NASA Technical Reports Server (NTRS) 20050169984: Thermo-Chemical Convection in Europa's Icy Shell with Salinity

0.82 MB·English

by  NASA Technical Reports Server (NTRS)

Lunar and Planetary Science XXXVI (2005) 1465.pdf
THERMO-CHEMICALCONVECTIONINEUROPA’SICYSHELLWITHSALINITY.L.Han,A.P.Showman,Department
ofPlanetarySciencesandLunarandPlanetaryLaboratory,UniversityofArizona,Tucson,AZ,85721,USA([email protected]).
Summary: Europa’sicysurfacedisplaysnumerouspits, where(cid:10) istemperature,(cid:10)(cid:29)’ ismeltingtemperature,and(cid:7) (cid:26)0/
uplifts, and chaos terrains thathave been suggested toresult (cid:0)1(cid:1) (cid:3)(cid:6)2 Pa sec is the viscosity at the melting temperature. We
from solid-state thermal convection in the ice shell, perhaps adopt #3(cid:14)5476 , corresponding to an activation energy of6 (cid:1)
aided by partial melting [1-4]. However, numerical simula- kJmole8 (cid:3) . Thecutoffviscosity,(cid:7) cutoff,rangesfrom(cid:0)1(cid:1)(cid:4)(cid:3):9 to
tions of thermal convection show that plumes have insuffi- (cid:0)1(cid:1)(cid:27); (cid:3) Pasec,implyingaviscositycontrastof(cid:0)1(cid:1)=< –(cid:0)1(cid:1) (cid:5) .
cient buoyancy to produce surface deformation [5-6]. Here
wepresentnumericalsimulationsofthermo-chemicalconvec-
tion to test the hypothesis that convection with salinity can
produceEuropa’spitsanddomes. Oursimulationsshowthat
domes (200-300 m)and pits(300-400 m)comparable to the
observations can be produced in an ice shell of 15 km thick
with5-10% compositional density variation ifthemaximum
viscosityislessthan(cid:0)(cid:2)(cid:1)(cid:4)(cid:3)(cid:6)(cid:5) Pasec.
Introduction: Europa’s surface displays two dominant
terraintypes,theridgedplains,whichconsistofmultiplegen-
erations of overprinted ridge pairs, and the mottled terrains,
which are distributed with chaos and numerous small (3–30
km-diameter) pits, uplifts, and irregularly shaped landforms
[1,4,7-9]. Ithasbeensuggested thatpits,domes, andchaos
resultedfromthermalconvectionintheiceshell,perhapsaided
bypartialmelting[1-4]. ShowmanandHan[5-6]performed
numericalsimulationsofthermalconvection,showingthatthe
upliftedtopographyproducedinthermalconvectionmodelsis
farlessthantheobservations.
Sofar,allthenumericalsimulationsofconvectioninEu-
ropa’s ice shell assume a pure-ice composition[10-11,5-6].
However,evidencesuggeststhatsaltsarepresentonEuropa’s
surface. GalileoNIMSspectracontainfeaturesthatarewellfit
byseveralhydratedsulfates[12]andsolidified,hydratedsulfu-
ricacid[13]. Motivatedbythedifficultyofexplainingdomes
withthermalconvectioninpureice,PappalardoandBarr[14]
suggested that the domes instead result from compositional
densitycontrastsinasaltyiceshell.
To date, no numerical simulations of convection in Eu-
ropa’s ice shell have included the effects of salinity. Here
wepresenttwo-dimensionalnumericalsimulationsofthermo-
chemical convection inEuropa’s icyshelltoevaluate theef-
fectsofsalinityonconvectionpatterns,icyshellstructure,and
surfacetopography. Figure1:Dynamictopography,composition,andtemperature
ModelandMethods. Weusedtheparticle-in-cell(PIC) forasimulationinadomain45kmwideand15kmdeep. In
finiteelementcodeellipsis3d[15]tosolvetheequationsgov- themiddlepanel,theblackmaterialrepresentsthelow-density
erningthermo-chemicalconvectioninEuropa’ssaltyiceshell icethatwasinitiallyinthebottom3kmofthedomain.
in2-Dcartesiangeometry. Thevelocityboundaryconditions
arereflectiveonthesidesandfree-sliponthetopandbottom. Results. Figure1displaystheresultsfromamodelina
Thetemperatureboundaryconditionatthebottomisfixed(270 domain45kmwideand15kmdeep,whichisinitializedwith
K),asrequiredbytheunderlyingocean,andthetopisheldat twodifferentlayers. Thetoplayer(from0–12kmdepth)hasa
95K.Forthesimulationspresentedhere,thelayerthickness densityof1100kg/m2 ,andthebottomlayer(from12–15km
is15km. Inthesepreliminarysimulations,notidalheatingis depth)hasadensityof1000kg/m2 .Theupperlayerrepresents
included. dense,salineice,whilethelow-densitymaterialinthebottom
Temperature-dependentNewtonianviscosityisimplemented. 3kmrepresentslow-salinityiceenvisionedtoresultfromloss
Thetemperature-dependentviscosityisrepresentedasfollows: ofsaltsduetopartialmeltingnearthebaseoftheiceshell[14].
The cutoff viscosity is (cid:0)(cid:2)(cid:1)(cid:4)(cid:3):9 Pa sec. The initial temperature
(cid:10)(cid:29)’
(cid:7)(cid:9)(cid:8)(cid:11)(cid:10)(cid:13)(cid:12)(cid:15)(cid:14)(cid:17)(cid:16)(cid:19)(cid:18)(cid:21)(cid:20)(cid:23)(cid:22)(cid:24)(cid:7) cutoff (cid:7)(cid:27)(cid:26)(cid:29)(cid:28)(cid:31)(cid:30)(cid:4) "!$#&% (cid:0)(cid:2)*,+.- (1) increaseslinearlyfromthetop(95K)tothebottom(270K)
(cid:25) (cid:10))(
Lunar and Planetary Science XXXVI (2005) 1465.pdf
with a small initial disturbance. In the simulation, the low- pleparameterizationforpartialmelting,andmeltdrainage,of
density material develops numerous small-scale ( few-km- icecontaininglow-eutectic-temperaturesalts. Partialmelting
(cid:0)
wide)convectiveinstabilitiesbeforeascendingenmasseinto is assumed to begin at temperatures of 220 K, leading to a
a 20 km-wide diapir, which splits in two as dense, salty ice reduction indensity ofup to 5%as temperatures riseto250
descends through itscenter (Fig.1). Uplifts250–300 m tall K or above. Uplifts 200 m in height and pits300–400 min
andpitsabout250–400mdeep,withdiametersofabout15km, depth,withdiametersof20km,areproducedonthethesur-
areformed. Thesefeaturesresultdirectlyfromthehorizontal face. Thesetopographiesarecomparabletotheobservedpits
buoyancycontrasts;thecombinedbuoyancyfromtemperature and domes on Europa. Ifthe compositional density contrast
andcompositionalvariationsfarexceedsthethermalbuoyancy is increased, then even taller domes and deeper pits can be
alone, which explains the large topography relative to that produced.
obtainedintheearlier,pure-icesimulations[5-6]. Eventually, Simulationsshowthattheviscositystructureplaysakey
the lower density material from the bottom mixes up with roleindeterminingwhethertopographycanoccur. Ina15-km-
thehigher density materialatthetop, andthestrong surface thickiceshellwith5–10%compositionaldensitycontrasts,the
topographydisappears. simulationscannotproducetheobservedpitsanddomesifthe
viscosityinthecoldregionexceeds(cid:0)(cid:2)(cid:1) (cid:3)(cid:2)(cid:1) Pasec.
ConclusionsandDiscussions:Thermo-chemicalconvec-
tioninEuropa’sicyshellcanproducepitsanddomescompa-
rabletotheobservationsunderappropriateconditions. Ifthe
compositional density variation is 5–10% and the maximum
viscosityislessthan(cid:0)1(cid:1) (cid:3)(cid:6)(cid:5) Pasec,thermo-chemicalconvection
ina15km-thickiceshellproducespitsandupliftswithtopo-
graphicamplitudesof200–400m. Butoursimulationsshow
thatconvectionunderthesameconditionscannotproducethe
observed pits and domes if the viscosity in the cold region
exceeds(cid:0)1(cid:1) (cid:3)(cid:2)(cid:1) Pasec. Conceivably,brittlebehaviorinthecold
near-surfaceicecouldallowthelithosphericdeformationnec-
essaryforupliftsandpitstoform[6].
FutureWork: Elasticitynearthesurfaceandtidalheat-
ing within the ice shell can strongly affect thermo-chemical
convection and its related features on Europa; these will be
consideredinthefuture. Althoughthepitsandupliftsinour
simulationsaretransientfeatures,itispossiblethatmelt-filled
fracturesmaycontinuetoinjectsaltintotheiceshellandthat
partial melting near the base of the shell (perhaps driven by
tidal heating) would produce a continuing supply of fresh,
low-density ice there. Our future models will include these
sourcesandsinkstodeterminewhetherpitsandupliftscanbe
continuallygeneratedonEuropa.
ThisworkissupportedbytheNASAPG&Gprogram.
References: [1]. Pappalardo, R.T.,etal. (1998)Nature
391, 365-368. [2]. Head, J.W. and R.T. Pappalardo (1999)
JGR104,27143-27155. [3]. Collins,G.C.,etal. (2000)JGR
105,1709-1716.[4]. Spaun,N.A.(2002)Ph.D.Thesis,Brown
University. [5]. Showman,A.P.andL.Han(2004a)JGR109,
E01010,doi:10.1029/2003E002103. [6]. Showman,A.P.and
L. Han (2004b) Lunar Planet. Sci. Conf. XXXV, [abstract
Figure2:Dynamictopography,composition,andtemperature
1466]. [7]. Greeley, R.,etal. (1998) Icarus135, 4-24. [8].
forasimulationinadomain45kmwideand15kmdeep. In
Riley,J.,etal.(2000)JGR105,22599-22615.[9]. Greenberg,
the middle panel, the black material represents the material
R.,etal. (2003) Icarus161, 102-126. [10]. Sotin,C.,etal.
thatwasinitiallyinthebottom3kmofthedomain.
(2002) GRL 29, 10.1029/2001GL013884. [11]. Tobie, G.,
et al. (2003) JGR 108, doi:10.1029/2003JE002099. [12].
McCord, T.B., et al. (1999) JGR 104(E5), 11,827-11,851.
Figure 2 gives the results from a model in a domain 45
[13]. Carlson,R.W.,etal. (1999)Science286, 97-99. [14].
kmwideand15kmdeep. Theinitialcondition issimilarto
Pappalardo, R.T. and A.C. Barr (2004a) GRL 31, L01701,
thatinFig.1exceptthatthebottomdensityisonly5%smaller
doi:10.1029/2003GL019202. [15]. Moresi,L.,etal. (2003)
thanthetopdensity. However,thissimulationincludesasim-
J.Comput. Phys. 184,476-497.