Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 87, NO. Btt, PAGES 9236-9246,NOVEMBER t0, 1982 Mechanismsfor LithosphericHeat Transport on Venus' Implications for Tectonic Style and Volcanism SEAN C. SOLOMON Departmentof Earth and PlanetarySciences, Massachusetts Institute of Technology Cambridge,Massachusetts 02139 JAMES W. HEAD Departmentof GeologicalSciences, BrownUniversity,Providence,RhodeIsland02912 The tectonicand volcaniccharacteristicsof the surfaceof Venus are poorly known, but thesecharacteristicsmust be closelyrelated to the mechanismby which Venus rids itself of internal heat. On the other solidplanets andsatellites of thesolarsystem, lithospheric heat[ransport is dominated by oneof three mechanisms: (1) plate recycling,(2) lithosphericconduction,and (3) hot spot volcanism.We evaluateeach mechanismas a candidatefor the dominantmode of lithosphericheat transferon Venus,and we explore the implicationsof eachmechanismfor the interpretationof Venussurfacefeatures.Despiteclaimsmade to the contrary in the literature,plate recyclingon Venus cannot be excludedon the basisof either theoreticalargumentsor presentobservationson topographyand radar backscatter.Landformsresulting from plate convergenceand divergenceon Venus would differ substantiallyfrom those on the earth becauseof the high surfacetemperatureand the absenceof oceanson Venus,the lack of free or hydrated water in subductedmaterial,the possibilitythat subductionwould more commonlybe accompaniedby lithosphericdelamination,and the rapid spreadingratesthat would be requiredif plate recyclingremoves a significantfractionof the internalheat.If plate recyclingoccurson Venus,the rolling plainsand lowlands provinceswouldbe approximateanalogsto terrestrialoceanbasinsin termsof age,igneousrock type,and formativeprocess;highlandson Venuswould be roughlyanalogousto terrestrialcontinents.The hypothesisthat lithosphericconductiondominatesshallowheat transferon Venusleadsto the predictionthat the lithosphereis thin. If Venus has a global heat lossper massequal to that for earth, then temperatures marking the baseof the thermal lithosphereon earth would be reachedon Venusat an averagedepth of about 40 km. Unlessthe mantleconvectiveplanformcan maintainlithosphericregionsof persistentlylow heat flow or unlessthe presentatmosphericgreenhouseon Venus is geologicallyrecent,then such a lithosphericthicknessleadsto the conclusionthat the topographicfeaturescontributingto the 13 km of relief on the planet must be geologicallyyoung. The hypothesisthat hot spot volcanismdominates lithosphericheat transferon Venusleadsto the predictionthat the surfacemustbe coveredwith numerous activevolcanicsources.In particular,if a typicalVenushot spothas a volcanicflux equal to the average fluxfor theHawaiianhot spotfor thelast40 m.y.,then10'• suchhot spotsare necessary to removethe Venusinternalheat by volcanism.Sucha numberwould produceenoughvolcanicmaterialto resurfacethe entire planet to a depth of 1 km every 2 m.y.; few areas of the planet would escaperesurfacingfor geologicallylong periodsof time. We find that noneof the mechanisms for lithosphericheat transferon Venus can be excludedas unimportant at present;it is likely that, as on earth, a combinationof mechanisms operateson Venus.The strongestconclusionto emergefrom this evaluationis that most of the major topographicfeaturesand probablymany of the surfacegeologicalunitson Venusare youngby comparisonwith the surfacesof the smallerterrestrialplanets. INTRODUCTION mon, 1981]. For Jupiter'smoon Io, eruptions at individual volcaniccenters,or 'hot spots,'dominatethe lithosphericheat flux [Matson et al., 1981; Reynoldset al., 1980; O'Reilly and Davies,1981] and resultin geologicallyrapid rates of global resurfacing[Johnsonet al., 1979]. For Venus, the availableradar imaging and altimetry data examplesof bodiesin which one of three distinct mechanisms hasdominatedheatloss(Figure1),and the resultinggeological [Goldsteinet al., 1976,1978;Campbellet al., 1976,1979;Camphistoriesfor thesebodiesdiffer profoundly.For the earth, the bell and Burns,1980; Pettengillet al., 1980; Masurskyet al., majority of mantle-derived heat is delivered to the surface 1980] are at too coarse a horizontal resolution to allow the throughplate recycling:the processes of creation,cooling,and tectonic and volcanic featuresdiagnosticof one or more of subductionof oceaniclithosphere[Sclateret al., 1980].For the thesemechanismsof lithosphericheat transferto be discerned smaller terrestrialbodies,includingMars, Mercury, and the unequivocally.As a result,debateshaveensuedoversuchissues earth'smoon, heat transporthas occurredprincipallyby con- as whetherthe surfaceof Venusdisplaysfeaturesindicativeof ductionthrougha globallycontinuouslithosphericshell[Solo- lithosphericrecyclinganalogousto terrestrialplate tectonics mon,1978],and thelevelsof bothtectonicand volcanicactivity [Masursky et al., 1980; Phillips et al., 1981; Arvidson and have been much more limited than on earth [Head and Solo- Davies,1981;Head et al., 1981;Kaula andPhillips,1981;Arvidsonand Guinness,1982; Brassand Harrison, 1982; Phillipsand Malin, 1982]. Rather than simplycontinuingsucha debate,we Copyright1982by theAmericanGeophysicalUnion. takein thispapera differentapproachto the coupledquestions of lithosphericheattransportand volcanicand tectonicactivity Paper number2B1314. The mechanismby which a solid planet transportsheat acrossthe outer 100 km of its interior playsa pivotal role in determiningthe stylesand magnitudesof tectonicand volcanic activityat theplanet'ssurface.Amongtheplanetscanbefound on Venus. 0148-0227/82/002B-1314505.00 9236 SOLOMONAND HEAD: LITHOSPHERICHEAT TILANSPORTON VENUS LITHOSPHERIC CONDUCTION '//•Moon, Mercury MECHANISMS OF / / •Mars • remainingfrom such ancient processesas planetary accretion and core-mantledifferentiation.A recentreviewof theseaspects of the thermal history problem is given by Solomonet al. [1981]. Of the currentlyoperativeprocesses, radioactivedecay contributesthe vastmajority of the heat generation,with minor incrementscontributedfrom the motion of phaseboundaries (e.g.,inner corefreezing)and tidal dissipation.To determinethe partitioning of global heat loss betweenpresentlygenerated and ancient heat, it is necessaryto solvefor the earth'sthermal evolution,includingparticularlythe effectof solidstateconvection on mantle IO PLATE RECYCLING HOT SPOT VOLCANISM Fig. 1. Schematic ternary diagram showing the relative contributionsof eachmechanismof lithosphericheat transfertoward global heat loss on solid planetsand satellites.For the moon, Mercury, and Mars, thermal conductioncarriesall of the lithosphericheat flow. The position of the earth, on which the majority of lithospheric heat transport occursby plate recycling,is from Sclater et al. [1980] and does not include the contribution to global heat loss from crustal radioactivity. Hot spot volcanismdominatesheat loss on Iv, but the relative fraction contributed by conductionis not known precisely [Reynoldset al., 1980; O'Reilly and Davies, 1981]; at the position shownon the diagram,the conductiveheat flow on Iv is equal to the earth'saverageheat flux. We pose three hypothesesfor lithosphericheat transfer on nate heat flux on other planetary bodies:(1) that lithospheric heat transportoccursprimarily throughplate recycling,(2) that lithosphericconductiondominatesthe heat loss,and (3) that heat is transportedthrough the lithosphereprincipally by hot spot volcanism.These three hypothesesshould be regardedas end-members;clearly, combinationsof mechanismsmay also serveto deliver internal heat to the planetary surface.We test eachhypothesisagainstthe known propertiesof Venus,and we explore the implicationsof each for the character of tectonic and volcanicactivity at the Venussurface.We find that none of the individualhypotheses can currentlybe excluded.Although specificpredictionsof the three hypotheses for surfacegeological evolution differ substantially,each hypothesisleadsto the conclusionthat many,if not most,of the physiographicfeatures of the Venussurfaceare geologicallyyoung. VENUS HEAT FLOW To quantify discussionsof specificmechanismsfor lithosphericheattransporton Venus,it is necessary to adopta value for the rate of globalheat lossfrom the planet.Sincethe surface heat flux on Venus has not been measured,we assumeas a basis for an estimatethat the heat lossper masson Venusis identical to that on earth. The presentheat lossfrom the solid earth is 4.2 x 10•3 W, for a globallyaveraged heatflowof 82mW/m2 [-Sclateret al., 1980-[.The massof Venus is 0.815 earth masses [-e.g.,Ash et al., 1971-[,and the surfacearea of Venus is 4.60 x 108km2 rPettengillet al., 1980-]. The earth-scaled valuesfor global heat loss and surfaceheat flux for Venus are therefore 3.4 x 10•3W and74mW/m•, respectively. The heat flux from the earth consistsof two parts: heat generatedby currently operativeinternal processes and heat heat transfer. A number of such calculations have recently been performed using one of several parameterizationschemesfor heat transportby mantleconvection [Sharpeand Peltier, 1978, 1979; Schubertet al., 1979a, b, 1980; Schubert,1979; Sleep,1979; Stevensonand Turner, 1979; Turcotte et al., 1979; Daly, 1980; Davies, 1980; McKenzie and Weiss,1980; Stacey, 1980; Turcotte, 1980; Cook and Turcotte, 1981; McKenzie and Richter, 1981]. Thesecalculationsindicate that some fraction between 50 and 90% of the earth's heat loss is contributedby presentlyoperativeheat generationprocesses, with the specific value dependent on the type of parameterizationadopted and the assumptionsmade about heat sources,the temperature dependenceof viscosity,and radial zoning in the mantle. The remainderof the earth'sheat lossis a consequence of the secularcoolingof the planet. There are several observationaland theoretical arguments that provide somesupportfor the assumptionof equal ratesof heat loss per massfor Venus and earth. Severalindependent cosmochemical Venus, each based on one of the mechanisms known to domi- 9237 models for the formation of the terrestrial planetsyield roughlycomparablebulk abundancesby massfor heat-generatingelementsin the two bodies [see Wood et al., 1981]. The measured abundancesof U, Th, and K at the Venera 8, 9, and 10 landing sites [Vinogradov et al., 1973; Surkov et al., 1977; Surkov, 1977] and the measuredK20 abundancesat the Venera 13 and 14 landing sites[Barsukov, 1982] are generally comparable to values in terrestrial tho- leiitesand alkali basalts.The total amountof radiogenic '•øAr in the Venus atmosphereis within a factor of 3 of that in the earth'satmosphere[Hoffman et al., 1980-1,though whetherthe lower amount on Venus is attributable to differences in out- gassinghistoryor in the bulk K abundanceis not known. Parameterized-convection models of the thermal history of Venus [Sharpeand Peltier, 1979; Turcotte et al., 1979; Solomon et al., 1981; Phillipsand Malin, 1982] indicatethat for the same assumptionsas thoseusedin modelsof the thermal history of the earth, the evolution of the two planetshas been broadly similar. In particular, secularcooling of Venus contributesto the predictedpresentheat loss of Venus by an amount comparableto that for the earth. We concludethat a scalingof heat loss by mass from the earth to Venusis a reasonableworking hypothesis.The valueof the Venusheat loss,however,evenundersucha hypothesis,has considerableuncertainty.The earth'sheat flux may have been 25 to 50% greaterat timesof enhancedratesof seafloorspreading in the Phanerozoic[Hays and Pitman, 1973; Pitman, 1978; TurcotteandBurke, 1978; Sleep,1979; Harrison, 1980; Sprague and Pollack, 1980-1,so that the terrestrial heat lossdetermined for the presentearth [Sclater et al., 1980] may be lessthan the averageglobalheat losson time scalesof 108 years.If the difference in the amountof '•øArbetweenthe atmospheres of Venusand earth representsa differencein the planetaryabundanceof potassium,the Venus heat losswould be lessthan the value assumedhere. The difference,however,would probably 9238 SOLOMON AND HEAD: LITHOSPHERIC HEAT TRANSPORT ON VENUS be small.The presentcontributionof '•øK to the heat productionin terrestrialmaterialwithK/U = 10'• andTh/U = 3.7 by weight [Wasserburget al., 1964] is 15%; thus if the abundancesby weightof refractoryelementssuchas U and Th are comparableon Venusand earthbut the abundanceof the more volatile element K is lower by a factor of 3, the overall heat productionper masswouldbe only 10% lessfor Venusthan for the earth. PLATE RECYCLING The recyclingof lithosphereinto the underlyingmantleis the major governorof terrestrialvolcanicand tectonicactivity and of the formationof the large-scalephysiographicfeaturesof the earth'ssurface.Creation of new lithosphereat mid-oceanridges is the dominant volcanicprocesson the planet.Spreadingand coolingof the oceaniclithosphere,the thermal boundary layer of mantle convection,producesthe characteristicrelationship betweenoceandepth and seafloorage [e.g., Parsonsand Sclater, 1977] as well as the distinctivepattern of ridge axis and fracturezone physiography.Trenches,great thrust faults,and volcanic island arcs mark the loci of subduction of oceanic lithospherebeneathanother oceanicplate. Island chainsand many oceanicaseismicridges are productsof the motion of surfaceplatesover mantle sitesof persistenthot spotvolcanism [e.g., Wilson, 1965]. Passivemarginsof continentsmark the sitesof former intracontinental rifting where new oceanswere created.Tectonicallyactivemountainbeltsform wheresubduction occursbeneath overriding continental lithosphere,as in the Andes,or where two continental massescollide after closure of an interveningocean,as in the Himalayas [e.g.,Deweyand Bird, 1970]. Plate recyclingis also the dominant mechanismfor heat transfer across the outermost 100 km of the solid earth. The most carefulanalysisof the partitioning of terrestrialheat flow among the contributiong processesis that of Sclater et el. [1980]. Theseworkersestimatethat about 65% of the present heat loss of the earth is associatedwith the plate recycling process,including the creation and cooling of oceaniclithosphereand continentalorogenyassociatedwith plate interactions. Another 20% of the heat loss occurs by conduction throughoceanicand continentallithosphere,and 15% is generated by radioactivedecaywithin the continentalcrust.As noted earlier, there have been extendedperiodsduring the Phanerozoic when the volumesof mid-oceanridges,and by inference the globalratesof plate creation,havebeensubstantiallygreater than they are at present [Hays and Pitman, 1973; Pitman, 1978; Turcotte and Burke, 1978]. During suchtimes both the total rate of global heat lossand the fraction of that heat loss contributed by the plate recyclingprocesswere greater than their presentvalues [Turcotte and Burke, 1978; Sleep, 1979; Harrison,1980;SpragueandPollack,1980]. Becausesuch a large fraction of the earth'sheat is lost by platerecycling,the typicalthermalgradientsin old oceanicand continentallithosphereare muchlessthan theywouldbe in the absenceof plate tectonics[e.g.,Sclateret el., 1980]. That conduction through the continental lithospherehas contributed only modestlyto the earth'sglobal heat lossas far back as the Archcanis stronglysuggested by the observationof Burkeand Kidd [1978] that productsof crustalremeltingare not observed on a regionallyextensivebasisin exposedsectionsof lower continentalcrustin the Archcanterrain of the SuperiorProvince. A clear implication of the processof lithosphericheat transport on earth is that terrestrial continentaland oceanic geothermscannotbe extrapolatedto Venus without making internally consistentassumptionsabout both the global heat lossand the fraction of that heat lossdeliveredby lithospheric conduction.This point has not been addressedin severaldiscussionsof the thermal structureof the Venus interior [e.g., Warner, 1979;McGill, 1979; Kaula andPhillips,1981]. BecauseVenus and the earth have similar masses,bulk densities,and positionsin the solar system,and on the basisof the assumedsimilaritiesin globalheat loss,a reasonablehypothesis is that Venus, like the earth, losesmuch of its heat by lithospheric recycling.We shall evaluate this hypothesisand its implications for the evolution of the Venus surface in this section. The hypothesisthat plate recycling presently occurs on Venushasbeenchallengedin the literature on severalgrounds: (1) that topographicfeaturesindicativeof terrestrialplate tectonicscannotbe 'seen'on Venus[Masurskyet al., 1980;Phillips et al., 1981; Arvidson and Davies, 1981; Kaula and Phillips, 1981; Arvidsonand Guinness,1982; Schaber,1982], (2) that most of the Venus surface is 'ancient,' as indicated by the distributionof inferredimpact cratersand basins[Schaberand Boyce,1977;Masurskyet al., 1980;Phillipset al., 1981],(3) that becauseof the high surfacetemperatureon Venus, the lithosphereis less likely to subduct than oceaniclithosphereon earth [Anderson,1981; Phillips and Malin, 1982], and (4) that becauseof the high surfacetemperatureand from the observed characteristicsof 'ridges' on Venus, plate recyclingis a less 'significant'processfor removing heat than on earth [Kaula and Phillips,1981]. As we discussbelow,however,none of these argumentsis sufficientlystrong to rule out plate recyclingon Venuson the basisof presentevidence. On TopographicFeaturesDiagnosticof Plate Recycling The global topographyof Venus as measuredby the Pioneer Venus orbiter [Pettengill et al., 1980] stimulatedconsiderable interestin the questionof whether the nature of tectonicand volcanic activity on that planet can be discernedfrom the large-scalephysiography.In particular, the statementhas been made repeatedlyin the literaturethat Venuslackstopographic featuresindicativeof terrestrial-styleplate tectonics[Masursky et al., 1980; Phillips et al., 1981; Arvidsonand Davies, 1981; Arvidsonand Guinness,1982; Schaber,1982]. There are severalreasonswhy a searchfor featuresindicative of lithosphericrecyclingon Venus may be more difficultthan has been generallyappreciated.The horizontal resolutionof the Pioneer Venus topographicdata is about 100-200 km [Pettengill et al., 1980], so that features of lesserhorizontal dimensionsthan thesevaluesare not resolvable.A topographic map of the earth at comparableresolutionwould not resolve such features,for instance,as many oceanictrenchesand the axial valleys on slowly spreadingmid-oceanridges[Head et al., 1981]. Further, oceanictopographyon earth is capableof approximately50% greater relative relief becauseof the presenceof the water layer. In addition,the rate of deepeningof the terrestrialoceaniclithosphere,or thermal boundarylayer, with ageis proportionalto the differencein temperaturebetweenthe bottom and top of the layer [Parsonsand Sclater, 1977]; on Venus,with its higher surfacetemperature,the rate of deepening of the surfaceof any analog to oceaniclithosphereshould be correspondinglyless.Both of theseeffectswould thusact to reducethe reliefof Venusanalogsto spreadingridges,trenches, and submarinehot spot tracescomparedwith their terrestrial counterparts[Phillips et al., 1981; Head et al., 1981; Arvidson SOLOMON AND HEAD' LITHOSPHERIC HEAT TRANSPORT ON VENUS ,•ndDavies,1981; Kaula and Phillips, 1981]. While this reduced relief would still permit the identification of large-scaleplate tectonicfeatures(suchasmid-oceanridges,mountainbelts,and hot spottraces)on earthusingonly altimetryat PioneerVenus resolution,at issueis whether suchfeaturescould be correctly interpretedif no other information were available [Head et al., 1981;ArvidsonandDavies,1981]. Any tectoniccomparisonof the topographyof Venus and earth shouldalsobe made in recognitionof the fact that many of the physiographiccharacteristicsof terrestrial features of plate tectonic origin are a direct result of the presentdistribution of continentsand ocean basins,the presentrange of ridge spreadingrates,and the presenceof surfacewater. On a planetwith a highersurfacetemperatureand negligiblesurface water, many of thesecharacteristics might be expectedto differ. For example,the orthogonal pattern of mid-oceanridge segmentsand transformfaultsis held to be a diagnosticfeatureof terrestrialplate tectonicsthat is visiblein the topographicdata for some oceanic regions even at Pioneer Venus resolution [ArvidsonandDavies,1981]. On the basisof laboratoryexperiments,however,it appearsthat this pattern may be a function of materialproperties.Oldenburgand Brune[1975] foundthat in wax analogsof spreadingcentersystems,establishmentand maintenanceof an orthogonal ridge-transformpattern occurred only if the resistivestressesalong transform faults are less than the shear vtrength of the lithosphere.If, compared with the earth, the shearstrengthat the elevatedsurfacetemperature of Venus is reducedrelative to transform-faultresistance,then Venusmight displaylithosphericrecyclingwithout the familiar orthogonalpatternat spreadingcenters. Convergencezones on a Venus with lithosphericrecycling may also differ substantially from analogous zones on the earth. Island arc volcanism,thoughincompletelyunderstood,is thought to result from partial melting initiated by the dehydration at depth of subductedoceaniccrust [e.g.,Lofgrenet al., 1981]. If subduction occurred on Venus without the entrainment of water or hydrousphasesin the subductedmaterial, then dehydrationat depth would not occurand the Venus equivalent of island arc volcanismneed not be present.The deep-seatrenchesmarking the locus of subductionof oceanic lithosphereare known to owe their physiographiccharacteristics,suchas relative depth and width, to the thicknessof the elasticor viscoelasticportion of the subductedplate, the velocity of convergence, and the nature of any adjacentsourcesof sediment [e.g., Grellet and Dubois, 1982]; these parameters would be expectedto differ for Venus. Becauseof the high surfacetemperatureof Venus,a bettermodelfor someconvergencezoneson that planet might be basedon the notion of lithosphericdelaminationas proposedfor zonesof terrestrial continentalcollision[Bird, 1978,1979;Reutteret al., 1980;Bird and Baumgardner,1981; Fleitout and Froidevaux,1982]. According to this model, a low-strengthzone at the base of the crustpermitsdelaminationof the lithosphere,subductionof the mantle portion, and deformation without subductionof the more buoyant crustalmaterial. Such a model, applied to some subductionzones on Venus, would lead to the prediction of regions of intensesurfacedeformationsurroundingconvergencezonesbut little recyclingof surficial or crustal material back into the underlyingmantle. It is worth noting that there are topographicfeatureson the Venussurfacewhichdiffer from any seenon the smallerterrestrial planets and which resemblefeatureson earth of plate tectonic origin. These featuresinclude linear mountain belts 9239 suchas Akna and Freyja montes,continentalsizeplateaussuch as Lakshmi Planum,and arcuateridgesand troughssuchas in the Ut and Vesta Rupes region on the southwestmargin of Ishtar Terra [Masursky et al., 1980; Head et al., 1981; McGill et al., 1981, 1982; Campbellet al., 1982; Phillipsand Malin, 1982]. We thereforeregard the tectonicinterpretation of the topographic featuresof the Venus surfaceas uncertain at the resolution of our current information.Both an improvementin the resolution of surface features on Venus and detailed investi- gationsof the effectsof Venussurfaceconditionson the characteristiclandformsproducedby variousgeologicalprocesses are neededto improveour understandingin this area. On the Age of the VenusSurface A number of approximatelycircular featureson the Venus surface,ranging in diameter from 20 to over 1000 km, have beenidentifiedon the basisof variationsin radar reflectivityor topography [Goldsteinet al., 1976, 1978; Schaberand Boyce, 1977; Campbellet al., 1979; Campbelland Burns, 1980]. The hypothesisthat many or all of these featuresoriginated by impact has led to the inferencethat much of the surface of Venusis geologicallyancient.In particular,if the quasi-circular regionsof low radar backscatter200 km or greaterin diameter have beencorrectlyidentifiedas impact basinsdating from the heavy bombardment of the inner solar system,then Venus containsmajor geologicunits with surfaceagesapproaching4 b.y. [SchaberandBoyce,1977;Masurskyet al., 1980]. Becauseof the high surfacetemperature,topographicfeatures on Venus are subjectedto viscousrelaxation at rates much higher than for the other terrestrialplanets[e.g., Weertman, 1979]. In particular, if the surfacetemperatureof Venus has been comparable to its present value for much of the planet'shistory,then a 4-b.y.-oldimpact basinon Venusshould havenegligibletopographicrelief [Solomonet al., 1982]. Large circulardepressions on Venus,such as Atalanta Planitia, are substantiallyyoungerthan 4 b.y. and are unlikely to be impact basins.Such topographicdepressions may owe their origin to one of the tectonicmechanismsproposedfor intracontinental basinson earth, i.e., cooling and subsidencefollowing lithosphericextensionor thermaluplift and erosion[SleepandSnell, 1976; McKenzie, 1978]. The large quasi-circular,radar-dark featureson the Venus surfacemay still be impact basins,degraded by essentiallycompleterelaxation of long-wavelength topography.Suchfeaturesmay also be older versionsof circular lowland regionsof Venus,however,and by implication also of tectonicorigin. The question of the age of the Venus surfaceshould be regardedas open.There are alternativetectonicinterpretations of the large circular featuresvisiblein radar images.Even the population of smaller 'craters' 20-200 km in diameter may contain featuresof volcanicas well as impact origin [Saunders and Malin, 1976; Campbell and Burns, 1980; McGill et al., 1982]. Theseuncertaintiesdo not precludethe possibilitythat portions of the Venusiansurfaceare billions of years old, nor do they permit the rejectionof the hypothesisthat much of the Venusian surface, as with the solid surface of the earth, is geologicallyyoung. On LithosphericBuoyancyon Venus A necessarycondition for subductionis that the subducted material be of greaterdensitythan the surroundingmantle. The negative buoyancy of the subductedslab of lithosphere on 9240 SOLOMON AND HEAD: LITHOSPHERIC HEAT TRANSPORT ON VENUS earth is a consequence of its relativelycoolertemperatureand also a result of phasechangeswhich occur at shallowerthan normal depth in the cooler slab [Toks& et al., 1971; Turcotte and Schubert,1971]. The argumenthas been advancedthat a Venus lithospheresimilar in crustalthicknessand other physical propertiesto oceaniclithosphereon earth would have less negativebuoyancybecauseof the higher surfacetemperature and smaller lithosphericthicknesson Venus [Anderson,1981; Phillipsand Malin, 1982]. Anderson[1981] has arguedfurther that an additional factor which might make the Venus lithospherepositivelybuoyant is the enhanceddepth range of the basalt stability field on Venus compared with that on earth, thereby openingthe possibilityfor crustalthicknesses substantially greaterthan in terrestrialoceanbasins. Thesebuoyancycalculationsare sensitiveto the assumptions made about parameterswhich are at best poorly known for Venus,includingthe thicknessof a crustallayer, the locationof the mantle residuumremainingafter extractionof partial melt to producethe crust,and the partitioningof planetaryheat loss among plate recyclingand conduction.Thus the uncertainty attachedto an individualresultis large,and both positiveand negativevaluesfor net lithosphericbuoyancyare possibleresultsat our presentstageof understanding[Phillipsand Malin, 1982]. That the Venusian lithosphereis less negatively buoyant than old oceaniclithosphereon earth is likely, yet even this conclusionis not a strong argument against some form of subductionon Venus. On the earth, oceaniclithospherewith seaflooras youngas 10 m.y. is subducted[Menard, 1978], yet such lithosphereis certainly lessnegativelybuoyant than unsubductedoceaniclithosphere150 m.y. old. Parsons[1982], in fact, has demonstratedthat the distribution of ocean floor area versusage on the earth is consistentwith the hypothesisthat plate consumptionis uniformlydistributedwith age exceptfor seaflooryoungerthan 10 m.y. old. Subductionof young seafloor may, of course,be sustainedby a finite-amplitudeinstability; i.e., 10 m.y. old oceaniclithospheremay be only marginally unstable or even stable gravitationally yet continue to subductif attachedto a negativelybuoyantslab.Sucha finiteamplitudeinstabilitymight equallywell be invoked to sustain subductionon Venus. Once subductionof an aging piece of Venusianlithosphereis initiated,by this argument,the subduction processmay be sustainedby thermal anomaliesand elevated phasetransitionsin the subductedslab,and the characteristic age of subductedmaterial may be substantiallyyounger than that initially subducted.If, as suggestedearlier, lithosphericdelaminationoccurson Venus,then the mantleportion of the lithospherewould be negativelybuoyant even if the crustalportion is not. Thus a thickercruston Venuscompared with terrestrial oceanbasins[Anderson,1981] need not restrict subductionof the mantleportion of the lithospherewheresuch delamination occurs. To the extent that the mantle residuum complementaryto a basalticcrustresidesin the lithosphere,in fact,the net lithosphericbuoyancymay be largelyinsensitiveto crustal thickness. On the Kaula-Phillips Argumentson Heat Removal by LithosphericRecyclingon Venus Kaula and Phillips [1981] estimatedan upper bound on the fraction of the global heat losson Venus contributedby plate tectonics.While not an argument against plate recyclingon Venus per se, their conclusionthat the heat transported by plate tectonicsis 15% or lessof the global heat flux on Venus compared with 65% on earth [Sclater et al., 1980] would, if true,providean importantcontrastbetweenthe two planets. We do not regard the 15% result of Kaula and Phillips [1981], however, as necessarilyvalid. Their calculation was based on an identificationof candidatespreadingridges on Venus, a measurement of the rate of decreasein elevation with distancefrom each segmentof the ridge, an application of thermalboundarylayertheoryto scalethe physicalparameters of mantle convection in the earth to those of Venus, and a conversionof the measuredrates of change in elevation to estimatesfor the heat delivered along each spreadingridge segment.Further elaboration of the boundary layer scaling argumentsis givenby PhillipsandMalin [1982]. The candidatesfor ridges chosen by Kaula and Phillips [1981] includeBeta Regio and Aphrodite Terra. Both of these features contain large regions which, together with Ishtar Terra, composethe highlandsprovinceof Venus [Masurskyet al., 1980]; terrestrial continentsrather than oceanridgesmay provide a preferableanalog to much of the Venus highland terrain [e.g., Head et al., 1981]. Having made this selectionof ridges, Kaula and Phillips then note that the Venus ridge heightsdo not showa narrow distributionabout a mode as do terrestrial ocean ridgeswell removedfrom hot spots [Parsons and $clater, 1977] and that theseridgesdo not form an interconnectedglobal system. While these observationsmay be sufficientto rule out either Beta Regio or Aphrodite Terra as Venusiananalogsto terrestrialmid-oceanridges[eft $chaber, 1982], neither observationis a compellingargument against lithosphericrecyclingon Venusin general. Having posedthe hypothesisthat plate recyclingdominates lithosphericheat transport on Venus,we shouldconsiderthe expectedform of Venusian spreadingcenters.A reasonable guide to the rate of heat loss per length of spreadingridge followsfrom the spreading-platemodel of oceaniclithosphere [Williams andyonHerzen, 1974; Kaula andPhillips,1981]: q, = pC•,LvAT (1) wherep andCparethedensityandspecific heatof thelithosphericplate, L is the lithosphericthickness,v is the average half spreadingrate, and AT is the differencein temperature betweenthe bottom and top of the lithosphere.Subductionof the lithospherebeforeit has reachedthermal equilibrium will reduceq, from that givenby (1). Becauseof the highersurface temperatureon Venus, AT will be lessfor Venus than for the earth'soceanbasins,thoughboundarylayer scalingarguments suggestthat the differencein A T betweenVenusand earth may be somewhatless than the differencein surfacetemperatures [Kaula and Phillips, 1981; Phillips and Malin, 1982]. For a spreadingridge systemon Venus to deliver to the surfacean amount of heat comparable to that delivered by ridges on earth,eitherthe characteristic spreadingrate or the total length of ridges(or both) must be greater on Venus than on earth. Becauseeither possibilitywould likely be accompaniedby a lessercharacteristicage of subductedlithosphere on Venus than on earth, the product of ridgelengthand mean spreading rate would exceedthat quantityfor the earth by a ratio greater than simplythe ratio of the valuesof A T for the two bodies. Theseconsiderations suggestthat the mostlikely candidates for spreadingridgeson Venusshouldbe characterizedby rapid ratesof spreadingand correspondingly smallratesof changein topographicheightwith distancefrom the ridge axis.Phillips and Malin [1982] have shownthat a fast spreadingridge, such as the East Pacific Rise, would have only modesttopographic SOLOMON AND HEAD: LITHOSPHERIC HEATTRANSPORT ON VENUS 'Convergence'Continent ' Zone' Spreading /Center 'Ocean H•ghlands ____ Basin' Rolling Pla•__ns._. • r LowLands Plate Recycling Hypotnes•s Fig. 2. A schematic illustrationof theplaterecyclinghypothesis for lithosphericheat transporton Venus.The rolling plains(0 to 2 km elevationwith respectto the planetarymodal radius)and lowlands(0 to -2 kin) arc analogousto terrestrialocean basinsand include spreading centersandconvergence zonesnot currentlyresolvable from altimctryor imagingdata. The highlands(2 to 11 km elevation)arc analogousto terrestrialcontinents. relief at Venusconditionsand that sucha ridge on Venuscould not be detectedfrom PioneerVenus altimetry given the 200-m standard error for thesedata [Pettengill et al., 1980]. By the plate recyclinghypothesisfor Venus, the rolling plains and lowlands provinces (as defined by Masursky et al. [1980]) should be the Venusian analogs to terrestrial ocean basins. Altimetric and imaging data from rolling plains and lowlands areasmust be obtainedat resolutionssuperiorto thoseof data currently availablein order to searchfor the predictedextensive systemof fast spreadingridgesand thereby to provide a rigoroustestof the platerecyclinghypothesisfor Venus. Assessment and Implicationsof the Hypothesis We find no convincingargument to reject at presentthe hypothesisthat someform of plate recyclingdominateslithosphericheat transferon Venus.This conclusiondoesnot mean that plate tectonicscan presentlybe demonstratedon Venus-or that we may not at somefuture date obtain improved information on the Venusian surface or interior 9241 extensiveby this hypothesisand should be concentrated at divergent plate boundaries.Intraplate volcanic activity may also be present.Volcanismat subductionzonesmay be minor at presentif no free water or hydrousphasesare subducted. Tectonic activity shouldbe widespreadand dominated by the large-scalehorizontal motions and mutual interactions of the plates. As noted above, however, the specificphysiographic characteristicsof many tectonicfeaturesat plate boundaries may differ from those on earth becauseof the greater surface temperatureand negligiblesurfacewater on Venus.Theoretical models are neededfor the expectedform of such featuresat Venus-surface conditions. By the plate recycling hypothesis,the rolling plains and lowlands provincesare analogsto terrestrial ocean basinsin terms of formativeprocess,crustalcomposition,and geologically youthful age. The spreadingcentersare expectedto be characterizedby rapid spreadingrates and modest relief in comparisonto earth. The Venushighlandsmay be analogsto terrestrial continents, though whether theanalogy extends to compositionor simply to crustal thicknessis uncertain. The surfaceagesof highland geologicunits on Venus may be considerablygreater than the agesof units in the plains and lowlands. As with terrestrial continents, highland lithosphere should have greater buoyancy than the lithosphere beneath plains and lowlandsand shouldthereforebe more difficult to subduct.Mountainous terrain should,by this hypothesis,form by the collision of highland blocks (Himalayan analog) or at the locusof subductionbeneathhighland lithosphere(Andean analog).Both highlandregionsin generaland mountain belts in particular will be modified as they age. Becauseof the high surface temperature and negligible water on Venus, viscous relaxation may be a more rapid processthan weatheringand erosionfor reducinghighlandrelief,in contrastto earth. LITHOSPHERIC CONDUCTION that will allow the The dominant mechanismfor lithosphericheat transport on questionto be firmly resolved.Rather,the conclusiondemands that at our presentlevel of understanding,the hypothesisthat the smallerterrestrialplanetsis conductionthrough the single plate recycling is important for heat flow on Venus should global lithosphericshell. That lithosphericconduction domicontinue to be rigorously tested against available geological natesthe planetary heat losson Venus has been an implicit or information. explicit elementof a number of recentdiscussions of the interThe lithosphericrecyclinghypothesisfor Venusis illustrated nal and surfaceevolution of Venus [Schaberand Boyce, 1977; schematicallyin Figure 2, and some of the implicationsof the Warner, 1979; Anderson,1981; Phillips et al., 1981; Arvidson hypothesisfor the surfacecharacteristics of the planetare listed andDavies,1981]. If the global heat loss on Venus is deliveredto the surface in Table 1. As on earth, volcanic activity on Venus should be TABLE 1. Implicationsof End-Member Hypothesesfor LithosphericHeat Transporton Venus Surface Characteristics Plate Recycling LithosphericConduction Hot Spot Volcanism Volcanic activity extensive;activity dominantly at divergent boundaries minor extensive; active centersnearly Tectonic widespread;dominated by plate interactions possiblyextensive primarily vertical features cover surface deformation of tectonics thin lithosphere Agesof surfaceunits rolling plains and lowlandsgeologically young( < 10s years) Nature of mountain belts productsof plate convergence Nature of highlands analogsto terrestrial continents Nature of rolling plains and lowlands analogs to terrestrial ocean basins unconstrained; ancient impact featuresmay be preserved much of surface young(<• 107 years) anomalouslythick crust and lithosphere volcanic thickened crust and thickened lithosphere thinned crust and lithosphere constructs volcanic crust 'normal'-thickness volcanic crust 9242 SOLOMON AND HEAD' LITHOSPHERIC HEAT TRANSPORT ON VENUS entirely by lithospheric conduction, then the conductive geothermmay be estimatedfrom the averagesurfaceheat flow of 74 mW/m2 derivedearlier.This heat flow is very closeto L•thospher•c L•t hospher•c Thickening • • H•ghlands Roll•ng Plains Lowlands twice the heat flux in old oceanic lithosphere near thermal equilibrium,givenas 38 + 4 mW/m2 by Sclateret al. [1980]. Thus if the Venuslithospherehas a thermal conductivitysimilar to the averagevalue of 3.1 W/m K adoptedfor the terrestrial oceaniclithosphere[Parsonsand Sclater,1977], then the thermal gradientin the Venus lithosphereis 24 K/km, or approximately twice that in old ocean basins. Geotherms for oceaniclithosphereat equilibrium [Sclater et al., 1980] and for average lithosphere on Venus, assumingthat heat transfer occurssolelyby lithospheric conduction, areshownin Figure3. A consequefice oftheVenusgeotherms shown in Figure3 is that the lithospherewould be substantiallythinner than on earth for comparablematerial propertiesof the mantle. The baseof the thermallithospherein oceanbasinsis 1350 + 275øC [ParsonsandSclater,1977].A temperatureof 1350øCwouldbe reachedat a depth of about 40 km on Venusaccordingto the geothermsshown in Figure 3. The base of the elasticlithospherein oceanicregionsis well approximatedby the position of the 500 + 150øC isotherm [Watts et al., 1980]. An elastic lithosphereon Venuslimited by this sametemperature,a value governedby the ductile flow law for dry olivine [Goetze and Evans,1979], would be at most 10 km in averagethicknessfor a purelyconductiveVenus. On both of thesegrounds,it is difficult to envisionmechanismsfor supportingthe 13 km of relief on Venus[Pettengillet al., 1980] for geologicallylong periodsof time exceptperhaps throughsheartractionsassociatedwith mantle convectiveflow. A likely implication,therefore,of the hypothesisthat conduction dominatesheat transferon Venusis that all surfacetopographicrelief on scalessmallerthan the characteristichorizontal scalesof mantle convectionis geologicallyyoung. Preservation of topographicrelieffor extendedperiodsof time might occur more readily if elevated regions on Venus are 500 i000 • 5O 1500 Venus: L•thospher•c "-,.•o .,.., nOnly z• Base of Elastic ' characterizedby a mantle heat flux that is lower than average. That such regionally low values of mantle heat flux could persistfor hundredsof millionsto billionsof yearsfor a planet with lithosphericheat transferdominatedby conduction,however, is unlikely. Alternatively,sincethe history of the surface temperature is uncertain [Pollack, 1971, 1979], topographic relief may have persistedfor geologicallylong periodsif the characteristictime for viscousrelaxation is comparable to or lessthan the time sinceformation of the presentatmospheric 'greenhouse.'The viscousrelaxationtime is unlikely to exceeda few hundredmillion years [Solomon et al., 1982],however,so this possibilitywould requirea geologicallyrecent greenhouse. The Venus geotherm shown as a solid line in Figure 3 is basedon the simplificationthat all of the heat lost by the planet is generatedbeneaththe lithosphere.The lithosphericthermal gradientwould be lessenedif concentrationof radioactiveheat sourcesin the Venus crust has occurred,thereby reducingthe heat flux from the mantle [e.g.,PhillipsandMalin, 1982].Since, by the conductionhypothesis,plate recyclingand attendant remelting of basalticcrust at subductionzoneswould not have progressedon Venus to the current stageof concentrationof heat sources in continental crust on the earth, a reasonable T• øC 0 Lithospheric Conduction Hypothes•s Fig. 4. A schematicillustration of the conductionhypothesisfor lithosphericheat transport on Venus. On average,the lithosphereis only about 40 km thick and may be readily deformedby tractions associatedwith convectionin the underlyingasthenosphere. Modestly elevatedregionsin the rollingplainsmay be areasof recentlyextended and thinnedcrustand lithosphere, areaswhichwill subsideto lowland elevationsduring lithosphericcoolingand thickening[e.g.,McKenzie, 1978]. The more elevatedhighlandsmay be areas of thickenedcrust and lithosphereresultingfrom lithosphericcompression. upper bound to the fraction of Venus heat flux contributedby crustalradioactivityis the terrestrialvalue of about 15% [Sclater et al., 1980]. A greater concentrationof radioactive heat sources into the crust on Venus than on the earth would also be difficult to reconcile with the lower '•øAr abundance in the Venus atmosphere[Hoffman et al., 1980]. With the 15% value assumedfor the fraction of global heat loss generatedin the - -- L•thosphere crust, the mantle heat flux on a Venus with the same heat loss Earth: • Equ•l•br•urn per massasthe earthwouldbe29 x 10•2 W, and the temperaturegradientin the Venuslithospherewould,as shownby the dashedcurve in Figure 3, be only slightly modified from that I00 150 discussed above. I I Fig. 3. Average lithosphericgeothermson Venus assumingthat conductionis the only modeof lithosphericheat transfer.The solidline showsthe casewhenall of the heat lossfrom Venusis generatedbelow the lithosphere;the short-dashedcurveindicatesthe casewhen 15% of the Venusheat lossis generatedby radioactivitydistributeduniformly in a crust 30 km thick. Also shownare the terrestrialgeothermfor an old ocean basin in thermal equilibrium [Parsonsand Sclater, 1977; Sclater et al., 1980] and the range of isothermsinferred to define the baseof the elasticlithospherein oceanbasins[Watts et al., 1980]. The hypothesisthat lithosphericheat flux on Venus occurs principally by conductioncannot be rejectedon the basisof presentlyavailableinformation.A schematicillustrationof this hypothesisis given in Figure 4, and a summary of the implicationsfor the surfacegeologyof Venusis givenin Table 1.The averagelithosphericthermalgradientsare predictedto be substantially greater, and the lithosphericthicknesscorrespondingly less, than for volcanic mechanismsof heat transport, including both plate recyclingand hot spot volcanism.The lithosphericstrengthand resistanceto deformationshouldgenerally be lessfor this hypothesisthan for the hypothesisthat either plate recyclingor hot spot volcanismdominateslithosphericheat transport on Venus.As a result,lithosphericand SOLOMONAND HEAD' LITHOSPHERIC HEAT TRANSPORTON VENUS crustalthinning and thickeningmay occurin responseto tractions exertedon the baseof the lithosphereby mantle convective flow (seeFigure 4). The agesof surfacegeologicalunits are not constrainedby the conductionhypothesisand may spana greatrange.Becauseof the expectedhigh rate of viscousrelaxation, however,topographicrelief on Venus should be either geologicallyyoungor dynamicallymaintained.Sincethe planform of mantle dynamicsis not likely to be steady on time scalesapproachingthe ageof theplanet,it wouldbe reasonable to concludeby this hypothesisthat none of the topographic featuresof Venus is likely to date from the first half of the planet'shistory. HOT SPOT VOLCANISM Individual volcanichot spotstransportheatfrom a planetary interior by the ascentand coolingof magmaand by conduction through a locally thinned lithosphere.Phillips and Malin ['1982] have proposedthat hot spot activity on Venus may providethe dominantmechanismof lithosphericheattransfer. The implicationsof sucha hypothesismay be quantifiedby consideringfirst the archetypicalterrestrialhot spot,the source of the Hawaiian-Emperor island-seamountchain ['Wilson, 1965;Mor•Ian, 1971].The rate of heatlossQvdue to volcanism at a hot spotis givenapproximatelyby 9243 every200-km squareof the Venussurface.Whether the mantle convection system on Venus could supply magma to this number of volcaniccentersessentiallysimultaneouslyon geologicaltime scalesis an openquestion. Two effectsmay increasesomewhatthe contributionof volcanismat a Hawaiian-typehot spot to the total heat flux of Venus.The eruption temperatureof typical hot spot magmas on Venus may be greater than on earth, due either to a lower FeO/(MgO + FeO) ratio [Goettel et al., 1981; Phillips and Malin, 1982] or to lower volatile abundances[McGill, 1979; Anderson,1980] in the Venusmantle comparedwith that of the earth.The rate of heat lossQofor a Venus'Hawaii' increasesby 0.2 x 109 W for every100K increase in AT in equation(2),so that even a 500 K differencein eruption temperaturebetween Venus and earth would raise Q,by no more than 40%. A second effect, also noted in the last section, is that concentration of radioactiveheat sourcesin the Venus crust may reduce the heat flux from the mantle and thus the number of individual hot spotsneededto remove that heat. Adopting, as above,the 15% value [Sclater et al., 1980] as an upper bound to the fraction of Venusheat lossgeneratedby crustalradioactivity, the contribution of volcanismat a Hawaiian-type hot spotto thetotalheatlosswouldstillbeabout1partin 10'•. A planetwith 10'• hot spotsasactiveasHawaiiwouldhavea total volcanic flux sufficientto affect the planetary surface profoundlyon geologicallyshorttime scales.If the global rate Qv= p(C•, AT+ Amf)dt (2) of heatlossQ is contributedentirelyby hot spotvolcanism,the wherep andC•,arethedensity andspecific heat,respectively, of planetaryvolcanicflux is thevolcanic material,AHœis theheatof fusionof themagma, dV* Q dV ATisthediffere.!•, between theeruption temperature ofthe magma and the•,,ii•i•i•ent surface temperature, and dV/dt isthe volumetricflux of magmawith time.Shaw[-1973]hasestimated thattheHawaiianhotspothasproduced 8.5 x 105km3 of lava between the time of formation of the bend in the Hawaiian- Emperorchainand the present.Adoptingthe figure42 m.y. for the age of the bend I'Dalryrnpleand Cla•tue, 1976] gives an averagefor dV/dt of 2 x 10-2 km3/yr,thoughvalueslargerby up to a factor of 5 havebeenappropriatefor time.intervalsof a few million years [Shaw, 1973]. To estimatethe averagevolcanic heat loss contributedby the Hawaiian hot spot for the last42m.y.,weassume p = 2.8g/cm 3,C•,= 1.2J/gK, AHœ= 400 J/g, and AT = 1300 K [Solomonet al., 1981]. Then from (2),Qo- 3.5 x 109W, or 0.008%oftheearth'sglobalheatloss. The total contributionof hot spot volcanismto the overall heat lossof the earth is minor [Sclater et al., 1980]. Of all the proposedhot spots [e.g., Morgan, 1971; Burke and Wilson, 1976], Hawaii hashad probablythe greatestvolumetricflux of magmain recentgeologichistory.Thus whetherthe important terrestrialhot spotsnumber20 [Morgan, 1971] or closerto 100 [Burke and Wilson, 1976], the summed contribution of volcanicallydeliveredheatfrom suchhot spotsto globalheat flow is less than 1%. If a hot spotwith the samemagmaticflux as Hawaii were on Venus, the rate of heat loss due to volcanism would be some- what lessthan on earth becauseof the higher surfacetemperatureand thereforethe lesserextentof cooling.For an eruption temperatureand other parametersequal to thoseassumedfor theearth,equation (2)givesQ,= 2.5 x 109 W, or 0.007%of the global Venus heat lossif the heat lossper massof Venus and the earth are equal.Thus to removethe total heat flux from Venusby volcanismat hot spotswith ratesof volcanicactivity similarto thatof Hawaii,a totalof 10'• suchhotspotswouldbe needed.This figureamountsto an averageof one 'Hawaii' on dt - p(C•, AT + AHœ) (3) For Q = 34 x 10•2 W andotherparameters asadoptedabove, dV*/dt = 280km3/yr.Averagedoverthesurfaceof Venus,this fluxamountsto a rateof surfaceburialof 6 x 10-7 km/yr,or an additional 1-km thickness of new volcanic material over the entireplanetevery2 m.y. This globalburial rate is comparable to the estimatedlower bound on the burial rate for Io [Johnson et al., 1979]. Terrestrial hot spotscontribute heat to the lithosphereby mechanisms other than extrusivevolcanism.Lithosphericthinning beneathhot spotsleadsboth to an enhancedconductive heatflux and to modestelevationsin surfacetopography,such as the Hawaiian swell associatedwith the hot spot beneath Hawaii [Detrick and Crough,1978]. This lithosphericthinning occurson time scalesshort compared with the characteristic time for lithosphericheatingor coolingby conduction[Detrick andCrough, 1978];asthenoSPheric stoping, or in situdelamination and sinkingof the lower lithosphere,may be involvedin the rapid thinningprocess[Turcotte, 1982].The lithosphereis apparently thinned by about a factor of 2 to 3 beneath the Hawaiian hot spot [Detrick and Crough, 1978; Crough, 1978; Detrick et al., 1981;Sandwell,1982;yonHerzen et al., 1982]. Ifa large number of hot spots are operative on Venus, then the conductivegradient would be lessenedand the effectivelithospheric thicknessincreasedin those areas distant from hot spotscomparedwith the global averagescalculatedin the last section. Theseideasmay be quantifiedas follows.Let • be the globally averagedheat flux, let q0 be the averageheat flux in areas distantfrom hot spots,letf be the fractionalarea of planetary surfaceoccupiedby hot spots,and let r/q0 be the conducted heat flux in the vicinity of a typical hot spot.The quantity r/ may be regarded as the factor by which the lithosphere is 9244 SOLOMON ANDH•D: LITHOSPHERIC HEATTRANSPORT ONVENUS characteristicallythinned by someunspecifiedprocessat a hot spot.Clearly, Large Volcanic Construct Pervasive Volcanic Centers H•ghlands Roll•ng O =fnqo + (1 -f)qo + qv (4) Lowlands Plains •,•' • /// whereq• is the contributionto the globalheat flux deliveredby volcanism. If volcanicallydeliveredheat is smallcomparedwith conducted heatat hot spotsand elsewhere (i.e.,q, ,• •), thenq0 is given by o qo= 1--f +fil (5) Under the assumptionthat the Hawaiian hot spotis a guideto the characteristicsof possiblehot spots on Venus, we take r/•_ 2-3 [Detrick and Crough, 1978; Crough, 1978; Detrick et al., 1981; Sandwell, 1982; van Herzen et al., 1982]. For r/= 2, q0 •> •/2. That is, the lithosphericthicknessfar from hot spots can be increasedby a factor only as large as 2 comparedwith the global averagederivedin the last section,with the factor of 2 approachedonly for the casewhere hot spotscover most of the surfaceof Venus (f_• 1). For r/= 3, q0 •> •/3, again with near equality only if f •_ 1. Lithosphericthicknesses approaching typical terrestrial valueswould be possible,but only in a few areasof very limited spatialextent. If, in contrast, lithosphericheat transfer on Venus occurs dominantly by magma transport at individual hot spots(i.e., q• -• 0 in equation(3)),the averagelithosphericthicknessis not constrainedby global heat flow [cf. O'Reilly and Davies,1981] and can be considerablygreaterthan estimatesobtainedin the last section.In particular, the averagelithosphericthickness can be greaterthan the 100-kmdepth of isostaticcompensation inferred from the gravity anomaliesover a number of topographicallydistinctfeatureson the planet [Phillipset al., 1979, 1981; Reasenberget al., 1981]. Thus large topographicrelief couldbe supportedpassivelyand on nearly a globalbasiseither by local isostaticcompensationor by regional compensation and lithosphericstrength. The hypothesisthat hot spot volcanism dominates lithosphericheat transferon Venus(Figure 5) cannotbe excludedon the basisof our presentknowledgeof the surface.The hypothesiscarriesimportant implicationsfor the interpretatonof surface physiographicfeatures (Table 1). The Venus surface,if most of the heat loss occursby volcanism,should be densely Hot Spot Volcanism Hypothesis Fig. 5. A schematicillustrationof the hot spot volcanismhypothesisfor lithosphericheat transport on Venus. Volcanismcarriesthe majority of the heat throughthe lithosphere,and activevolcaniccenters on the surface of Venus should be widespread.The highland terrain consistsof large volumes of volcanic depositsand probably thickenedcrust.The surfacematerial is generallyvolcanicand geologically young. sideredin this paper, and the implicationsof their model involve a combination of aspectsof the scenariosdepicted in Figures4 and 5. CONCLUSIONS Without more detailed information on the Venus surface,all surface material should, in either case, consist of volcanic de- of the mechanismsfor lithospheric heat transfer considered here (lithospheric recycling, conduction, and hot-spot volcanism)shouldbe regardedas potentiallyimportant for Venus. Though eachof thesemechanismshasbeentreatedindividually in this paper, weightedcombinationsof mechanismsmust, of course,also be considered[e.g., Phillipsand Malin, 1982]. The possibilitythat the dominant mechanismearly in the history of the planetwasdifferentfrom that at present[e.g.,Phillipset al., 1981] shouldbe recognizedas well. Each of the possibleend-membermodels for lithospheric heat transfer on Venus carries different implicationsfor the detailed characteristicsof Venus landforms. The predicted characteristics (Table 1) shouldserveasa guideto the interpretation of existingand future imaging and topographicdata on the Venussurface.For each of the hypotheticalmodelsfor heat transfer,however,there is a clear implication that either many of the surfacegeologicalunits or much of the surfacetopography is geologicallyyouthful. A common presumptionin discussionsof comparativeplanetary evolutionhas beenthat planetary sizeplaysa crucial role in controlling the internal heat budget and the volcanic and tectonicresponses to global heat loss.The comparisonof Venus and earth holds a key spot in the test of this presumption [Phillipset al., 1981;Head andSolomon,1981]. Establishingthe natureof the tectonicand volcanichistoryof the Venussurface and inferring the dominant mechanismsfor lithosphericheat transferon that body remain elusivegoalsof the highestpriority for our understandingof the evolution of the terrestrial planets,includingour earth. posits.In particular, the Venus highlandswould most likely have beenformedby volcanicconstruction.Tectonicactivity in the absenceof large-scalehorizontal motions should be principally restrictedto that associatedwith verticalmotionsof the lithosphere. The Phillips and Malin [1982] model for hot spot tectonics on Venus includesboth volcanictransport of heat and lithosphericthinning beneathmajor volcaniccenters,with the latter processmaking a greatercontributionto the global heat loss. This modelmay be viewedas a combinationof the mechanisms of lithosphericconduction and hot spot volcanismas con- Acknowledgments.Discussionsat the VenusTectonicsWorkshop (April 20-21, 1982),sponsoredby the Lunar and Planetary Institute and organizedby Roger Phillips,were helpfulin sharpeningthe argumentspresentedhere.We appreciateconstructivereviewsof an earlier draft of this manuscript by David Armbruster, Kevin Burke, Rob Comer, Ken Goettel, Bill Kaula, Mike Malin, and GeorgeMcGill. We thank Ray Arvidson,Bob Detrick, Bill Kaula, George McGill, Roger Phillips,JerrySchaber,and Don Turcottefor preprintsof papersprior to publication,and Dorothy Frank and Jan Nattier-Barbara for their assistancein the preparation of this manuscript.This researchwas supportedby NASA grants NSG-7081 and NSG-7297 at MIT and grantsNGR-40-002-088 and NGR-40-002-116 at BrownUniversity. covered with thousands of distinct centers of current or recent volcanic activity. Most of the physiographicfeatureswould have surfaceslessthan 10 m.y. old, thougha few areasof older terrain may have escapedrecentresurfacing.If the majority of the heat deliveredat hot spotsis due to lithosphericthinning, then hot spotswould nearly have to coverthe surfaceof Venus for the lithosphericthicknessfar from hot spotsto be substantially greater than the globally averagedvalue. Much of the SOLOMON AND HEAD: LITHOSPHERIC HEAT TRANSPORT ON VENUS REFERENCES Anderson,D. L., Tectonicsand compositionof Venus, Geophys.Res. Lett., 7, 101-102, 1980. Anderson, D. L., Plate tectonicson Venus, Geophys.Res. Lett., 8. 309-311, 1981. Arvidson, R. E., and G. F. Davies, Effects of lateral resolution on the identification of volcanotectonicprovinces on earth and Venus, Geophys.Res.Lett., 8, 741-744, 1981. Arvidson, R. E., and E. A. Guinness,Global topography of earth, Venus,Mars: Cluesto tectonicstyles,J. Geol.Educ.,30, 86-92, 1982. Ash, M. E., I. I. Shapiro, and W. B. Smith, The systemof planetary masses,Science,174, 551-556, 1971. Barsukov, V. L., New Venera results,paper presentedto the 13th Lunar and Planetary ScienceConference,Lunar and Planet. Sci. Inst., Houston, Tex., March 15-19, 1982. Bird, P., Initiation of intracontinentalsubductionin the Himalaya, J. Geophys.Res.,83, 4975-4987, 1978. Bird, P., Continental delamination and the Colorado Plateau, J. Geophys.Res.,84, 7561-7571, 1979. Bird, P., and J. Baumgardner,Steady propagation of delamination events,d. Geophys.Res.,86, 4891-4903, 1981. Brass, G. W., and C. G. A. Harrison, On the possibility of plate tectonicson Venus, Icarus, 49, 86-96, 1982. Burke, K., and W. S. F. Kidd, Were Archean continentalgeothermal gradientsmuch steeperthan thoseof today?, Nature, 272, 240-241, 1978. Burke, K. C., and J. T. Wilson, Hot spotson the earth's surface,Sci. Am.,235(2),46-57, 1976. Campbell,D. B., and B. A. Burns,Earth-basedradar imageryof Venus, J. Geophys.Res.,85, 8271-8281, 1980. Campbell,D. B., R. B. Dyce,and G. H. Pettengill,New radar imageof Venus, Science,193, 1123-1124, 1976. Campbell,D. B., B. A. Burns,and V. Boriakoff, Venus: Further evidenceof impact crateringand tectonicactivity from radar observations, Science,204, 1424-1427, 1979. Campbell, D. B., J. K. Harmon, A. A. Hine, and J. W. Head, Venus radar images(abstract),LunarPlanet.Sci.,XIII, 83, 1982. Cook, F. A., and D. L. Turcotte, Parameterized convection and the thermal evolution of the earth, Earth Planet. Sci. Lett., 75, 1-17, 1981. Crough,S. T., Thermal origin of mid-platehot-spotswells,Geophys.J. R. Astron. Soc.,55, 451-469, 1978. Dalrymple, G. B., and D. A. Clague, Age of the Hawaiian-Emperor bend, Earth Planet. Sci. Lett., 31, 313-329, 1976. 9245 Head, J. W., and S.C. Solomon,Tectonicevolutionof the terrestrial planets,Science, 213,62-76, 1981. Head, J. W., S. E. Yuter, and S.C. Solomon,Topographyof Venusand earth:A testfor the presence of platetectonics,Am.Sci.,69, 614-623, 1981. Hoffman, J. H., R. R. Hodges,T. M. Donahue,and M. B. McElroy, Compositionof the Venus lower atmospherefrom the Pioneer Venusmassspectrometer, J. Geophys. Res.,85, 7882-7890,1980. Johnson,T. V., A. F. Cook II, C. Sagan,and L. A. Soderblom,Volcanic resurfacingratesand implicationsfor volatileson Io, Nature, 280, 746-750, 1979. Kaula, W. M., and R. J. Phillips,Quantitativetestsfor plate tectonics on Venus,Geophys.Res.Lett.,8, 1187-1190, 1981. Lofgren,G. E., A. E. Bence,M. B. Duke, M. A. Dungan, J. C. Green,S. E. Haggerty,L. A. Haskin,A. J. Irving,P. W. Lipman,A. J. Naldrett, J. J. Papike,A.M. Reid,J. M. Rhodes,S. R. Taylor, and D. T. Vaniman, Petrologyand chemistryof terrestrial,lunar and meteoriticbasalts,in Basaltic Volcanismon the Terrestrial Planets, pp. 1-437, Pergamon,New York, 1981. Masursky,H., E. Eliason, P. G. Ford, G. E. McGill, G. H. Pettengill,G. G. Schaber,andG. Schubert,PioneerVenusradarresults: Geologyfromimagesandaltimetry,J. Geophys. Res.,85, 8232-8260, 1980. Matson, D. L., G. A. Ransford,and T. V. Johnson,Heat flow from Io (JI), J. Geophys.Res.,86, 1664-1672,1981. McGill, G. E., Venus tectonics: Another earth or another Mars?, Geophys.Res.Lett., 6, 739-742, 1979. McGill, G. E., S. J. Steenstrup, C. Barton,and P. G. Ford, Continental riftingand the origin of Beta Regio,Venus,Geophys. Res.Lett., 8, 737-740, 1981. McGill, G. E., J. L. Warner, M. C. Malin, R. E. Arvidson, E. Eliason, S. Nozette, and R. D. Reasenberg,Topography, surface properties,and tectonic evolution,in Venus,edited by D. M. Hunten, L. Colin, and T. M. Donahue, in press,University of Arizona Press,Tucson, 1982. McKenzie, D., Some remarks on the developmentof sedimentary basins,Earth Planet. Sci.Lett., 40, 25-32, 1978. McKenzie, D., and F. M. Richter,Parameterizedthermalconvectionin a layeredregionand the thermalhistoryof the earth, J. Geophys. Res.,86, 11667-11680, 1981. McKenzie, D., and N. Weiss,The thermal history of the earth, The ContinentalCrust and Its Mineral Deposits,editedby D. W. Strangway,Spec.Pap. Geol.Assoc.Can.,20, 575-590, 1980. Menard, H. W., Fragmentationof the Farallon plate by pivoting subduction,J. Geol.,86, 99-110, 1978. Daly, S. F., Convectionwith decayingheat sources:Constantviscosity, Geophys. J. R. Astron.Soc.,61,519-548, 1980. Morgan, W. J., Convectionplumesin the lower mantle,Nature, 230, Davies, G. F., Thermal histories of convectiveearth models and con42-43, 1971. straintson radiogenicheat productionin the earth,J. Geophys. Res., Oldenburg,D. W., and J. N. Brune,An explanationfor the orthog85, 2517-2530, 1980. onality of oceanridgesand transformfaults,J. Geophys. Res.,80, 2575-2585, 1975. Detrick, R. S., and S. T. Crough, Island subsidence,hot spots and lithosphericthinning,J. Geophys. Res.,83, 1236-1244,1978. O'Reilly,T. C., and G. F. Davies,Magma transportof heat on Io: A Detrick, R. S., R. P. von Herzen, S. T. Crough, D. Epp, and U. Fehn, mechanismallowinga thick lithosphere,Geophys. Res.Lett., 8, 313Heat flow on the Hawaiian Swell and lithosphericreheating,Nature, 316, 1981. 292, 142-143, 1981. Parsons,B., Causesand consequences of the relationbetweenarea and Dewey, J. F., and J. M. Bird, Mountain belts and the new global ageof theoceanfloor,J. Geophys. Res.,87, 289-302,1982. tectonics,J. Geophys.Res.,75, 2625-2647, 1970. Parsons,B., and J. G. Sclater,An analysisof the variation of ocean Fleitout, L., and C. Froidevaux,Tectonicsand topography for a lithofloor bathymetry and heat flow with age, J. Geophys.Res., 82, 803-827, 1977. spherecontainingdensityheterogeneities, Tectonics,1, 21-56, 1982. Goettel, K. A., J. A. Shields,and D. A. Decker, Density constraintson Pettengill,G. H., E. Eliason,P. G. Ford, G. B. Loriot, H. Masursky, and G. E. McGill, Pioneer Venus radar results:Altimetry and surthe compositionof Venus,Proc.Lunar Planet.Sci.,12B, 1507-1516, 1981. faceproperties, J. Geophys. Res.,85, 8261-8270,1980. Goetze, C., and B. Evans, Stress and temperature in the bending Phillips,R. J., and M. C. Malin, The interior of Venusand tectonic lithosphere as constrained by experimental rock mechanics, implications,in Venus,editedby D. M. Hunten,L. Colin and T. M. Geophys.J. R. Astron.Soc.,59, 463-478, 1979. Donahue,in press,Universityof ArizonaPress,Tucson,1982. Goldstein,R. M., R. R. Green, and H. C. Rumsey,Venusradar images, Phillips,R. J.,W. L. Sjogren,E. A. Abbott,J. C. Smith,R. N. Wimberly, J. Geophys.Res.,81,4807-4817, 1976. and C. A. Wagner,Gravity field of Venus:A preliminaryanalysis, Goldstein,R. M., R. R. Green, and H. C. Rumsey,Venus radar brightScience,205, 93-96, 1979. nessand altitude images,Icarus,36, 334-352, 1978. Phillips,R. J.,W. M. Kaula,G. E. McGill, and M. C. Malin, Tectonics Grellet, C., and J. Dubois, The depth of trenchesas a function of the and evolution of Venus,Science,212, 879-887, 1981. subductionrate and age of the lithosphere, Tectonophysics, 82, Pitman, W. C., III, Relationshipbetweeneustacyand stratigraphic 45-56, 1982. sequences of passivemargins,Geol.Soc.Am. Bull., 89, 1389-1403, 1978. Harrison,C. G. A., Spreadingratesand heat flow, Geophys.Res.Lett., 7, 1041-1044, 1980. Pollack, J. B., A nongreycalculationof the runaway greenhouse: Hays, J. D., and W. C. Pitman III, Lithosphericplate motion, sea-level Implicationsfor Venus'pastandpresent,Icarus,14,295-306,1971. changes and climatic and ecological consequences,Nature, 246, Pollack, J. B., Climatic changeon the terrestrialplanets,Icarus, 37, 18-22, 1973. 479-553, 1979. 9246 SOLOMONAND HEAD: LITHOSPHERICHEAT TRANSPORTON VENUS Reasenberg,R. D., Z. M. Goldberg, P. E. MacNeil, and I. I. Shapiro, Venus gravity: A high-resolutionmap, J. Geophys.Res., 86, 71737179, 1981. Reutter, K.-J., P. Griese, and H. Closs, Lithosphericsplit in the descendingplate: Observationsfrom the northern Apennines, Tectonophysics, 64, T1-T9, 1980. Reynolds,R. T., S. J. Peale,and P. Cassen,Io' Energy constraintsand plumevolcanism,Icarus,44, 234-239, 1980. Sandwell,D. T., Thermalisostasy'Responseof a movinglithosphereto a distributedheat source,J. Geophys. Res.,87, 1001-1014,1982. Saunders,R. S., and M. C. Malin, Venus' Geologicalanalysisof radar images,Geol.Rom.,15, 507-515, 1976. Schaber,G. G., Venus' Limited extensionand volcanismalong zones of lithosphericweakness,Geophys.Res.Lett., 9, 499-502, 1982. Schaber,G. G., and J. M. Boyce,Probabledistributionof largeimpact basins on Venus: Comparison with Mercury and the moon, in ImpactandExplosionCratering,editedby D. J. Roddy,R. O. Pepin, and R. B. Merrill, pp. 603-612,Pergamon,New York, 1977. in The Earth: Its Origin, Structureand Evolution,edited by M. W. McElhinny, pp. 227-263, Academic,New York, 1979. Surkov, Yu. A., Geochemicalstudiesof Venus by Venera 9 and 10 automaticinterplanetarystations,Proc. Lunar Sci. Conf, 8th, 26652689, 1977. Surkov, Yu. A., F. F. Kirnozov, V. N. Glazov, and G. A. Fedoseyev, Investigations of Venusian gamma-radiation by Venera 9 and Venera 10,SpaceRes.,17, 659-662, 1977. Toks6z, M. N., J. W. Minear, and B. R. Julian,Temperaturefield and geophysicaleffectsof a downgoingslab,J. Geophys.Res.,76, 11131138, 1971. Turcotte, D. L., On the thermal evolution of the earth, Earth Planet. Sci. Lett., 48, 53-58, 1980. Turcotte, D. L., Mechanisms of crustal deformation, J. Geol. Soc. London,in press,1982. Turcotte, D. L., and K. Burke, Global sea-level changes and the thermal structure of the Earth, Earth Planet. Sci. Lett., 41, 341-346, Schubert, G., Subsolidus convection in the mantles of terrestrial Turcotte, D. L., and G. Schubert,Structureof the olivine-spinelphase boundaryin the descending lithosphere,J. Geophys.Res.,76, 7980- planets,Annu.Rev.Earth Planet.Sci.,7, 289-343, 1979. Schubert, G., P. Cassen, and R. E. Young, Subsolidusconvective coolinghistoriesof terrestrialplanets,Icarus,38, 192-211, 1979a. Schubert,G., P. Cassen,and R. E. Young, Core coolingby subsolidus mantle convection,Phys.Earth Planet.Inter., 20, 194-208, 1979b. Schubert,G., D. Stevenson,and P. Cassen,Whole planet coolingand the radiogenic heat source contents on the earth and moon, J. Geophys.Res.,85, 2531-2538, 1980. Sclater, J. G., C. Jaupart, and D. Galson, The heat flow through oceanic and continental crust and the heat loss of the earth, Rev. Geophys.SpacePhys.,18, 269-312, 1980. Sharpe,H. N., and W. R. Peltier, Parameterizedmantle convectionand the earth'sthermalhistory,Geophys.Res.Lett., 5, 737-740, 1978. Sharpe,H. N., and W. R. Peltier, A thermal historymodelfor the earth with parameterizedconvection,Geophys.J. R. Astron.Soc.,59, 171203, 1979. Shaw,H. R., Mantle convectionand volcanicperiodicityin the Pacific; evidencefrom Hawaii, Geol. Soc.Am. Bull., 84, 1505-1526, 1973. Sleep,N.H., Thermal historyand degassingof the earth: Somesimple calculations,J. Geol.,87, 671-686, 1979. Sleep, N.H., and N. S. Snell, Thermal contraction and flexure of mid-continentand Atlantic marginalbasins,Geophys.J. R. Astron. Soc.,45, 125-154, 1976. Solomon, S.C., On volcanism and thermal tectonicson one-plate planets,Geophys. Res.Lett., 5, 461-464, 1978. Solomon, S.C., T. J. Ahrens, P.M. Cassen,A. T. Hsui, J. W. Minear, R. T. Reynolds,N.H. Sleep,D. W. Strangway,and D. L. Turcotte, Thermal historiesof the terrrestrialplanets,in BasalticVolcanismon the TerrestrialPlanets,pp. 1129-1234,Pergamon,New York, 1981. Solomon, S.C., S. K. Stephens,and J. W. Head, On Venus impact basins'Viscousrelaxationof topographicrelief,J. Geophys.Res.,87, 1978. 7987, 1971. Turcotte, D. L., F. A. Cooke, and R. J. Willemann, Parameterized convectionwithin the moon and the terrestrialplanets,Proc.Lunar Planet.Sci.Conf, loth, 2375-2392,1979. Vinogradov,A. P., Yu. A. Surkov,F. F. Kirnozov, and V. N. Glazov, The content of natural radioactive elements in the Venusian Geochem.lnt., 10, 1-8, 1973. von Herzen, R. P., R. S. Derrick, S. T. Crough, D. Epp, and U. Fehn, Thermal origin of the Hawaiian Swell: Heat flow evidenceand thermalmodels,J. Geophys. Res.,87, 6711-6723, 1982. Warner, J. L., A model for the lithosphereof Venus (abstract),Lunar Planet. Sci., X, 1295-1297, 1979. Wasserburg,G. J., G. J. F. MacDonald, F. Hoyle, and W. A. Fowler, Relative contributionsof uranium, thorium, and potassiumto heat productionin the earth, Science,143,465-467, 1964. Watts, A. B., J. H. Bodine, and M. S. Steckler, Observations of flexure and the stateof stressin the oceaniclithosphere,J. Geophys.Res.,85, 6369-6376, 1980. Weertman, J., Height of mountainson Venus and the creepproperties of rock,Phys.Earth Planet.Inter., 19, 197-207, 1979. Williams, D. L., and R. P. yon Herzen, Heat loss from the earth: New estimate,Geology,2, 327-328, 1974. Wilson,J. T., Evidencefrom oceanislandssuggesting movementin the earth, Philos. Trans. R. Soc.London,Set. A., 258, 145-165, 1965. Wood, J. A., W. R. Buck, E. Anders,J. W. Morgan, E. Stolper,D. L. Anderson,W. M. Kaula, G. J. Consolmagno,A. E. Ringwood,and H. Wanke, Geophysicaland cosmochemicalconstraintson properties of mantlesof the terrestrialplanets,in Basaltic Volcanismon the TerrestrialPlanets,pp. 633-699, Pergamon,New York, 1981. 7763-7771, 1982. Sprague, D., and H. N. Pollack, Heat flow in the Mesozoic and Cenozoic, Nature, 285, 393-395, 1980. Stacey,F. D., The cooling earth: A reappraisal,Phys. Earth Planet Inter., 22, 89-96, 1980. Stevenson,D. J., and J. S. Turner, Fluid models of mantle convection, rock accordingto the data obtainedby the automatic probe 'Venera 8,' (ReceivedApril 29, 1982; revisedAugust 11, 1982; acceptedAugust18, 1982.)