meters,which implies that they would not be detected in theMagellan radar data. However, the possiblepresence of associatedpyroclastic flow and fine-grained ashfall deposits may constitute aidsto the identificationof sitesof vulcanianeruptions on Venus.
1. Introduction the formationof pyroclasticflows is morelikely [Sugitaand Matsui, 1993; Thomhill, 1993]. It may be expectedtherefore that The veryhigh atmospheric pressure on Venus(ranging from steadyexplosive (plinian or Hawaiian)activity on Venusis -10 MPa to -5 MPa over the -10 km rangeof planetaryeleva- relativelyuncommon, with effusiveactivity dominating. Indeed, tion) will tendto suppress,or at leastreduce, the exsolutionand analysesof the huge amountof high-resolutionradar data expansionof volatilescontained in ascendingmagmas [Wood, returnedby the Magellanmission to Venushave demonstrated 1979; Wilson and Head, 1983; Head and Wilson, 1986]. the widespreadoccurrence of vastvolcanic plains and immense Uncommonlyhigh (by terrestrialstandards) total magmatic lava flows [e.g., Head et al., 1991], with relatively few volatilecontents are requiredfor subsurfacemagma fragmenta- indicationsof pyroclasticactivity [Head et al., 1991;Guest et al., tion (and hence the initiation of explosiveactivity in which a 1992; Iranov, 1992; Moore et al., 1992; Wenrichand Greeley, steadydischarge of gasand pyroclasts takes place) to occurat all 1992]. Nevertheless,intermittent, transient explosive eruptions [Garvin et al., 1982;Head and Wilson,1986]. It is by no means arepossible at quitemodest magmatic volatile contents on Venus clear that these volatile contentsare achievable in magmas on if gasconcentration occurs in a slowlyascending or stationary Venus. Furthermore, exsolved volatile contentsin excessof intrudedmagma. The resultmay be Strombolianactivity in low- severalweight percent, together with favorablecombinations of viscosity,mafic magmas[Garvin et al., 1982] or vulcanian high ventaltitude and eruption temperature, are necessaryfor activityin moreviscous melts [Fagents, 1994]. eruptionsfeeding high, convectingeruption clouds to be Fagentsand Wilson[1993] developeda modelfor transient maintained.Otherwise the collapseof the eruptioncolumn and vulcanianexplosive eruptions on Earthbased on the scenarioin whichmagma intrudes close to the surfaceof theplanet but fails 1Now at Centerfor Earthand Planetary Studies, National Air and to erupt(most likely because the rise rate is sufficientlyslow that SpaceMuseum, Smithsonian Institution, Washington, D.C. excessivecooling intervenes). Exsolving volatiles may accumu- late at the top of sucha magmacolumn, possibly as a foambut Copyright1995 by theAmerican Geophysical Union. later, if the foam collapses[Jaupart and Vergniolle,1989], as a Papernumber 95JE03202. gas pocket. This magmaticgas may effectively be trapped O148-0227/95/95JE-03202505.00 beneatha rigid "lid" if the pathwayto the surfaceis sealed
26,327 26,328 FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS
(possibly by magma invading any near-surface fractures). 1992]. Whether this high-viscositymorphology is indicativeof a Alternatively, if potentially volatile compoundsexist within the moresilicic magma, or is simplya cooling-inducedfeature, or the country rocks, these may be evaporated and trapped as high- effect of a highly inflated bubble-richmagma, remains open to pressuregases if their path is not connectedefficiently to the debate [McKenzie et al., 1992; Pavri et al., 1992; Sakimotoand surface.The failure of part of the lid overlying an accumulation Zuber, 1993]. of high-pressuregases leads to the expansionof locally released Considerationof volatile solubilitieswith respectto the high gasesand the accelerationof the overlying rocks. Furthermore, atmosphericpressure on Venus causesdoubts as to whether local decompressionleads to the propagationof an expansion steadyexplosive activity can occur at all. For typical magmatic wave into the surroundings,and this can triggerfailure of moreof volatile contents,the high pressurewill act to inhibit or at least the lid, leadingquickly to the catastrophicdisruption of all of the reducethe exsolutionand expansionof magmaticgases [Wood, pressurizedregion in a vulcanianstyle explosion[Wilson, 1980]. 1979; Garvin et al., 1982; Wilson and Head, 1983; Head and The Fagents and Wilson [1993] model incorporatesa detailed Wilson,1986]. In the shallowlithosphere, lithostatic pressures are treatment of the mutual aerodynamic interactions between greaterthan on the Earth as a result of the high atmosphericsur- volcanicejecta and the gasflow field aroundthe eruptivesite and face pressure.As a result,the pressuresat which magmafrag- thusrepresents a significantimprovement over previousefforts to mentationtakes place (when the magmaticgas bubble volume model the dispersalof coarse volcanic material. Analysesof a fractionexceeds some critical value of-•0.75 [Sparks,1978]) are number of terrestrial eruptionshave shown that, by comparing always reached at shallower depths on Venus [Garvin et al., computed travel distances of large clasts ejected in such 1982;Head and Wilson,1986] and muchlarger volatile contents explosionswith their observedpositions in the field, it is possible are requiredif magmafragmentation is to take place at all. Table to constrainthe rangesof initial conditionsthat would explainthe 1 givesthe minimumtotal H20 andCO 2 contentsrequired for observeddispersal of ejecta[Fagents and Wilson,1993; Fagents, disruption of a basaltic magma at various vent altitudes.For 1994]. Geologicallyplausible values of theseparameters (excess rhyolitic magmas,Head and Wilson [1986] show that exsolved gas pressure,gas mass fraction, and sourceregion radius) are volatile contentsof severalweight percentare neededfor plinian obtained.Having testedthis model with terrestrialcase studies, it eruptioncolumns to form. Clearly, the minimumvolatile contents now seems appropriate, in view of the questions raised by requiredfor the disruptionof Venusianmagmas (i.e., to produce Magellan about Venus still being volcanically active, to employ continuousexplosive eruptions) lie aboveor at the upperend of the model to predict the extents of eruptive deposits from the range of typical magmatic volatile contentsof terrestrial transient volcanic explosions on Venus, given plausible magmas[Basaltic Volcanism Study Project, 1981]. preemptionconditions. Even if a magmaon Venus containssufficient volatiles for the A number of candidate sites for explosive vc•le•ni•m have been identified on Venus [e.g., Head et al., 1991, 1992; Ivanov, will also reducethe amountof gas expansion.This reducesthe 1992; Moore et al., 1992; Wenrichand Greeley,1992; Campbell, energyavailable to drive the eruption,which is a functionof the 1994; Campbelland Rogers;1994], with evidencefor quitelocal- ratio of the initial (preexpansion)and final (atmospheric)gas ized activity in somelocations [e.g., Bulmer, 1994]. The nature pressures,relative to that on Earth. The velocity reachedby the and modesof formation of thesedeposits are poorly understood. erupting fluid on reaching the surface will thereforebe rather The studypresented in this paperrepresents an investigationinto modest compared with terrestrial conditions. Thornhill [1993] how the productsof one well-known style of explosiveeruption showsthat steadyexplosive plinian activity,forming a high con- would be manifestedin the Venusian environment.Initially, we vectingeruption cloud, can only be sustainedfor unlikelycombi- review the reasonsfor anticipatingthat vulcanian-typeexplosive nationsof high volatile contents,high vent altitudesand high eruptionsare possibleon Venus andgive detailsof how the theo- eruptiontemperatures; fountain-fed pyroclastic flows are more retical model of Fagents and Wilson [1993] can be used to likely to occur in an explosive scenario.It seemslikely that describethis explosionprocess under Venusian physical environ- steady explosive ashfall activity, while not necessarilyabsent mental conditions.We then presentthe predictionsthat can be (tentativeidentification has been made of plinian ashfalldeposits made about the dimensionsand grainsize characteristicsof the in Bell Regio [Barsukov et al., 1986; Campbell and Rogers, depositsthat might be expectedto be produced. 1994] and Guinevere Planitia [Head et al., 1991]), is rare on Venus. 2. Background However, transient activity may be rather more common. Garvin et al. [1982] arguedthat Strombolianactivity is possible Volcanism on Venus in low-viscosityVenusian basalts if they ascendthrough the crest The analysis of Magellan synthetic aperture radar (SAR) sufficiently slowly. The low rise speedallows time for bubble images of Venus has revealed a startling array of volcanic growth by gas diffusion; then, as the buoyancy-drivenupward features, including shield volcanoesand lava flows on many accelerationof bubbles becomes more important, there are an scales,the morphologiesof which are consistentwith a basaltic composition[Head et al., 1991]. Venera Lander geochemical data also indicatepossible tholeiitic or alkali basaltsurface rock Table 1. Minimum Volatile Contentsin Weight PercentRequired compositions[Surkov et al., 1984, 1987]. There is someevidence for Magma Fragmentationon Venus that more evolvedmagmas exist on Venus:Venera 8 gammaray spectrometerdata suggestan intermediateto silicic composition Vent Altitude in km above mpr [Surkov et al., 1976, 1977; Nikolaeva, 1990] at one site. Volatile Species 0 5 10 Furthermore,the morphologiesof somesteep-sided, flat-topped "pancake" domes observed in plains areas of Venus are reminiscent(albeit on a larger scale)of rhyolite-dacitedomes on H20 2.8 2.1 1.5 CO2 5.3 3.9 2.8 Earth [Head et al., 1991, 1992; Guest et al., 1992; Pavri et al., FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS 26,329 increasingnumber of coalescenceevents and furthergrowth by ties saw the introductionof an expressionfor drag resistance, decompression.The end result is large, rapidly rising bubbles which is a function of velocity, atmosphericdensity, and block representing significant local volatile enrichment, which radius. Fudali and Melson [1972] and Steinbergand Steinberg ultimatelyburst through the surfaceof the melt, disruptingit into [1975] useda constantdrag coefficient, which is not in generalan a fine sprayof pyroclasts[Blackburn et al., 1976]. adequateassumption: given that the dragcoefficient is a function In a similar fashion,vulcanian activity can occuron Venus as of the Reynoldsnumber [Hoerner, 1965], the drag force will vary a resultof gasexsolution and accumulationunder a retaininglid as gas decompressionprogresses and thereforeas the relative wherea magmais risingslowly or hasstalled near to thesurface. velocityof theclast with respectto theatmosphere varies. Wilson Given sufficienttime for gasto accumulate,even low magmatic [ 1972]described a computationalscheme that allowed the trajec- volatile contents(of any species)can producea significantcon- tory of a pyroclastlaunched from a given point with a chosen centration beneath the caprock. Whereas the dense Venusian velocityvector to be followedby takingdetailed account of the atmospherewill rapidlysuppress the flight of therelatively finer- variation of aerodynamicdrag forces acting on the clast with grained Strombolianejecta, leading to rangesof only a few varying Reynoldsand Mach numbers.This was subsequently meters [Garvin et al., 1982], the larger piecesof fragmented used to infer the initial velocitiesof clastsejected in vulcanian caprockejected during vulcanian events, by virtueof theirgreater eruptionsby Nairn [1976], Steinberg[1977], Steinbergand inertia, will be able to attainsignificantly greater distances. Babenko[1978], andSelf et al. [1980]. The pressuresobtained from use of the MBE and GBE were all unexpectedlyhigh (e.g., 360 to 500 MPa for Bezymianny Transient Vulcanian Explosions 1955 [Gorshkov,1959]; 300 to 500 MPa for Bezymianny 1955 Vulcanian eruptionsare defined as discrete,intermittent vol- and Arenal 1968 [McBirney, 1973]; 490 MPa for Arenal 1968 canic explosionsseparated by intervalsof minutesto hoursor [Melsonand Saenz,1973]; 200 MPa for Ngaumhoe1975 [Nairn days[Wilson, 1980] and on Earth are typically associatedwith and Self, 1978]). The overpressurerequired for failureto occurin intermediatecomposition stratovolcanoes. Dense slugs of gasand a rock is no more than twice its tensilestrength [e.g., Tait et al., solid material are ejected at velocitiesthat may be supersonic 1989]. Touloukian et al. [1981] measuredtensile strengthsof relativeto the speedof soundin the surroundingatmosphere (as 8.6+1.4 and 13.8+2.1 MPa for pristine samplesof basalt and evidencedby the observationof shockwaves [Steinbergand granite,respectively; values for sedimentaryrocks are lowerthan Steinberg, 1975; Nairn, 1976; Livshitts and Bolkhovitinov, thoseof basaltby a factorof 2 to 3 [Tait et al., 1989].In volcanic 1977]), with pyroclasticejecta usually consisting of_>50 wt % terrains where fractured, scoriaceous,or mbbly overburdenis nonjuvenilematerial. A proportionof the ejectedmaterial con- present,tensile strengths are likely to be significantlyless than sistsof largelithic andjuvenile blocks projected on near-ballistic for laboratoryspecimens. Furthermore, field examinationof hot paths. Finer material is entrained into a convectingeruption dacitic lava blocks has indicatedthat their strengthsmay be low: cloud, which typically reachesheights on Earth of <5 to 20 km the blocks were easily shatteredby blows from a hammer [Cas and Wright, 1988] in single isolated events.On Earth, [Mellors et al., 1988; Sato et al., 1992]. These lower strengths small-volume pyroclastic flows are commonly, but not maybe attributedto thermalstresses, trapped residual gas or high necessarily,associated with vulcanianeruptions, as a result of straindue to the presenceof a largeproportion of crystalswithin rapid fall back of ejecta onto the summit area of the volcano. the viscousmelt. It follows that caprockconsisting of hot juve- However, the observed ran-out distances and inferred flow nile materialmay be expectedto displaysimilar properties of velocitiesare relatively small, which is consistentwith both the failure. Thus it may be arguedthat maximumtensile strengths limited volumes of material emitted and the low degree of may be of order 10 MPa, leadingto a maximumpossible excess fluidization occurring as a result of the high proportion of gas pressureof-20 MPa; this is at least 1 order of magnitude nonjuvenile(and hencecooler) rock. smallerthan the pressurescalculated from the MBE andGBE. The earliest work on modelingterrestrial vulcanian eruptions In a reappraisalof the fluid mechanicsof volcanicexplosive used the so-called modified Bernoulli equation (MBE) processes,Wilson [1980] showedthat the MBE and GBE are in [Matuzawa, 1933] or the gun-barrelequation (GBE) [Minakami, fact inappropriatefor describingany type of eruptionand pro- 1950; Decker and Hadikusumo, 1961] to give a relationship poseda treatmentwhich took explicit accountof the gasexpan- betweenthe velocity with which clastswere expelledfrom an sionduring explosions. This analysiswas readilyshown to lead explosionsite andthe pressurein the trappedgas responsible for to lower preemptionpressures [Self et al., 1979]. However, an the explosion.No very accurateestimates of eruptionvelocity assumptionemployed in settingup this model was that clasts have been obtained from direct observationsof vulcanian erup- were ejectedinto an atmosphereat rest. Fagents and Wilson tionson Earth. Estimatesderived from photographicmethods (in [1993] pointedout that the initiationof a transienteruption will the few caseswhere images with high enoughspatial resolution cause the atmosphereoverlying the vent to be displaceden have been obtained:Chouet et al. [1974], Nairn and Self[1978], masse,such that the initial speedof the displacedatmosphere will Steinbergand Babenko[1978], and Ripepeet al. [1993]) suffer be essentiallythe sameas that of the ejectedcaprock. As the from inadequateframing ratesto follow the rapid accelerations caprockraptures into fragments,individual blocks are launched and decelerationsat the onsetof the explosions.It is thus very into atmospheremoving at approximatelythe samespeed. The difficult to obtain the peak velocity (after expansionout of the implication is that initial drag forces acting on the clastsare vent) that is requiredto be linked with the excesspressure in the essentiallyzero, only becomingsignificant as the gas velocity vent. In early studies,initial clastvelocities, and hencepressures, decays,a strikingcontrast to the earlierwork in whichthe gas- were inferredfrom the observedranges of clastsusing the classi- clast relative velocity (and hencedrag force) would have been cal ballistic equationwhich neglectsatmospheric drag forces assumedto be initially very high. Thus the velocity (and [Gorshkov,1959; Decker and Hadikusumo,1961; Gorshkovand thereforeinitial excesspressure) required for a block to reacha Bogoyavlenskaya,1965; H•dervtiri, 1968; Steinberg,1977; Self given distance from the vent would have been greatly et al., 1980]. Attempts at accountingfor the atmosphericproper- overestimated.Fagents and Wilson[1993] showedthat, with due 26,330 FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS
allowancefor this effect, preexplosiongas pressures completely compatible with rock tensile strengthswere predicted for a (a) caprock numberof well-documentederuptions. Turcotteet al. [1990] employeda treatmentof the unsteady flow of a "pseudogas"through an expansionfan to model vul- canianeruptions. However, the assumptionof a pseudogas(in this case approximating finely divided magmatic material thermally and physically coupled with the exsolvedjuvenile phase)is inconsistentwith scenariosin which a largeproportion of the eruptiveproducts consists of blocky, nonjuvenilematerial driven ahead of the expandinggas phase.Field observationsof depositsproduced from transienteruptions in whichat leastsome juvenile gas was involved supportthe premisethat the erupted material is not a homogeneousmixture of gas and fine (b) particulates,but rather a highly nonuniform mixture of gas, juvenile material, and countryrock. The model of Turcotteet al. [ 1990] may thereforebe moreapplicable to transienteruptions in ' decaysas which material is much more finely fragmented,for example, phreatomagmaticevents in which a large proportion of the blocks'".
This radial variation of u representsthe continuityequation for magmasdo not containat leastsome H20 [Headand Wilson, the displacedatmosphere (treated as incompressible).The time 1986]. As an alternativeto the driving gasesbeing providedby a constantr is given by magmaticsource, the ascendingmagma may encountera volatile reservoir contained within the lithosphere.Water and carbon ,= t'- to, (3) dioxide cannot exist in solid or liquid phasesunder the present where t' is the durationof the whole "gasthrust" phase, at the end high-temperature conditions on Venus, but thermodynamic of which, if convective motion were not to take over, the gas calculations show that sulfur, if present, would be stable as a would have deceleratedto zero velocity (i.e., after the initial liquid almost anywhereon the Venusiansurface or at depthsup kinetic energyof the gas has beenexpended). This is foundby to several hundred meters [Head and Wilson, 1986]. The continuingthe integrationof equation(1) until the conditionu=0 possibilityof CO2 releaseby thermaldecomposition of carbonate is reached.The form of r representsan improvementover that sediments has also been considered [Head and Wilson, 1986]. given in Fagentsand Wilson [1993], in that equation(3) more Although it is not clear that country rock volatiles may be satisfactorilyaccounts for the fasteranticipated velocity decay on encounteredduring magma ascent on Venus, there is every Venus relative to the Earth, which would result from the much reason to believe that at least some magmatic volatiles will lower ratio of vent excesspressure to externalatmospheric pres- commonlybe liberated as magmasapproach the surface[Head sure. and Wilson, 1992], thusensuring that a mechanismfor generating This model allows for lateral expansionof gas abovethe vent, volatilesthat can accumulatein stalledmagma bodies does in fact whichis mostlikely to occurwhen there is no vent/conestructure exist. However, there is no simple way of predicting an upper to channelthe materialvertically upwards. Previous models only limit on the gas/caprockmass ratio (n), sinceit is dependenton treatedone-dimensional gas expansion,which would lead to the many factors: the dimensionsof both the caprock and the gas predictionof a greatervelocity for a givenpressure because none pocket; the sourceof the gas (magma or countryrock); and the of the energyfrom the expandinggas would be usedin accelerat- residencetime of the magma in the near-surfacecrust. An upper ing the eruptingmaterial laterally. value of 30 wt % was suggestedby Head and Wilson[1986] as The model is implementedin two FORTRAN computerpro- being representativeof the possible amount of gas that could grams:the first describesthe initial gasexpansion phase of the accumulateon any terrestrialplanet, and this is used as an upper explosion,to findthe launch velocity u 0, launchposition R 0, and limit here, but is the subjectof further discussionbelow. gas velocity decay constantr for any chosencombination of initialgas pressure Pgz' gas mass fraction n,and gas region radius Size of Gas-Filled Region r l; thesecond computes the pyroclast trajectory and landing posi- The high Venusianatmospheric pressure can have a significant tion subjectto the atmosphericmotion given the parametersu0, effect on the densitystructure of growingvolcanoes and so on the R0, r, andthe clast density and radius. ascentbehavior of magmas,determining whether the magma is likely to reach the surface directly or to stall as an intrusion, possiblyleading to accumulationin a shallowmagma chamber. 4. Application of Model to Venus Head and Wilson [ 1992] show that there appearto be two possi- In order to predict the likely rangesof ejecta in vulcanian bilities: in areas of intermediate to high altitude, ascending explosionson Venus it is necessaryto chooseplausible ranges of magma encountersa zone of neutral buoyancyat somedepth in valuesfor the parameterscontrolling the explosionprocess: the the crust.The magma then stallsand will only continueits ascent initialgas pressure Pgz, the gas mass fraction n and the gas region if gas exsolutiontakes place to lower the bulk densityand thus radiusrl; andalso for theproperties of theatmosphere through increasemagma buoyancy relative to the surroundingcountry which the ejectedpyroclasts travel. We addressthese factors in rock. A prolongedintrusive history may lead to formationof a turn. shallow magma reservoir, and if dikes propagatefrom such a reservoirthey are likely to have similar widthsto dikes on Earth Initial Gas Pressure sincethe planetarygravity and the total rangeof depthsat which neutralbuoyancy zone (NBZ) reservoirsreside are similar [Head It seemslikely, given that the rangesof compositionsof ter- and Wilson, 1992]. However, in areasof low elevation the high restrial and Venusian rocks are taken to be broadly similar, that atmosphericpressure has a greaterinhibiting effect on volatile their tensile strengthswill also be similar. It may be argued, exsolution and on the formation of NBZs [Head and Wilson, however, that the pervasive fracturing of volcanic terrains 1992] and thereforeprecludes the formationof shallowmagma observed on Venus from Magellan images may cause likely reservoirs.Magma may thereforerise directly from the mantle caprockstrengths (and hencethe correspondingexcess pressures partial melt zones at depthsof--,10 to 20 km, leading to wider which accumulatein trappedgases) to fall in the lower end of the near-surfacedikes than are commonly seen on Earth. Elasticity range of expectedterrestrial values. However, for the sake of theory [Pollard, 1987] suggeststhat the width factorwill increase comparison,the entire range of vent pressuresexpected on the only as the squareroot of the depth, as a first approximation, basisof analyses ofterrestrial vulcanian explosions, Pgz= 0.01to suggestingdikes 2-3 times wider. This may have some conse- 20 MPa, hasbeen investigated. quencefor thesize of theinitial gas region r 1 (sincerl = w/sino•, where w is the dike half-width, see Figure lb), thoughit is not Gas Mass Fraction immediatelyobvious to what extentit will be affectedsince dike The high abundanceof CO2 (andpaucity of H20) in the width is not the only factorcontrolling the gasregion radius; the present-dayVenusian atmospheresuggests that CO2 may solidangle subtended by thegas region, 12 (or = cos-l(1-l]t2zr); currentlybe the main volatile speciesin magmasource regions. seeFigure 1) is alsoimportant. However, we explorethe possibil- However, despite the fact that some geochemicalmodels of ity that the region occupiedby gas may be systematicallylarger Venuscall for anextreme H20-depleted interior [e.g., Goettel et on Venusthan on Earth,and use values of r 1 in therange 10 to al., 1981], there appearsto be no directevidence that Venusian 200 m. 26,332 FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS
Atmospheric Properties In addition,the velocitiesattained for CO2 are significantly The necessaryparameters describing the Venus atmosphere lessthan when H20 is usedas the driving gas. This is a resultof (variationsof pressure,density and temperaturewith elevation the greatermolecular weight of CO2, whichhas the effectof increasingthe gasdensity in the vent for any given gaspressure. above mean planetary radius) were taken from Kliore et al. [ 1985]. The viscosityof the atmosphericgas was computedfrom The gasmass is thusgreater than for H20, whichensures a greater caprock mass for any chosengas mass fraction n. The the temperaturetaking the compositionto be pureCO 2 and a valueof 8.8m s-2 was adopted for the acceleration due to gravity consequenteffect is a significantdiminution of the accelerations near the Venus surface. of the caprock and displaced atmosphere(the denominatorin equation(1) is muchlarger for CO2 thanfor H20), andhence much lower velocities are attained. The velocities (and hence 5. Results clastranges) produced if sulfur(likely to be in theform of S2) is taken as the driving volatile are further reducedwith respectto In accordancewith the aboveconsiderations, the two computer thosefor CO2, as a resultof its evengreater molecular weight programscomprising the model were run repeatedlyfor all per- (64). mutationsof the following ranges of input parameters:excess The differencein clast rangevalues between the planetsis not pressureinthe vent, Pgz = 0.01to 20 MPa; gas mass fraction, n = just a consequenceof the differing ejection velocitiesobtained 0.01 to 0.3; gasregion radius, r 1= 10 to 200 m. BothH20 and for similar initial conditions,but is alsothe resultof the differing CO2 wereemployed as driving gases for the eruption simulations. atmosphericproperties and hencethe amountof retardationof the Additionally,explosions were modeledas occurringat a rangeof clast during flight. A comparisonof the variation of projectile topographicelevations, varying from the mean planetaryradius travel distancewith initial pressureand gas mass fraction is (mpr=6051.84 km [Ford and Pettengill, 1992]) up to the maxi- shownin Figure 3, for variousdriving gasesand vent elevations mum possiblealtitude (mpr+10 km). Each setof figuresshown is on Venus and Earth. In each case the models were run for a clast the resultof modelingthe motion of an ejectedblock of radius 1 of radius1 m and an ejectionangle of 45ø. As can be seen,the m anddensity 2600 kg m -3. differingplanetary environments act to producevery largediffer- Figure 2 presentsa comparisonof the ejection velocity as a ences in clast range, which is typically several kilometerson functionofinitial gas pressure Pgz' and gas/caprock mass ratio n, Earth and a few hundred meters on Venus. for variouscombinations of driving gas and elevationon Venus Again, the effect of vent altitudeand the choiceof drivinggas and on Earth. The solid curves show the results of employing isclearly seen: for n = 0.10and Pgz = 10MPa, Figure 3 gives, for initial gas temperatureof 1200 K (i.e., approximatelymagmatic H20 andCO 2 respectively,clast ranges of 200 and135 m at the temperature)in the model, whereasthe dashedcurves represent mpr and 350 and 270 m at a vent elevation of mpr+10 km on an initial gastemperature of 1000 K (allowingfor the fact that the Venus. This increasein range attainedby blocks launchedfrom gasmay have cooledsomewhat during the accumulationperiod). explosionsat higher altitudes is not merely a reflection of the In the case of magmatic gasesbeing involved in driving the increasedvelocities achieved by suchclas. ts, but is alsoa resultof explosion, it is not anticipatedthat the temperaturewill drop the smaller retarding drag forces imposedon the clastsby the significantlyduring accumulation,as they will be insulatedto thinneratmosphere at high elevations. some extent by the overlying caprock. However, if external A thresholdvalue for the gaspressure is apparentin Figures2 volatiles are involved, which is more likely on Earth, gas and 3, which is greaterfor lower valuesof n and lower vent alti- temperaturesmay be significantly lower, since the gas would tudes.This represents theminimum value of Pgzrequired for have had to be heatedup from much lower temperatures.As can explosiveactivity to occur: the explosionis suppressedby the be seen,the effect of advocatinga lower initial gastemperature is heavyVenusian atmosphere for lower gaspressures. to reducethe ejectionvelocity attained,which is a resultof less Finally, Figure 4 illustratesthe increasein clast range that energybeing availableto drive the gasexpansion. occursas a resultof invokinga greatersize of regionr l, in which It is also apparentfrom Figure 2 that the ejection velocities gas accumulatesprior to the eruptionfor Venus than for Earth. obtainedfor Venus are systematicallysignificantly lower thanfor Forgreater values of rl, largermasses of gasare involved in the the Earth. Taking a gascontent of 0.10 and an initial gaspressure explosionfor any chosenvalue of n. Thus the gas expansion of 10MPa, it canbe seen that u 0 = 95m s -1 (H20) and 38 m s -• takesplace over a greaterdistance, and hence R 0 is larger.For (CO2)on Venus,compared with 440 m s'l and275 m s-1, Venus (Figures4a and 4b), wherethe clastranges in "free"flight respectively,on Earth. This difference in velocity is due to the (i.e., subsequentto the initial expansionphase when they are still reduction of gas expansion by the oppressive atmospheric locked to the gas motion) are only of the order of 100 m, owing pressureenvironment of Venus. to the large decelerationsimposed by the retardinginfluence of It is also evident that the ejection velocity for any one set of the denseatmosphere, the additionaldistance contributed by the initial parameters(gas massfraction, initial gaspressure, choice greaterR 0 meansthat significantlylarger final rangescan be of volatile species)is significantlygreater for highervent eleva- obtained. For Earth (Figure 4c), the increase in clast range tions.On Venus, for n = 0.10and P gz = 10MPa, an increase in obtainedfor largerr 1 is a muchlower proportion of the total ejectionvelocity of severaltens of m s-1 results from the decrease range owing to the lesser degree of retardationof the clast in in atmosphericpressure from -10 MPa to 5 MPa over a 10-km flight. The final velocitiesattained do not differ significantlyfor elevation difference.In terms of the explosionmodel, the higher a largevariation in r•, providedn is keptconstant. atmosphericpressure at low elevationsrestricts the amountof gas expansion (the driving pressureterm in equation (1) is much 6. Discussion reduced)and the massof the atmosphereto be displacedby the explosionis much greater.The accelerationsand henceejection The dominantfactor determiningthe limited extentto which velocitiesattained are thereforemuch smallerthan they would be material may be ejected in Venusian explosionsis the dense for the higherelevations. atmosphere.This restrictsthe exsolutionand expansionof gas FAGENTSAND WILSON: EXPLOSIVEVOLCANISM ON VENUS 26,333
Initial Gas Pressure, Psz in MPa
(a) Venus, I-I•O, 0 km (b) Venus, I-I20, 10 km
3OO
200
0.10
100 0.02
0.02
o 5 lO 15 20 0 5 10 15 20
250 200(½) Venus• CO• 0 km (d) Venus, C02, 10 km
2OO
150
0.10 lOO . ,,,.,,,,,,,.. ,,'.."'"' "' '
0.05 5o
0 5 10 15 20 o 5 lO 15 20 800 (e) Earth, (f) Earth, CO•
6OO
0.10
300 0.10 0.05 2OO 0.05
0.01 100 0.01
0 5 10 15 20 0 5 10 15 20
Initial Gas Pressure, P•z in MPa Figure2. Ejectionvelocity uo as a functionofinitial excess pressure inthe vent P-z and gas/caprock mass ratio n fortransient explosive eruptions. (a)Venus: driving volatile isH20, vent elevation •s0 km above mpr, (b) Venus: H20at an elevation of 10km, (c) Venus: CO 2 at0 km,(d) Venus: CO 2 at10 km, (e) Earth: H20 at mean sea level (msl),and (f) Earth:CO 2 at msl.Initial gas region radius r i = 50 m forVenus, 25 m forEarth. Solid lines represent resultsof modelfor aninitial gas temperature of 1200 K; dashedcurves represent an initial gas temperature of 1000 K. Curvesare labeled with values of thegas/rock mass ratio. andretards the motionof clastsin flight.The questionof whether variationin the size of the initial gasregion can significantly there exist crustal reservoirs of volatiles that can contribute to the affect pyroclast ranges on Venus, where the extra distance relativelyhigh gas mass fractions required for Venusianexplo- afforded by the larger sourcesize may make a significant sionsremains open to debate.However, it is possibleto envisage contributionto the final clast range. However, despitethe sufficientgas accumulationby invokinga periodof sustained possibilityof widerdikes on Venus,the resultingdeposits would sub-surfacedegassing prior to eruption.This mustcontinue for a still be very localizedwith respectto theEarth. sufficienttime thatthe lowerlimits on pressureand gas concen- We will now considerthe likely maximumranges of frag- trationare exceededso that, on caprockfailure, a significant mentsejected from transienteruptions on Venus. The greatest explosionmay occur; otherwise, no explosion takes place. distanceis attainedwhen each of the initial modelparameters is The discussionabove suggested that larger dikes may occur on the maximum possible [Fagents and Wilson, 1993]. Table 2 Venusthan on Earth. The resultsof the modelingshow that a presentsthe modelingresults for simulationsat the mpr and the 26,334 FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS
Initial Gas Pressure, Pg• in MPa
(a) Venus, H20, 0 km (b) Venus, H20, 10 km
250 4OO
3OO
2OO lOO
0.01 50
o 5 lO 15 20 0 5 10 15 20
3OO (c) Venus, CO2, 0 km (d) Venus, CO•, 10 km 0.3 250
0.1 3OO
0.05
50
o 5 lO 15 20 0 5 10 15 2o I I I (e) Earth, I•O o.3o (f) Earth, CO•
0.20 '
O. lO
0.05 0.05
0.01 O.Ol
0 5 10 15 20 0 5 10 15 20
Initial Gas Pressure, Ps• in MPa
Figure3. Clasttravel distance asa functionof initialexcess pressure in the vent P-z and gas/caprock mass ratio n fortransient explosive eruptions. (a) Venus:driving volatile is H20, ventelevation is 0 km abovempr, (b) Venus: H20 at anelevation of 10km, (c) Venus:CO2 at 0 km,(d) Venus:CO2 at 10km, (e) Earth:H20 at msl,and (f) Earth:CO2 at msl. Initial gas region radius rl=50 m forVenus, r 1 = 25m forEarth, clast density is 2600kg m -3, clastradius is 1 m. Curvesare labeledwith valuesof the gas/rockmass ratio.
maximumplanetary elevation, taking the mostextreme values of from Venusian eruptionsare expectedto be far more localized block size (5 m, basedon field measurementsof the largest than those on Earth. ejectedblocks on Earth), initial pressure(20 MPa, basedon con- Whereasthe ejectionvelocities calculated by the modelare in siderationsof the maximumpossible caprock tensile strength), reasonableagreement with thosecalculated by Wilsonand Head gas mass fraction (0.1, from vent geometricalconsiderations), [1986], who employedthe treatmentof Self et al. [1979] for vul- sizeof pressurizedgas region (200 m), anda launchangle of 45ø. canian explosions,their correspondingclast rangesare signifi- Takingthe results for themaximum elevation and H20 asthe cantlylarger than thosegiven above.They cite a typicalrange of driving gas,material is expectedto be ejectedto maximumdis- between1 and5 km for an H20 gasmass fraction of 0.1 andan tancesfrom the vent of order 1 km on Venus,an orderof magni- initial pressureof 10 MPa, which compareswith --200 to 350 m tude less than the maximum likely distancecalculated for the (taken from Figure 3), dependingon vent altitude.A maximum Earth [Fagentsand Wilson, 1993]. More commonly,deposits rangeof 15 km given by Wilsonand Head [1986] doesnot com- FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS 26,335
Initial Gas Region Radius,r• in m sourcewould lead to a greaterlikelihood of buildingup an angle- I , I I of-restcone, which could then grow outwardsby debrisrolling or 600(a) Venus, H20 shortpyroclastic flow deposition. Alternatively,fields of closelyspaced craters may be produced 500 •- by impactingblocks (similar to thosedocumented at someterres- trial volcanoes,e.g., Arenal [Fudali and Melson, 1972]), pro- vided they have sufficientkinetic energy.It is debatablewhether 400• Pgz=l0MPa the depositsor crater fields may be detectablein currentimage 3OO data,however. The bestspatial resolution of Magellanradar data (-120 m/pixel) may just allow depositsfrom the largestpossible explosionsto be resolved, though it must be stressedthat the requiredcombination of initial conditions (n, Pgz' ri' altitude, etc.) is unlikely. If depositsof this size cannotbe identified,this confirms the improbability of occurrence of the extreme 0 50 100 150 2o0 combinationsof theseparameters. Smaller depositsof the order I I I I of a few tensto a few hundredsof metersin diameterare unlikely 5oo• , to be detectedother than via their effect on sub-pixel surface roughness. It is anticipated that ongoing work on the '(b)Venus,CO• _•• deconvolutionof surfaceroughness, orientation, and electrical properties from the radar signal may provide a tool for the 300- .J Pgz=10MPa_ identification of such deposits in regions where explosive volcanismmay have taken place. :zoo-85,. . The possible presence of pyroclastic flow deposits may increasethe chanceof detectionof sitesof transientexplosions. Though they commonly(but not always) accompanyvulcanian eruptionson Earth (e.g., at Ngauruhoein 1975 [Nairn and Self, 1978] and at Galeras[Calvache and Williams, 1992]), they also accompanya rangeof other stylesof volcanism,so carefulinter- 0 ,• • • 0 50 1• 150 2• pretation would be required. Thornhill [1993] concludedthat l, ,i, • ! pyroclastic flows would occur more commonly than buoyant (c)Earth convectingcolumns in the case of high-volume flux sustained explosiveeruptions (assuming that theseoccurred); it is not clear if this would be the case for transienteruptions. The tentative identificationof a pyroclasticflow deposit[Moore et al., 1992] on plainscentered about 165øE, 37øS implies that Magellan data are adequate for detecting larger flows, whatever their • 5- P•z=IMPa' mechanism of emplacement. However, pyroclastic flows associatedwith terrestrialvulcanian eruptions are generallymuch shorterthan thoseaccompanying more violent,sustained activity • 3- _•j.•.•.,- owing to the limited volumesof material emittedand the lower 17=9...... degreeof fluidizationafforded by the inclusionof large amounts
2 ' I ' 'l I of nonjuvenilematerial at temperaturesconsiderably lower than 0 50 100 150 200 magmatic.Simple energyconservation arguments imply smaller flow lengthson Venus than on Earth, sincethe distancetraveled Initial Gas RegionRadius, r• in m is proportionalto the squareof the vent velocity [Wilsonet al., Figure 4. Clast range as a functionof initial gas region radius, 1982], which is lesson Venus. However, the ability to ingestand r 1. (a) Venus,H20 as the drivinggas; (b)Venus, CO 2 as the heat denseatmospheric gas may enhancethe fluidizationof the driving gas. Solid lines representexplosions at the mean plane- flow and hence causegreater run-out distancesthan on Earth. tary elevation;dotted lines representexplosions at an altitudeof Suchpyroclastic flows would haveto be of order 1 km in length 10 km. (c) Earth, explosionsat meansea level, solid lines repre- to be readilydiscernible in currentMagellan data sets. sentH20 asthe driving gas, dashed lines represent CO 2. In each case,results for twoinitial gas pressures (Pgz) are shown, togetherwith the correspondingvalues for ejectionvelocity in ms -1. Table 2. Comparisonof the PredictedMaximum DispersalRadii (in meters)of Material Ejectedfrom TransientExplosions on Venus pare favorablywith an absolutemaximum of- 1 km takenfrom
Table 2; this discrepancyis a consequenceof their having VENUS EARTH neglected the considerable drag effects of the Venusian Vent Altitude = Vent Altitude = atmosphereon even very large clasts: they calculatedranges Gas 0 km 10 km Mean Sea Level usingthe classicalballistic equation for the largestclasts. The localized nature of Venusianvulcanian deposits implies CO2 475 650 6800 that the ejectedmaterial may build up a significantthickness in H20 640 1010 10800 the vicinity of the vent. Repeated explosionsfrom a central 26,336 FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS
Finally, anotherindication of a vulcaniandeposit might be the predominantlyresponsible for controlling the ashfall deposit presenceof a fine ashfalldeposit resulting from the convecting width, the prevailing wind conditionsexert a strongercontrol eruptioncloud associated with suchan event,extending to greater over the dimensionof the depositalong the dispersalaxis, so that distancesthan the ejectedblocky material.Sites for severalsuch for strongwinds, ash depositsmay have lengthsthat far exceed depositshave been proposed on Venus[Head et al., 1991, 1992; their widths. Wenrich and Greeley, 1992; Campbell,1994; Campbelland It therefore appearspossible that transientexplosive activity Rogers, 1994]. Again, ashfall depositsare associatedwith a providesa mechanismfor distributingpyroclastic products in the variety of volcanic styles,but those from vulcanianexplosions Venusianenvironment. While not wantingto precludethe possi- are expectedto be of more limited areal extent than ashfall bility of morevigorous explosive (plinian) activity, which may be depositsof plinian origin owing to the smaller volumes of required to emplace ash over greater distances(in excessof material emitted; however, taken in context with other indications severaltens of kilometers[e.g., Head et al., 1992]), the problem this may lead to the positiveidentification of depositsassociated of the high volatilecontents necessary for suchactivity is circum- with transient eruptive activity. Indeed, a category of small vented if vulcanian-style activity, in conjunction with the volcaniccones rangingfrom <2 to 15 km in diameterhave been atmosphericwind r6gime, is able to disperseash-sized material identified [Guest et al., 1992], some of which have more over distances somewhat greater than the 5 km theoretical localized ashfall depositsextending to several km [Bulmer, maximumdeposit width. 1994]. It has been shown that the downwind width of an ashfall 7. Summary and Conclusions depositis approximatelyequal to the heightof the eruptioncloud and that this relationshipis essentiallyindependent of the plane- 1. Vulcanian-style eruptionsmay provide a mechanismfor tary atmosphericcharacteristics [Head and Wilson,1986]. Cloud dispersalof pyroclasticmaterial associated with magmashaving rise height, h, for discreteexplosions is relatedby a fourth root volatile contentslower than the minimum required for steady relationshipto the thermalenergy available to drive convection explosiveeruptive activity (hawaiian or plinian)on Venus. [Morton et al., 1956] and henceto the massof volcanicmaterial, 2. The velocity of ejection and the flight distanceof clasts M, injectedinto the cloud[Settle, 1978; Wilsonet al., 1978].As a expelledin suchexplosions are sensitiveto boththe initial excess result of the differing atmosphericstructures, buoyant rise is pressurein the driving gas and the massratio of the gasto solid reducedon Venus:eruption cloud rise heightsare expectedto be ejecta.In addition,clast range is more sensitiveto the size of the around0.6 times as large as thosefrom the equivalentterrestrial regionsoccupiedby gasprior to the explosionon Venusthan on eruptions[Esposito, 1984]. For transientexplosions on Venus Earth. the relationshipis 3. Ejectionvelocities ranging up to a maximumof-200 m s-1 h = 0.98(c AT F M)1/4, (4) (H20)and -100 m s -1 (CO 2) are predicted forplausible combina- tions of gas content, excesspressure and gas region radius. where c is the specific heat capacity of the ash particles Depositsof large, blocky ejecta are likely to be very localized, (essentiallythat of rock), AT is the temperaturedecrease experi- with a predictedmaximum extent of 1 km on Venus (cf. >10 km encedby the particlesfrom vent level (-1200 K) to their final on Earth). In most casesthe distributionof blocky ejectawill be temperature(assumed equal to thatof the atmosphere),and F is a below the limits of detectionin Magellan SAR data. factordescribing efficiency of heatusage [Wilson et al., 1978]. 4. Ashfall depositsor pyroclasticflows associatedwith tran- If only 50% of the ejectedmaterial is juvenile and hencehot sienteruptions may extendto greaterdistances than largeblocky and able to drive convection (based on terrestrial field observa- ejecta and thereforebe identifiablein Magellan data. Eruptions tions [Nairn and Self, 1978]), and of this proportiona further analogousto vulcanianexplosions on Earth couldbe responsible 50% comprises large blocky ejecta which falls essentially for the formationof the possibleashfall deposits having widths immediately from the column [Self et al., 1979], a value of somewhat in excess of 5 km observed near small volcanic F=0.25 must be assumed.It is found from equation (4) that edifices on Venus. Greater dispersal is only possible from massesof 1.8x108kg to 6.6x109kg are requiredto produce discreteexplosions with a higher(>0.25) thermalefficiency, or cloudsrising to between2 and 5 km and henceproduce deposits from maintained plumes arising from plinian activity or ranging up to 5 km in diameter, which is consistentwith the frequentlyrepeated vulcanian explosions. observationsof Bulmer [1994]. Massesof this order can readily be ejected accordingto the explosion model presentedabove, Acknowledgments.S.A.F. thanksthe UK Scienceand Engineering although the larger massesrequire the more extreme of the ResearchCouncil for a researchstudentship and gratefully acknowledges possiblecombinations ofmodel initial parameters (high Pgz' large the SmithsonianInstitution for a PostdoctoralResearch Fellowship during r 1, low n). It is possibleto obtaingreater cloud heights with these whichthe manuscriptwas completed.L.W. acknowledgespartial support massesby advocatinga greaterthermal efficiency factor, F. This from the Royal Society through a Leverhulme Senior Research would imply either a larger or a more fragmentedjuvenile Fellowship.Both authorsare gratefulto Mark Bulmer,who madeavail- able his dataon Venusiancones. Helpful reviewswere providedby L.S. componentin the ejecta [Wilsonet al., 1978]; eithercase would Crumplerand an anonymousreviewer. ensure that more heat would be available to drive convection. The extreme (and unrealistic)case would be when F-I, in which References casecloud heights (and hencedeposit widths) of-8 km mightbe expected.Alternatively, repeatedexplosions may serveto create Barsukov,V.L., et al., The geologyand geomorphology of theVenus sur- greaterdeposit widths, either if they are separatedby intervalsof faceas revealedby Veneras15 and 16, Proc. Lunar Planet.Sci. Conf less than a few minutes [Wilson et al., 1978], causing a 16th,Part 2, J. Geophys.Res., 91, suppl.,D378-D398, 1986. BasalticVolcanism Study Project, Basaltic Volcanism on the Terrestrial maintained plume to be established,or by occurring under Planets,Pergamon, New York, 1981. varying wind conditionswhich may disperseash along different Blackburn,E.A., L. Wilson, and R.S.J. Sparks,Mechanisms and dynam- axes.It is also worth pointingout that, while the cloudheight is ics of Strombolianactivity, J. Geol.Soc. London, 132, 429-440, 1976. FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS 26,337
Bulmer, M.H., An examinationof small volcanoesin the plainsof Venus McBirney, A.R., Factorsgoverning the intensityof explosiveandesitic with particularreference to the evolutionof domes,Ph.D. thesis, Univ. eruptions,Bull. Volcanol.,37, 443-453, 1973. of London,England, 1994. McKenzie, D., P.G. Ford, F. Liu, and G.H. Pettengill,Pancakelike domes Calvache V., M.L., and S.N. Williams, Lithic-dominated pyroclastic on Venus,J. Geophys.Res., 97, 15,967-15,976,1992. flows at Galerasvolcano, Colombia- An unrecognizedvolcanic haz- Mellors, R.A., R.B. Waitt, and D.A. Swanson,Generation of pyroclastic ard, Geology,20, 539-542, 1992. flows and surgesby hot-rockavalanches from the domeof Mount St. Campbell,B.A., MergingMagellan emissivity and SAR datafor analysis Helens volcano, USA, Bulk Volcanol., 50, 14-25, 1988. of Venussurface dielectric properties, Icarus, 112, 187-203, 1994. Melson, W.G., and R. Saenz,Volume, energyand cyclicity of eruptions Campbell,B.A., and P.G. Rogers,Bell Regio, Venus: Integrationof of Arenal volcano, Costa Rica, Bulk Volcanol., 37, 416-437, 1973. remote-sensingdata and terrestrialanalogs for geologicanalysis, J. Minakami, T., On explosiveactivities of andesiticvolcanoes and their Geophys.Res., 99, 21,153-21,171, 1994. forerunningphenomena, Bull, Volcanol.,10, 59-87, 1950. Cas,R.A.F., andJ.V. Wright, VolcanicSuccessions, Modern and Ancient, Moore, H.J., J.J. Plaut, P.M. Schenk, and J.W. Head, An unusualvolcano Unwin Hyman, Boston,Mass., 1988. on Venus,J. Geophys.Res., 97, 13,479-13,494,1992. Chouet, B., N. Hamisevicz, and T.R. McGetchin, Photoballisticsof vol- Morton,B.R., G. Taylor, andJ.S. Turner, Turbulent gravitational convec- canicjet activityat Stromboli,Italy, J. Geophys.Res., 79, 4961-4976, tion from maintained and instantaneoussources, Proc. R. Soc. London 1974. A, 234, 1-23, 1956. Decker, R.W., and D. Hadikusumo,Results of the 1960 expeditionto Nairn, I.A., Atmosphericshock waves and condensationclouds from Krakatau,J. Geophys.Res., 66, 3497-3511, 1961. Ngauruhoeexplosive eruptions, Nature, 259, 190-192, 1976. Esposito,L.W., Sulfur dioxide:Episodic injection shows evidence for Nairn, I.A., and S. Self, Explosiveeruptions and pyroclasticavalanches active Venus volcanism, Science, 223, 1072-1074, 1984. from Ngauruhoein February1975, J. Volcanol.Geotherm. Res., 3, 39- Fagents,S.A., The analysisof someexplosive volcanic processes on the 60, 1978. Earth, Venus and Mars, Ph.D. thesis, Lancaster Univ., Lancaster, Nikolaeva,O.V., Geochemistryof the Venera 8 materialdemonstrates the England,1994. presenceof continentalcrust on Venus, Earth, Moon and Planets, Fagents,S.A., and L. Wilson, Explosivevolcanic eruptions, VII, The 50/51,329-341, 1990. rangesof pyroclastsejected in transientvolcanic explosions, Geophys. Pavri, B., J.W. Head, K.B. Klose, and L. Wilson, Steep-sideddomes on J. Int., 113, 359-370, 1993. Venus:Charactefistics, geologic setting, and eruptionconditions from Ford, P.G., and G.H. Pettengill,Venus topographyand kilometer-scale Magellan data,J. Geophys.Res., 97, 13,445-13,478,1992. slopes,J. Geophys.Res., 97, 13,103-13,114,1992. Pollard, D.D., Elementaryfracture mechanics applied to the structural Fudali, R.F., and W.G. Melson, Ejecta velocities,magma chamber pres- interpretationof dykes,in Mafic Dyke Swarms,edited by H.C. Halls sure and kinetic energy associatedwith the 1968 eruptionof Arenal and W.F. Fahrig, Geol. Soc.Can. Spec.Pap., 34, 5-24, 1987. volcano, Bull. Volcanol., 35, 383-401, 1972. Ripepe,M., M. Rossi,and G. Saccorotti,Image processingof explosive Garvin, J.B., J.W. Head, and L. Wilson, Magma vesiculationand pyro- activity at Stromboli,J. Volcanol.Geotherm. Res., 54, 335-351, 1993. clastic volcanismon Venus, Icarus, 52, 365-372, 1982. Sakimoto, S.E.H., and M.T. Zuber, Venus pancakedome formation: Goettel,K.A., J.A. Shields,and D.A. Decker,Density constraints on the Morphologic effects of a cooling-inducedvariable viscosityduring compositionof Venus, Proc. Lunar Planet. Sci. Conf 12th, 1507- emplacement(abstract), Lunar Planet. Sci. Conf., XXIV, 1233-1234, 1993. 1516, 1981. Sato, H., T. Fujii, and S. Nakada, Crumbling of dacite dome lava and Gorshkov,G.S., Gigantic eruption of the volcano Bezymianny,Bull. Volcanol., 20, 77-109, 1959. generationof pyroclasticflows at Unzen volcano,Nature, 360, 664- 666, 1992. Gorshkov, G.S., and G.E. Bogoyavlenskaya,Bezymyanny and its Self, S., L. Wilson, and I.A. Nairn, Vulcanian eruption mechanisms, Eruption1955-1963 (in Russian),Akad. Nauk SSSR,Moscow, 1965. Nature, 277, 440-443, 1979. Guest,J.E., M.H. Bulmer, J. Aubele, K. Beratan,R. Greeley,J.W. Head, Self, S., J. Kienle, and J.-P. Huot, Ukinrek Maars, Alaska, II, Deposits G. Michaels, C. Weitz, and C. Wiles, Small volcanic edifices and vol- and formation of the 1977 craters,J. Volcanol. Geotherm. Res., 7, 39- canismin the plains of Venus, J. Geophys.Res., 97, 15,949-15,966, 65, 1980. 1992. Settle,M., Volcanic eruptionclouds and the thermalpower outputof ex- Head, J.W., and L. Wilson, Volcanicprocesses and landformson Venus: plosiveeruptions, J. Volcanol.Geotherm. Res., 3, 309-324, 1978. Theory, predictionsand observations,J. Geophys.Res., 91, 9407- 9446, 1986. Sheridan,M.F., and K.H. Wohletz, Hydrovolcanicexplosions: The sys- Head, J.W., and L. Wilson, Magma reservoirsand neutral buoyancy tematics of water-pyroclastequilibration, Science,212, 1387-1389, 1981. zoneson Venus:Implications for the formationand evolutionof vol- Sheridan,M.F., andK.H. Wohletz,Hydrovolcanism: Basic considerations caniclandforms, J. Geophys.Res., 97, 3877-3903, 1992. and review, J. Volcanol. Geotherm. Res., 17, 1-29, 1983. Head, J.W., D.B. Campbell,C. Elachi, J.E. Guest,D.P. McKenzie, R.S. Sparks,R.S.J., The dynamicsof bubbleformation and growth in magmas: Saunders,G.G. Schaber, and G. Schubert, Venus volcanism: Initial A review and analysis,J. Volcanol. Geotherm. Res., 23, 147-188, analysisfrom Magellan data,Science, 252, 276-288, 1991. 1978. Head, J.W., L.S. Crumpier,J.C. Aubele, J.E. Guest,and R.S. Saunders, Steinberg,G.S., On the determinationof theenergy and depth of volcanic Venus volcanism: Classification of volcanic features and structures, explosions,Bull. Volcanol.,40, 1-5, 1977. associations,and global distributionfrom Magellan data,J. Geophys. Steinberg,G.S., andJ.I. Babenko,Experimental velocity and density de- Res., 97, 13,153-13,198, 1992. terminationof volcanicgases during eruption, J. Volcanol.Geotherm. H6dervftri,P., Volcanophysicalinvestigations on the energeticsof the Res., 3, 89-98, 1978. Minoan eruptionof Volcano Santofin,Bull. Volcanol.,32, 439-461, Steinberg, G.S., and A.S. Steinberg, On possiblecauses of volcanic 1968. tremor,J. Geophys.Res., 80, 1600-1604, 1975. Hoerner, S.F., Fluid-DynamicDrag, S.F. Hoerner,New York, 1965. Sugita,S., and T. Matsui, Are pliniantype eruptionspossible on Venus? Ivanov, M.A., Explosivevolcanism on Venus?(abstract), Lunar Planet. (abstract),Lunar Planet.Sci. Conf.,XXIV, 1375-1376,1993. Sci. Conf, XXIII, 577-578, 1992 Surkov, Yu. A., F.F. Kirnozov, V.N. Glazov, A.G. Dunchenko, and L.P. Jaupart,C., and S. Vergniolle, The generationand collapseof a foam Tatsil, The content of natural radioactive elements in Venusian rock as layer at the roof of a basalticmagma chamber, J. Fluid Mech., 203, determinedby Venera9 andVenera 10, Kosmich.Issled., 14, 704-709, 347-380, 1989. 1976. Kienle, J., P.R. Kyle, S. Self, R.J. Motyka, and V. Lorenz, Ukinrek Surkov, Yu. A., F.F. Kirnozov, V.N. Glazov, and G.A. Fedoseyev, Maars, Alaska, I, April 1977 eruptionsequence, petrology and tectonic Investigationsof venusiangamma-radiation by Venera 9 and Venera setting,J. Volcanol.Geotherm. Res., 7, 11-37, 1980. 10, COSPARSpace Res., 17, 659-662, 1977. Kliore, A.J., V.I. Moroz, and G.M. Kesting, The Venus International Surkov,Yu .A., V.L. Barsukov,L.P. Moskalyeva,V.P. Kharyukova,and ReferenceAtmosphere, COSPAR Rep. JPL-D-2216, Adv. SpaceRes., A.L. Kemurdzhian,New dataon the composition,structure and prop- 5, 1985. erties of Venus rock obtainedby Venera 13 and 14, Proc. Lunar Livshitts,L.D., and L.G. Bolkhovitinov,Weak shockwaves in the erup- Planet.Sci. Conf.14th, Part 2, J. Geophys.Res., 89, suppl.,B393- tion column, Nature, 267, 420-421, 1977. B402, 1984. Matuzawa,T., Gasdruckbei einigenvulkanischen Ausbruchen in Japan, Surkov, Yu. A., F.F. Kirnozov, V.N. Glazov, A.G. Dunchenko, L.P. Bull. EarthquakeRes. Inst. Univ. Tokyo,11,347-349, 1933. Tatsy, and O.P. Sobornov,Uranium, thorium and potassiumin the 26,338 FAGENTS AND WILSON: EXPLOSIVE VOLCANISM ON VENUS
Venusian rocks at the landing sites of Vega 1 and 2, Proc. Lunar martianvolcanic eruption mechanisms, Eur. SpaceAgency Spec. Planet. Sci. Conf 17th, Part 2, J. Geophys.Res., 92, suppl.,E537- Publ., ESA SP-185, 107-113, 1982. E540, 1987. Wilson, L., R.S.J. Sparks,T.C. Huang, andN.D. Watkins,The controlof Tait, S.R., C. Jaupart,and S. Vergniolle,Pressure, gas content and erup- volcanic column heights by eruption energeticsand dynamics.J. tive periodicity of a shallow crystallizing magma chamber,Earth Geophys.Res., 83, 1829-1836, 1978. Planet. Sci. Lett., 92, 107-123, 1989. Wohletz, K.H., Mechanismsof hydrovolcanicpyroclast formation: Size, Thornhill,G.D., Theoreticalmodeling of eruptionplumes on Venus,J. scanningelectron microscopy and experimentalstudies, in Explosive Geophys.Res., 98, 9107-9111, 1993. Volcanism, edited by M.F. Sheridan and F. Barberi, J. Volcanol. Touloukian,Y.S., W.R. Judd,and R.F. Roy, PhysicalProperties of Rocks Geotherm. Res., 17, 31-63, 1983. and Minerals, Data Ser. on Mater. Properties,vol. 1-2, McGraw-Hill, Wohletz, K.H., and M.F. Sheridan, Hydrovolcanic explosions,II, New York, 1981. Evolution of basaltictuff rings and tuff cones,Am. J. Sci., 283, 385- Turcotte,D.L., H. Ockendon,J.R. Ockendon,and S.J.Cowley, A mathe- 413, 1983. maticalmodel of vulcanianeruptions, Geophys. J. Int., 103, 211-217, Wood, C.A., Venus volcanism:Environmental effects on styleand land- 1990. forms, NASA Tech. Memo., TM-80339, 244-246, 1979. Wenrich,M.L., and R. Greeley,Investigation of Venusianpyroclastic volcanism (abstract), Lunar Planet. Sci. Conf, XXIII, 1515-1516, 1992. S.A. Fagents,Center for Earth and PlanetaryStudies, National Air Wilson,L., Explosivevolcanic eruptions, II, The atmospherictrajectories and SpaceMuseum, SmithsonianInstitution, Washington, DC 20560. (e- of pyroclasts,Geophys. J. R. Astron.Soc., 30, 381-392, 1972. mail: sfagents@ ceps.nasm.edu) Wilson,L., Relationshipsbetween pressure, volatile content and ejecta L. Wilson, Environmental Science Division, Institute of velocityin threetypes of volcanicexplosion, J. Volcanol.Geotherm. Environmentaland BiologicalSciences, Lancaster University, Lancaster Res., 8, 297-313, 1980. LA1 4YQ, England.(e-mail: [email protected]) Wilson,L., andJ.W. Head,A comparisonof volcaniceruption processes on Earth, Moon, Mars, Io and Venus, Nature, 302,663-669, 1983. (ReceivedMarch 17, 1995; revisedJuly 20, 1995; Wilson, L., J.W. Head, and P.J. Mouginis-Mark,Theoretical analysis of acceptedOctober 17, 1995.)