Icarus148, 226–265 (2000)

doi:10.1006/icar.2000.6471, available online at http://www.idealibrary.com on

Europa’s Crust and Ocean: Origin, Composition, and the Prospects for Life

Jeffrey S. Kargel

U.S. Geological Survey, 2255 North Gemini Drive, Flagstaff, Arizona 86001E-mail: jkargel@usgs.gov

Jonathan Z. Kaye

School of Oceanography, University of Washington, Box 357940, Seattle, Washington 98195

James W. Head, III

Department of Geological Sciences, Brown University, Box 1846, Providence, Rhode Island 02912

Giles M. Marion

Desert Research Institute, Earth and Ecosystem Sciences, 2215 Raggio Parkway, Reno, Nevada 89512

Roger Sassen

Geochemical and Environmental Research Group, Texas A&M University, 833 Graham Road, College Station, Texas 77845

James K. Crowley

U.S. Geological Survey, Reston, Virginia 20192

Olga Prieto Ballesteros

Departemento de Quimica Inorganica y Materiales, Facultad CC, Experimentales y Tecnicas, Universidad San Pablo-CEU,Urb. Monteprincipe km. 5,300, Boadilla del Monte, Madrid 28668, Spain

Steven A. Grant

U.S. Army Cold Regions Research and Engineering Laboratory, 72 Lyme Road, Hanover, New Hampshire 03755-1290

and

David L. Hogenboom

Department of Physics, Lafayette College, Easton, Pennsylvania 18042

Received January 7, 2000; revised May 1, 2000

We have considered a wide array of scenarios for Europa’s chem-ical evolution in an attempt to explain the presence of ice and hy-drated materials on its surface and to understand the physical andchemical nature of any ocean that may lie below. We postulate that,following formation of the jovian system, the europan evolutionarysequence has as its major links: (a) initial carbonaceous chondriterock, (b) global primordial aqueous differentiation and formationof an impure primordial hydrous crust, (c) brine evolution and in-tracrustal differentiation, (d) degassing of Europa’s mantle and gasventing, (e) hydrothermal processes, and (f) chemical surface al-teration. Our models were developed in the context of constraints

provided by Galileo imaging, near infrared reflectance spectroscopy,and gravity and magnetometer data. Low-temperature aqueous dif-ferentiation from a carbonaceous CI or CM chondrite precursor,without further chemical processing, would result in a crust/oceanenriched in magnesium sulfate and sodium sulfate, consistent withGalileo spectroscopy. Within the bounds of this simple model, awide range of possible layered structures may result; the final statedepends on the details of intracrustal differentiation. Devolatiliza-tion of the rocky mantle and hydrothermal brine reactions couldhave produced very different ocean/crust compositions, e.g., anocean/crust of sodium carbonate or sulfuric acid, or a crust con-taining abundant clathrate hydrates. Realistic chemical–physical

226

0019-1035/00 $35.00All rights of reproduction in any form reserved.

EUROPA’S

evolution scenarios differ greatly in detailed predictions, but theygenerally call for a highly impure and chemically layered crust.Some of these models could lead also to lateral chemical hetero-geneities by diapiric upwellings and/or cryovolcanism. We describesome plausible geological consequences of the physical–chemicalstructures predicted from these scenarios. These predicted conse-quences and observed aspects of Europa’s geology may serve asa basis for further analysis and discrimination among several al-ternative scenarios. Most chemical pathways could support viableecosystems based on analogy with the metabolic and physiologicalversatility of terrestrial microorganisms.

Key Words: Europa; ocean, Europa; cryovolcanism, Europa; spec-troscopy, Europa; Europa, prebiotic chemistry.

seu

n

a

y

lvu7sn

ivl

y

s

i

lyrdo

re-re-

i-floweachand

es ae of

de-s se-cean-

d

phicand

itingeen

eta-ere

teinand

paere)) if

1990,lidar-ight

ngtheno-ter-chybrit-uterandaker

r foron-

1. INTRODUCTION

Galileo radio-tracking/gravity data require that Europa hadifferentiated internal structure, including a dense core of mor metal sulfides, a rocky mantle, and a low-density icy cror ice-crusted ocean 80–170 km thick (Andersonet al. 1998).Analysis ofGalileoSSI images and other data offer compellievidence of interior ductile deformation in Europa, lateral lithspheric mobility measured in several tens of kilometers,local partial melting in geologically recent times (Beltonet al.1996; Carret al.1998; Greeleyet al.1998a,b; Greenberget al.1998, 1999; Spaunet al. 1998; Sullivanet al. 1998; Fagentset al. 2000; Fanaleet al. 1999; Headet al. 1999a; Head andPappalardo 1999; Pappalardoet al.1999, Prockteret al.2000).These results confirm suspicions raised earlier by scrutinVoyagerimages (Smithet al. 1979, Helfenstein and Parmentier 1980, Schenk and Seyfert 1980; Lucchitta and Soderb1982, Schenk and McKinnon 1989, Pappalardo and Sulli1996) and also support thermal modeling showing that sactivity is possible (Lewis 1971, 1972; Consolmagno 19Consolmagno and Lewis 1978; Cassenet al.1979, 1980; Squyreet al.1983; Ross and Schubert 1987; Ojakangas and Steve1989; Grassetet al.1996; Yoder and Sjogren 1996). Processinferred from geology include (1) local effusive and exploseruptions of icy liquids or slushes (e.g., Lucchitta and Soderb1982, Greeleyet al. 1998a, Headet al. 1998, Greenberget al.1999); (2) more widespread upwelling and shallow intrusionicy solids or liquids and consequent doming of the surface (ePappalardoet al. 1998); (3) ubiquitous fracturing of a thin iclithosphere due to tidal disruption (e.g., Helfenstein and Pmentier 1980, 1985, Lucchitta and Soderblom 1982, Geiset al.1998a, Greenberget al.1998, Hoppa and Tufts 1999); (4translation and extension of lithospheric blocks and upwellof material that fills in the voids thus created (e.g., Sullivanet al.1998, Tufts 1996, Tuftset al.2000, Prockteret al.1999b); and (5)complete local lithospheric disruption and formation of chaoterrain (e.g., Spaunet al. 1998, Williams and Greeley 1998Greenberget al.1999). Most visually compelling, chaos region

appear to consist of either (a) masses of icebergs, now frozeplace, that once floated and drifted in a frosty sea jammed w

OCEAN 227

atalst

go-nd

of-omanch5;

sonese

om

of.g.,

ar-ler

)ng

tic,s

pack ice (Greeleyet al. 1998a,b; Carret al. 1998; Spaunet al.1998; Greenberget al.1999) or (b) lithospheric material utterdisrupted by convective solid-state ice diapirism (Pappalaet al.1998, Spaunet al.1998).

Europa’s surface is geologically young—its mean cratertention age is around 10–100 Myr—and Europa probablymains active (Shoemaker 1996, Chapmanet al. 1998, Zahnleet al.1998, Pappalardoet al.1999). It is even possible that evdence of rapid ongoing resurfacing and high regional heatduring the era of spacecraft exploration is in hand, althoughtype of evidence is speculative or ambiguous (HelfensteinCook 1984, Domingue and Lane 1998, Geissleret al. 1998b,Spenceret al. 1999). In any case, Europa clearly possessgeologically young surface and complex history suggestivoceanic influences.

Geologic evolution, including a sequential stratigraphicvelopment of the lithosphere, has been documented; thiquence might be related to progressive freezing of an o(Headet al.1999b, Prockteret al.1999a), although alternate interpretations exist (Greenberget al. 1999). Recently describeevidence for an ancient mega-chaos region (Greenberget al.2000a), if correct, is consistent with the proposed stratigrasequence if an allowance is made for repetition of the cyclepreservation of remnants from a previous cycle (a very excprospect). A sequential evolution of lineaments has also bidentified (Geissleret al.1998b).

Regardless of the controversy for each geologic interprtion of Europa, it is agreed that Europa has an icy lithosphthat is mechanically isolated from a rocky mantle (Helfensand Parmentier 1980; Schenk and Seyfert 1980; LucchittaSoderblom 1982; Schenk and McKinnon 1989; Carret al.1998;Geissleret al. 1998a; Greeleyet al. 1998a,b; Pappalardoet al.1998, 1999; McKinnon 1999). The solid icy shell on Euroshould have a plastic, convecting lower part (asthenosphand a thermally conductive, brittle upper part (lithospherethe solid crust is thicker than 10–30 km (Squyreset al. 1983,Ojakangas and Stevenson 1989, Golombek and BanerdtGhail 1998, McKinnon 1999). If thinner than 10–30 km, the soicy shell would be entirely thermally conductive, except foreas that receive anomalous local heating from below that mtrigger solid-state diapirism. Possibilities allowed byGalileodata include (i) a thin, brittle, conductive icy shell overlyia deep global ocean, or (ii) almost complete solidity ofH2O-rich crustal layer with a warm, convecting icy asthesphere beneath a thin, brittle icy lithosphere. Something inmediate is also possible, such as (iii) a thin global or patliquid ocean beneath a thick, convecting ice crust and atle lithosphere, or (iv) a deep liquid ocean between an othick convecting ice crust and surface brittle lithosphere,an inner layered sequence below the ocean floor. Shoem(1996) estimated Europa’s crater retention age as 30 Mycraters>10 km, and found that the cratering record is c

n inithsistent with a 10-km-thick ice crust overlying a liquidocean.

lpo

gs,

en

e

hmg

fm

sg

-

t

oe

lna

nod

a

ott

MSsseserva-ag-“salt

ri-

rvedintedate-ue tolab

lfateshlyese

ear-umuire

bleials,ui-sere

an-) to, any

m),

n-

ofllede

usg

, sul-buticyy-edl–

228 KARGEL

As modeled by McKinnon (1999) and others, it is actuavery difficult to freeze a Europan ocean entirely due simto the fact that a stagnant, nonconvecting icy shell is mtidally dissipative than a convecting one. Even without conseration of any tidal heating, Grassetet al. (1996) modeled thecontinuous existence of a liquid ocean since Europa’s oriVery high levels of tidal heating and thinner ice shells are poble in some models, according to Yoder and Sjogren (1996)they also highlighted the difficulty of ruling out an icy shell ssoft that it remains efficiently convective and solid to the basthe crust. However, a salty Europan ocean (Kargel 1991, Faet al.1998) would reduce the solidus temperature to less thanlikely convective adiabatic temperature (Kargel 1996). Hena present-day Europan ocean appears to be a robust theorresult supported also by geologic observations.

Electrical induction in a hypothetical Europan ocean was fipredicted by Colburn and Reynolds (1985). Electrical curreflowing in a salty ocean should induce a magnetic field twould possess certain measureable diagnostic attributes ofnitude, orientation, structure, and temporal oscillation (Karand Consolmagno 1996). The discovery by theGalileomagne-tometer team of what appears to be an induced magnetic fieEuropa with the predicted properties suggests that a largetion of the H2O layer is a briny, liquid ocean around 100 kdeep (Kivelsonet al. 1997, Khuranaet al. 1998). If the oceanlacked appreciable solutes, or if it was thinner than a few tenkilometers or frozen completely, it would not conduct enoucurrent to explain the observed magnetic signature. Uncertain the total thickness of the H2O-rich layer—gravity data permit 80 to 170 km (Andersonet al.1998)—still allows for eithera thick (up to∼70 km) or a thin (few kilometers) frozen crusCloser scrutiny of the magnetic field data and possible generamechanisms has supported the brine ocean induction mechawithout revealing strongly compelling alternatives (Kivelset al. 1999). However, a comparable induced field has bdiscovered at Callisto (Stevenson 1998, Kivelsonet al. 1999),where there is scant geologic evidence of an ocean. Calcould have an ocean, especially if ammonia is present (Pre1996), but the icy lithosphere would have to be very stablecompetent. Possible alternative explanations for Callisto’s mnetic signature—that the electrically conducting layer mightsome other substance, for instance, graphite—have not beeveloped in detail. In sum, Europa’s magnetic signature suppthe existence of a brine ocean, but this support is tempereCallisto’s similar magnetic signature.

Hydroxylated phyllosilicate minerals that are so abundon Callisto and Ganymede (McCordet al. 1998b) exhibit ab-sorption features not seen on Europa (McCordet al. 1998a).However, nonice hydrated substances are regionally abunon Europa.GalileoNIMS reflectance spectra of many regionsEuropa exhibit highly distorted water absorptions indicatingpresence of bound water—and little or no ice—wherever

deep interior has been exposed by deep impact or riftingthe lithosphere or by extrusion of liquids or warm icy mass

ET AL.

lylyre

id-

in.si-butoofale

anyce,tical

rstntsatag-el

ld atrac-

ofh

inty

.tivenismnen

istoticend

ag-be

de-rtsby

nt

dantf

hehe

(McCordet al. 1998a, 1999; Carlsonet al. 1999; Fanaleet al.1999). Consistent with earlier results from telescopic data, NIdata for other europan terrains reveal nearly pure ice. Two claof related materials have been suggested to explain obstions of Europa’s prevalent nonice material: (1) hydrated mnesium and sodium sulfates and/or sodium carbonates (thehypothesis,” McCordet al.1998a, 1999a; Prietoet al.1999), or(2) hydrated sulfuric acid with traces of radiolytic sulfur imputies (the “acid hypothesis,” Carlsonet al.1999). Dalton and Clark(1999) have emphasized ambiguities in attributing the obsedistorted water bands to specific hydrates, and they have poout slight spectral mismatches between the Europa nonice mrial and proposed substances. These mismatches likely are ddifferences between the conditions or grain size pertaining todata and those of Europa. Each of the proposed hydrated suhas a strong theoretical backing from cosmochemistry, is higwater soluble, and has an electrolytic nature; therefore, thsolutes can explain the electrically conductive nature of a nsurface liquid layer required by magnetic field data. Sodicarbonates are also reasonable for Europa but would reqa complex chemical evolution. Below we explore (a) possiphysical–chemical scenarios for the origin of these mater(b) brine evolution from the viewpoint of chemical phase eqlibria and chemical compatibility, and (c) relationships of thematerials with the physical geology of Europa’s icy lithospheand ocean and possible biochemistry.

We adopt a modified version of the chemical and mechical/structural terminology of Pappalardo and Head (1999describe the layering of Europa. As may be deduced abovecentral segregation of metallic or sulfide phases is termed acore.A rocky layer surrounding the core is termed themantle; if themantle is further differentiated (e.g., due to silicate magmatisthe ultramafic residue of partial melting is still termed theman-tle and a differentiated outer rocky layer is termed thesilicatesubcrust, since it will underlie a volatile crust. The outer layer eriched in H2O and other volatiles is termed thecrust. Followingformation of a primordial hydrous crust but before separationa core, the rock-rich material beneath the crust is simply cathe rocky interior. The volatile crust may further differentiatinto (i) a solid upper crust, which we refer to as theicy shell;(ii) a liquid interior layer, which we refer to either as theoceanor thehydrosphere(the latter term may include the ocean plany liquid inclusions or liquid-filled pods or conduits intrudininto the icy shell or leading onto the surface); and (iii) avolatilesubcrust, which encompasses any layered sequence of saltsfur, or other volatile-bearing solids deposited on the seafloorresting above the silicate mantle or silicate subcrust. Theshell could further differentiate into various compositional laers. If a surficial layer of ice-free salty material exists, it is termthe evaporite layer. The icy shell is divided on a mechanicastructural basis into a cold, brittle top layer termed thelitho-sphereand a warm, solid, convecting interior layer termed theas-

of

esthenosphere. It is possible that no ocean exists (Pappalardoet al.1999) or that convection/diapirism is unimportant in the icy shell

’S

d

pnHe

v;e

eun

rem

nn(iu

erm

cn

udui

ts

3pe

p

d onis

tileof

used,on-histhe

altsre,Is

iteset-M

s ofns ifentes,slyitesndandtionf CI

usthe

90,8,

an-al-maytiono-

sidetedelyr-neb-

, con-hasendwa-n byensi-

EUROPA

(Greenberget al. 1999, 2000). It seems to us that the evidenstrongly supports the ocean/asthenosphere/lithosphere mo

2. FRAMEWORK FOR EUROPA’S CHEMICAL EVOLUTION

What relations exist between the landforms and geologiccesses of Europa, the structure of its crust, the compositioany ocean that may exist, and the makeup of its surface?is Europa’s present state related to initial conditions? We sereasonable chemical evolutionary framework, compatible wthe latestGalileo data, upon which possible answers to thequestions may be based.

The Mg–Na–sulfate and/or Na–carbonate minerals beliepresent on Europa (Kargel 1991; McCordet al. 1998a, 1999aFanaleet al. 1998) also are widespread in terrestrial contintal evaporites (Eugster and Jones 1979, Eugster 1980),gesting that evaporation and concentration of aqueous pritates occurred on Europa. The alternative, hydrated sulfacid (Carlsonet al. 1999), raises the possibility that differeaqueous equilibria were at work, but still evaporative procesand sulfate chemistry may have been partly responsible.ropa’s surface, like the salts of dry lakes, represents an exttype of water-based chemical fractionation. The surfaceterials are just the last link in a complex chain of processWe postulate that, following formation of the jovian systeand accretion of the major satellites, the europan chain haits major links: (a) initial conditions represented by carboceous chondrite rock, (b) global primordial aqueous differeation and formation of an impure primordial hydrous crust,brine evolution and intracrustal differentiation, (d) degassof Europa’s mantle and gas venting into the icy ocean/cr(e) hydrothermal reactions of brine with the rocky mantle arocky subcrust, and (f) alteration in the surface environmThese are not strictly sequential stages—some are concuand others ongoing. These events are further described andeled in Section 3.

This conceptual framework is illustrated in Fig. 1, whishows identical chondritic starting material but different estates that depend on details of the chemical pathway. Noteeach scenario starts with identical low-temperature incongrmelting of hydrated salts in a CI or CM chondrite (the moof Kargel 1991) and ends up with precipitation of sulfurocompounds on the seafloor; the scenarios differ in compositof the ocean, crust, and seafloor deposits. The process imcations of these and several other possible crust composiare summarized in a different form in Table I. The compotions and some properties of selected liquids that could aby these processes are summarized in Table II. In Sectionconsider some of these processes in more detail. For comson of our new results and those of Kargel (1991), the readreferred to the recent work of Spaun and Head (2000), who hmodeled the fractional crystallization of a hypothetical euro

ocean and formation of a layered crustal sequence in the sysMgSO4–H2O.

OCEAN 229

ceel.

ro-of

owk aithse

ed

n-sug-cip-rictsesEu-mea-

es.ms asa-ti-c)ngst,

ndnt.rentod-

hd

thatentels

onspli-

ionsi-

riseweari-r isavean

2.1. Initial Conditions

We assume that Europa started as hydrated rock formethe warm side of the Jupiter system’s “snow line.” Europaassumed initially to have had solar elemental ratios of involaand slightly volatile rock-forming elements and an absencevolatile ices. Following Kargel (1991), we take carbonaceoCI chondrites (the most volatile rich, most aqueously alterand most compositionally primitive class of carbonaceous chdrite) as a reasonable analogue to Europa’s initial material. Tis merely a convenient assumption that provides (a) roughlycorrect amount of water needed for a volatile crust, (b) sthought to exist on Europa’s surface, (c) sulfides for a coand (d) approximately the correct bulk density. Actually, Cprovide a bit too much water, as we shall see; CM chondr(another volatile-rich class of chondrite) may be slightly bter in this regard but are a bit too dry. Whereas CI and Cchondrites closely bracket the ideal composition, other typecarbonaceous chondrites might also satisfy these conditiosome additional water was provided in the form of ice and thaqueous alteration was allowed (e.g., CV and CO chondriboth of which are much less volatile rich and less aqueoualtered than CIs and CMs). Ordinary and enstatite chondr(which contain abundant free metal and very little water acarbonaceous matter) lack sufficient water, are too reduced,would yield too large a core, although heterogeneous accreof some such material with late accretion of some amount ochondrites or comets might work.

The use of hydrated rock similar to CI or CM carbonaceochondrites is a long-standing assumption by modelers ofGalilean satellites (Ransfordet al. 1981, Squyreset al. 1983,Crawford and Stevenson 1988, Golombek and Banerdt 19Kargel 1991, Prentice 1996, Mueller and McKinnon 198Hogenboomet al. 1995, Fanaleet al. 1998, McKinnon 1999).Some who have modeled these objects with initial ice plushydrous rock (e.g., ordinary chondrites) did so to simplify cculations but acknowledged that carbonaceous chondritesbe a better analogue (Lupo and Lewis 1979). This assumpis rooted in the concept from meteoritics and nebula thermchemistry that rocky nebula condensates just on the warmof the “snow line” lacked volatile ices but possessed a hydramineralogy and solar proportions of refractory and moderatvolatile rock-forming elements. There were important diffeences in nebula processes in the jovian nebula and the solarula (Cameron and Pollack 1976, Pollack 1985, Stevensonet al.1986, Prinn and Fegley 1989). Gas pressures, temperaturesdensation, and accretion time scales were different. Gas-pequilibrium (or lack of) depends sensitively on nebula P–T anebula lifetime, so assemblages of ices more volatile thanter ice depend on where condensation took place. As showLewis (1971, 1972), condensation temperatures depend stively on nebula pressures, but condensationsequencesfor solar

temcomposition gases are fairly insensitive to nebula pressure. Thereis relatively little difference in the sequence of condensation

230

KARGEL ET AL.

FIG

.1.

Sch

emat

icch

emic

alev

olut

ion

ofth

ecr

ust/

ocea

nof

ach

ondr

itic

Eur

opa.

(a)

For

mat

ion

ofa

crus

t/oc

ean

enric

hed

inm

agne

sium

and

sodi

umsu

lfate

s.(b

)F

orm

atio

nof

chlo

ride

ocea

nby

exte

nsiv

ecr

ysta

lliza

tion

ofth

em

agne

sium

-sul

fate

-ric

hbr

ine

repr

esen

ted

in(a

).(c

)M

odifi

edpa

thw

ayth

atin

clud

esad

ditio

nof

CO 2,

whi

chm

ight

resu

ltin

form

atio

nof

acr

ust/

ocea

nen

riche

din

sodi

umca

rbon

ate

and

sulfa

te.

(d)

Mod

ified

path

way

that

incl

udes

addi

tion

ofS

O2

orS

O 3an

dre

sults

info

rmat

ion

ofan

ocea

n/c

rust

enric

hed

insu

lfuric

acid

and

cert

ain

salts

,suc

has

mag

nesi

umsu

lfate

.(e)

Mod

ified

path

way

that

incl

udes

addi

tion

ofC

H4,w

hich

resu

ltsin

acl

athr

ate-

rich,

salty

crus

t.

e

). Se-tes

s in

e

e

-

za-

e,

ents

EUROPA’S OCEAN 231

TABLE IAlternative Surface Compositions and Implications for Europa’s Hydrospheric Evolution

Thermal state of ocean(if it exists) and

Major components of surface icy crust Allowed by NIMS data? Implications for evolution of hydrosphere

(a) H2O, MgSO4, Na2SO4 Warm ocean Yes Simple evolution by incongruent melting of chondritThin icy crust salts—Fig. 3, models A5 and B; Fig. 1a. Minimal

hydrothermal interaction with basalt subcrust.

(b) H2O, Na2SO4 Warm ocean Yes Incongruent melting of chondrite salts as in case (aThin icy crust questration of Mg in carbonates or Mg-phyllosilica

and partial reduction of sulfate by hydrothermalcirculation in basaltic subcrust.

(c) H2O, Na2SO4, Na2CO3 Warm ocean Yes Melting of chondrite salts and sequestration of Mg aThin icy crust case (b). Addition of CO2 by pyrolysis of mantle

carbonates and organics—Fig. 1c.

(d) H2O, Na2CO3 Warm ocean Yes Similar to case (c), but sequestration of Na by albit-Thin icy crust ization of silicate subcrust due to high-T hydro-

thermal reaction or fractional crystallizationof Na2SO4.

(e) H2O, NaCl Cold ocean No? Sulfate reduction and sequestration of Mg as in casThick icy crust (b), and fractional crystallization of Na2SO4

and Na2CO3.

(f ) H2O, NaCl, MgCl2 Cold ocean No? Extensive fractional crystallization. Minimal hydro-Thick icy crust thermal interaction with basalt subcrust—Fig. 4.

(g) H2O, MgSO4, Na2SO4, H2SO4 Warm ocean Yes Similar to case (a) but large additions of S gases duThin icy crust to breakdown of sulfates in mantle or lower crust.

Minimal hydrothermal reaction with basalt.

(h) H2O, H2SO4, NaCl, MgCl2 Cold ocean Yes? Similar to case (g) but extensive fractionalThick icy crust crystallization of Mg–Na-sulfates.

(i) H2O, Na2SO4, H2SO4 Warm ocean Yes Similar to case (g) but sequestration of Mg in carboThin icy crust nates or Mg-phyllosilicates due to low-T hydro-

thermal circulation (little sulfate reduction).

( j) H2O, H2SO4, NaCl Cold ocean Yes? Similar to case (h), but extensive fractional crystalliThick icy crust tion of Mg salts, or sequestration of Mg by hydro-

thermal processes or carbonate precipitation.

(k) H2O No constraint Allowed regionally, Efficient fractional crystallization and floatation of icon ocean but not globally or diapiric instabilities of ice in icy shell

with complete ice resurfacing.

(l) H O, CH or organic Depends on salts Depends on salts Deep mantle pyrolysis of carbonaceous compon

2 4

tholins,+ or− salts of mantle or reduction of carbonates.

tidrooC

r

ledcar-saltCMiticas-ili-ursig-areM

and expected mineralogy down to the point of ice condensaTherefore, we conclude that although carbonaceous chonmeteorites are solar nebula materials, basically identicalshould have been formed in the jovian nebula, and that folling aqueous alteration it should contain abundant Mg–Na–sulfate-rich soluble material.

Closed-system,in situ aqueous alteration at low to modeate temperatures of high-temperature nebula condensatesing much free metal produces the mineral assemblages ofbonaceous chondrite meteorites (DuFresne and Anders 1Bostrom and Fredrikkson 1964, Richardson 1978, Fredriks

and Kerridge 1988, Anders and Grevesse 1989, Zolenskyet al.1989, Burgesset al. 1991, Brearley and Jones 1998). Zoloto

on.riteckw-a–

-lack-car-962,son

and Shock (2000a) have confirmed this result; they modethe aqueous alteration assemblages resulting from an initialbonaceous CV chondrite (nearly anhydrous) and obtainedassemblages somewhat similar to those occurring in CI andchondrites. Experimental hydrothermal alterations of chondrmaterial have reproduced part of the CI and CM chondritesemblages, including formation of carbonates and phylloscates (Scottet al.2000). However, to make the reactions occon laboratory time scales, the experimentalists found thatnificant reaction occurred at temperatures much higher thanthought to have been involved in the alteration of CI and C

vchondrites; sulfate formation was not reported but probably per-tains to the particular oxidation state of the experiments. The

system

232 KARGEL ET AL.

TABLE IIDensities, Compositions, and Solid–Liquid Equilibria of Some Key Liquids Likely on Europa

Derived by melting The frozen solid yieldsFreezing/melting pt. Density at 1 atm: or incongruent Mineralogy what liquid upon

(at 1 atm solidus) solid (s) liquid (l) Composition melting of this when frozen incongruentSubstance (K) (kg m−3) (wt.%) assemblage (wt.%) remelting?

Water 273.15 917 (s) 100.0 H2O ice ice 100.0 H2O1000 (l)

Eutectic E 268.2 1144 (s) 81.0 H2O ice 48.9 ice Liquid Ein system 1208 (l) 16.0 MS MS12 44.7 MS12H2O–MS–NS 2.8 NS mir 6.4 mir

Peritectic P1 275 1235 (s) 75.5 H2O MS12 33.6 ice Liquid Ein system 1278 (l) 20.5 MS eps 57.3 MS12H2O–MS–NS 4.0 NS mir 9.1 mir

Peritectic P2 319 1444 (s) 65.4 H2O eps 6.0 ice Liquid Ein system 1361 (l) 29.4 MS hex 82.2 MS12H2O–MS–NS 5.2 NS blo 11.8 mir

Peritectic P3 334 1517 (s) 61.9 H2O hex 78.0 MS12 Liquid P1in system ∼1465 (l) 33.1 MS blo 10.6 epsH2O–MS–NS 5.0 NS MSNS2.5 11.3 mir

Peritectic P4 337 1530.1 (s) 60.7 H2O hex 68.6 MS12 Liquid P1in system ∼1485 (l) 34.0 MS kie 19.4 epsH2O–MS–NS 5.3 NS MSNS2.5 12.0 mir

Peritectic P5 344 1869.2 (s) 64.6 H2O blo 17.6 MS12 Liquid P1in system ∼1380 (l) 15.5 MS MS·3NS 35.3 MS7H2O–MS–NS 19.9 NS MSNS2.5 47.1 blo

Eutectic 211 1259 (l) 65.1 H2O HS6 73.4 HS6 Eutecticin system ∼1245 (s) 34.9 HS ice 26.6 iceH2O–HS

Note. The liquid and solid mixtures indicated by P1, P2, P3, P4, and P5 refer to the series of peritectics known for the ternary system H2O–MgSO4–Na2SO4

(International Critical Tables1928, Kargel 1991); note that these are not exactly the same peritectics as designated by Kargel (1991) for the binaryMgSO4–H2O. “E” represents the ternary eutectic in this same system (Kargel 1991). Chemical components are designated in short-hand; i.e., water is H2O, MgSO4

is MS, Na2SO4 is NS, and H2SO4 is HS. Minerals are epsomite, eps; hexahydrite, hex; kieserite, kie; mirabilite, mir; thenardite, the; bloedite, blo; MgSO4·12H2O,

MS12; MgSO4·Na2SO4·2.5H2O, MSNS2.5; and MgSO4·3Na2SO4, MS·3NS. The densities (in kg m−3) of solid phases are MS12, 1500; eps, 1677; hex, 1745;kie, 2571; blo, 2274; mir, 1490; and the, 2680. Other solid phase densities are not known and were guessed on the basis of data in Hogenboomet al. (1995).

pnltob

.9

ws

tu

avese-

gen-. CItem-91,Mgffec-lu-mallh astioni-

water needed for the aqueous alteration of actual chondriteent bodies could have been provided either by dehydratiohydrous silicates condensed from the solar nebula or by meof a small amount of ice accreted with an initial anhydrous rassemblage. Our specific model starts with the salt assemknown from CI chondrites—closely approximated as 8.8%mass epsomite, 4.9% by mass bloedite, 1.5% anhydrite, 2dolomite, and 0.7% magnesite (Bostrom and Fredriksson 1Fredriksson and Kerridge 1988, Burgesset al. 1991, Kargel1991). The salt assemblages of CM chondrites are somesimilar. What is most crucially different between CIs and CMthe amount of water—about 6 wt.% in CMs and up to 18 wt.%CIs. Our model water content, 18%, is from the H content lisby Anders and Grevesse (1989) for the CI chondrite, OrgeAlthough details of any model will depend on the initial salt asemblage and the amount of water that is available, the more

portant results of any such models are fairly independent of thdetails.

ar-of

ingcklageby4%64,

hatisinedil.

s-im-

2.2. Primordial Aqueous Differentiation

Progressive heating to temperatures of 268–380 K would hthermally dissociated hydrated salts in a definable stagedquence known from the relevant phase equilibria. Fluids soerated would then have formed a primordial hydrous crustand CM chondrite brines are well represented in the sysH2O–MgSO4–Na2SO4–97% of the highly water-soluble fraction of CI chondrites is represented by this system (Kargel 19Fredriksson and Kerridge 1988). Calcium sulfate and Ca–carbonates, although abundant in CI chondrites, are not etively leached during this initial low-temperature stage of evotion because of the low solubility of these phases and the swater : rock ratios involved. Some other components, sucNi– and Mn–sulfates, can be treated to a first approximaas a part of the MgSO4 component due to chemical similar

eseties. Chlorides would have been present, and although they arehighly soluble, they are not nearly as abundant as sulfates in

233

84

03

7

drite leachatewhich would

EUROPA’S OCEAN

TABLE IIIC-Chondrite Leach Compositions Compared to Seawater (Mass Fractions)

Major cations (sum= 1.00) Major anions (sum= 1.00)

Mg Na+K Ca SO4 Cl CO3

CI chondrites, entire salt 0.438 0.107 0.455 0.603 0.013 0.3assemblagea

CI chondrites, entire salt 0.381 0.208 0.411 0.979 0.021 0assemblagea with carbonatessubtracted

CI chondrites, salt assemblagea 0.653 0.351 0.005 0.968 0.032 0with carbonates and CaSO4

subtractedb

CM chondrite leach obtained 0.274 0.433 0.293 0.965 0.035 0by boilingc

CM chondrite leachc with 0.386 0.609 0.005 0.952 0.048 0CaSO4 subtracted

“La Mancha” playa 0.386 0.605 0.009 0.806 0.191 0.0evaporitic brined

Earth’s seawatere 0.100 0.868 0.032 0.125 0.868 0.00

a From Kargel (1991); small amounts of Ni and Mn included with Mg.b All CaSO4 subtracted except for enough to saturate the solution.c Fanaleet al. (1998).

d “La Mancha” brine from de la Penaet al. (1982).

rlaf

yp

t

d

c

o

o

ndvem-ur-tionoom

c-tict ofuldningcedd ar-r by

ouldionolidlt–lly

e Seawater composition from Krauskopf (1979).

these aqueous solutions due to the low cosmic Cl : S ratiolow chloride : sulfate ratios of carbonaceous chondrites (Ka1991, Fanaleet al.1998). The aqueous phases generated bytemperature dissociation of salts would have been hypersand the solutes dominated by magnesium and sodium sulcorresponding to the series of peritectics in this ternary sys(International Critical Tables1928, Kargel 1991). A summarof the eutectic and peritectic compositions in this system isvided in Table II. The salt assemblages in carbonaceous cdrites (Table III) would have yielded the initial liquids indicatein Table II; CIs yield a liquid corresponding to the peritecP2 that is approximately 65.4% H2O, 29.4% MgSO4, and 5.2%Na2SO4 (there would also have been just over 1% of chlorinot listed in Table II). This is the initial brine composition usedall of our models. A rough validation of the salt assemblageculated by Kargel (1991) is provided in the CM chondrite leaexperiments done by Fanaleet al.(1998, 1999). By boiling CM2Murchison they obtained a leachate resembling that used inmodeling here, but this leachate even more closely approximthe ternary peritectic P5 in the system H2O–MgSO4–Na2SO4 (ifCaSO4 is subtracted from the leach), which is not surprising csidering the temperature of the experiment. However, theleach differs most dramatically in that it also contains abundCaSO4, which is difficult to explain in terms of equilibrium dissolution, but it might be due to use of a large water : rock in thleach experiments. The high opacity reported for the CM ch

probably was due partly to precipitation of CaS4,have been grossly oversaturated in that mixtur

andgelow-lineatestem

ro-hon-dic

e,inal-ch

theates

n-CMant-se

on-

2.3. Brine and Crust Evolution

Beginning with the first exhalations of aqueous fluids aprobably still continuing, Europa’s impure crust would haundergone further aqueous differentiation into multiple copositional layers. The development of density stratification ding cryomagmatism depends on details of the melt generaprocesses and eruption mechanisms (Kargel 1991, Hogenbet al. 1995, Wilsonet al. 1997, Headet al. 1998, Kadelet al.1998, Wilson and Head 1999, Fagentset al.2000). Equilibriumcrystallization—or quenching with glass formation—might ocur in some rapidly cooled cryovolcanic flows and cryoclasdeposits, resulting in frozen layers that would tend to consiseutectic or peritectic compositions. Layers of pure salts coform by rapid water boil-off of brine lavas, by slow sublimatioof ice on the surface, or by fractional crystallization and settlof salt crystals. Layers of nearly pure ice could be produby recondensation of water vapor at the poles, in shadoweeas, or on cold surfaces adjacent to active brine vents, oflotation of ice in brine magma bodies (Hogenboomet al.1995,Wilson and Head 1999). Salt–water systems on Europa shhave a strong tendency to evolve with fractional crystallizatdue to the combination of large density differences between sand liquid phases and the low viscosity of the liquid in sawater systems (Hogenboomet al. 1995). Idealized fractionacrystallization might occur in most situations involving slow

Oe.cooled, deep-seated magma bodies (brine plutons) or freezingof an ocean on Europa, resulting in massive sills, lenses, and

L

aitraefyf

aeb

aau

sdn

eo

p

iait

c

igll

am

lde

i

fidepre-

t thexi-

untson-) thatsen-the

ldhasof

blelfateandess

int innt

hene al-pos-ticticur-sd

s ofme-

tionow

has

ag-ouslethethatccurto-

234 KARGE

dikes and thick crustal layers of water ice and segregatedmasses. Interlayering of salts and ice would be gravitationunstable in the crust. Diapiric instabilities might arise, withbodies ascending and salt bodies descending. At one exall the ice within Europa’s crust eventually could segregateform a pure-ice upper crust, leaving the salts confined to delayers. However, this extreme model does not seem to satisNIMS data, which indicate concentrations of salts on the sur(McCordet al.1998a, 1999a). Various scenarios of brine evotion and crustal differentiation are developed below. NIMS dwhich seem to require an abundance of some type of sulfatpossibly sodium carbonate), tend to rule out models wherehigh degree of crystallization produces a sulfate-depleted cride ocean (e.g., model ‘b’ in Fig. 1).

2.4. High-Temperature Degassing of Europa’s Mantleand Fluid Reaction in the Hydrosphere

As Europa’s deep mantle heated further, hydrous silicwould have undergone progressive metamorphic dehydrbetween about 500 and 1100 K (Winkler 1979). Aqueous flphases would have been generated at each stage. Thesewould have migrated toward the surface, further leaching thecate interior and carrying additional dissolved salts, thus adto an already impure crust. Over a similar temperature raother volatile-bearing phases in Europa’s mantle would havecomposed, producing additional fluid constituents. Thvolatile-bearing phases would likely include Ca–Mg carbates, Ca sulfate, organic compounds, and sulfides, whichfairly insoluble in water at low temperatures; hence, theymarily remained in the mantle during low-temperature aquedifferentiation. The fluids so generated would have ventedand been trapped by the ice-capped hydrosphere. This situis different from the Earth, where gases are vented directlythe atmosphere or into the ocean, which is in communicawith the atmosphere. Once in the europan ocean and icy sgases such as SO2, CO2, and CH4 are then available to reaand form other types of brines and solid phases not predictethe basic low-temperature chondrite evolution model descrabove and modeled below. Some implications of degassinocean and crust composition are briefly explored below, incing possible formation of a sulfuric acid ocean. Possibly althese gases (and others besides) are added to Europa’salthough in this paper we consider only some possible chemeffects due to separate addition of SO2, CO2, and CH4.

With further heating to temperatures just below the silicsolidus, sulfides would have melted and separated from thetle, resulting in core formation. As the deep interior of Eurobecame hotter yet, partial melting of silicates may have yiebasaltic or komatiitic liquids, which then would have forma basaltic or ultramafic silicate subcrust either composingseafloor or underlying a volatile subcrust. The emphasis ofwork is on the petrological fate of Europa’s impure crust, whwe model in detail. We make a simple accounting of the s

aration of Europa’s core and do not investigate silicate diff

ET AL.

saltlly

ceemendperthe

acelu-ta,(ory a

hlo-

testionid

fluidsili-ingge,de-sen-areri-

ousntotion

ntoionhell,td bybedfor

ud-ofcrust,ical

tean-

paeddthethischep-

entiation at all, other than the removal of aqueous and sulphases. Note that our choice of a carbonaceous chondritecursor means that little or no free metal was available, so thacore would be made of pure metallic sulfides, which are appromated here as FeS. Although crude, this approximation accofor most of the sulfide component present in CI and CM chdrites (Tomeoka and Buseck 1985, Brearley and Jones 1998would exist following dehydration. According to Prior’s Rulethe rocky mantle and any silicate melts would be extremelyriched in FeO component (hence, significantly denser thanterrestrial planets’ mantles for equivalent pressures).

2.5. Hydrothermal Processes in the Seafloor Environment

Hydrothermal circulation of brine in a silicate subcrust coucause chemical reequilibration of solutes. Earth’s seawaterevolved a chloride-rich composition partly due to reactionsulfates with crustal ferroan silicates to form nearly insolumetal sulfides; magnesium from a dissolved magnesium sucomponent reacts with the rock to form serpentine, chlorite,other phyllosilicates (Von Damm 1995). A comparable proccould occur on Europa under the right conditions, resultingdepletion of the brine in magnesium sulfate and enrichmensodium chloride. A thick layer of seafloor salts might circumveor modify any such reactions between silicates and brines.

2.6. Surface Alteration

After all the processes above have had their influence, tmaterial erupted onto the surface or otherwise exposed will btered by the space environment. Surface alteration includessible thermal or radiation-induced dehydration and radiolyprocessing. Charged-particle implantation and UV radiolychemistry are likely to have major effects on salt stability and sface composition (e.g., Carlsonet al. 1999). Surface processewould not be too effective in altering interior chemistry anstructure, and they would not explain geologic associationnonice material, unless alteration products are buried and sohow recycled deep into the crust. However, surface alteradoes affect our view of the surface, which is our best windinto the interior.

3. CHEMICAL PATHWAYS AND MODELS

3.1. Chondrite Evolution Models

The salt interpretation of nonice materials on Europaa short list of suspected components (McCordet al. 1998a,1999a,b). Among them are the types of salts—hydrated mnesium and sodium sulfates—predicted from simple aquedifferentiation of a CM or CI chondrite (Kargel 1991; Fanaet al.1998, 1999). The short list of candidate materials hasunpredicted addition of hydrated sodium carbonate. We notesodium carbonate and magnesium sulfate are unlikely to otogether, since they are highly reactive and do not occur

er-gether on Earth (Eugster and Hardie 1978), although both occur

’S

a

r

e

o

d

oa

nnv

rMd

i

ch

o

g

i

icalom-

c-of alaints to-rust

areriticallyly be

ouselt-aven ofse-

opa.tureof

cean.lveoweduteda

ntedonalaryb-blelay-de-

rent.lib-theel-morethatt as-t for

zeds thelita-

adetingcar-argelm-therong

EUROPA

on Earth in separate evaporite deposits. Sodium carbonatesodium sulfate minerals occur together, as do magnesium suand sodium sulfate (Eugster and Hardie 1978). Other proceno doubt have affected the composition of Europa, but it is woconsidering the chondrite evolution models in some detailstarting point. According to a widely held opinion among plantary scientists, Europa is probably unique among icy satellitethat ice was not initially present or at least not abundant. Eurmay have started as a purely rocky object made of hydroussimilar to carbonaceous chondrite meteorites (FredrikssonKerridge 1988; Kargel 1991; Fanaleet al.1998, 1999). The suiteof models considered here implies either that Europa accrcarbonaceous chondrite material or that some other chondmaterial was accreted but soon underwent aqueous chemthat resulted in a mineral assemblage similar to that of carbceous chondrites—specifically CI or CM chondrites.

Some candidate Europan salts suggested both by chonevolution models (Kargel 1991; Fanaleet al. 1998, 1999) andby the interpretation of NIMS spectra (McCordet al. 1998a,1999a,b) include epsomite (MgSO4 · 7H2O), bloedite(Na2Mg(SO4)2 · 4H2O), and mirabilite, Na2SO4 · 10H2O. In ad-dition, the NIMS data admit possible sodium carbonates, enatron (Na2CO3 · 10H2O). These are common terrestrial evapite minerals. The unusual compositions of chondritic leach(compared to terrestrial brines) are highlighted in Fig. 2. Ctain rare terrestrial closed-basin brines, such as that of a pof the La Mancha region of Spain (de la Penaet al. 1982), arefairly close approximations to a CM or CI chondrite brine. Itextremely rare to have Cl : SO4 ratios on Earth as low as thosexpected from chondrite brines. The other main distinctiothe extremely magnesium-rich composition of chondrite bricompared to more sodic terrestrial brines, but there are serare examples here on Earth that come close to the hyposized europan Mg : Na ratio. The “La Mancha” playa evapodeposits from Spain are dominated by the same types ofNa–sulfate minerals (de la Penaet al. 1982) that are preferrefor Europa (McCordet al.1998a, 1999a).

An important sulfate constituent of carbonaceous chondrcalcium sulfate (anhydrite, CaSO4; and gypsum, CaSO4 · 2H2O),is not highly water soluble in most naturally occurring waters aso is not expected to be a major constituent of a europan oor an icy-salty crust. Calcium sulfate might be leached bydrothermally circulated water affecting chondritic rock if theis large water : rock ratio (analogous to the CM chondrite bing experiments done by Fanaleet al.1998), but gypsum wouldreadily precipitate upon reaching the ocean, possibly formextensive deposits near hydrothermal vents.

Chlorides are minor constituents of chondrite leacha(Kargel 1991, Fanaleet al. 1998) but are very important because they are highly effective freezing-point depressantswould be concentrated in the coldest brines during freezinan ocean or brine magma body. Chlorides likely would befective in preventing complete freezing of the ocean.

It is not necessary to think solely in terms of ice conta

ing minor salty contaminants that accumulate only upon sev

OCEAN 235

andlfatessesrths ae-s inopaockand

tedriticistryna-

rite-

.g.,r-teser-laya

iseiseseralthe-iteg–

tes,

ndeany-

reil-

ing

tes-andof

ef-

n-

evaporation or freezing. For instance, one structural–chemmodel produced by Kargel (1991) has a europan crust cposed of a peritectic mixture of 84% MgSO4 · 12H2O+ 11%Na2SO4 · 10H2O+ 5% ice—ice is the minor constituent. Acording to that particular model, a 20-km-thick layer madeperitectic mixture of hydrated salts and ice would be underby a much thicker layered sequence of less-hydrated saltally lacking ice. In another model, Europa has a eutectic cof 50% ice+ 44% MgSO4 · 12H2O+ 6% Na2SO4 · 10H2O. Themain points of this previous modeling were that: (a) saltsmajor constituents of the crust and ocean derived by chondevolution, (b) magnesium and sodium sulfates are specificpredicted as the major salts, and (c) an ocean would probabhypersaline, with solutes exceeding 17% by mass.

Kargel (1991) modeled the first, coldest stages of aquedifferentiation without evaluating logical progressions of ming and further differentiation at higher temperatures. We hcomputed many additional models; Fig. 3 shows a selectioalternative chemical structures (or possibly an evolutionaryquence) that might represent a differentiated chondritic EurThe models all assume that a gravitationally stable strucis achieved. The models differ in the extent of dehydrationthe mantle and of aqueous phase segregation in the crust/oThose having an icy shell made of pure ice implicitly invosome process, such as segregation of large ice bodies follby solid-state ice diapirism. Figure 3 also shows the compvalues ofC/MR2 of each structure;C/MR2 was not used asconstraint on the models, but rather was computeda posteriori.A similar set of Europa structure models was recently preseby Spaun and Head (2000); their work is based on fracticrystallization of a putative ocean approximated in the binsystem MgSO4–H2O and took full advantage of Europa’s oserved value ofC/MR2 as a constraint. Their models resemsome of those presented here in that they show multipleers of various hydration states and ice content, although intail the models differ since the modeled processes are diffeModeling results by Zolotov and Shock (2000b) of the equirium crystallization of Europa’s ocean have a problem withCaSO4 parameterization added to the FREZCHEM 2 moding software, and so those results are erroneous. However,useful is the important result by Zolotov and Shock (2000a)aqueous alteration of an initial CV chondrite produces a salsemblage similar to that we have used as the starting poinour models. Their model points out that almost any oxidichondritic meteorite assemblage could have been used astarting point for our models and would have yielded a quatively similar end result; details, of course, will differ.

Each of the models in Fig. 2 has an identical dense core mof iron sulfide. We approximated the core’s mass by calculathe amount of FeS representing the content of sulfide in CMbonaceous chondrites (Bostrom and Fredriksson 1964, K1991). In reality, chondritic sulfides are chemically more coplex, but it is not our purpose to investigate in any detailchemical nature of the core. However, considering the st

ereinfluence of the existence of a core onC/MR2, it was necessary

b,dala4aa

d9

pt

o

s

esa

uenn

s

ef-re-) isweral

ur-del

latileonaltion

’s

ofng.Hs

thehavestill

edsthe

s

eritic

. We

236 KARGEL

FIG. 2. Triangular plots showing the compositions (mass ratios) of solusalts of carbonaceous chondrites and terrestrial brines: (left) Major anions(right) major cations. Rectangles numbered 1–3 are data for the CI chonOrgueil–number 1 is the total salt content, including sulfates, chlorides,carbonates; 2 is with carbonates subtracted; 3 is with carbonates and casufate subtracted. Rectangles numbered 4 and 5 are for the experimental leobtained by Fanaleet al.(1998) from the CM chondrite Murchison—numberis the actual “original leach solution,” and 5 is that solution with calcium sulfsubtracted. The most directly comparable points for these two meteoritesand 5. Seawater and a closed-basin playa brine from the “La Mancha” regSpain (de la Penaet al. 1982), are shown as the rectangles numbered 6 anrespectively. In both plots dots are closed-basin brines (data from Eugster 1

to take a first-order look at how a carbonaceous chondritecursor could also yield a dense core potentially satisfyingGalileogravity data.

The specific models illustrated in Fig. 3 correspond to the flowing scenarios, all computed using Europa’s observed m(Andersonet al. 1998), a carbonaceous chondrite precur(Fredriksson and Kerridge 1988), and low-pressure phase elibria and phase densities in the system H2O–MgSO4–Na2SO4

(Kargel 1991). The model scenarios are briefly described he

Model A5. Sequential steps of perfect fractional incongruemelting of hydrated salts at a series of peritectics as energadded to the system. Each peritectic liquid (represented byfrozen crustal layers P2, P3,. . .) is saltier and denser than thsolid peritectic assemblage that preceded it (Table II), andstable density stratified crust is developed by sequential stof brine intrusion and underplating. It is assumed that thetrusive layers solidify rather than form an ocean. The amoof water and salts available to the crust/ocean system is dmined by phase equilibria during dehydration of salts contaiin the primordial europan chondritic assemblage. The maof this model, after differentiation, still contains phyllosilicateand the water of hydration of salts that are fairly insoluble astable at low temperatures, including gypsum (CaSO4 · 2H2O)and kieserite (MgSO4 ·H2O). The implication is that the mantlehas not been heated very strongly, an unlikely situation (conering that Europa has a core).C/MR2 is far too high compared

to Europa’s actual value (0.346± 0.005; Andersonet al.1998).

ET AL.

leandritend

ciumchate

tere 3ion,7,

80).

re-he

l-assorqui-

re:

nty isthe

o agesin-ntter-edtlesnd

id-

Model B. This structure results from multiple stages officient incongruent remelting of the hydrated salt layers repsented in Model A5—eventually a eutectic crust (or oceanformed and all excess salts exist in anhydrous form as a locrust.C/MR2 still is far too high compared to Europa’s actuvalue.

Model C. Taking aqueous differentiation yet another step fther, it is assumed that the eutectic crustal material of MoB segregates into a buoyant crust of ice and a deeper vosubcrust of hydrated salts. Segregation may occur by fracticrystallization of an ocean or by solid-state phase segregaand ice diapirism.C/MR2 still is too high compared to Europaactual value.

Model D′′. More heat is injected into the hydrated saltsModel C, resulting in multiple stages of incongruent remeltiAt each stage of remelting the residual salts contain less2Othan before and the amount of H2O in the uppermost icy layer iincreased. Also, all salts that had remained in the mantle inmodels above, including carbonates, gypsum, and kieserite,now been driven out and added to the crust, but the mantleretains phyllosilicates. As differentiation of the crust proceto completion, the upper crust consists entirely of ice andsubcrust of anhydrous salts.C/MR2 is much closer to Europa’actual value but still is a bit high.

Model D′. Similar to Model D′′, but the amount of water in thcrust is increased due to complete dehydration of the chondrock’s phyllosilicate minerals and all other hydrated phases

FIG. 3. Five model structures developed from a chondritic Europa(described in text).

’S

a

o’sa

mtiseiasf

adre

tu

p

sie

eapn

befrth

o

l

i

ithern.

oresalts.ul-en-p-he

ar-s toute

onical

99).lu-

y ofeez-olve

n re-is

hatbse-bein

le for

inec-

. 4ateved

mite

al12

hlo-sul-outpu-

pa’sater.al-ix-as

EUROPA

assume a toal water content equal to that of CI chondrites clated from H abundance given by Anders and Grevesse (19with the aproximation that all H is in the form of water—H assciated with organics is relatively minor. With total dehydratiof Europa’s interior,C/MR2 is far too low compared to Europaactual value (0.347). The radius of the differentiated Europalso too large and the mean density too low.

The models illustrated in Fig. 3 do not include any coplexities caused by addition of other volatiles during heaof Europa’s rocky mantle. None of the computed structureFig. 3 hasC/MR2 very close to the observed value, but modD’ and D” bracket the observed value (0.347), showing that fasimple models can evolve into structures consistent with grtational constraints. Considering that the amount of solubleand water in chondrites varies widely (roughly a factor of 3CI and CM chondrites), the results would indicate that an idsolution could be forced with iteration of the amount of saltswater initially present or by iteration of the degree of interiorhydration. Boato (1954) gave a drier estimate of CI chondcomposition (∼12% H2O by mass) than the 18% value we us(from Anders and Grevesse 1989). Had we used the lowerter content, complete aqueous differentiation in model D′ wouldhave resulted inC/MR2 much closer to Europa’s actual value

Our purpose with Fig. 3 is not to present an ideal strucof Europa. Anyway, there is a difference between producinmodel that satisfies available constraints (very easy) andducing a model that is uniquely satisfactory (not possible).showing models that bracket Europa’s observedC/MR2 andmean density, and also presenting a range of other modeldo not do so well, we merely wish to point out that a great varof structures can result from relatively simple chemical modstarting with identical chondritic material.

We note that equilibrium petrology would not allow devopment of an upper crust of pure ice or of pure salts (hydror not). An evaporite layer of pure-salt surface would imsublimation of ice. Any chondritic model that has ice formithe only major constituent of the uppermost crust must invo(a) efficient fractional crystallization of brine flows (wheresalts sink and ice floats), (b) some form of solid-state segrtion of phases and diapirism, or (c) water evaporation andrecondensation. There is strong geologic evidence of andretical backing for diapirism (Pappalardoet al. 1998, Rathbunet al.1998, Spaunet al.1999, McKinnon 1999) and frost condensation (Squyreset al.1983, Domingueet al.1991), and frac-tional crystallization of brines is all but inevitable (Hogenboet al.1995).

The propensity for solid-state diapirism or cryovolcanismoccur, and also the stability of a chemically layered crustaquence, is greatly dependent on the relative densities ofsolid and liquid phases. Table II compiles the densities ofportant solid phases, solid assemblages, and liquids in the sy

MgSO4–Na2SO4–H2O. This table also indicates the compostion of the liquids formed by lowest temperature melting of ea

OCEAN 237

lcu-89),o-n

is

-nginlsrlyvi-altsorealnde-ited

wa-

.re

g aro-

By

thatty

els

l-tedlyglveyga-osteo-

-

m

tose-keym-stem

solid phase assemblage. The listed data do not consider ecompression due to elevated pressure or thermal expansio

3.2. Formation of a Chloride Ocean by Partial Freezingof Chondrite Brines

It is well understood that common chloride salts are meffective as freezing-point depressants than common sulfateIn multicomponent solutions containing both chlorides and sfates it is generally true that partial freezing results in conctration of chlorides and precipitation of sulfates. (A key excetion is sulfuric acid, if it is present, as discussed below.) TFREZCHEM 5 model (modified after Spenceret al. 1990 byMarion and Grant 1994, Mironenkoet al. 1997, and Marionand Farren 1999, with modifications including addition of cbonate minerals and improved parameterizations of sulfatehigh ionic strengths) was used without carbonate to compthe thermal evolution of chondritic brines. An earlier versiof this computational tool was used to investigate the chemevolution by freezing of Earth’s seawater (Marionet al. 1999)and a hypothetical martian ocean (Morse and Marion 19The FREZCHEM models select the compositions of the sotions and solid phases that minimize the Gibbs free energthe system. Solid phases formed during evaporation or fring are assumed to have the potential to subsequently dissor react as solution properties change (e.g., the solids cacrystallize and add water of hydration, or lose water). Thisequilibrium crystallization, which represents an extreme in tall precipitated phases are assumed to be available for suquent reactions. Fractional crystallization, which also cansimulated with FREZCHEM 5, represents another extremethat precipitated phases are buried and thus are unavailabsubsequent reactions.

Figure 4 shows the modeled evolution of a chondrite brduring progressive freezing alternately by equilibrium or frational crystallization. The initial high-temperature brine in Figis that calculated by Kargel (1991) for the Mg–Na–Ca–sulfcomponents with addition of chlorides in the amounts obserby Fanaleet al. (1998) in their leachate. In the equilibriumfreezing model, magnesium sulfate first precipitates as epso(MgSO4 · 7H2O) but then is completely replaced by MgSO4 ·12H2O by reaction with the brine as it cools. In the fractioncrystallization model MS7 is preserved and then buried by MSand other salts. The calculations show the enrichment of crides at lower temperatures and a crossover in chloride andfate abundances at about 266 K, which coincidentally is abthe temperature where ice begins to precipitate. These comtations do not consider effects of elevated pressures in Eurocrust; some pressure effects are considered qualitatively lFigure 4a shows the thickening of the icy shell (with twoternative assumptions that it is entirely ice or is a cotectic mture of ice and salts) and thinning of the liquid brine layer

i-chthe system cools. Figure 4d shows the difference between frac-tional and equilibrium crystallization in terms of the final mineral

perature.

FIG. 4. Freezing evolution of a CI-chondrite-derived brine ocean on Europa. (a) Thickness of liquid brine layer and icy shell versus ocean tem(b) Composition of the brine as a function of temperature, calculated with FREZCHEM 5 with the assumption of equilibrium crystallization. The top three curvesfor liquid water, ice, and water in hydrated salts represent the percentage of the total water in the system that is in those three forms, which add up to 100%. The othercurves show the concentrations of each ionic or anionic species. Also shown is the initial temperature at which each solid phase begins to precipitateas the brineis cooled. (c) Same, but fractional crystallization. (d) Completely frozen mineral assemblage produced, alternately, by equilibrium and fractional crystallization. The bar plots show relative abundances of solid phases, not vertical sequences of minerals.

238

239

solidified as hydratedcentrated in the resid

EUROPA’S OCEAN

C

c

u

hsf

e

inrustbe

resit iszenof am-er,hereted

FIG. 4—

assemblages that are produced. Fractional and equilibriumtallization produce slightly different lines of brine evolution, bthe key point about chloride enrichment at lower temperatremain qualitatively the same in the two scenarios.

If the initial model brine of Fig. 4 reflects the real Europone implication is that any europan ocean in equilibrium wan icy shell would likely have concentrations of chlorides higthan previously believed by Kargelet al. (1991). The reason ithat the brine starts out with very high solute (especially sulsalt) contents, and by the time the temperature reaches awhere ice can precipitate, much of the brine’s mass has alr

sulfates, causing chlorides to becomeual solution. The brine does not comple

ontinued

rys-utres

a,ither

atepointady

freeze until 238.55 K, at which point the sulfates are mostlysolid form already. Considering pressures deep in Europa’s c(to nearly 200 MPa), the actual solidus temperature wouldcloser to 216 K. This is likely much colder than temperatuof convective adiabats in Europa’s icy shell, meaning thathighly improbable that a europan ocean would be totally froor even partly frozen to such an extent that it would consisteutectic chloride solution (cf. McKinnon 1999 who uses a teperature of a convecting icy shell of about 260 K). Howeveven if the temperature is so cold as the chloride eutectic, tis enough chloride in chondritic material to produce a satura

con-telyeutectic chloride ocean over 4 km deep (Fig. 4a). For likelier sit-uations having crustal convective adiabats warmer than 216 K

L

uritipis

e

te

sae

ee

p

lvf

tyd

t

bno

-n

hcti

ityro-onsas

es oc-peron

thising

atesrel-d S

tove

lity

pon

verewisgh-

lidus

(3)po-any

s intexis,es

-atears

en-crgel,yetas-in-

O

240 KARGE

(or >238.55 K if not considering pressure effects), the liqocean would be thicker, the icy shell thinner, and the chlocontent lower, as Fig. 4 shows. In any event, if the identificaof magnesium and sodium sulfates on the surface of Eurocorrect (McCordet al. 1998a), and if the surface material issome way reflective of the ocean composition, then it suggefairly warm ocean (to keep sulfates in solution). An ocean teperature of 266 K would allow a sulfate-rich ocean to exist astill allow ice to precipitate and form part of the crust (neededexplain surface ice observed spectroscopically). This high tperature would also mean that the liquid mass would be falarge, the ice mass fairly small (meaning a thin icy shell), andocean quite deep—consistent with interpretations of theGalileomagnetometer data. Nevertheless, it is difficult to have bofrozen ice-bearing shell and a liquid brine ocean that is quitdeep as the magnetometer team would like; at least 30 km isily possible so long as the icy shell is very thin. Of course, thijust one chemical model of the ocean/crust. A thin ice crustdeep bring ocean up to 100 km thick (favored my magnetomresults) would require lower initial solute contents than modehere.

3.3. Addition of SO2 or SO3

Stevenson (1982) and Crawford and Stevenson (1988)considered possible roles of dissolved gases, including C2,SO2, and CH4, in explosive water volcanism on small icy satlites. An important further advance of our understanding ofplosive cryovolcanism has been made by Fagentset al. (2000),who have examined in detail the explosive venting of euroaqueous liquids due to exsolution of CO2, SO2 and other dis-solved gases. On Earth sulfuric acid—an important dissospecies in volcanic crater lakes and acid rain—normally ismed by the introduction of SO2 gas into aqueous systems (Karget al.1999). The fact that SO2 is so abundant on Io suggests thamay also be vented into Europa’s crust, a process that wouldH2SO4 (Kargel 1998, 1999a). SO2 also occurs on Callisto anappears to be endogenic or partly endogenic in origin (Hibbet al.2000). This endogenic mechanism is interesting in ligh(1) spectroscopic evidence of traces of SO2 (Laneet al. 1981,Noll et al.1995) on Europa, (2) correlation of Europa’s UV asorbers (now thought to be a combination of S–O compouelemental sulfur, sulfates, peroxides, and ozone) with geolfeatures in addition to a longitudinal pattern related to exogeprocesses (Hendrix 1998, Hendrixet al.1998), and (3) new interpretations by Carlsonet al.(1999) that Europa’s predominanonice hydrated material is H2SO4, which was thought by themto have a mainly radiolytic origin but clearly has some typepartial control by endogenic processes.

Sulfur flux from the volatile subcrust and mantle should genate acidity, which may tend to be neutralized by reactions wsilicate rocks. With appreciable sulfur input, a portion of tcations in solution would be charge-balanced by sulfur sperather than by carbonate. In Earth’s continental evaporite set

(strongly oxidizing), this scenario naturally leads to neutral N

ET AL.

iddeona isnts am-ndtom-

irlythe

h aas

eas-isndtry

led

firstOl-x-

an

edor-elitield

ittsof

-ds,gicnic

t

of

er-itheiesngs

(Mg)–SO4–Cl brines. The exception of course is when acidis not neutralized (nonreactive wall rocks) and builds up to pduce acid brines. It is valid to ask whether Europan conditiwould be sufficiently oxidizing to produce sulfate. Oxygen hbeen observed in Europa’s tenuous atmosphere (Hallet al.1995)but has been attributed to magnetospheric plasma processcurring on Europa’s surface and might say little about deeconditions. On the other hand, all nonice materials identifiedEuropa are highly oxidized compounds—SO2, CO2, H2O2, andone or another type of sulfate. We primarily assume thatoxidized condition persists to deep levels. If we err in makthis assumption, then maintaining sulfur in reduced forms (H2S,metallic sulfides, and native sulfur) should favor Na-carbonand chlorides as the major “evaporite” products. Thus, theative amounts and oxidation state/molecular form of C ancompounds are fundamental questions to explore.

Production of copious quantities of SO2 fits into the sulfate salscenario just described. The chondrite-evolution models abomit CaSO4 as a major constituent because of its low solubiin water. The bulk of any CaSO4 that would be leached fromEuropa’s interior would promptly saturate and precipitate ureaching the ocean. Some of Europa’s CaSO4 would have set-tled into a thick layer on the seafloor, but most would neeven have been leached from the mantle. We note that L(1982), in evaluating the geochemistry of Io’s sulfur under hitemperature, anhydrous conditions, considered CaSO4 to be amajor repository of sulfur and a possible source of SO2. Whensubjected to temperatures approaching silicate mantle sotemperatures, CaSO4 may break down and yield free SO2 orSO3 fluid by reactions such as these:

CaSO4↔ CaO+ SO3 (1)

CaSO4+ 3FeO↔ CaO+ Fe3O4+ SO2 (2)

CaSO4+ FeS↔ CaO+ FeO+ SO2+ (1/x)Sx. (3)

FeO and CaO produced by the forward reactions (2) andwould further react with silicates and thus represent comnents of silicate minerals; the FeS in reaction (3) representssulfides still retained in the mantle. A similar process occurEarth’s subduction zones and regions of basal crustal anaexplaining SO2-rich silicate magmas erupted from volcanosuch as Mount Pinatubo, Lascar, and El Chich´on. These volcanoes’ S-rich magmas (0.1 to 1 wt.% S in the forms of sulfand SO2) is explained by anhydrite saturation at several kilob(Luhret al.1984, Luhr 1990, Bernardet al.1991, Mathewset al.1994, Fournelleet al.1996, Rutherford and Devine 1996).

The sulfurous volatiles produced by reaction (3) coincidtally can explain the SO2 and S2 gases venting from Io’s volcaniplume, Pele (John Spencer, personal communication to KaNovember 1999). Zolotov and Fegley (1998a,b, 1999), notaware of Spencer’s findings, modeled the sulfurous volatilesemblage expected from Pele and concluded that volatileteractions with silicate magmas would tend to produce S2

a–and S2 gases under some likely conditions. On Europa, thick

’S

faego

b

n

t

rer

sh

r

ti

e

rd

iri

Hhdat

srac

s

n-

re as-

rous

theite

ofd-oth

on-O

sub-bovel

ed

rd-

lglemr-,l-oraveofted;e,

the

ions

if

icthe

nini-on

rate

EUROPA

sedimentary accumulations of magnesium and sodium sulin the volatile subcrust also could be subjected to tidal hing; they could dehydrate, and with further heating they miyield SO3 by reactions analogous to (1), where the metalides would form a highly refractory, unsilicated substance. S2

or SO3 would tend to migrate upward and eventually wouldvented into the ocean or icy shell. SO3 would immediately re-act with water or ice to form sulfuric acid, and SO2 might alsoform sulfuric acid under oxidizing conditions; relevant reactiomight be

SO3+ H2O↔ H2SO4 (4)

3SO2+ 2H2O↔ 2H2SO4+ (1/x)Sx. (5)

Reaction (5) is important in many terrestrial volcanic cralakes, as reviewed by Kargelet al. (1999). The point is that, byone means or another, copious quantities of SO2 or SO3 can be re-leased into Europa’s ocean and icy shell from its deeper inteRadiolytic formation of H2SO4 at the surface also is inevitabland the process apparently works either starting with sulfuvirtually any form, e.g., indigenous Europan sulfate saltsIogenic sulfur implanted by plasma bombardment (Carlet al.1999). Hence, there is no shortage of mechanisms by wsulfuric acid can be made.

Unreacted SO2 in the ocean would dissolve up to a pressudependent saturation limit imposed by SO2 clathrate formation,which occurs by the reaction

4SO2+ 23H2O↔ 4(SO2 · 53/4H2O). (6)

Sulfuric acid also could accumulate until it reaches saturaand forms a hydrate phase:

H2SO4+ 6.5H2O↔ H2SO4 · 6.5H2O. (7)

Dissolved SO2 and H2SO4 would be available to drive explosivaqueous volcanism, which might also occur by interactionliquid water with frozen SO2 hydrate (Stevenson 1982, Crawfoand Stevenson 1988, Kadelet al.1998, Fagentset al.2000).

Whatever the precise chemistry behind production of SO2, ifwe look to Io’s plumes as a model of what may be happeninside Europa, then the amount of SO2 vented and reacted ovetime could be substantial. Io is losing mass, apparently maSO2, to the Io torus at a rate of about 1600–3200 kg s−1 (fromplasma mass loading calculated from plasma corotation lag;1979, Pontius and Hill 1982, Brown 1994), a quantity rougconsistent with optical and UV observations of escaping S anions (Cheng 1984, Brownet al.1983, Shemansky 1987, Bagen1994, Wilson and Schneider 1994). There is evidence for inannual variability of this mass loss rate over two decadeobservations; there is no way presently to know how thismay have varied over geologic time. We assume that the rerate of Io’s SO2 loss is characteristic, which we sum over timto yield the amount of SOthat Io has lost and a minimum e

2

timate of the SO2 with which Io was originally endowed. We

OCEAN 241

tesat-htx-Oe

s

er

ior.,in

oronich

e-

on

of

ng

nly

illlyO

ler-ofteente-

calculate approximately 3.6× 1020 kg. If Io’s rate of SO2 lossis scaled to Europa’s smaller mass (with H2O subtracted), andif this rate of SO2 venting from Europa’s interior has been costant through time, then in 4.5 Byr 2× 1020 kg of SO2 wouldhave been vented into the ocean. This is what we refer to hethe Io model of Europa’s SO2 production. It is simply one estimate of the possible magnitude of SO2 venting, not an estimateof the precise amount, nor a statement regarding this sulfumaterial’s present chemical form.

An independent calculation based on reaction (2) andamount of CaSO4 contained in typical carbonaceous chondrmaterial yields a similar magnitude of SO2—4.8× 1020 kg. Wecall this calculation the chondritic calcium sulfate modelEuropa’s SO2 production. Considering the crudity of the moels, the results are in surprisingly good agreement. To put bestimates into perspective, the Io model of Europa’s SO2 pro-duction yields an amount of SO2 equivalent to 0.4% of Europa’smass or 2.5% of its total water mass (according to the CM chdrite model); the chondritic calcium sulfate model produces S2

equivalent to 6% of Europa’s total water mass. These arestantial amounts, although as several aqueous reactions asuggest, much of the vented SO2 now would be in other chemicaforms, such as sulfuric acid and elemental sulfur.

We calculate the amount of sulfuric acid potentially producby taking the amount of SO2 from the chondritic calcium sulfatemodel and assume that it reacts virtually to completion accoing to reaction (5). The resulting calculation is 4.9× 1020 kgof sulfuric acid plus 8× 1019 kg of elemental sulfur. Elementasulfur would be deposited on the seafloor; if it formed a sinlayer of solid rhombic sulfur, it would be approximately 1300thick. The calculated quantity of sulfuric acid might react futher to form as much as 1.08× 1021 kg of a solid hydrate phaseH2SO4 · 6.5H2O, according to reaction (7). This quantity of sufuric acid would be structurally and chemically significant fthe evolution of Europa’s ocean and crust, but to date we hnot explicitly included it in our models. Of course the amountsulfuric acid might be nowhere near as substantial as calculasome vented SO2 may have been lost from Europa into spacand some may not have reacted to form H2SO4; some of themantle’s CaSO4 may not have decomposed, and some SO2 orsulfate may have reacted to form metal sulfides instead. Onother hand, some amount of other sulfate salts, e.g., MgSO4, mayhave decomposed, thus adding to the SO2 contributions madeby CaSO4; hence, the amount of SO2 and H2SO4 could be evengreater than calculated. The point is that endogenic contributof SO2 and H2SO4 could be enormous.

If this model of SO2 degassing is correct in concept evennot in magnitude, then the amount of endogenic SO2 and H2SO4

in all likelihood would vastly exceed radiolytic and photolytproduction of these materials. However, we do not disputelikelihood that radiolytic/photolytic equilibrium would maintaia high surface abundance of sulfuric acid regardless of thetial form of sulfurous molecules on Europa’s surface (Carlset al.1999). There is good evidence for a global pattern in hyd

distribution consistent with exogenic modulation of sulfuric acid

L

id.etis

t

hu,itteaet

taeelfa

aphhvhc

ooemmh

rtrrrset

ld

d

ley

tedta

nowcetantthatareseies.ngthetinges.li-usete

ichffi-nt,

thert aning

ede,

n-

ous,toeink

ibutethedpHw-):

)

e-nates

242 KARGE

abundance, but pronounced geologic controls are also evand require an endogenic process (McCordet al.1998a, 1999a)The oceanic site of sulfuric acid production as modeled abovfers a ready explanation for this geologic control. We believethe evidence supports a prevalent geologic control on the dbution of nonice materials on Europa, but that exogenic comsitional patterns are superposed over this endogenic distribu

The proposed existence on Europa of hydrated salts orfuric acid makes sense cosmochemically, and each hypotalone is capable of explaining the chief spectroscopic featof Europa’s nonice terrains as seen in NIMS data. Howeversalt and acid hypotheses are not mutually exclusive from ea cosmochemical or a spectroscopic perspective. As a matdetail, one would not expect sulfuric acid to coexist with bsalts, such as sodium carbonate, but the more general idcoexistence of dissolved sulfuric acid or sulfuric acid hydrawith some salts, such as magnesium sulfate, is fully accepand perhaps inevitable if SO2 vents into Europa’s ocean. Hencit would not be surprising, and it would be entirely consistwith all available observations, if the nonice hydrate materiaEuropa consists of hydrated sulfuric acid, magnesium suland other salts, with variable admixtures of ice.

A highly acidic ocean would likely drive much other chemistry. A low-pH regime suggests that a variety of metals, aespecially iron and aluminum, would be leached from silicrocks, especially under hydrothermal conditions, and transted in solution to the ocean and crust. One place on Earth wacid evaporative conditions exist is Lake Tyrrell, Australia. Tacidity there has consequences for the stability relations of eoritic salts; e.g., alunite (an aluminous sulfate containingdroxyl groups) and jarosite (an hydroxylated iron sulfate) ocin addition to the more usual halite and gypsum (Longet al.1992a,b). This creates a rather different situation from thmost common on Earth, where hydrothermal reactions inving neutral-to-basic solutions and ferroan silicates normally lto sulfate reduction. By analogy, an acidic europan oceanprecipitate extensive deposits of alunite and jarosite. Theseerals could occur along hydrothermal circulation paths witthe silicate subcrust. They might also form on Europa’s sface by freezing and evaporative concentration of erupted bHigh acidity might be maintained in the ocean despite neuizing tendencies of reactions with silicates if the potentiallyactive silicate rocks are sealed off by alteration materials oaccumulation of thick salt deposits on the seafloor. Fracturetems within such layers might permit sulfur gases to continube transmitted into the ocean from mantle degassing andmal devolatilization of evaporites beneath the ocean. The recould be runaway acidification. Further consequences courather bizarre, e.g., partial digestion of incoming asteroidsEuropa’s acid ocean! The indigestible residue—includingsilient interstellar/supernova condensates such as diamonsilicon nitride—-might settle onto a sulfur-encrusted seafloo

Alunite and jarosite and some other expected acid alteratminerals contain spectroscopically active hydroxyl and wou

ET AL.

ent

of-hattri-

po-ion.sul-esisrestheherr of

sea ofesble,ntonte,

-ndteor-ereeap-y-ur

selv-adayin-in

ur-ine.al-e-byys-to

her-sult

bebyre-and

r.

be clearly evident in near-infared reflectance spectra (Crow1993, Crowley and Hook 1996, McCordet al. 1998a). Alumi-nous hydroxyl minerals, including alunite, have been detecrecently in AVIRRIS images (a NIMS-like remote sensing daset) of Mount Shasta despite 50% subpixel coverage by spatches (Crowleyet al.1999). The fact that there is no evidenof these minerals in NIMS Europa spectra may be an imporclue to ocean or crustal chemistry. Unlike some mineralsare readily destroyed by solar ultraviolet, hydroxyl phaseshighly stable (Yenet al. 1999). The apparent absence of theminerals on Europa might be due to any of three possibilit(a) Acidic brine is not in contact with aluminum and iron-bearisilicate rocks; perhaps a thick accumulation of salts coversseafloor completely, as modeled above, so that any circulahydrothermal brines lack opportunities to react with silicat(b) Efficient neutralizing reactions of oceanic brine with sicates may effectively prevent buildup of high acidity and caany alunite and jarosite to remain in solid form in the silicasubcrust. (c) The ocean is intrinsically neutral or alkaline, whprobably would imply that sulfur gases are not vented in sucient amounts to overwhelm sources of alkalinity. In any eveit appears on the basis of nondetection of hydroxyl anion eithat Europa does not possess a highly acidic ocean or thaunknown mechanism prevents hydroxyl phases from formor from being exposed at the surface.

3.4. Addition of CO2

CO2 is among the materials detected on Europa, Ganymand Callisto, albeit in trace amounts (McCordet al. 1998b,Hibbittset al.2000), and sodium carbonate is on the list of cadidates for Europa (McCordet al.1998a). Although radiolysisof carbon-bearing ices might produce a variety of carbonacespecies on Europa, including CO2 (Delitsky and Lane 1998)there is evidence that CO2 is at least partly endogenic on Callis(Hibbitts et al. 2000). CO2 is widespread and abundant in thSolar System from Venus out to Triton, so it is reasonable to ththat Europa may also possess abundant endogenic CO2. We shallconsider how endogenic processes on Europa might contrto CO2 surface frosts and also yield sodium carbonate. Inbinary system H2O–CO2 the relevant equilibria are describeby three reactions, each of which dominates over a specificrange (given here for equilibria at 298 K, representative of lotemperature aqueous carbonate equilibria, Krauskopf 1979

CO2+ H2O↔ H2CO3 at low pH (<6 or so) (8)

CO2+ H2O↔ H+ + HCO−3 at medium pH (7–10 or so) (9

CO2+ H2O↔ 2H+ + CO2−3 at high pH (>11 or so). (10)

If initial aqueous differentiation yields a Mg–Na-sulfatdominated crust as considered above, then Ca–Mg-carbo

ionld(also contained in carbonaceous chondrites) would have beenleft behind in the mantle due to relatively low solubility. This

S

T

fh

O

d

ss

nta

a

li

7ei

sbhs

ra

ro

dgr

n-

andtilized

ali

ironandsys-anicrialbyby

al

ndinto

nder-en-

ablyat

as

.1).uldo-

h-atesc-

with

EUROPA’

is similar to the case for calcium sulfate described above.carbonates may have formed in the first place by oxidationorganics. At some point, intense heating of the mantle wohave decomposed the carbonates, and caused degassing o2.This no doubt occurred in Earth, too, but in Europa, the icy sand possibly an ocean may tend to trap the CO2, which would beavailable to drive formation of carbonates, bicarbonates, or C2-clathrate. How much CO2 might this be? There is enough CaMg-carbonates in chondritic material to supply 1.0× 1021 kg ofCO2 to Europa’s crust, if the carbonates entirely decomposThis compares to a total water mass of 8.1× 1021 kg in the CIchondrite model of Europa. But there would be some formchemical sequestration of CO2. Once the amount of dissolveCO2 exceeds about 3 wt.% (the exact amount dependingpressure, temperature, and salinity), then CO2 clathrate hydratewould sequester additional CO2. However, sodium carbonateand bicarbonates may start to form at much lower amountdissolved CO2. Furthermore, the icy shell over the ocean migbe leaky (e.g., fractures through which dissolved gases cacape during eruptions; Fagentset al.2000), in which case mosof the CO2 that goes into the crust ends up being vented to sp

In any case, the amount of CO2 potentially available to formclathrates, carbonates, and/or bicarbonates is quite substand perhaps coequal with SO2. However, only traces of CO2have been detected on Europa (McCordet al.1998b). The chem-ical nature of aqueous solutions formed by alteration of silicain the presence of high activities of CO2 depends very much onthe type of rocks. If the rocks contain abundant alkalis and aline earth metals (especially K, Na, Mg, and Ca), such as silundersaturated mafic volcanic rocks, hydrolysis and carbonareactions generate weak alkaline solutions (Krauskopf 19This type of situation is frequently encountered on Earth in arof rift volcanism and intraplate hotspot volcanism. Evaporatconcentration of dilute alkaline river input in closed basinssuch regions characteristically results in concentrated alkabrines enriched in sodium carbonate or bicarbonate and posing very high pH. A somewhat similar type of brine couldobtained on Europa, particularly if undersea volcanic rocksa highly silica undersaturated and alkaline character, such amafic feldspathoidal rocks common along the East African rthe southwestern United States, and many late-stage volcrocks capping oceanic island hotspot volcanoes. A controvebut highly regarded hypothesis for the origin of Earth’s oceinvolves this type of chemistry and an original sodic carboate composition, which only gradually evolved into a sodiuchloride composition (Kempe and Degens 1985).

Under other conditions, with hydrolysis and carbonationactions acting on acidic rocks, e.g., silica-rich granite, solutitend to have neutral to acidic compositions (Krauskopf 197Extensive hydrothermal reactions involving such silicic, acirocks, and chemical sedimentation on the seafloor under hianoxic conditions could produce rocks similar to banded i

formations (composed of iron oxides, Fe–Mg–Ca-carbonaand silica phases). Input of excess CO2 from undersea volcan-

OCEAN 243

heof

uldCOell

–

ed.

of

on

ofhtes-

ce.

ntial

tes

ka-ca-tion9).as

veinlineess-eadthe

ift,anicsialnsn-m

e-ns9).ichlyon

ism or venting of metamorphic fluids could result in a concetrated solution of carbonic acid and formation of CO2-clathrate-hydrate.

Carbon dioxide might also be consumed by other abioticbiogenic processes, such as the methanogenesis reaction uby some microorganisms:

CO2+ 4H2↔ CH4+ 2H2O. (11)

An important reaction on Earth leading to the formation of alklakes (Drever 1997) is

8SO2−4 (aq)+ 2Fe2O3(c)+ 15C+ 7H2O

↔ 4FeS2(c)+ 14HCO−3 + CO2−3 . (12)

When dissolved carbonate is abundant, this type of sulfate–oxidation–reduction reaction removes S as sulfide and CaMg as carbonates, leaving a predominantly Na–K carbonatetem. Carbon in the above reaction is either graphite or orgcarbon, which on Europa could be primordial organic mateinitially present in carbonaceous chondrite rock, acquiredimpacts of comets or P- or C-class asteroids, or producedbiological activity. Graphite could have originated from thermmetamorphism of original carbonaceous matter.

Other reactions could occur among oxidized crustal aoceanic materials when reduced gases from below percolatethe ocean, for instance,

8SO2−4 (aq)+ 2Fe2O3(c)+ 15CO2(g)+ 30H2(g)

↔ 4FeS2(c)+ 14HCO−3 + CO2−3 + 23H2O. (13)

The equilibrium constant for this reaction is extremely large apositive, which would drive bicarbonate formation. The minals that precipitate from alkaline systems are very much depdent on the assumedPCO2 and brine composition. AtPCO2 =0.1 atm for an alkaline carbonate brine (Na–CO3–SO4–Cl–H2O)(G. M. Marion, in preparation), NaHCO3 (nahcolite) precipi-tates across all temperatures from 298 to 273 K (and problower also). The pH of this system is 8.15 at 298 K and 7.87273 K. At PCO2 = 3.5× 10−4 atm (current Earth value), tron(NaHCO3 ·Na2CO3 · 2H2O) precipitates at 298 K, which giveway to natron (Na2CO3 · 10H2O) below 290 K. The calculatedpH of this system at 298 K is 10.15 (cf. measured value is 10If minerals such as natron precipitate on Europa, this woimply a low PCO2, high pH, and a high alkalinity system; nahclite would imply more concentrated CO2-rich systems at lowerpH. Since CO2 is potentially so plentiful on Europa—enougto generatePCO2 exceeding tens of bars if it were not for reactions such as these and for clathrate formation—bicarbonare likely. In the event that hydrothermal circulation of brine ocurs in a basaltic crust, then reduced iron should also react

tes,available carbonate ions to yield siderite, except under condi-tions of moderate to high acidity, in which case iron compounds

w

ply

ru

oirec

xea

,or

at

9n

a

s

Nu

emu

m

e

es

ts—xide,

ithan atnicbeces

t or-ass

nesse

e ofrustnicrouts—andthe

uldpedoor.beed inhy-rateper-

ig-lab

icaler-floorba-

ocktem-

riatealtic

975,mmcom-hy-tra-

ofex-of

tine,ses;ich

244 KARGEL

would likely be dissolved in the brine (and in which caseought to see them at the surface).

Eruption of carbonate-saturated brines could result in exsive activity due to exsolution of carbon dioxide. Alternativeinteraction of liquid water with CO2 hydrate (Stevenson 1982Crawford and Stevenson 1988, Fagentset al. 2000) or hydratediapirism could also lead to decomposition of the clathrate stture and explosive cryovolcanism.

3.5. Addition of Hydrocarbon Gases

Hydrocarbon gases may have been introduced to Europatime by comet and asteroid impacts or may have been derfrom Europa’s primordial carbonaceous chondrite mate(Anderset al.1973) by thermal metamorphism and hydrothmal processes. The CI chondrite Orgueil contains 3.5 wt.%bon, and CM Murchison contains 2.5 wt.%. The mass ratioorganic carbon compounds to bound water (including hydroexceeds 1 : 10 and may approach 1 : 5. Carbon in chondritmainly organic, but also forms carbonates, a graphitic phand other substances. The perspective of Anderset al. (1973)that chondritic carbonaceous materials, including organicsthe abiogenic products of solar nebula and possible presprocesses and subsequent aqueous chemistry is still the plent view. Anderset al. (1973) reviewed studies of CI and CMchondrites and gave some new analyses that revealed alkaromatic hydrocarbons, fatty acids, porphyrins, and at leasamino acids. The more recent review by Croninet al. (1988)tabulates 411 soluble organic compounds, including 74 amacids, in carbonaceous chondrites. Shock and Schulte (1showed that many amino acids and other water soluble orgain carbonaceous chondrites can be produced by low-temper(373 K) hydrothermal chemistry involving the reaction of simpgas mixtures (O2, CO2, and NH3) with pyrene and other organicbelieved to occur in interstellar clouds. Reaction of CO2, H2O,and N2 with basalt at 473 K achieves the partial reduction of2

and CO2 and abiotic formation of a complex mixture of aqueoorganic compounds (Shock 1990).

Carbon in chondrites is classified as a highly labile elemcomparable to argon, the alkali metals, and copper; most C isbilized and lost by heating between 973 and 1273 K (Lipschand Woolum 1988), but some is far more readily mobilizeCronin et al. (1988) indicates that about 30% of the organcarbon is in relatively labile forms ranging from C1 to C20 com-pounds and 70% is in highly resilient macromolecular forsometimes likened to kerogen. The most labile forms of orgacarbon are mobilized by water leaching, and these include amacids (60 ppm), dicarboxylic acids (>30 ppm), hydroxycar-boxylic acids (15 ppm), alcohols, aldehydes, and nitrogen hrocycles. The remaining labile organic compounds are releaby crushing or freeze–thaw disaggregation (trapped methaneheavier alkanes), or by leaching with acids, bases, benzenmethanol (Croninet al. 1988). A small amount of organics i

released by heating at 423 K and much more at 573 K whpyrolysis occurs. Pyrolysis of terrestrial sedimentary kerog

ET AL.

e

lo-,

,

c-

vervedialr-ar-ofyl)s isse,

arelar

eva-

nes,18

ino90)ics

turele

s

nto-tz

d.ic

snicino

te-sedand, or

and bitumen, without added oxygen, produced similar resullarge yields of methane and higher alkanes, carbon monohydrogen, and carbon dioxide between 473 and 573 K, wabout an order of magnitude more gas released at 573 K th473 K (Tannenbaum and Kaplan 1985). Much but not all orgacarbon would be volatilized during pyrolysis; some wouldgraphitized. Heating under oxidizing conditions also producarbonates.

On the basis of data just summarized, we estimate thaganic compounds amounting to about 100 ppm of Europa’s mwould have been dissolved in the initial low-temperature briduring primordial differentiation. This low-temperature releaof organics from Europa’s interior would produce an averagabout 1200 ppm of organic compounds in the primordial crepresented by model A5 in Fig. 3. This highly soluble orgamatter corresponds to about 4.8× 1018kg, and if condensed ovethe whole surface of Europa it could form an organic layer ab100 m thick. Far greater quantities of carbonaceous fluideasily two orders of magnitude more—would be producedreleased into the crust later as Europa’s interior warmed. Bytime of core formation, most of Europa’s juvenile carbon shohave vented into the crust, where it generally would be trapin the ice, dissolved in the ocean, or precipitated on the seafl

A comparatively small fraction of Europa’s organics woulddissolved in any ocean that exists; most would be sequestervarious solid or immiscible liquid phases. Methane and otherdrocarbon gases (up to isopentane) form solid ice-like clathhydrates under a wide range of high pressures and low tematures (Sloan 1998).

3.6. Hydrothermal Processes in the Seafloor Environment

If hydrothermal circulation through a basaltic subcrust is snificant on Europa, then terrestrial oceanic field data andexperiments offer a good basis for prediction of the chemchanges in the brine. Lab experiments with artificial hydrothmal systems and studies of altered basaltic rocks and seahydrothermal vent fluids are in good agreement as to thesic processes. For hydrothermal interaction at low water : rratios near unity, tens to a couple hundred megapascals, andperatures between 473 and 773 K—i.e., conditions appropto europan seafloor hydrothermal processes driven by basvolcanism—some expected trends (Bischoff and Dickson 1Mottl and Holland 1978, Seyfried and Janecky 1985, Von Da1995, Resing and Sansone 1999) are (1) rapid and almostplete loss of dissolved sulfate due to (a) precipitation of andrite or gypsum and (b) reduction of sulfate (with sequestion of sulfur as metal sulfides), and concomitant leachingcalcium (and reprecipitation as anhydrite or gypsum); (2)tremely rapid loss of Mg from solution due to a combinationsulfate reduction and sequestration of Mg by talc, serpentremolite, and other hydrated or hydroxylated alteration pha(3) much slower loss from solution of dissolved sodium, wh

enenreacts to form sodium feldspars and feldspathoids; (4) leachingand gradual enrichment in the solution of Si, K, CO2, and various

’S

r

t

nsidmraa

vr

i

scfd(her

ntp

e

mnttwtsn

u

b

p

driteasesor-

nt-ts ofon-

mite,idgeis

quiv-

nicsnt ofajor

ater;atoon-

ritesCOl, butater

wewithichust.noks,of79)resofili-nd isra-tion

3 and-ownbe-

rreak-K.

eldsres-isa’s

onsa-iteswa-

EUROPA

trace constituents such as B and Ba; and (5) marked decof pH (greater acidity). At low levels of reaction, magnesiuis removed from solution and balanced by calcium uptake;potential exists to produce a Na–Ca–SO4–Cl brine, where themajor cation by far is Na. Calcium increases in heavily reacbrines, but its abundance is sharply limited by anhydrite/gypssolubility; however, if sulfate is eliminated from the system acalcium continues to be leached, then the possibility existdevelop Na–Ca–Cl brines. Many transition metals—mostportantly iron—also are mobilized, then redeposited as sulfiunder conditions characteristic of Earth’s seafloor hydrothersystems; as dissolved species the transition metals are neveabundant unless there is extraordinary acidity. The abundof silica is also sharply limited by solubility unless acidity creach extraordinary levels. Hence, iron, calcium, and silicaexpected to be minor dissolved constituents—but possiblyimportant in the volatile subcrust—unless brine evolution pgresses in ways not encountered on Earth.

Except at very high degrees of reaction and high salinsodium, potassium, and chlorine are nearly conserved indrothermally reacted brines. Thus, Europa’s ocean could reble Earth’s if magnesium and sulfate are eliminated or redufrom levels in typical chondrite brines. There is no evidencea sodium chloride ocean on Europa, but rather there is goodirect evidence for abundant sulfate on Europa’s surfacepresumably the interior), suggesting that terrestrial styledrothermal processes do not occur or at least do not exdominating chemical influence. There are several possiblesons for this. (a) Europa might lack the degree of heat flneeded to drive hydrothermal circulation; considering evidefor recent geologic activity on Europa and perhaps a veryicy shell, and also considering tidal heat generation, this exnation seems unlikely. (b) The seafloor is not assuredly madrocks containing ferroan silicates; a thick accumulation of bded salts or sulfur on the seafloor, as modeled below, wouldallow the sulfate reduction process and uptake of magnesiuoperate as it does on Earth even if hydrothermal convectiovery strong. (c) Perhaps hydrothermally driven sulfate reducdoes occur on Europa, but the large quantity of sulfate and oalteration reactants effectively filled cracks and pore spacesalteration products, thus sealing the rock from further alterawithout consuming all the sulfate. (d) Europa might possesulfur cycle—inorganic or biologically mediated—that returprecipitated sulfide to sulfate form.

3.7. Silicate Dehydration

We considered above thermal dehydration and incongrmelting of sulfates, production of sulfur and carbon gasesthermal metamorphism and decomposition of sulfates, carates and organic substances, and hydrothermal processesseafloor environment. Another important high-temperature

cess is dehydration of hydrated and hydroxylated silicate merals in Europa’s rocky interior (Bunch and Chang 1980, Barb

OCEAN 245

easemthe

edumdto

m-esal

veryncenareeryo-

ty,hy-em-edorin-

andy-rt aea-

owce

hinla-

e ofd-not

tois

ionherith

ions as

entbyon-in thero-

1985, Zolensky and McSween 1988). Carbonaceous chonmeteorites contain some of the same hydrous silicate phknown in shales and other sedimentary clay-rich and metamphic rocks on Earth—tremolite, talc, chlorite, serpentine, momorillonite, etc.—although in detail there are peculiar aspecsome phyllosilicate phases in chondrites. In addition, they ctain large amounts of heavily hydrated salts, such as epsohexahydrite, bloedite, and gypsum (Fredriksson and Kerr1988). The total amount of H in the CI chondrite Orgueillisted by Anders and Grevesse (1989) as 2.02% by mass, ealent to just over 18% by mass H2O if all H was in the form ofwater. Some H is in the form of organics, but as the orgaare heated, they, too, produce some water. This large amouwater can be understood considering the water content of mphases that make up Orgueil (e.g., septechlorite, 9.7% wserpentine, 13%; montmorillonite, 22%; epsomite, 51%). Bo(1954) gave a somewhat drier estimate of 12% water in CI chdrites. CM chondrites are less heavily altered than CI chondand are drier—typically about 6% water by mass. CV andchondrites—even less altered—contain much less water stilif aqueous reactions are run fully, they would also have wcontents similar to Orgueil’s (Zolenskyet al.1989).

As Europa’s interior is heated to metamorphic conditions,expect that a stepped sequence of dehydration will occurformation and expulsion of high-temperature fluid phases, whultimately should be discharged into the ocean or frozen crAlthough details of the metamorphism of chondritic rocksdoubt would be different than metamorphism of terrestrial rocqualitatively the problem is exactly similar. Metamorphismterrestrial carbonate-bearing ultramafic rocks (Winkler 19is a fair analogy to metamorphism of chondrites. At pressuin the vicinity of 100–500 MPa, i.e., in the range of mostEuropa’s rocky mantle volume, significant dehydration of scate phases in ultramafic systems begins around 573 K avirtually complete by 1073 K. For example, in terrestrial ultmafic metamorphic systems, the first major step in dehydrais represented by the breakdown of serpentine between 57773 K, whereupon very H2O-rich and CO2-poor fluids are generated. Another major step is that represented by the breakdof talc anywhere between 820 and 1000 K for pressurestween 100 and 500 MPa and CO2-bearing fluids. The final majostage of dehydration of these rocks is represented by the bdown of tremolite, typically between about 1000 and 1073In general, metamorphism on Earth results in large water yiat low metamorphic temperatures and fluids that are progsively more enriched in CO2 at higher temperatures. The samelikely to be true in the metamorphism of chondrites and Europrocky mantle.

The total amount of water liberated by metamorphic reactiof Europa’s interior is likely to be prodigious, if the carbonceous chondrite petrologic model is any guide. CI chondrcontain more water (including hydroxyl, which is released as

in-erter during metamorphism) bound to silicate phases than is boundto salts; CM chondrites do not contain so much silicate-bound

L

r

iim

hth

nnc

a

ensc

i;n

s.

lo

p

bate

nif5urtu

th

etteorast-

ordandar-lab-va-

ates

246 KARGE

water, but it is still considerable. Hence, the late addition of laamounts of metamorphic water to the salty–icy crust and occan be chemically and geophysically very significant. Ventof carbon and sulfur gases, discussed above, occurs over smetamorphic temperature intervals. Hence, these late-stageadditions are apt to be multicomponent aqueous fluids, ware probably enriched also in silica, iron, aluminum, and oconstituents.

3.8. Alteration in the Surface Environment

Materials exposed on Europa’s surface must be affected benvironment. Although the resulting chemical alteration is osuperficial, it has importance for the interpretation of reflectaspectra. Surface pressure and temperature conditions maydehydration of salts and amorphization of ice. Europa’s tenuatmosphere does not impede ionizing ultraviolet radiationenergetic charged magnetospheric particles, thus allowingmation of free radicals and appearance of new chemical spthat otherwise would not exist. Alteration products mightbe only superficial; Carlsonet al. (1999) have suggested posible burial and endogenic recycling of exogenically produalteration products.

As mentioned before, NIMS data indicate a high hydratstate of Europa’s nonice minerals (McCordet al.1998a, 1999a,bCarlsonet al. 1999), which suggests that evaporation hasbeen very effective except perhaps in removing or reducingamount of ice. When hydrated salts were first suggested, Daand Clark (1998) countered that the low atmospheric presof the surface may induce dehydration of those mineralsresponse, several researchers (McCordet al. 1999a,b; Karge1999a; Prietoet al.1999; Zolotov and Shock 2000c) have prduced theoretical arguments and lab data indicating a stabof hydrated salts greater than that first considered by DaltonClark (1998). Nevertheless, hydrate stability and dehydrationmains an important issue. Salts appear on Earth with severadration states, depending on temperature and water vaporsure, and cycling of hydration state as relative humidity chanis sometimes observed (Kelleret al.1986). There are some pulished data on vapor pressure of both water ice and hydrsalts (Wilson 1921, Cliffordet al. 1923, Carpenter and Jet1923, Baxter and Cooper 1924, Matsuiet al.1927, Kelleret al.1986). These data and estimates of surface conditions on Eushow that the low temperature ought to inhibit the dehydratiothe hydrated salts. Accurate vapor pressure data are not avafor salts at europan temperatures, and so consideration oissue requires extrapolation of high-temperature data (Fig.

At ordinary temperatures the water in hydrated salts is bomore tightly than water molecules in ice and liquid water,sulting in vapor pressures lower than that of ice and liquid waThis situation probably persists to low temperatures, althoit is not assuredly the actual case for all salts (Fig. 5). Thepected situation, wherever ice is present on Europa, is that

should exist a local water vapor pressure that would tendcause salts to exist in their maximum hydration states. Howe

ET AL.

geeanng

ilarfluidicher

y itslyceauseousndfor-ciesot-ed

on

ottheltonureIn

-ilityandre-

l hy-res-

ges-ted

ropaof

lablethis).nde-er.gh

ex-ere

FIG. 5. Water–vapor pressures of MgSO4 · 7H2O, MgSO4 · 6H2O,Na2SO4 · 10H2O, water ice, and liquid water. (a) Data of Carpenter and J(1923) and Kelleret al. (1986). (b) Extrapolations to low temperatures; errenvelopes determined by selectively eliminating outlier points from the lesquare linear fits.

ice is not present everywhere on Europa, according to McCet al.(1998b, 1999a); the nonice terrains have a low albedo,thus are warmer than bright ice terrains. Dehydration or ptial dehydration of salts in those regions is possible. Newdata presented by McCordet al. (1999b) indicate that even under ultrahard vacuum conditions—where the slightest waterpor is continuously pumped away—magnesium sulfate hydr

to

ver,would remain metastable for millions of years at temperaturesless than 160 K, i.e., over the entire surface of Europa, except

’S

afpe

tuti

en

udwqmi

1

nam

tolt

ilc

Tv

i

rio

rip

di-ium

n in. Sul-ex-

ui-r

a-s—calses,

ionlor

lsonandiricsitsdis-

then

de ofstent

dinge IIPaice-

ust),larly.

d theh isunt

her

t-urever,ex-haseoly-

EUROPA

where heated by very shallow brine intrusions (Fagentset al.2000); natron remains hydrated for millions of years at tempatures less than 130 K, i.e., over much but not all of Europsurface; and mirabilite remains hydrated for millions of yeonly at temperatures less than 100 K. Thus, magnesium sulgenerally should remain hydrated across nearly all of Eurosurface, but mirabilite generally would dehydrate rapidly excin Europa’s polar regions.

Amorphous water ice may be produced by low-tempera(<110 K) quenching from the liquid state, by rapid condensafrom water vapor, or by severe radiation damage to crystalline(Hobbs 1974, Letoet al.1996, Delitsky and Lane 1998, Hansand McCord 2000). Water ice normally exists in its hexagoIh crystalline phase above about 150 K and as cubic Ice Ic be150 K. Europa’s surface temperature is mainly between aro80 and 130 K (depending on latitude, time of day, and albeso amorphous and cubic ice phases are possible whenis frozen on the surface. Interior temperatures and local etorial surface temperatures in low-albedo terrains are warand freezing may occur slowly in the interior, so hexagonalis also possible on Europa. Rapid cooling of brines also pduces metastable supercooled and glassy states (KargelHogenboomet al.1995). The reflectance spectra of amorphobriny solid materials are not known but may be different frothose of well-crystallized salt hydrates—e.g., absorption badue to H2O may be even broader than in the crystalline sulfhydrates, and fine structural details of the absorption bandsbe obscured.

Plasma effects on Europa’s surface, especially implantaof energetic magnetospheric sulfur ions derived originally frIo, were thought to be the cause of longitude-correlated uviolet albedo variations inVoyagerand IUE data (Laneet al.1981, Johnsonet al. 1988). Latitude-dependent variationsspectral reflectance may be associated with ultraviolet photoalterations. Plasma bombardment and ultraviolet irradiationinduce important changes in salt chemistry and stability.chemical changes caused by radiolysis and photolysis aresimilar, except that the depths of penetration of these ionizradiations differ greatly (Gerakineset al.2000). Analysis of im-proved spectra from the Hubble Space Telescope added a wrto the radiolysis/photolysis explanation of Europa’s reflectanaccording to Nollet al. (1995), who favored an internal control of SO2 venting and deposition. Hendrix (1998) and Hendet al.(1998), using Galileo UVS spectra, confirm a combinatof exogenic and geologic controls on the distribution of UV asorbers but interpret Europa’s UV absorptions as due to a mcomplex set of volatiles that include S–O compounds. The Scompounds may include the sulfates inferred by McCordet al.(1998a) and do not necessarily include SO2.

Electron bombardment at europan doses effectively destsalts of singly valent alkali metals, e.g., sodium and potasssulfates; hydroxides, free metals, and sulfur dioxide are

duced in place of sulfates (Orlandoet al.1999). Electron excita-tion is orders of magnitude less capable of generating equiva

OCEAN 247

er-a’srsatesa’spt

reonicenal

lownd

o),aterua-er,cero-991,usmdsteay

ionmra-

nyticanheery

ing

nklece,-ixn

b-ore–O

oysumro-

changes in salts of doubly valent magnesium and calcium (ibid.).The relative vulnerability of sodium and potassium salts to raolytic destruction is cited as the explanation for Europa’s sodand potassium atmosphere (Brown 1999).

Plasma and UV radiation also may cause amorphizatiowater ice, and doubtless radiation affects hydrated salts alsofuric acid hydrate, according to one interpretation, owes theiristence to a radiolytic sulfur cycle (Carlsonet al.1999). The aciditself is also affected by radiation, but the yields favor a high eqlibrium abundance of H2SO4 and minor amounts of many othespecies, including O, O2, O−2 , H2O2, OH, HO2, SO, S(OH)2,SO2, SO3, and Sx (Carlson 2000). Besides producing disloctions and color centers in ice, radiation acting on sulfatewhether sulfuric acid or sulfate salts—may produce free radiand result in formation of elemental sulfur, polysulfur oxidand other colored substances (Carlsonet al.1999). Short-chainelemental sulfur and polysulfur oxides produced by irradiatof sulfur-bearing ice is the cause of the reddish/brown coof nonice hydrated terrains on Europa, according to Caret al.(1999). The high concentration of colored compoundshydrated material along fractures, rifts, craters, shallow diapstructures, cryomagmatic intrusions, and cryovolcanic depois best explained by endogenic mechanisms that control thetribution of salts and other sulfurous compounds that canyield the colored photolytic and radiolytic products.

3.9. Influences of Elevated Pressure on Phase Equilibriaand Brine Evolution

The pressure at the base of a frozen Europan crust mapure ice or a pure water ocean some 70–170 km thick, consiwith theGalileo gravity data (Andersonet al.1998), would bebetween 84 and 205 MPa. The higher pressure (corresponto the thickest possible ice crust) is comparable to the ice I/icor ice I/ice III transition pressure, which is around 207 M(Bridgman 1912). The pressure at the base of Europa’ssaturated layer (the ice crust or frozen eutectic/peritectic craccording to the models in Fig. 3, varies with the particumodel, but it is around 140 MPa if this layer is frozen completeWe do not presume that our chemical models or theGalileograv-ity models are so well constrained that a higher pressure anpossible existence of a stable high-pressure ice polymorpruled out, but this seems unlikely unless it is a small amoat the very bottom of the ice layer. The system H2O–MgSO4

does not exhibit high-pressure phase transitions until far higpressures are attained (Bridgman 1948a,b; Livshitset al.1963;Hogenboomet al.1995; Sotinet al.1999). Thus, it seems doubful that Europa’s icy shell would contain stable high-presspolymorphs of ice or magnesium sulfate hydrates. Howepressures in the deep interior of Europa would have greatlyceeded the minimum pressures required for high-pressure ptransitions in solid salt hydrates. Hence, high-pressure p

lentmorphs may have been important during Europa’s primordialdifferentiation. As yet there are not enough high-pressure phase

L

t

s

x

mon

ey

suiit

gloi

y

(v

sgn

on-

kmand

in at

areturesf

over

s ofk-

ans-in-ergs-

-staleveel,

esthussim-

g the

anderalvan

ru-go-ngthlardopartundand

res

dsece;in

ar

248 KARGE

equilibrium data to make useful constraints on the effectsthis may have had on the course of differentiation.

High-pressure polymorphs of ice and salt hydrates might emetastably in impact crater ejecta due to shock effects, muchigh-pressure shocked silica phases occur in terrestrial imcraters. In general, the lower pressures required for solid-phase transitions in ice and salt hydrates compared to silicmeans that impact-shocked high-pressure phases might bmore abundant on Europa than on Earth.

Although high-pressure crustal phases of ice and saltsnot likely to exist stably on Europa today, the pressures eing within Europa’s crust are sufficient to induce considerachanges in solid–liquid phase equilibria. Even in Earth’s ocewhere maximum pressures barely exceed 100 MPa andtypical seafloor pressures are half that, pressure has a contrinfluence on carbonate equilibria. Calcium carbonate is strosupersaturated in surface water of Earth’s oceans but stroundersaturated in deep ocean water (Anderson and Mala1977), a fact that has profound consequences for the dedependent mineralogy of marine sediments (Ackeret al.1987).Hogenboomet al. (1995, 1999) found large pressure-inducshifts in the eutectic liquid compositions for the binary stems H2O–MgSO4 and H2O–Na2SO4; they predicted that theternary system H2O–MgSO4–Na2SO4 would exhibit a eutecticwith a Na : Mg ratio increasing rapidly with pressure at leastto 200 MPa. If so, this would have consequences for the intechemical structure of Europa’s crust and ocean and for itsface composition. For instance, ternary eutectic brines prodnear the base of Europa’s crust would possess Na : Mg ratto 3 times greater than those of low-pressure eutectic br(Hogenboomet al.1999). Such effects are not considered inmodels presented here.

4. GEOLOGIC IMPLICATIONS

The high-resolution image data obtained byGalileo for manydifferent areas on the surface of Europa reveal a host oflogical features and structures. High-spatial-resolution andspectral-resolution imaging data is complemented by hspectral-resolution and low-spatial-resolution NIMS daTogether, this information can be used to constrain the proceoperating in the crust of Europa and to distinguish amongseveral plausible paths for the evolution of the outer ice laThis information is twofold. First, such features provide insiginto the nature of processes in the recent geological pastdoes tectonism indicate expansion, does volcanism involbuoyant and/or volatile phase, do circular features represenapirism, does chaos involve a very thin brittle layer overlyingocean, are impurities in the upper crust associated with speprocesses?). Secondly, the sequence of these features andtures can provide information on the evolution of these proce(e.g., is the crust thickening with time, has there been a chanstyle of convection and resurfacing with time, is there evideof compositional evolution?). One purpose of this paper is to

the stage for further detailed discussions of these issues.

ET AL.

hat

xisth aspacttateatese far

areist-blean,orellingglynglyhoffpth-

ds-

uprnalur-ced

os 2neshe

eo-w-

gh-ta.ssestheer.hte.g.,e at di-ancificstruc-sese inceset

Initial results fromGalileo(Greeleyet al.1998a,b; Headet al.1999a) reveal the following features and stratigraphic relatiships:

(1) Lenticulae (elliptical pits, dark spots, and domes 7–15in diameter and spaced about 5–20 km apart) are commonmake up much of the texture that characterized mottled terraVoyagerresolution. Diapirism (Pappalardoet al.1998), as wellas melt-through of a very thin ice layer (Greenberget al.1998),have been proposed to explain these features. Lenticulaeobserved to cross-cut most other geologic features and struc(e.g., Spaunet al.1999, Prockteret al.1999a) and are some othe youngest features on Europa.

(2) Chaos occurs as irregularly shaped regions tens toa hundred kilometers in diameter (e.g., Carret al.1998, Spaunet al. 1998, Greenberget al. 1999, Kortzet al. 2000), formedby the destruction of preexisting lineated plains, and consista distinctive fine-textured matrix and polygonal blocks of bacground lineated plains that have been rotated, tilted, and trlated. Two end-member models for the formation of chaosvolve crustal melt-through from an underlying ocean (Greenbet al.1999) or disaggregation of a thin, brittle lithosphere by acending solid-state warm-ice diapirs (Pappalardoet al. 1998).According to Collinset al.(2000), a liquid or partly molten substrate is probably required to form the chaos features. The crumelt-through model appears to lack sufficient energy to achimelt-through according to the conventional melt-through modsuch that partially molten diapirs may be required (Collinset al.2000). Most chaos regions modify virtually all other featur(only a few post-chaos grooves have been observed), andis one of the youngest feature types on Europa, apparentlyilar in age to lenticulae. Greenberget al.(2000a) interpret somedegraded chaos regions as much older than others, raisinpossibility of a more complex or recurrent geological cycle.

(3) Bands and wedges are regions in which separationtotal replacement of background plains have occurred by latmovement in a manner similar to seafloor spreading (Sulliet al.1998, Prockteret al.1999b, Tuftset al.2000); strike-slipmovment is common (Hoppaet al. 1999, Tufts 1996, Prockteet al. 2000), and internal structure is often related to distribtion of stress due to accomodation to irregular and nonorthonal separation. Recent evidence indicates that long-wavelecompressional folds affect some bands (Prockter and Pappa2000). Bands and wedges often occur in the intermediateof exposed europan history, cutting and separating backgroplains but being crossed by ridges and disrupted by chaoslenticulae (e.g., Prockteret al.1999a).

(4) Ridges consist of several different types of featu(Greeleyet al.1998b, Greenberget al.1998, Pappalardoet al.1998, Headet al.1999b) that extend along-strike for hundreof kilometers, show different orientations as a function of timin individual areas, and may indicate an evolutionary sequentheories of origin for the positive topography (see reviewPappalardoet al.1999) include constructional volcanism, line

diapirism, flexure, and intrusion-related deformation. Associated

’S

viaid

ly-ain

nmuh-

ete

eoos

iTa

io

s

n

a

io

wo

c

fol-insaosism.

ddssting

heo-andannotiquidinrac-y beds

weded).

ouldd ontive?

seenpro-testatas-thispa’srittleposi-

tureandryo-elythe

ust

si-ar-uldce.

ay bete

od-ick-thick-rsityhe-thato we

antustal/

EUROPA

reddish deposits have been explained by rift-controlled cryocanism, by brine mobilization, or by metamorphic dehydratrelated to heating due to shallow brine intrusions. A reviewnew evaluation of eruptive mechanisms was recently provby Fagentset al. (2000).

(5) The morphology of cryovolcanic flows varies wideSmooth plains (Greeleyet al.1998b) consist of relatively lowalbedo regions appearing to be either emplaced in a fluid fion (Thomas and Wilson 2000) or partly mantling preexistterrain (e.g., Headet al. 1998). In part these appear to be crovolcanic in origin and are observed to postdate plains, bawedges, and ridges; at least in one area smooth plains areified by (hence, predate) chaos and lenticulae. One pecmitten-shaped lenticula- or chaos-like feature appears toa cryovolcanic/upwelling origin involving flow of a very viscous liquid or partly molten slurry that formed a towering etruded mass comparable to venusian pancake domes and ttrial dacite lava domes (Figueredoet al.2000). The very fact thamost of Europa’s surface is criss-crossed by tectonic lineamtraceable for hundreds to thousands of kilometers shows thacryovolcanic flooding has not been a dominant process sinctime of formation of Europa’s icy shell (2) or most original cryvolcanic surfaces have been thoroughly overprinted by tectfeatures. The relative rarity of cryovolcanism probably attestthe buoyancy of the icy shell.

(6) Background ridged plains consist of terrain highly drupted by younger ridges, wedges, lenticulae, and chaos.occur virtually everywhere and form a background of terrcomposed of bands, wedges, ridges, and other features inious states of degradation. These plains are the oldest unthe stratigraphic sequence in each area mapped. Unknowwhether background ridged plains define a specific and validologic unit or whether they simply represent our inability to dconvolve the complex record in the early history of each reg

(7) Impact craters, which are relatively uncommon on Eur(Mooreet al. 1998), can serve as stratigraphic markers (Pwand its extensive rays) and as probes to the interior (TyreMannanan). Pwyll rays are superposed on Conamara Chaono examples of highly dissected impact craters have yet breported, although Mannanan deposits are cut by several rid

(8) Geologic units: Mottled terrain appears to owe its disticolor and texture to disruption and discoloration accompanythe formation of chaos and lenticulae at the expense of bround ridged plains (Greeleyet al.1998b, Greenberget al.1999,Headet al.1999a, Pappalardoet al.1999).

(9) Surface age: Impact crater size–frequency distributhave been interpreted in two very different ways such thatsurface of Europa is typically 0.7–2.8 Gyr (Neukumet al.1998)or <0.1 Gyr (Zahnleet al. 1998). On the basis of the reviein Pappalardoet al. (1998) the latter interpretation seems mlikely.

In summary, these features, their stratigrapic realtionshand their ages, indicate that the crust of Europa has been

pletely resurfaced in the recent geological history, and that

OCEAN 249

ol-onnded

.

sh-g

y-ds,od-

liarave

x-rres-

ntst (1)the-nicto

s-heyinvar-it inn isge-e-on.payllandand

eenges.ctingck-

nsthe

st

ips,om-

period appears to consist of three phases, during which thelowing types of activity dominated: (1) background ridged plaformation, (2) band, wedge, and ridge formation, and (3) chand lenticulae formation, with some associated cryovolcanInterpretation of a large plains area by Greenberget al.(2000a)as an ancient region of heavily modified chaos, if correct, aa further wrinkle to this three-phase history, perhaps suggethat remnants of a prior cycle of resurfacing may exist.

Surface geologic features, their interpreted origins and rlogical implications, strongly suggest a thin ice lithospherea steep near-surface thermal gradient, but at present we cdistinguish between the presence of (a) a near-surface llayer and (b) warm, ductile ice and only local partial meltinga thicker layer. In addition, the apparent change in the chateristics of surface features with time suggests that there maevidence for thickening of the icy shell with time (e.g., banand wedges during times of earlier, thinner lithosphere, folloby isolated fractures, and by diapirism as the layer thickenIn addition, or alternatively, heat loss and layer thickness cvary regionally. Present geological analyses are thus focuseseveral questions: (1) Are these three stages indeed distinc(2) If so, how much overlap occurs? (3) Is this sequenceglobally? (4) Do such stages represent a linear geologicgression, or could there be repetitive cycles? (5) Did the laprogression of events, and perhaps each cycle, begin with ctrophic global resurfacing, and if so, what is the nature ofprocess? (6) What occurred during the first 50–98% of Eurocrustal history? (7) Does the apparent thinness of the outer blayer imply a present ocean, and what is its depth and comtion today?

How do our models link to these observations and what fuwork might be productive in further establishing these linksdistinguishing among evolutionary models? Evidence of cvolcanic activity is not common on the surface; this can largbe accounted for by the buoyancy of solid water ice anddifficulty of erupting eutectic brines through eutectic solid crwithout involving exsolved gases (Wilsonet al. 1997, Fagentset al. 2000). But some models may favor production of potively buoyant magma. For example, in model A5 (Fig. 3), ptial remelting of the bottom of the pile of hydrated salts coform a eutectic brine that is buoyant all the way to the surfaThe absence of common and widespread cryovolcanism man argumentagainstthis specific model or may simply indicathat the solidus temperature is rarely achieved.

Regional variation in Europa’s geologic styles of surface mification might suggest lateral variations in heatflow and/or thness of the frozen layer. Stevenson (2000) argued that suchness variations could not be very significant due to the propeof ice to flow glacier-style. However, he did not consider rological effects of an impure crust or of an impure oceanwould cause colder temperatures and a stiffer ice rheology, sconsider lateral thickness variations to be an open issue.

Our models support the possible formation of abundclathrate hydrates and gas-charged brines due to subcr

thissuboceanic venting of CO2, SO2, and hydrocarbons. Stevenson

lsss

4b

fnhthee

i

r

ttlta

ua

oid

h

is

s

h

b

williginceslelow-sity,

esets

riplei-ed

owege-ithtwoolid-ice

ing

rustuoy-nti-eco-rateticrbonpiricplexate-Gulf97;w,

tedflu-sidet of

onbonsdro-nt ofrgelr onll.

f thet theicezensionor

on-. Ifratesith

250 KARGEL

(1982), Crawford and Stevenson (1988), and Fagentset al.(2000)have considered ways in which clathrates might interact withuid water on Europa and other icy satellites to produce explocryovolcanic eruptions. Might clathrates be a common crumaterial on both Mars and Europa (and also certain areaEarth), and thereby play similar geologic roles in the formtion of a wide variety of features, including chaos (Milton 197Kargelet al.2000)? What similarities and differences wouldexhibited by clathrate-related processes on Europa, Mars,Earth?

Another aspect of our modeling that merits further investigtion is the degree of homogeneity of Europa’s surface salt assblages. Early interpretations (McCordet al.1998a, Fanaleet al.1999) emphasized the apparent homogeneity as deducedlack of much variation in the shape and depth of hydrate bain NIMS spectra of nonice regions. More recent analysispointed out possible subtle variations in the fine structurecould relate to compositional differences. Is the salty surfacEuropa representative of a homogeneous peritectic or eutinterior layer, or do compositional gradations exist that migrelate to compositional brine evolution produced by progressfreezing of an ocean or of briny diapirs? The possible formatof metastable salt phases of low hydration state (Kargel 19Prietoet al.2000) allows for possible subsequent hydrationactions and attendant volume changes that may drive crutectonics (Hogenboomet al. 1997). How might such volumechanges be reflected in the geology of Europa?

There are likely to be several mechanisms for the segretion of ice and salt phases and the production of interesfeatures and structures. For example, fractional crystallizaof an ocean or of smaller intrusive brine bodies could invo“squeeze-out” of interstitial brines during crystallization. Atendant process might be metamorphic recrystallization (angous to formation of mineral bands in gneiss and formationcoarse crystalline glacier ice from snow) or boiling of brine erutions with recondensation of water vapor. Heating could cabrine mobilization, collapse of near-surface ice (e.g., HeadPappalardo 1999), and dehydration of surface and interiorposits (Fagentset al.2000). Vapor–solid fractionation could alsoccur, with long-term, low-temperature evaporation producevaporitic salt accumulations in some areas and frost reconsation elsewhere.

Solid-state diapirism might accompany the evolution of tcrust. This might involve the ascent of segregated ice anddescent of salts. Further development of our models can ypredictions on the sizes of ice masses or salt layer thickneand predictions about the development of Rayleigh–Taylorstabilities in these layers on geologic time scales. Predictionvariations in temperature and composition will provide impotant information in which to assess observational data anddistinguish among the several models (e.g., Pappalardoet al.1998, Rathbunet al.1998, Spaunet al.1999, McKinnon 1999).In addition, crustal thermal and compositional layering will cetainly influence the style of convection and whether it might

layered or inhibited, for example. Compositional informatio

ET AL.

iq-ivetalof

a-,

eand

a-em-

romdsasatofctichtiveon91,e-stal

ga-ingionve-lo-ofp-sendde-

ngen-

etheeldsesin-

ofr-elp

r-e

predicted by models and interpreted from measurementsalso have a major influence on distinguishing among the orof features and different paths of evolution. For example, traof sulfuric acid occurring as a solid solution within ice singcrystals (even less than 11.5 ppm) can induce a drasticering of the power-law stress exponent, the effective viscoand other rheologic properties (e.g., Bakeret al.1999, Trickettet al. 1999). Many other naturally occurring trace impuritihave similar softening effects on ice. In Earth’s polar ice shesulfuric acid and salts accumulate at grain boundaries and tjunctions (Fukazawaet al. 1998), where the impurities stablize low-temperature liquid films. These films permit enhancmigration of water molecules and theoretically should allincreased rates of strain recrystallization, thus increasing rlation and flow rates. Thus, soluble impurities (consistent wthose predicted by our models) can soften ice by at leastdistinct mechanisms. This is expected to cause enhanced “sstate” deformation rates, making diapirism and solid-stateextrusions more likely and more rapid than without softendue to soluble impurities.

Many gases expected to percolate from the mantle to the care clathrate-hydrate-forming substances. On Earth, the bancy of hydrocarbon gases results in their accumulation in aclinal traps created by salt diapirs, an aspect that has greatnomic importance. In seafloor sediments methane-rich clathhydrates commonly form by the combination of high hydrostapressures, low temperatures, high concentrations of hydrocagases, and existence of abundant water (Sloan 1998). Diastructures, mud volcanoes, and chaotic terrains with comfaulting, slump, and slide structures are typical of some clathrrich seafloor terrains on Earth, such as those in the northernof Mexico (Roberts and Aharon 1994; Roberts and Carney 19Sassenet al.1998, 1999a). A spectacular glacier-like mass flosome 1500 km2, on the Mediterranean seafloor, might be relaeither to deformation of clathrate hydrate or to sedimentidization related to breakdown of the hydrate phase (Woodet al. 1998). Extrusion of clathrate hydrates or developmenclathrate mounds around sites of fluid venting are not uncommoccurrences in some areas of the seafloor where hydrocarvent to the deep, cold sea floor. Clathrate hydrates and/or hycarbon venting also appear to be associated with developmepingo-like broad-based mounds in terrestrial permafrost (Kaand Lunine 1998). Similar processes and features may occuEuropa’s seafloor and perhaps on the surface of the icy she

On Europa the buoyancy of free hydrocarbon gases or ohydrate phases could result in their concentration in traps alower surfaces of the icy shell, in anticlinal traps created byplutons, salt hydrate domes, or fold structures within the frocrust, or in subcrust beneath the ocean. Alternatively, ascenof gases through fractures may result in formation of dikesveins of gas hydrates cutting through the icy shell, possibly ctributing to the formation of structures such as triple bandsgases are trapped at the base of the icy shell, then gas hydmay develop a lower chemical layer of a stable icy shell, w

nordinary ice I dominating the upper layer. However, if the icy

S

s

da

ocvhhin

ersm)

n

ma

in

a

nra

ct

n

es

eoa

rtialue,ticAl-od-ncy

ded

de asesed

uterays

lifed byhur,

a-eseistedlso

evo-ase

dis-arthand

withatese IIthex-

driveoundportov-sur-

lm-6c).sulfur

ford

erials for

EUROPA’

shell is enriched in salts, then gas hydrate diapirism is poble. Methane hydrate has a low density, 910 kg m−3 at 273 K(Davidson 1983), so that hydrate diapirism could occur reain the icy shell. If the hydrate phase contains a significant frtion of high molecular weight gases such as CO2 and SO2, thenthe hydrate has a much higher density (e.g.,∼1100 kg m−3 forpure CO2 hydrate; Davidson 1983), so that diapirism involvinthese hydrates should be inhibited or may occur only in arof the icy shell that are composed of extremely salty ice, e.geutectic ice–salt mixture. Decomposition of clathrates becaof crustal heating or depressurization in ascending diapirs cresult in enormous volume changes, attendant formation oflapse structures, and possible explosive cryovolcanic actiEscape and evaporation of hydrocarbon gases near clatdomes may produce residues of heavy hydrocarbons, wcould yield detectable organic tholins by solar UV radiolysIt is possible that such processes contribute to the formatiolenticulae (pits, spots, domes, and microchaos).

In this and all prior work it has been assumed that anyropan ocean is homogeneous and in thermodynamic equilibthroughout—not an implausible situation, but also not necesily correct. Experience with terrestrial surface brines in warand cold-climate desert playas (Last and Schweyen 1983ice-covered Antarctic lakes (Toriiet al.1975), in coastal marinesabkhas (Brantleyet al. 1984), and in stratified brine pools othe seafloor (Wallmannet al. 1997) indicates that inhomogeneous brines are very common. Strong vertical gradientsdevelop in salinity, anion and cation ratios, temperature,density. Stratified brine pools can be amazingly stable agamixing, because normal thermally driven convection is inhited by density stratification. Heat transfer and compositiohomogenization in such situations may occur by diffusionby forced convection due to tidal or wind-driven water curents or crystal flotation/settling. Since heat input to the bof the brine pool may initially exceed the rate at which heat cbe transferred upward, a strong vertical temperature gradiecommonly achieved. In situations where the bottom of the bpool is covered by salts, any rise in the temperature of the blayer generally causes further dissolution, thus accentuatingdensity stratification. In some ice-covered and ice-free Antarlakes of the Dry Valleys, heating is by solar radiation andgeothermal heat flux; temperature characteristically rises wdepth—the dense bottom water of Lake Vanda reaches 29(Torii et al.1975). Density-stratified brine lakes in Africa, sucas Lake Nyos, accumulate volcanogenic CO2 to extreme levels.Destabilization and degassing of these stratified lakes cacatastrophic, with exceedingly violent and deadly limnic erutions (Ladbury 1996).

On Europa, the possibility exists for far more extreme degrof density and thermal stratification. Any process that can debilize Europa’s oceanic brine column—especially if it produca strong convective plume or catastrophic turnover—could lto a heat pulse at Europa’s surface or at the base of the frcrust. Destabilizing processes could include lithospheric fr

turing and sudden local degassing, comet impacts, or ab

OCEAN 251

si-

ilyc-

geas., auseuldol-ity.rateichs.of

u-iumar--

, in

-ayndinstb-alorr-seant is

inesalthetic

heith8 Kh

bep-

esta-

esadzenc-

termination of heating at the bottom of the ocean and pacrystallization of bottom water. Geologic activity could enspossibly including solid-state convective diapirism, cryoclaseruptions, crustal melt-through, or even global resurfacing.though the conventional cold-ocean crustal melt-through mels of Europa’s chaos lack the thermal energy transfer efficieneeded to achieve melt-through (Collinset al. 2000), a high-temperature density-stratified ocean might supply the neeenergy.

In summary, the concepts and models offered here proviframework in which to consider chemical and physical procesinvolved in the formation of surface features. The continuanalysis of data from theGalileo mission will help to chooseamong the several evolutionary pathways for Europa’s olayers outlined here or to develop new concepts and pathwthat may better match the observations.

5. ASTROBIOLOGICAL IMPLICATIONS

Speculation that Europa’s putative ocean might harboror prebiotic conditions has increased since first suggesteConsolmagno (1975) and developed in fictional form by ArtC. Clarke (Reynoldset al. 1983, Lunine and Lorenz 1997Jakosky and Shock 1998, Chybaet al.1999, Gaidoset al.1999,McCollom 1999, Chyba 2000). Life on Earth requires liquid wter, chemical disequilibria, and elemental building blocks; thstipulations may indeed be met on Europa and may have excontinuously or sporadically since Europa’s origin. Europa aprovides insights regarding the geological and biochemicallution of early Earth during a hypothetical ice-covered ph(Bada 1996, Gaidos,et al.1999).

Although hydrocarbons are generally considered to be apersed biologic resource in ice-covered oceans on E(Gaidoset al.1999), in some cases it appears that microbialmacrofaunal chemosynthetic assemblages are associatedlocalized hydrocarbon vents and concomitant clathrate hydr(Carney 1994). Molecular and isotopic studies of structurgas hydrates in lightless chemosynthetic communities ofGulf of Mexico provide evidence that bacteria can directly oidize hydrate-bound methane (Sassenet al. 1999b). Moreover,hydrocarbons from decomposing gas hydrates appear tocomplex biogeochemical processes in sediments that surrstructure II gas hydrate extrusion features, helping to supcomplex chemosynthetic microorganisms (Fig. 6). The discery of high concentrations of a new polychaete species on theface of exposed gas hydrate in the Gulf of Mexico at∼540 m wa-ter depth (Sassenet al.1999b, Fisheret al.2000) has additionaimplications with respect to direct association of relatively coplex life (annelids), gas hydrate, and saturated brines (Fig.

Possible metabolic processes, such as methanogenesis,reduction, and iron oxide reduction, have been suggestedEuropa (Gaidoset al. 1999, Chyba 2000) and Mars (Fisk anGiovannoni 1999). Excreted byproducts and decaying matfrom chemoautotrophic organisms could serve as the basi

rupta europan ecosystem by supplying carbon and energy sources

yn

y

ado

w

rv

t

c

rsenm

h

leiar

sehHiods

te

h

f

lugi

c-oseans

, sots.nr-

n-and

uchth’sar

notepndible

s

tsco-mi-thed

r ex-orts

esn-

eenulf

ishto

inman,a

n toalt)erdes

rob-fu-en-

r-

ly,

252 KARGEL

to heterotrophic life forms at higher trophic levels. Photosthesis is unnecessary in this scenario. The potential amoubiomass on Europa has been calculated to be low (JakoskyShock 1998), and Gaidoset al.(1999) have reinforced this viewon the basis of a perceived deficiency of oxidants. Our analabove suggests otherwise. The availability of chemical eneneeded to drive development of a biomass may be far grethan that considered by Jakosky and Shock (1998) accorto some geologic interpretations of Europa. The calculatiabove also allow for the possibility that organic carbon maymuch more abundant than that on Earth, such that a dense,developed ecosystem might exist. Even in the unlikely abseof endogenic sources of organic carbon, impacts can deliveganics, including amino acids, a fraction of which will surviimpact and be retained by Europa (Pierazzo and Chyba 20A europan ocean may or may not support a lush ecosys(or any ecosystem), but it is illustrative to examine the phyological stresses that past or present europan organismsexperience.

The geology of Europa indicates that liquid is probably psent and widely distributed; i.e., an ocean probably does exian ocean exists one thing is certain and another extremely likfirst, the top of any ocean must be in equilibrium with ice; secosolute abundances in the ocean are almost certainly extrehigh and saturated, suggesting that the seafloor should beered with precipitated salts. Thermodynamic disequilibria agradients comparable to those in Earth’s ocean and seafloorexist on Europa and may be exploited as energy sourcesprovide varied niches for the origin and evolution of life. Tocean’s temperature at the base of the frozen layer is constraby these features to lie between about 216 and 266 K. Theact temperature depends on the pressure and solutes; localtemperature could be much warmer near a hydrothermal vThe pressure—up to about 200 MPa—depends on the thness of the icy shell. Other conditions in Europa’s oceanpoorly known: the ocean’s pH may be extremely acidic, neutor extremely alkaline; concentrations of organic carbon mayextremely low or high; the seafloor could be made of materialdiverse as basalt, elemental sulfur, or salts; the types of solutEuropa’s ocean are uncertain. There is some basis to favorsulfate concentrations, but the chief cations are debated.flow on Europa due to tidal dissipation may be a small fractof or somewhat greater than Earth’s total heat flow. High tiamplitudes indicated by Europa’s pattern of fractures (Geiset al.1998a, Hoppa and Tufts 1999) and suggested theoretic(Greenberget al.1998) would drive tidal currents and circulanutrients (Greenberget al.2000b).

Where europan and terrestrial environments overlap in tphysicochemical nature, examples from Earth reveal thatcan cope and grow under conditions as extreme as most orange of conditions modeled for Europa’s ocean. Possibleropan conditions which are not represented on Earth incfar colder temperatures (according to some models) and hipressures, but neither of these extremes is inherently prohib

to life as long as liquid water is present to allow organisms

ET AL.

n-t ofand

sisrgyteringnsbeell-

nceor-e00).emsi-ould

e-t. Ifly:d,ely

cov-ndmayandeinedex-y thent.

ck-re

al,beass inigheatnallerally

eirlifethe

eu-dehertive

grow. Diffusion rates will be slow at cold temperatures, and acordingly life processes may proceed sluggishly relative to thon Earth. Sea ice, hydrothermal systems, and evaporitic salt pon Earth may provide direct or partial analogues to Europawe consider briefly some aspects of life in these environmen

A leading hypothesis for the origin of Earth’s life, based ogenetic classifications of microbial forms pointing to a hypethermophilic ancestor to all life, is that it occurred in an evironment such as a deep-sea hydrothermal vent (BarossHoffman 1985, Baross and Holden 1996, Baross 1998). Senvironments not only provide the conditions needed by Earmost primitive existing life forms to thrive, but they also appeto provide a range of gases and catalysts, such as H2S, CO, and(Ni,Fe)S particles, that are effective in the activation of amiacids and formation of peptides, which may be a critical sneeded for life’s origin according to one hypothesis (Huber aWachterschauser 1998). Since similar conditions are posson Europa, the same hypothesized origin of life may apply.

5.1. Salt, pH, and Metal Tolerances of Terrestrial Organism

Halotolerant microorganisms thrive in all briny environmenon Earth, including hypersaline lakes, soils, playas, marineastal lagoons, and marine sabkhas (Javor 1989). Halophiliccrobes even require high concentrations of salt for grow(Ventosaet al. 1998). These ecosystems, frequently composof relatively few species (Benllochet al. 1995), are productivedespite being saturated in various salts. The Dead Sea, foample, is a hypersaline lake saturated with halite, and it suppthriving populations of heterotrophic archaea (HalococcusandHalobacterium), cyanobacteria, and sulfate-reducing microb(Ventosaet al.1999, Evans and Kirkland 1988). One cubic cetimeter of Dead Sea brine contains about 107 red halophilicarchaea and 4× 104 red algae (Sonnenfeld 1984).

Comparable bacterial and cyanobacterial densities have bfound in hypersaline marine bodies along the coasts of the Gof Elat and Gulf of Suez. One hypersaline body, the GavSabkha, has a lateral salt gradient ranging from about 5>35% (w/v); yet microorganisms thrive over this entire rangea succession of salinity-dependent sedimentary zones (Friedand Krumbein 1995, Gerdeset al.1985a, Ehrlich and Dor 1985Kesselet al.1985). The biological community of Gavish Sabkhincludes 30 species of halotolerant cyanobacteria in additioother microbes and a wide variety of halotolerant (up to 14% smetazoans: crustaceans, nematodes, and many others (Get al.1985b).

Extremes in pH pose adaptational challenges but are not plematic in terrestrial ecosystems. Prolific communities of sulric acid-producing bacteria are found in acid mine drainagevironments with a pH of 0.3–0.7 (Schrenket al.1998), and onearchaeon in pure culture can even grow at pH−0.06 (Schleperet al. 1995), equivalent to 1.2 M HCl solution. This hypeacidophile species,Picrophilus oshimae, is also moderately ther-mophilic and occurs naturally in solfatara fields. Remarkab

toit grows optimally at pH 0.7 and 333 K.P. oshimaeappears

S

il

n

ea

ra

teia

p

teaolslina

du

dlc

ip

aa

iticlly

panet-edalso

ex-e

ina-

eldia,nvi-

ithavyrobeat

4).tsarsrdi-eshend

the

reshees.n-di-

ta-deich-

3 K,e

mod-r in0).

ry

mendces

edlyort

EUROPA’

to be phylogenetically related toThermoplasma acidophilaandother prokaryotes known from burning coal heaps and Yellostone acidic thermal vents (Darlandet al. 1970). Microbes un-doubtedly reside in a peculiar acid chloride hypersaline br(pH 3–4) in Lake Tyrrell and some other southeastern Austrasalt lakes and playas (Longet al.1992a,b). It is not known howthe most extremely acidophilic organisms cope; Schleperet al.(1995) speculate that either they possess a strong proton puma low proton permeability. Alkaline waters also contain thrivimicrobial communities; one of the most basic extremes knoto allow growth is pH 12.5 (Duckworthet al.1996).

Microbial activity is also robust inside Palaeozoic hydrothmal metallic ores of the Tinto River basin, southwestern SpThis activity has driven the formation of a complex geomcrobiological system that lowers the pH of solutions betwe1.7 and 3.5 along the river up to its mouth. The major eneand mass transfers in the system involve sulfide oxidationmobilization of Fe3+ in acidic solution (Fernandez-Remolaet al.2000).

Viable microbial life at the limit of salt saturation—even inside halite and mirabilite crystals (Wei-Dongyanet al.1998)—isa geologically ancient adaptation. Microbial stromatolibuilding communities were often formed in association whalite and imply high levels of ancient biological productionsalinities exceeding halite saturation (Friedman 1980). Evaitic rocks, including halite and potash salts, some over 500 Mold, normally contain fossilized microbes (Knoll 1985). Sulfareducing microbial activity is belived to have been an importpart of Earth’s sulfur cycle at least back to the Early Proteroz(2.1–2.5 Ga), when such activity appears either to have evoor to have finally overwhelmed abiogenic means of sulfer itope fractionation (Farquharet al.2000). More recent microbiaactivity is implicated in the formation of elemental sulfurthe sulfate caprocks of salt domes along the U.S. Gulf CoBiocatalyzed reduction of evaporitic sulfates byDesulfovibrioprobably occurs first by reduction to H2S and then by oxidationof H2S to native sulfur caused by remaining sulfate (Feely aKulp 1957, Sassenet al.1989).

Additional brine environments on Earth include those in sice (Bowman et al. 1997), oceanic crust (Bischoff anRosenbauer 1989), and continental crust (Vreeland and H1991). In sea ice, temperatures usually vary between 258272 K, and salt concentrations span 0.1–15% (w/v) totalsolved salt; these habitats harbor diverse assemblages of haerant bacteria (Bowmanet al.1997). Halotolerant and halophilimicroorganisms have also been cultured from coastal oil-wand terrestrial subsurface brines (Huuet al. 1999, Orphanetal. 1998, Adkinset al.1993, Vreeland and Huval 1991). Brinethat exist in the oceanic crust beneath the global mid-ocean-rsystem (Bischoff and Rosenbauer 1989) have likewise beenposed as a suitable habitat for halotolerant microbes (KayeBaross 1998, 2000). Extremely halotolerant microbes havebeen found in low-temperature Antarctic hypersaline lakessubglacial environments (Franzmannet al. 1987, Franzmann

1991, Jameset al.1990).

OCEAN 253

w-

neian

p orgwn

r-in.i-engyndr

-

-tht

or-yr-ntic

vedo-

st.

nd

ea

valandis-otol-

ell

sdgero-

andlsond

The characteristic high abundance of nutrients in evaporsettings often comes at the price of having to cope with normatoxic levels of heavy metals (Lyons and Gaudette 1985). Eurobrines are likely to be comparably enriched in toxic heavy mals. Millimolar levels of cadmium and cobalt are easily toleratby some prokaryotes, and frequently these organisms arehalotolerant (Ventosaet al.1998).Pseudomonas aeruginosa, adeep-sea vent microorganism, can precipitate and removetraordinary levels of soluble cadmium in a short period of tim(Wanget al. 1997). Adaptations to heavy-metal-rich brinesevaporative lakes are thought to underlie the biogenic formtion of certain types of metal sulfide ore deposits (Sonnenf1984, Eugster 1985, Krumbein 1985). Lake Tyrrell, Australis viewed as a modern analogue for ancient depositional eronments of stratiform metal sulfide ore deposits (Feganet al.1992, Giblin and Dickson 1992). Some species associated wformation of organometallic complexes and sedimentary hemetal sulfide ore deposits, e.g., the sulfate-reducing anaeDesulfovibrio, occur in hypersaline marine brines such as thof Gavish Sabkha (Schidlowskiet al.1985), in the sulfur–sulfidecaprocks of salt domes, and in oil-field brines (Sonnenfeld 198

Hypersaline, acidic, alkaline, and metal-rich environmenharbor tenacious microbial communities, and it thus appethat organisms are able to exist and thrive under an extraonary range of conditions wherever fresh liquid water or brinare present. Similarly, brine-filled cracks in sea ice or in tlithosphere provide a suitable habitat for microorganisms, athese types of environments may offer close parallels withice and rocky interior of Europa.

5.2. Temperature and Pressure Tolerancesof Terrestrial Microorganisms

Conditions in Europa’s ocean might include temperatucolder than anywhere on Earth, although it is more likely tocean may be comparable to Antarctic and Arctic saline lakBecause life survives (and likely grows) in all of the coldest evironments on Earth where liquid is present, even colder contions may permit biological growth, as long as liquid water is sbilized by solutes. Microorganisms isolated from high-latituenvironments have been found to grow at temperatures whdip to 268 K (Franzmannet al.1987, Franzmann 1991), and additional bacterial species grow at temperatures as low as 26albeit slowly (Neidhardtet al. 1990); these temperatures arcomparable to those of the sulfate-saturated europan oceaneled above. Microbial activity has been hypothesized to occupolar sea ice brine pockets at 243 K (Deming and Huston 200The permanent ice cover of briny lakes of the Antarctic DValleys is considered an ecological microbial “oasis” (Priscuetal. 1998, Psenner and Sattler 1998). Low temperatures coat the cost of a large reduction in metabolic rates (Demin aHust 2000). Low temperatures and the high solute abundanneeded to stabilized brines at low temperatures should markincrease the liquid’s viscosity and thus reduce solute transp

rates in the solution.

254 KARGEL ET AL.

EUROPA’S OCEAN 255

ssociated

g page)he fl. P

FIG. 6. Photos of chemosynthetic faunal community on the seafloor of the Gulf of Mexico in association with structure II gas hydrate buildup and ahydrocarbon venting. (a, preceding page) A hydrate mound in Green Canyon at about 540 m depth partly covered by orange mats ofBeggiatoa, a sulfide oxidizingbacterium, is shown. Scale of hydrate features is about 1 to 2 m diameter. The leading edge of the hydrate mound reveals exposed gas hydrate. (b, precedinBacterial mats associated with small sediment pockmarks from which hydrocarbons and highly saline, saturated brines issue at about 650 m depth. Toors ofpockmarks are partly filled with turbid, saturated brine. (c, above) Hydrate mound with associated tube worms and trains of hydrocarbon gas bubbleshotos by

Roger Sassen.

a

Lz

ca

li

u

ilni

93),hert or-ated

dercon-s ofnosim-itiesomeiles,an

cel-uso

ngiesutzen

Temperatures below the freezing point of a brine act to pserve microbes. Though not in a metabolically active stmicroorganisms are routinely maintained at 193 K in freezor at 78 K in liquid nitrogen in the laboratory and subsequenrevived after storage for several years. In a natural setting,crobes recovered from an ice core taken above subglacialVostok, Antarctica, remain viable despite being trapped in frolake water for over 100,000 years (Abyzovet al.1998). Brinesmay freeze on Europa over long time scales as the frozenthickens, as brines intrude into cold near-surface ice, or as pmolten briny ice diapirs ascend into the lithosphere (80–120(Pappalardoet al. 1998, Rathbunet al. 1998), or freezing mayoccur rapidly due to cryovolcanism (Kargel 1991). The abiof microorganisms to withstand extended periods of extrecold may be integral to europan life and may provide for futsampling and culturing of surface ices.

At the high-temperature end, the current upper temperalimit for a microorganism in pure culture in the laboratory386 K (Stetter 1999). A sulfur-reducing hyperthermopharcheon isolated from a deep-seavent is capable of adaptivarious high temperatures (from 349 to 383 K) by adjust

re-te,

erstlymi-akeen

rustrtlyK)

tymere

tureisicg tong

the abundances of certain proteins (Holden and Baross 19Pledger and Baross 1991). Growth may occur at even higtemperatures in natural systems. As on Earth, thermotoleranganisms on Europa could inhabit subocean environments heby geothermal sources.

As for pressure, the deepest point in Earth’s ocean is un110 MPa of hydrostatic pressure (over 10,600 m depth) andtains viable microorganisms (Yayanos 1995). Pure culturecertain bacteria grow at 100 MPa and 275 K as well (Yaya1995). Hydrostatic pressure coupled with lithostatic pressureplies that subseafloor and subterranean microbial communlikely experience even higher pressures (Yayanos 1995). Shydrothermal vent archaea and bacteria, known as barophexhibit affinities for elevated pressure. They can even showincrease in their optimum temperature for growth and an aceration in growth rates at supraoptimum temperatures (Eraet al.1993, Pledgeret al.1994).

Among the most interesting biological explorations takiplace on Earth with pertinence to Europa is the drilling activitat Lake Vostok, Antarctica. The liquid part of the lake is abo600 m deep in places and is covered by up to 200 m of fro

te

a

nf

o

rop

i)

y

ss

eld

hiy

smutud

relyen-riginandinnottheffer.a ist all

pro-rmal

thelife.olv-heseex-n inur in

bac-,e iceion

sys-isms

ntosul-

bio-rup-eme-

biq-em-ma-ofim-is-ringmes

willonaturess

ntsomarl

256 KARGEL

lake water and about 3500 m of glacial ice above that (Jouet al. 1999). Pressure at the lake floor is 39 MPa. The icwater interface exhibits net melting in the northern and wesparts of the lake and net freezing in the south (Siegertet al.2000). Drill cores taken from the frozen lake water above the lhas produced evidence of a microbial community (Priscuet al.1999), part of which remains viable and has been cultured abeing trapped in the lake ice for thousands of years (Karlet al.1999). These results show that the lake maintains a functioecosystem despite being physically isolated from the rest oplanet’s biosphere for about a million years.

5.3. Implications for Europa

A review of the adaptive strategies of microorganisms frEarth reveals that most every physiological stress can be ocome so long as the environment contains liquid water. Fthis physiological perspective, model organisms for a euroecosystem include members of the versatile bacterial geHalomonas. A halotolerant group,Halomonasspp. grow overa wide range of salt concentration (0.01% w/v salt to saturabrine) (Ventosaet al.1998), tolerate high pH (Duckworthet al.1996), grow at temperatures at least between 268 and 31(Ventosaet al.1998, Franzmannet al.1987), and tolerate mil-limolar levels of heavy metals (Ventosaet al.1998).Halomonasspp. have been isolated from deep-sea, high-pressure envments at mid-ocean ridges (22 MPa) (Kaye and Baross 1998deep-sea sediments (34 MPa) (Takamiet al.1999), and they havebeen found in sea ice brines (Bowmanet al.1997) and in brinescontained within the continental crust (Huuet al.1999, Orphanet al.1998, Adkinset al.1993, Vreeland and Huval 1991). Moreover,Halomonasspp. and other oligotrophic bacteria can groon low concentrations of organic carbon (0.002% w/v) (Kaand Baross 1998, 2000; Schutet al.1997).

Some deep-sea vent archaea prefer elevated pres(Erausoet al.1993, Pledgeret al.1994). Accordingly, pressureup to 200 MPa underneath a thick icy shell are not likelyinhibit life processes on Europa. Even though Europa’s ocis in equilibrium with ice, and is therefore predominantly cohydrothermal activity in the suboceanic crust may sustaincal temperatures that are very high, and so hyperthermoptraits known from Earth’s most phylogenetically primitive mcroorganisms (Baross and Hoffman 1985, Baross 1998) marelevant to life on Europa, also.

The physicochemical conditions most likely in Europaocean are all surmountable by life based on analogy withphysiological adaptations of microorganisms on Earth. Hoever, one can imagine possible europan chemical evolutionnarios that would result in conditions that might be too extrefor even the most well-adapted life–an ocean of eutectic sulfacid solution, for instance, would exceed the low-temperaand low-pH limits known for terrestrial life. On the other han

Earth does not have such natural environments (though strspheric clouds come close), and there is no predicting how

ET AL.

zele–rn

ke

fter

ingthe

mver-mannus

ted

8 K

ron-and

-we

ures

toan,

lo-ilic-be

’sthew-ce-e

ricre,

might develop if such conditions were stable. It is yet an entidifferent matter whether life could originate in the extremevironments possible in Europa’s ocean. The independent oof life may involve different problems than the sustenanceevolution of life; unfortunately, we know little about the origof life even here on Earth. Despite uncertainties, we woulddiscount possibilities for life in Europa’s ocean even undermost extreme environments that the ocean might have to o

An important aspect of the chemical scenarios for Europthat a wide variety of carbon and sulfur gases and solids, nonecessarily chemically equilibrated with one another, areduced and mixed or layered in the crust. Chemical and thegradients are likely to exist on a range of spatial scales, andresulting disequilibria is apt to be exploited by any europanHydrocarbon clathrate hydrates and chemical gradients inving sulfur compounds are of special significance, because tchemical situations are similar to some on Earth that lifeploits very well. The process of bacterial methane oxidatiogas-hydrate-related environments on Earth appears to occthe absence of molecular oxygen, along with concomitantterial reduction of sulfate (Hinrichset al. 1999). On this basisif hyrocarbon-bearing gas hydrates are present beneath thof Europa, the probability of life capable of methane oxidatand sulfate reduction is increased.

The carbon and sulfur cycles relevant to Earth’s marine ecotems, and the physiology and adaptive strategies of organfound near cold hydrocarbon seeps (Sassenet al. 1993) andseafloor hydrothermal vents (Karl 1995) may offer insights ipossible europan ecosystems. Considering that biogenicfur formation is so ubiquitous on Earth, we speculate thatgenic sulfur might be produced on Europa, too, and that etions of brine-containing organisms or sulfur deposited by thcould contribute to the yellowish/reddish color of nonice dposits there.

It is conceivable, if Io once had an aqueous past (Fanaleet al.1974, Nash and Fanale 1977, Kargel 1996, 1999b), that its uuitous sulfur compounds might bear some relation to biochical processes in the very distant past. Although sulfurousterials seen inGalileo imaging and reflectance spectroscopyIo and Europa might represent chemical biomarkers, it isprobable thatGalileo data or other remote sensing could dtinguish biogenic and abiogenic sulfur, especially considethat any biogenic sulfur probably has been recycled many tiand affected by intense abiogenic processing.In situ studies,including stable isotope analysis, offer some hope that weultimately elucidate the origin of this material, especiallyEuropa, where complications due to repeated high-tempervolcanic recycling and fractionations due to global volatile lohave not occurred.

Planned drilling and sampling of the waters and sedimeof subglacial Lake Vostok should prove very interesting fra geochemical and exobiological perspective (Bell and K

ato-life1999). Lake Vostok is probably the closest physical analogueto Europa’s ocean that Earth has to offer, though its salinity is

’S

reu

aueemtaeaoihlao

evu

ic

nsat

sd

neevael

anin-

lateolu-ndheat areesefor

ow

edur-heto a

ofand

inkscopicand

d be

tic-n-ern-an.

edon-ro-hys-willndspe-cectordels

s-icyoredtheughthe

tingessheAd-ionngne-d to

EUROPA

probably much lower than Europa’s ocean. The flow of eneand organic matter and the particular adaptive strategies uslife in the Lake Vostok microbial community may be analogoto those of Europa’s ocean.

6. IMPLICATIONS FOR FUTURE EUROPA MISSIONSAND OTHER OBSERVATIONS

There are several areas where further observations andysis would help constrain the chemical composition, structand evolution of Europa and prospects for life there. First,isting NIMS data should be scrutinized for minor spectral ftures within the asymmetric water bands to glean any infortion on specific identity and hydration state of nonice hydraAdditional lab data on the spectral reflectivity of hydrated sand sulfuric acid at cryogenic temperatures are needed to bevaluate the existing spacecraft data. Evidence in NIMS datminor nonsalt constituents, especially tholins, and their geoldistribution also should be sought. A future orbit-based imagspectrometer should yield improved spatial resolution, whicneeded to understand the geological controls in the empment and modification of surface materials, but it should alsdesigned with improved spectral resolution to help distinguminor components and the physical state of major constitueMajor gaps in high-pressure physical chemistry data on swater systems relevant to Europa exist (Hogenboomet al.1995,Commission on Equilibrium Data 1999) and should be fillEspecially needed are precise phase volume data at elepressures relevant to the present state of Europa’s ocean (200 MPa) and the primordial differentiation of Europa’s ma(up to 2500 MPa).

The Europa Orbiter mission, with its subsurface soundradar will be invaluable in assessing prospects for an extant o(Chybaet al. 1998; see also http://www.jpl.nasa.gov/icefire//europao.htm). However, the likelihood of multiple reflectioand the electrolytic character and freezing-point deprescaused by the salty crust will make direct sounding of the oceinterface with the icy shell difficult (Winebrenner and Sylves1996). Dispersed brine inclusions will occur well above the mjor ice/brine interface and make the ice radar lossy. The posroles of impurities, including salts, in the interpretation of rasounding data have been considered by Mooreet al. (1992,1994), Corret al. (1992), and Chybaet al. (1998). (See alsohttp://outerplanets.larc.nasa.gov/outerplanets/.) In any evefuture Europa orbiting radar should produce data on the dto a partial melting isotherm. Renewed radar lab work toamine the roles of the specific types of impurities now belielikely, e.g., sulfuric acid, should be conducted. WinebrennerSylvester (1996) suggested use of Jupiter’s natural dekamradio emissions as a source of long-wave radio energy forlossy subsurface sounding.

The variety of morphologic features seen inGalileo imag-

ing provide information on the distribution and nature of tetonic and volcanic processes on Europa. Color data prov

OCEAN 257

gyd bys

nal-re,x-a-a-

es.ltstterforgicngisce-be

ishnts.alt-

d.atedp to

ss

ngean

sionn’s

era-iblear

t, apthx-edndtricess

high-spatial- and low-spectral-resolution information that cbe compared to the high-spectral- and low-spatial-resolutionformation from NIMS. These data are being used to formuand test models for Europa’s structural and geological evtion. Further acquisition of global high-resolution imaging aaltimetry from a Europa Orbiter will be very useful in testing tglobal significance of geologic processes and sequences themerging at the local and regional levels. Taken together, thdata may permit one to distinguish among several modelsthe evolution of Europa’s outer layers.

The composition and origin of Europa’s atmosphere, nknown to contain oxygen, sodium, and potassium (Hallet al.1995, Brown and Hill 1996, Brown 1999), should be examinfor any possible controls by sublimation or sputtering of sface materials or direct venting from Europa’s interior. Is tlack of magnesium in the atmosphere (Brown 1999) duelack of magnesium salts on the surface, or is it an indicationthe greater stability of magnesium salts against sputteringradiolysis? Additional lab data are needed to understand lbetween surface and atmosphere compositions. New telesobservations of Europa are needed to search for chlorineother minor atmospheric constituents. Consideration shoulgiven to flight of an orbiting mass spectrometer.

Additional observations of the europan magnetic field, parularly from orbit, would be especially valuable in reaching a cosensus on the origin of the europan magnetic field and in discing more specific constraints on a possible origin in an oceSimilarly, higher-resolution gravity data—particularly obtainfrom the Europa Orbiter—are needed to provide improved cstraints on three-layer models and to look for regional hetegeneities in Europa’s mass distribution. Surface-based geopical surveys, including electrical and seismic observations,also be very helpful in probing the crustal distribution of salts aelectrolyte solutions. Perhaps more than anything else, morecific compositional information, especially obtained by surfasample analyses by a landed spacecraft or by a flyby/impaspacecraft pair, is needed to narrow the range of viable moof Europa’s geochemical evolution.

Undoubtedly the most informative and exciting of all posible robotic missions to Europa would be one where theshell is penetrated and any ocean that exists would be expldirectly. The likelihood that salts are major constituents ofcrust poses formidable challenges to any plans to melt throthe shell, as some advance planning envisions. Althoughsalts are soluble in water, the tendency might be for a meldevice to induce partial (incongruent) melting, whereby a lhydrated solid salt is generated along with liquid brine. Tless-hydrated salt could tend to encrust the melting device.ditional heating would only continue the process of dehydratand partial melting but would not likely remove the encrustisalt residue, which may eventually accumulate into an impetrable mass. Extremely high temperatures might be require

c-idecompletely melt this encrustation. The complex thermochemi-cal dynamics of a melting device operating in a very salt-rich,

L

euiog

pcna

iad

icnays

vhondovicu

e

r

deet

thm

mtncra

ous

ex-onouty forogictheghtiallyn of-ectsy helpency

ringitled,rantSys-sionsom-

aker forll elsesiasm

the

987.

993.um

ic and

ete-

. L.four

th. In1996,rnia

sure-

n

258 KARGE

heterogeneous crust have not been studied to our knowledgmust be evaluated rigorously; testing of such a device in pice, polar ice, or even somewhat impure sea ice is not a sufficengineering test. It may be that a mechanical digger or a cbined melter/digger would be a better approach than meltinitself.

The optimism we have for existence of europan life, desextreme conditions possible in the ocean, also raises a conAs pointed out by a reviewer, there is a small but possiblyinsignificant risk of biological contamination of Europa’s oceif appropriate risk assessments and any needed countermeaare not taken for future exploration. It could be that the jovenvironment reduces this risk to nil, but this is not assureso. TheGalileo team should consider disposingGalileo, at endof mission, on Io or Jupiter, rather than take the very smrisk that a crash on Europa could result in eventual biologcontamination, which potentially could be more harmful to aendemic europan life and our ability to study it in the future ththe plutonium aboardGalileo. The contamination risk posed bthe Europa Orbiter and any future lander should be asseand risk mitigation strategies implemented if the risk is deemsignificant.

7. CONCLUSIONS

CI and CM carbonaceous chondrite meteorites may prothe best analogues for Europa’s primordial composition—tcan provide sulfides for a core, and salts, water, and other vtiles for the crust/ocean. Europa’s value for the gravitatioparameterC/MR2 is bracketed by some highly differentiatechemical–structural models based on CI chondrites. The cposition of Europa’s crust is very sensitive to the specific elutionary sequence: magnesium–sodium sulfates are predby the simplest evolution from CI and CM chondrites; a sfuric acid ocean is possible if there was extensive SO2 ventingfrom the rocky interior or if some sulfate salts thermally dcomposed; a sodium carbonate ocean is possible if thereextensive CO2 venting; and a crust containing many kilometeof clathrate hydrates (including some combination of CO2, SO2,CH4, higher hydrocarbons, or other gases) is possible accorto some likely scenarios. Many other possibilities can bevisioned considering Europa’s complex evolution, but genpredictions are that: (a) there should be extensive precipitaof sulfurous substances on the seafloor, (b) there shouldist a strong tendency for mineralogical density layering ofcrust, (c) brine volcanism and diapirism are allowed by sochemical–structural configurations of the crust but wouldmuch less likely in other configurations, and (d) most checal models would present good oceanic habitats, whereas ocould be fairly hostile to life as we know it. Foremost, we coclude that realistic chemical–physical evolution scenariosfor a highly impure and heterogeneous crust and a salt-satu

ocean. Definitive identification of the nonice substances on Eropa would greatly narrow the range of acceptable chemic

ET AL.

butre

entm-by

iteern.otnsuresnly

allalyn

seded

ideeyla-al

m-o-tedl-

-wass

ingn-ralionex-ee

bei-

hers-alltedu-

structural models and would help determine Europa’s aquechemical history and prospects for life.

In summary, further analysis of present data and futureploration will bring new and clearer insights into the evolutiof Europa and its partly liquid outer layer. Europa standsamong the planets and satellites as a distinctive laboratorthe study of the evolution of aqueous systems over geoltime and the study of environments potentially conducive toformation and evolution of life. Europa’s present state mioffer a good analogue for Earth’s earliest ocean. Io potentcould represent an end state after extreme devolatilizatioa Europa-like object (Kargelet al. 1999b). Comparative geological, geochemical, and exobiological studies of these objand others that once may have had oceans, e.g., Mars, maplanetologists develop a deeper understanding of the frequof and conditions needed by life in the Universe.

ACKNOWLEDGMENTS

Marion and Grant were partially supported by the U.S. Army EngineeResearch and Development Center Work Unit 61102/AT24/129/EE005 ent“Chemistry of Frozen Ground.” Head acknowledges support from NASA GNAG5-8144. Kargel was partially supported by a grant through the Jupitertems Data Analysis Program. Kaye thanks John Baross for useful discusand support. The authors thank Christopher Chyba for very helpful review cments. Kargel also expresses deepest gratitude to the late Eugene Shoemserving as mentor and helping to cultivate a special interest in Europa (and aunder the Sun); Gene’s continuous flow of insights and contagious enthuare sorely missed, but he remains an intellectual beacon.

REFERENCES

Abyzov, S. S., I. N. Mitskevich, and M. N. Poglazova 1998. Microflora ofdeep glacier horizons of Central Antarctica.Microbiology67, 547–555.

Acker, S. G., R. H. Byrne, Ben-Yaakov, R. A. Feely, and P. R., Betzer 1The effect of pressure on aragonite dissolution rates in seawater.Geochim.Cosmochim. Acta51, 2171.

Adkins, J. P., M. T. Madigan, L. Mandelco, C. R. Woese, and R. S. Tanner 1Arhodomonas aquaeoleigen. nov., sp. nov., an aerobic, halophilic bacteriisolated from a subterranean brine.Internat. J. Syst. Bacteriol.43, 514–520.

Anders, E., and N. Grevesse 1989. Abundances of the elements: Meteoritsolar.Geochim. Cosmochim. Acta53, 197–214.

Anders, E., R. Hayatsu, and M. H. Studier 1973. Organic compounds in morites.Science182, 781– 789.

Anderson, J. D., G. Schubert, R. A. Jacobson, E. L. Lau, W. B. Moore, and WSjogren 1998. Europa’s differentiated internal structure: Inferences fromGalileo encounters.Science281, 2019–2022.

Anderson, N. R., and A. Malahoff 1977.The Fate of Fossil Fuel CO2 in theOceans.Plenum Press, New York.

Bada, J. L. 1996. Europa as an analogue of an ice-covered primitive EarEuropa Ocean Conference, Capistrano Conference, No. 5, Nov. 12–14,San Juan Capistrano Research Institute, San Juan Capistrano, Califo,pp. 11–12. [Abstract]

Bagenal, F. 1994. Empirical model of the Io plasma torus: Voyager meaments.J. Geophys. Res.99, 11,043–11,062.

Baker, I., Y. L. Trickett, and P. M. S. Pradhan 2000. The effect of H2SO4 onthe stress exponent in ice single crystals. InThe International Symposium o

al–the Deformation of Glaciological Materials, London, Sept. 6–8, 1999. Ann.Glaciol., in press.

S

t

e

n

f

t

n

ln

gp

i

a

c

m

A

s

8

u

ls

b

ceouspar-

hur-com-

tem.

ro-lyticr

ropa

com-

drated

uropa.

pa: A

ropa?

ng in

of a

com-istry

ertain

val-

qui-nical

s.

riors

pu-

d the

car-

EUROPA’

Barber, D. J. 1985. Phyllosilicates and other layer-structured materials in smeteorites.Clay Miner.20, 415–454.

Baross, J. A. 1998. Do the geological and geochemical records of theEarth support the prediction from global phylogenetic models of a thmophilic ancestor? InThermophiles: The Keys to Molecular Evolution anthe Origin of Life?(J. Wiegel and W. W. Adams, Eds.), pp. 3–18. Taylor aFrancis.

Baross, J. A., and S. E. Hoffmann 1985. Submarine hydrothermal ventsassociated gradient environments as sites for the origin and evolution oOrigins Life15, 327–345.

Baross, J. A., and J. F. Holden 1996. Overview of hyperthermophiles andheat-shock proteins.Adv. Protein Chem.48, 1–34.

Baxter, G. P., and W. C. Cooper, Jr. 1924. The aqueous pressure of hydcrystals. II. Oxalic acid, sodium sulfate, sodium acetate, sodium carbodisodium phosphate, barium chloride.J. Am. Chem. Soc.46, 923–933.

Bell, R. E., and D. M. Karl 1999. Evolutionary processes, a focus of decade-ecosystem study of Antarctic’s Lake Vostok.Eos Trans. Am. Geophys. Unio80, 573–579.

Belton, M. J. S., and 33 colleagues 1996. Galileo’s first images of Jupiterthe Galilean satellites.Science274, 377–385.

Bennloch, S., A. J. Martinez-Murcia, and F. Rodr´ıguez-Valera 1995. Sequencinof bacterial and archaeal 16S rRNA genes directly amplified from a hysaline environment.System. Appl. Microbiol.18, 574–581.

Bernard, A., D. Demaiffe, N. Mattielli, and R. S. Punongbayan 1991. Anhydrbearing pumices from Mount Pinatubo: Further evidence for the existencsulfur-rich silicic magmas.Nature354, 139–140.

Bischoff, J. L., and F. W. Dickson 1975. Seawater–basalt interaction at 20◦Cand 500 bars: Implications for origin of sea-floor heavy-metal depositsregulation of seawater chemistry.Earth Planet. Sci.25, 385–397.

Bischoff, J. L., and R. J. Rosenbauer 1989. Salinity variations in submahydrothermal systems by layered double-diffusive convection.J. Geology97,613–623.

Boato, G. 1954. The isotopic composition of hydrogen and carbon in thebonaceous chondrites.Geochim. Cosmochim. Acta6, 209–220.

Bostrom, K., and K. Fredriksson 1964. Surface conditions of the Orgueilteorite parent body as indicated by mineral associations.Smithsonian Misc.Coll. 151.

Bowman, J. P., S. A. McCammon, M. V. Brown, D. S. Nichols, and T.McMeekin 1997. Diversity and association of psychrophilic bacteriaAntarctic sea ice.Appl. Environ. Microbiol.63, 3068–3078.

Brantley, S. L., D. A. Crerar, N. E. Moller, and J. H. Weare 1984. Geochemiof a modern marine evaporite: Bocana de Virrila, Peru.J. Sediment. Petrol.54, 447–462.

Brearley, A. J., and R. H. Jones 1998. Chondritic meteorites. InPlanetary Mate-rials, Reviews in Mineralogy(J. J. Papike, Ed.), Vol. 36, Chap. 3, pp. 1–39Mineralogical Society of America, Washington, DC.

Bridgman, P. W. 1912. Water, in the liquid and five solid forms, under pressProc. Am. Acad. Arts Sci.47, 441–458.

Bridgman, P. W. 1948a. Rough compressions of 177 substances to 40,00cm2. Proc. Am. Acad. Arts Sci.76, 71–87.

Bridgman, P. W. 1948b. The linear compression of various single crysta30,000 kg/cm2. Proc. Am. Acad. Arts Sci.76, 89–99.

Brown, M. E. 1994. Observation of mass loading in the Io plasma torus.Geophys.Res. Lett.21, 847–850.

Brown, M. E. 1999. Trace elements in the atmosphere of Europa as a prosurface composition.Eos Trans. Am. Geophys. Union80, F604.

Brown, M. E., and R. E. Hill 1996. Discovery of an extended sodium atmospharound Europa.Nature380, 229–231.

Brown, R. A., D. E. Shemansky, and R. E. Johnson 1983. A deficiency of Oin the Io plasma torus.Astrophys. J.264, 339.

OCEAN 259

ony

arlyer-dd

andlife.

heir

ratedate,

ong

and

er-

te-e of

0nd

rine

ar-

e-

.in

try

.

re.

0 kg/

to

e of

ere

Bunch, T. E., and S. Chang 1980. Carbonaceous chondrites—II. Carbonachondrite phyllosilicates and light element geochemistry as indicators ofent body processes and surface conditions.Geochim. Cosmochim. Acta44,1543–1577.

Burgess, R., I. P. Wright, and P. T. Pillinger 1991. Determinations of sulpbearing components in C1 and C2 carbonaceous chondrites by steppedbustion.Meteoritics26, 55–64.

Cameron, A. G. W., and J. B. Pollack 1976. On the origin of the Solar Sysand of Jupiter and its satellites. InJupiter(T. Gehrels, Ed.), pp. 61–84. Univof Arizona Press, Tucson.

Carlson, R. W. 2000. Abundance of sulfur, sulfur dioxide, sulfuric acid, hydgen peroxide, and carbon dioxide on Europa’s surface and their radioproduction and destruction processes. InPhotolysis and Radiolysis of OuteSolar System Ices (PROSSI) Workshop, p. 7. [Abstract]

Carlson, R., R. E. Johnson, and M. S. Anderson 1999. Sulfuric acid on Euand the radiolytic sulfur cycle.Science286, 97–99.

Carney, R. S. 1994. Consideration of the oasis analogy for chemosyntheticmunities of Gulf of Mexico hydrocarbon vents.Geo-Marine Lett.14, 149–159.

Carpenter, C. D., and E. R. Jette 1923. The vapor pressures of certain hymetal sulfates.J. Am. Chem. Soc.45, 578–590.

Carr, M. H., and 21 colleagues 1998. Evidence for a subsurface ocean on ENature391, 363–365.

Cassen, P., S. J. Peale, and R. T. Reynolds 1980. Tidal dissipation in Eurocorrection.Geophys. Res. Lett.7, 963–970.

Cassen, P., R. T. Reynolds, and S. J. Peale 1979. Is there liquid water on EuGeophys. Res. Lett.6, 731–734.

Chapman, C. R., W. J. Merline, B. Bierhaus, and S. Brooks 1998. Craterithe jovian system: Intersatellite comparisons.Lunar Planet. Sci.29, abstract1927.

Cheng, A. F. 1984. Escape of sulfur and oxygen from Io.J. Geophys. Res.89,3939–3944.

Chyba, C. F. 2000. Energy for microbial life on Europa.Nature403, 381–382.

Chyba, C. F., S. J. Ostro, and B. C. Edwards 1998. Radar detectabilitysubsurface ocean on Europa.Icarus134, 292–302.

Chyba, C. F., and 11 colleagues 1999. Europa and Titan: Preliminary remendations of the Campaign Science Working Group on Prebiotic Chemin the outer Solar System.Lunar Planet. Sci.30, abstract 1537.

Clifford, D., C. D. Carpenter, and E. R. Jette 1923. The vapor pressures of chydrated metal sulfates.J. Am. Chem. Soc.45, 578–590.

Colburn, D. S., and R. T. Reynolds 1985. Electrolytic currents in Europa.Icarus63, 39–44.

Collins, G. C., J. W. Head III, R. T. Pappalardo, and N. A. Spaun 2000. Euation of models for the formation of chaotic terrain on Europa.J. Geophys.Res.105, 1709–1716.

Commission on Equilibrium Data 1999. Influence of pressure on chemical elibria in aqueous systems—With particular reference to seawater. TechReport, IUPAC.Pure Appl. Chem.71, 871–890.

Consolmagno, G. J. 1975.Thermal History Models of Icy Galilean SatelliteM.S. thesis, MIT, Cambridge, MA.

Consolmagno, G. J., and J. S. Lewis 1978. The evolution of icy satellite inteand surfaces.Icarus34, 280–293.

Corr, H., J. C. Moore, and K. W. Nicholls 1992. Radar absorption due to imrities in Antarctic ice.Geophys Res. Lett.20, 1071–1074.

Crawford, G. D., and D. J. Stevenson 1988. Gas-driven water volcanism anresurfacing of Europa.Icarus73, 66–79.

Cronin, J. R., S. Pizzarello, and D. P. Cruikshank 1988. Organic matter inbonaceous chondrites, planetary satellites, asteroids and comets. InMete-

III orites and the Early Solar System(J. F. Kerridge and M. S. Matthews, Eds.),pp. 819–857. Univ. of Arizona Press, Tucson.

am

a

d

i

il

,

e

n

e

ic

9

9

k

e

terea,

rite

a,m-

iew

h’s

r

eria,

ul--

for

o,s:

r a

esin-

,on

ls,s

a.

on-

e-

ns

d

ence

pa:

260 KARGEL

Crowley, J. K. 1993. Mapping playa evaporite minerals with AVIRIS data:first report from Death Valley, California.Remote Sens. Environ.44, 337–356.

Crowley, J. K., and S. J. Hook 1996. Mapping playa evaporite mineralsassociated sediments in Death Valley, California, with multispectral therinfrared images.J. Geophys. Res.101, 643–660.

Crowley, J. K., D. R. Zimbelman, and J. C. Mars 1999. Mapping hydrothermaltered rocks with AVIRIS imagery of Mount Shasta, California.Eos Trans.Am. Geophys. Union80, F1151.

Dalton, J. B., and R. N. Clark 1998. Laboratory spectra of Europa candimaterials at cryogenic temperatures.Eos Trans. Am. Geophys. Union79,F541.

Dalton, J. B., and R. N. Clark 1999. Analysis of the Europan surface composbased on Galileo NIMS spectra.Eos Trans. Am. Geophys. Union80, F604.

Darland, G., T. D. Brock, W. Samsonoff, and S. F. Conti 1970. A thermophacidophilic mycoplasma isolated from a coal refuse pile.Science170, 1416–1418.

Davidson, D. W. 1983. Gas hydrates as clathrate ices. InNatural Gas Hydrates:Properties, Occurrence, and Recovery(J. L. Cox, Ed.), pp. 1–16. ButterworthStoneham, MA.

Delaney, J. R., D. S. Kelley, M. D. Lilley, D. A. Butterfield, J. A. BarrosW. S. D. Wilcock, R. W. Embley, and M. Summit 1998. The quantum evof oceanic crustal accretion: Impacts of diking at mid-ocean ridges.Science281, 222–230.

de la Pena, J. A., J. M. Garcia-Ruiz, R. Marfil, and M. Prieto 1982. Growfeatures of magnesium and sodium salts in a recent playa lake of La Ma(Spain).Estudios Geol.38, 245–257.

Delitsky, M. L., and A. L. Lane 1998. Ice chemistry on the Galilean satellitJ. Geophys. Res.103, 31,391–31,403.

Deming, J. W., and A. L. Huston 2000. An oceanographic perspective on mbial life at low temperatures. InCellular Origins and Life in Extreme Habitats(J. Seckbath, Ed.), Kluwer, Academic, Dordrecht/Norwell, MA, in press.

Domingue, D. L., and A. L. Lane 1998. IUE views Europa: Temporal variatioin the UV.Geophys. Res. Lett.25, 4421–4424.

Domingue, D. L., B. W. Hapke, G. W. Lockwood, and D. T. Thompson 19Europa’s phase curve: Implications for surface structure.Icarus90, 30–42.

Drever, J. I. 1997.The Geochemistry of Natural Waters, 3rd ed. Prentice Hall,Upper Saddle River, NJ.

Duckworth, A. W., W. D. Grant, B. E. Jones, and R. van Steenbergen 1Phylogenetic diversity of soda lake alkaliphiles.FEMS Microbiol. Ecol.19,181–191.

DuFresne, E. R., and E. Anders 1962. On the chemical evolution of the carbceous chondrites.Geochim. Cosmochim. Acta26, 1085–1114.

Ehrlich, A., and I. Dor 1985. Photosynthetic microorganisms of the GavSabkha. InHypersaline Ecosystems: The Gavish Sabkha(G. M. Friedmanand W. E. Krumbein, Eds.), pp. 296–321. Springer-Verlag, New York.

Erauso, G., and 10 colleagues 1993.Pyrococcus abyssi sp. nov., a new hy-perthermophilic archeon isolated from a deep-sea hydrothermal vent.Arch.Microbiol. 160, 338–349.

Eugster, H. P. 1980. Geochemistry of evaporitic lacustrine deposits.Annu. Rev.Earth Planet. Sci.8, 35–63.

Eugster, H. P. 1985. Oil shales, evaporites, and ore deposits.Geochim. Cos-mochim. Acta49, 619–635.

Eugster, H. P., and L. A. Hardie 1978. Saline lakes. InLakes—Chemistry,Geology, Physics(A. Lerman, Ed.), pp. 237–293. Springer-Verlag, New Yor

Eugster, H. P., and B. F. Jones 1979. Behavior of major solutes during clobasin brine evolution.Am. J. Sci.279, 609–631.

Evans, R., and D. W. Kirkland 1988. Evaporitic environments as a sourc

petroleum. InEvaporites and Hydrocarbons(B. C. Schrieber, Ed.), pp. 256–299. Columbia Univ. Press, New York.

ET AL.

A

ndal

lly

ate

tion

ic,

,nt

thcha

s.

ro-

ns

1.

96.

ona-

ish

.

sed-

of

Fagents, S. A., R. Greeley, R. J. Sullivan, R. T. Pappalardo, and L. M. Prock2000. Cryomagmatic mechanisms for the formation of Rhadamanthys Lintriple band margins, and other low albedo features on Europa.Icarus 144,54–88.

Fanale, F. P., T. V. Johnson, and D. L. Matson 1974. Io: A surface evapodeposit?Science186, 922–925.

Fanale, F. P., Y. H. Li, E. Decarlo, N. Domergue-Schmidt, S. K. SharmK. Horton, J. C. Granahan, and the Galileo NIMS Team 1998. Laboratory siulation of the chemical evolution of Europa’s aqueous phase.Lunar Planet.Sci.29, abstract 1248.

Fanale, F. P., and 22 colleagues 1999. Galileo’s multi-instrument spectral vof Europa’s surface composition.Icarus139, 179–188.

Farquhar, J., H. Bao, and M. Thiemens 2000. Atmospheric influence of Eartearliest sulfur cycle.Science289, 756–758.

Feely, H. W., and L. J. Kulp 1957. Origin of Gulf Coast salt-dome sulphudeposits.Bull. Am. Assoc. Petrol. Geol.41, 1802–1853.

Fegan, N. E., D. T. Long, W. B. Lyons, M. E. Hines, and P. G. Macumb1992. Metal partitioning an acid hypersaline sediments: Lake Tyrrell, VictorAustralia.Chem. Geol.96, 167–181.

Fernandez-Remolar, D., N. Rodriguez, R. Amils, and O. Prieto 2000. The sfuric chemistry in the terrestrial system of Tinto River and Europa: A comparison.Lunar Planet. Sci.31, abstract 1679. [CD-ROM]

Figueredo, P. H., F. C. Chuang, R. L. Kirk, and R. Greeley, 2000. Evidencea cryovolcanic origin of Europa’s “mitten” feature.Lunar Planet. Sci.31,abstract 1026. [CD-ROM]

Fisher, C. R., I. R. MacDonald, R. Sassen, C. M. Young, S. A. MackS. Hourdez, R. S. Carney, S. Joye, and E. McMullin 2000. Methane ice wormHesiocaeca methanicolacolonizing fossil fuel reserves.Naturwissenschaften87, 184–187.

Fisk, M. R., and S. J. Giovannoni 1999. Sources of nutrients and energy fodeep biosphere on Mars.J. Geophys. Res.104, 11,805–11,815.

Fournelle, J., R. Carmody, and A. S. Daag 1996. Anhydrite-bearing pumicfrom the June 15, 1991, eruption of Mount Pinatubo: Geochemistry, meralogy, and petrology. InFire and Mud, Eruptions and Lahars of MountPinatubo, Philippines(C. G. Newhall and R. S. Punongbayan, Eds.)pp. 845–864. Philippine Institute of Volcanology and Seismology, QuezCity, and Univ. of Washington Press, Seattle.

Franzmann, P. D. 1991. The microbiota of saline lakes of the Vestfold HilAntarctica. InGeneral and Applied Aspects of Halophilic Microorganism(F. Rodrıguez-Valera, Ed.), pp. 9–14. Plenum Press, New York.

Franzmann, P. D., H. R. Burton, and T. A. McMeekin 1987.Halomonas sub-glaciescola, a new species of halotolerant bacteria isolated from AntarcticInternat. J. Syst. Bacteriol.37, 27–34.

Fredriksson, K., and J. F. Kerridge 1988. Carbonates and sulfates in CI chdrites: Formation by aqueous activity on the parent body.Meteoritics23,35–44.

Friedman, G. M. 1980. Review of depositional environments in evaporite dposits and the role of evaporites in hydrocarbon accumulation.Bull. Cent.Rech. Explor. Prod. Elf-Aquitaine4, 589–608.

Friedman, G. M., and W. E. Krumbein (Eds.) 1985.Hypersaline Ecosystems:The Gavish Sabkha.Springer-Verlag, New York.

Fukazawa, H., K. Sugiyama, S. Mae, H. Narita, and T. Hondoh 1998. Acid ioat triple junctions of Antarctic ice observed by Raman scattering.Geophys.Res. Lett.25, 2845–2848.

Gaidos, E. J., K. H. Nealson, and J. L. Kirschrink 1999. Life in ice-covereoceans.Science284, 1631–1633.

Geissler, P. E., and 14 colleagues, and the Galileo Imaging Team 1998a. Evidfor non-synchronous rotation of Europa.Nature391, 368–370.

Geissler, P. E., and 15 colleagues 1998b. Evolution of lineaments on Euro

Clues from Galileo multispectral imaging observations.Icarus 135, 107–126.

S

p

a

hl

e

r

,

eo

as

9

D

e

inte

in

o

g

a

f a

le re-

esses

ul-

k-eter.

ryake

ng

siumuilib-

andstem.

etec-

ature-eon

ults

pa.

acids

ter

the

990.o-

ry

s of

EUROPA’

Gerakines, P. A., M. H. Moore, and R. L. Hudson 2000. Carbonaic acidduction in H2O : CO2 ices: UV photolysis vs. proton bombardment.Astron.Astrophys.357, 793–800.

Gerdes, G., W. E. Krumbein, and E. Holtkamp 1985a. Salinity and wactivity related zonation of microbial communities and potential stromalites of the Gavish Sabkha. InHypersaline Ecosystems: The Gavish Sabk(G. M. Friedman and W. E. Krumbein, Eds.), pp. 238–266. Springer-VerNew York.

Gerdes, G., J. Spira, and C. Dimentman 1985b. The fauna of the Gavish Saand the Solar lake—A Comparative study. InHypersaline Ecosystems: ThGavish Sabkha(G. M. Friedman and W. E. Krumbein, Eds.), pp. 322–34Springer-Verlag, New York.

Ghail, R. C. 1998. Ice crust thickness and internal composition of Europa.LunarPlanet. Sci.29, 1766.

Giblin, A. M., and B. L. Dickson 1992. Source, distribution, and economsignificance of trace elements in groundwaters from Lake Tyrrell, VictoAustralia.Chem. Geol.96, 133–149.

Golombek, M. P., and W. B. Banerdt 1990. Constraints on the subsurface sture of Europa.Icarus83, 441–452.

Grasset, O., J. Castillo, and C. Sotin 1996. Thermal evolution of Europa:plications for the existence of a deep ocean? InEuropa Ocean ConferenceCapistrano Conference No. 5, Nov. 12–14, 1996, San Juan Capistranosearch Institute, San Juan Capistrano, California, pp. 30–31. [Abstract]

Greeley, R., R. Sullivan, M. D. Coon, P. E. Geissler, B. R. Tufts, J. W. HIII, R. T. Pappalardo, and J. M. Moore 1998a. Terrestrial sea-ice morpholConsiderations for Europa.Icarus135, 25–40.

Greeley, R., and 20 colleagues 1998b. Europa: Initial Galileo geological obvations.Icarus135, 4–24.

Greenberg, R., P. Geissler, G. Hoppa, B. R. Tufts, D. D. Durda, R. PappalJ. W. Head, R. Greeley, R. Sullivan, and M. H. Carr 1998. Tectonic proceon Europa: Tidal stresses, mechanical response, and visible features.Icarus135, 64–78.

Greenberg, R., P. Geissler, B. R. Tufts, and G. V. Hoppa 2000b. Habitabof Europa’s crust: The role of tidal-tectonic processes.J. Geophys. Res.105,17,551–17,562.

Greenberg, R., G. V. Hoppa, B. R. Tufts, P. Geissler, J. Riley, and S. Kadel 1Chaos on Europa.Icarus141, 263–286.

Greenberg, R., J. Riley, G. V. Hoppa, B. R. Tufts, and P. Geissler 2000a.tribution of chaotic terrain on Europa.Lunar Planet. Sci.31, abstract 1936.[CD-ROM]

Hall, D. T., D. F. Strobel, P. D. Feldman, M. A. McGrath, and H. A. Weav1995. Detection of an oxygen atmosphere on Jupiter’s moon, Europa.Nature373, 677–679.

Hansen, G. B., and T. B. McCord 2000. Amorphous and crystalline ice onGalilean satellites: A balance between thermal and radiolytic processes.LunarPlanet. Sci.31, abstract 1630. [CD-ROM]

Head, J. W., III, and R. T. Pappalardo 1999. Brine mobilization durlithospheric heating on Europa: Implications for formation of chaosrain, lenticula texture, and color variations.J. Geophys. Res.104, 27,143–27,156.

Head, J. W., R. T. Pappalardo, L. M. Prockter, N. A. Spaun, G. C. CollR. Greeley, J. Klemaszewski, R. Sullivan, C. Chapman, and the GalileoTeam 1999b. Europa: Recent geological history from Galileo observatiLunar Planet. Sci.30, abstract 1404.

Head, J. W., R. T. Pappalardo, and R. Sullivan 1999a. Europa: Morpholocharacteristics of ridges and triple bands from Galileo data (E4 and E6)assessment of a linear diapirism model.J. Geophys. Res.104, 24,223–24,235.

Head, N. Sherman, R. Pappalardo, C. Thomas, R. Greeley, and the GSSI team 1998. Cryovolcanism on Europa: Evidence for the emplacem

of flows and related deposits in the E4 region (5N, 305W) and interpreeruption conditions.Lunar Planet. Sci.29, abstract 1491.

OCEAN 261

ro-

terto-a

ag,

bkha

5.

icia,

truc-

Im-

Re-

adgy:

ser-

rdo,ses

ility

99.

is-

r

the

gr-

s,SSIns.

icaland

lileoent

Helfenstein, P., and A. F. Cook 1984. Active venting of Europa? Analysis otransient bright feature.Lunar Planet. Sci.15, 354–355. [Abstract]

Helfenstein, P., and E. M. Parmentier 1980. Fractures on Europa: Possibsponse of an ice crust to tidal deformation.Proc. Lunar Planetary Sci. Conf.11th, 1987–1998.

Helfenstein, P., and E. M. Parmentier 1985. Patterns of fracture and tidal strdue to nonsynchronous rotation: Implications for fracturing on Europa.Icarus61, 175–184.

Hendrix, A. R. 1998. Galileo UVS observations of Europa: Variations intraviolet absorber with surface features.Eos Trans. Am. Geophys. Union79,F535.

Hendrix, A. R., C. A. Barth, C. W. Hord, and A. L. Lane 1998. Europa: Disresolved ultraviolet measurements using the Galileo ultraviolet spectromIcarus135, 79–94.

Hibbitts, C. A., T. B. McCord, and G. B. Hansen 2000. Distributions of CO2

and SO2 on the surface of Callisto.Lunar Planet. Sci.31, abstract 1908.[CD-ROM]

Hill, T. W. 1979. Inertial limit on corotation.J. Geophys. Res.84, 6554–6558.

Hines, M. E., W. B. Lyons, R. M. Lent, and D. T. Long 1992. Sedimentabiogeochemistry of an acidic, saline groundwater discharge zone in LTyrrell, Victoria, Australia.Chem. Geol.96, 53–65.

Hinrichs, K.-U., J. M. Hayes, S. P. Sylva, P. G. Brewer, and E. F. DeLo1999. Methane consuming archaebacteria in marine sediments.Nature398,802–805.

Hobbs, P. V. 1974.Ice Physics.Clarendon Press, Oxford.

Hogenboom, D. L., J. S. Kargel, J. P. Ganasan, and L. Lee 1995. Magnesulfate–water to 400 MPa using a novel piezometer: Densities, phase eqria, and planetolog ical implications.Icarus115, 258–277.

Hogenboom, D. L., J. S. Kargel, and P. V. Pahalawatta 1999. Densitiesphase relationships at high pressures of the sodium sulfate-water syLunar Planet. Sci.30, abstract 1793. [CD-ROM]

Hogenboom, D. L., J. S. Kargel, M. L. Reiter, and Y. N. Khor 1997. Volumchanges attending hydration ofquenched magnesium sulfate brine: Thetonics of Ganymede’s sulci.Lunar Planet. Sci.28, 579–580. [Abstract]

Holden, J. F., and J. A. Baross 1993. Enhanced thermotolerance and temperinduced changes in protein composition in the hyperthermophilic ArchaES4.J. Bacteriol.175, 2839–2843.

Hoppa, G. V., B. R. Tufts, R. Greenberg, and P. Geissler 1999. Strike slip faon Europa: Global shear patterns driven by tidal stress.Icarus, 141, 287–298.

Hoppa, G. V., and B. R. Tufts 1999. Formation of cycloidal features on EuroScience285, 1899–1902.

Huber, C., and G. Wachtershauser 1998. Peptides by activation of aminowith CO on (Ni,Fe)S surfaces: Implications for the origin of life.Science281,670–672.

Huu, N. B., E. B. M. Denner, D. T. C. Ha, G. Wanner, and H. Stan-Lot1999.Marinobacter aquaeoleisp. nov., a halophilic bacterium isolated froma Vietnamese oil-producing well.Internat. J. Syst. Bacteriol.49, 367–375.

International Critical Tables1928. Vol. 4, pp. 349–351 and Fig. 127.

Jakosky, B. M., and E. L. Shock 1998. The biological potential of Mars,early Earth, and Europa.J. Geophys. Res.103, 19,359–19,364.

James, S. R., S. J. Dobson, P. D. Franzmann, and T. A. McMeekin 1Halomonas meridiana, a new species of extremely halotolerant bacteria islated from Antarctic saline lakes.Syst. Appl. Microbiol.13, 270–278.

Javor, B. 1989.Hypersaline Environments: Microbiology and Biogeochemist.Springer-Verlag, New York.

Johnson, R. E., M. L. Nelson, T. B. McCord, and J. C. Gradie 1988. AnalysiVoyager images of Europa: Plasma bombardment.Icarus75, 423–436.

tedJouzel, J., J. R. Petit, R. Souchez, N. I. Barkov, V. Ya. Lipenkov, D. Raynaud,M. Stievenard, N. I. Vassiliev, V. Verbeke, and F. Vimeux 1999. More than

ue

s-

b

a

ea

u

a

r

ia

i

,

.r

s

n

,

o

s—

,

tion.

m-

y of

s ofB

ical

–

i-

a,

G.des

n.

on

ated

tionhe

Im-

ites.

entshalag,

g-he

ict. 94-

ys

ence

262 KARGEL

200 meters of lake ice above subglacial Lake Vostok, Antarctica.Science286,2138–2141.

Kadel, S. D., S. A. Fagents, R. Greeley, and the Galileo SSI Team 1998. Trobounding ridge pairs on Europa: Considerations for an endogenic modformation.Lunar Planet. Sci.29, abstract 1978. [CD-ROM]

Kargel, J. S. 1991. Brine volcanism and the interior structures of asteroidsicy satellites.Icarus94, 368–390.

Kargel, J. S. 1996. Europa’s crust and ocean: Origin, composition, and posimplications for Io’s early condition. InEuropa Ocean Conference, Capistrano Conference No. 5, Nov. 12–14, 1996, San Juan Capistrano ReseInstitute, San Juan Capistrano, California, pp. 42–43. [Abstract]

Kargel, J. S. 1998. The salt of Europa.Science280, 1211–1212.

Kargel, J. S. 1999a. Aqueous chemical evolution and hydration state of Eurosalts.Lunar Planet. Sci.30, abstract 1851.

Kargel, J. S. 1999b. Evolution of the icy Galilean satellites: Differentiatiooceans, and evaporative processes.Eos, Trans. Am. Geophys. Union80, F598.

Kargel, J. S., and G. J. Consolmagno 1996. Magnetic fields and the detectaof brine oceans in Jupiter’s icy satellites.Lunar Planet. Sci.27, 643–644.[Abstract]

Kargel, J. S., P. Delmelle, and D. B. Nash 1999. Volcanogenic sulfur on Eand Io: Composition and spectroscopy.Icarus142, 249–280.

Kargel, J. S., K. L. Tanaka, V. R. Baker, G. Komatsu, and D. R. MacAy2000. Formation and dissociation of clathrate hydrates on Mars: Polar cnorthern plains, and highlands.Lunar Planet. Sci.31, abstract 1891.

Karl, D. M. 1995. Ecology of free-living, hydrothermal vent microbial commnities. In The Microbiology of Deep-Sea Hydrothermal Vents(D. M. Karl,Ed.), pp. 35–124. CRC Press, Boca Raton, FL.

Karl, D. M., D. F. Bird, K. Bjorkman, T. Houlihan, R. Shackelford, and L. Tup1999. Microorganisms in the accreted ice of Lake Vostok, Antarctica.Science286, 2144–2147.

Kaye, J. Z., and J. A. Baross 1998. Salt-tolerant microbes isolated fhydrothermal-vent environments.Eos Trans. Am. Geophys. Union79, F59.

Kaye, J. Z., and J. A. Baross 2000. High incidence of halotolerant bacterPacific hydrothermal vent and pelagic communities.FEMS Microbiol. Ecol.32, 249–260.

Keller, L. P., G. J. McCarthy, and J. L. Richardson 1986. Mineralogy and stabof soil evaporites in North Dakota.Soil Sci. Soc. Am. J.50, 1069–1071.

Kempe, S., and E. T. Degens 1985. An early soda ocean?Chem. Geol.53,95–108.

Kessel, M., Y. Cohen, and A. E. Walsby 1985. Structure and physiologysquare-shaped and other halophile bacteria from the Gavish Sabkha. InHyper-saline Ecosystems: The Gavish Sabkha(G. M. Friedman and W. E. KrumbeinEds.), pp. 265–287. Springer-Verlag, New York.

Khurana, K. K., M. G. Kivelson, D. J. Stevenson, G. Schubert, C. T. RussR. J. Walker, and C. Polanskey 1998. Induced magnetic fields as evidencsubsurface oceans in Europa and Callisto.Nature395, 777–780.

Kivelson, M. G., K. K. Khurana, S. Joy, C. T. Russel, D. J. Southwood, RWalker, and C. Polansky 1997. Europa’s magnetic signature: Report fGalileo’s pass on 19 December 1996.Science276, 1239–1241.

Kivelson, M. G., K. K. Khurana, D. J. Stevenson, L. Bennett, S. Joy, C. T. RusR. J. Walker, C. Zimmer, and C. Polanskey 1999. Europa and Callisto: Induor intrinsic magnetic fields in a periodically varying plasma environmeJ. Geophys. Res.104, 4609–4625.

Knoll, A. H. 1985. A paleobiological perspective on sabkhas. InHypersalineEcosystems: The Gavish Sabkha(G. M. Friedman and W. E. Krumbein, Eds.)pp. 407–425. Springer-Verlag, New York.

Kortz, B. E., J. W. Head III, and R. T. Pappalardo 2000. Thrace Macula, EurCharacteristics of the southern margin and relations to background plains

Libya Linea.Lunar Planet. Sci.31, abstract 2052. [CD-ROM]

Krauskopf, K. B. 1979.Introduction to Geochemistry. McGraw-Hill, New York.

ET AL.

gh-l of

and

ible

arch

pa’s

n,

ility

rth

alps,

-

s

om

in

lity

of

ell,e for

J.om

el,cedt.

pa:and

Krumbein, W. E. 1985. Applied and economic aspects of sabkha systemGenesis of salt, ore, and hydrocarbon deposits and biotechnology. InHyper-saline Ecosystems: The Gavish Sabkha(G. M. Friedman and W. E. KrumbeinEds.), pp. 426–436, Springer-Verlag, New York.

Ladbury, R. 1996. Model sheds light on a tragedy and a new type of erupPhysics TodayMay, 20–22.

Lane, A. L., R. M. Nelson, and D. L. Matson 1981. Evidence for sulphur iplantation in Europa’s UV absorption band.Nature292, 38–39.

Last, W. M., and T. H. Schweyen 1983. Sedimentology and geochemistrsaline lakes of the Great Plains.Hydrobiologia105, 245–263.

Leto G., M. E. Palumbo, and G. Strazzula 1996. Structural characteristicirradiated and unirradiated ices.Nucl. Instrum. Methods Phys. Res. Sect116, 49.

Lewis, J. S. 1971. Satellites of the outer planets: Their physical and chemnature.Icarus15, 174–185.

Lewis, J. S. 1972. Low-temperature condensation in the solar nebula.Icarus16,241–252.

Lewis, J. S. 1982. Io, geochemistry of sulfur.Icarus50, 103–114.

Lipschutz, M. E., and D. S. Woolum 1988. Highly labile elements. InMeteoritesand the Early Solar System(J. F. Kerridge and M. S. Matthews, Eds.), pp. 462487. Univ. of Arizona Press, Tucson.

Livshits, L. D., Yu. A. Genshaft, and Yu. N. Ryabainin 1963. Equilibrium dagram of the crystal hydrates of MgSO4 at high pressures.Russ. J. Inorg.Chem.8, 676–678.

Long, D. T.,et al. 1992a. Geochemistry of acid brines: Lake Tyrrell, VictoriAustralia.Chem. Geol.96, 33–52.

Long, D. T., N. E. Fegan, J. D. Mckee, W. B. Lyons, M. E. Hines, and P.Macumber 1992b. Formation of alunite, jarosite, and hydrous iron oxiin a hypersaline system: Lake Tyrrell, Victoria, Australia.Chem. Geol.96,183–202.

Lucchitta, B. K., and L. A. Soderblom 1982. The geology of Europa. InSatellitesof Jupiter(D. Morrison, Ed.), pp. 521–555. Univ. of Arizona Press, Tucso

Lucchitta, B. K., L. A. Soderblom, and H. M. Ferguson 1981. StructuresEuropa.Proc. Lunar Planet. Sci. B12, 1555–1567.

Luhr, J. F. 1990. Experimental phase relations of water- and sulfur-saturarc magmas and the 1982 eruptions of El Chich´on Volcano.J. Petrol. 31,1071–1114.

Luhr, J. F., I. S. E. Carmichael, and J. C. Varekamp 1984. The 1982 erupof El Chichon volcano, Chiapas, Mexico: Mineralogy and petrology of tanhydrite-bearing pumices.J. Volcanol. Geotherm. Res.23, 69–108.

Lunine, J. L., and R. L. Lorenz 1997. Light and heat in cracks on Europa:plications for prebiotic synthesis.Lunar Planet. Sci.28, 855–856. [Abstract]

Lupo, M. J., and J. S. Lewis 1979. Mass-radius relationships in icy satellIcarus40, 157–170.

Lyons, W. B., and H. E. Gaudette 1985. Trace metal concentrations in sedimfrom the Gavish Sabkha. InHypersaline Ecosystems: The Gavish Sabk(G. M. Friedman and W. E. Krumbein, Eds.), pp. 346–349. Springer-VerNew York.

Marion, G. M., and R. E. Farren 1999. Mineral solubilities in the Na-K-MCa-Cl-SO4-H2O system: A re-evaluation of the sulfate chemiustry in tSpencer-Moller-Weare model.Geochim. Cosmochim. Acta63, 1305–1318.

Marion, G. M., and S. A. Grant 1994. FREZCHEM: A chemical-thermodynammodel for aqueous solutions at subzero temperatures, CRREL Spec. Rep18. USACRREL, Hanover, New Hampshire.

Marion, G. M., R. E. Farren, and A. J. Komrowski 1999. Alternative pathwafor seawater freezing.Cold Regions Sci. Technol.29, 259–266.

Mathews, S. J., A. P. Jones, and M. C. Gardeweg 1994. Lascar volacno: Evidfor steady-state disequilibrium.J. Petrol.35, 401–432.

Matsui, M., S. Fukushima, and S. Nakada 1927. Dissociation vapor pressure ofsodium sulfate decahydrate. II.J. Chem. Ind.30, 330–31.

’S

em

s

t

.

ov

tra

e

7sN

ri

l

e

t

mi

e

t

del

u

e

a

n

te

a gray

-state

urface

onaltem-vi-

onset intimal

sys-

ma

lites

mal

pec-of

tary,

.),

. P.8.

bove

olds

linsgion

, R.n of

ing

EUROPA

McCollom, T. M. 1999. Methanogenesis as a potential source of chemical enfor primary biomass production by autotrophic organisms in hydrothersystems on Europa.J. Geophys. Res.104, 30,729–30,742.

McCord, T. B., 11 colleagues, and the Galileo NIMS Team 1998a. SaltEuropa’s surface detected by Galileo’s near infrared mapping spectromScience280, 1242–1245.

McCord, T. B., 12 colleagues, and the Galileo NIMS Team 1998b. Non-waice constituents in the surface material of the icy Galilean satellites fromGalileo near-infrared mapping spectrometer investigation.J. Geophys. Res103, 8603–8626.

McCord, T. B., and 11 colleagues 1999a. Hydrated salt minerals on Eursurface from the Galileo near-infrared mapping spectrometer (NIMS) intigation.J. Geophys. Res.104, 11,827–11,851.

McCord, T. B., G. B. Hansen, T. M. Orlando, M. T. Sieger, G. Teeter, N. G. PeL. van Keulen, and J. K. Crowley 1999b. Hydrated sulfates and carbonand their stability on Europa’s surface.Eos Trans. Am. Geophys. Union80,F599.

McKinnon, W. B. 1999. Convective instability in Europa’s floating ice shGeophys. Res. Lett.26, 951–954.

Milton, D. J. 1974. Carbon dioxide hydrate and floods on Mars.Science183,654–656.

Mironenko, M. V., S. A. Grant, G. M. Marion, and R. E. Farren 199FREZCHEM2: A chemical-thermodynamic model for aqueous solutionsub-zero temperatures. CRREL Report 97-5. USACRREL, Hanover,Hampsire.

Moore, J. C., A. P. Reid, and J. Kipfstuhl 1994. Microstructure and electproperties of marine ice and its relationship to meteoritc ice and seaJ. Geophys. Res.99, 5171–5180.

Moore, J. M., J. Paren, and H. Oerter 1992. Sea salt dependent electricaduction in polar ice.J. Geophys. Res.97, 19,803–19,812.

Moore, J. M., and 17 colleagues 1998. Large impact features on Europa: Rpof the Galileo nominal mission.Icarus135, 127–145.

Morse, J. W., and G. M. Marion 1999. The role of carbonates in the evoluof early martian oceans.Am. J. Sci.299, 738–761.

Mottl, M. J., and H. D. Holland 1978. Chemical exchange during hydrotheralteration of basalt by seawater—I. Experimental results for major and mcomponents of seawater.Geochim. Cosmochim. Acta42, 1103–1115.

Mueller, S., and W. B. McKinnon 1988. Three-layered models of Ganymand Callisto: Compositions, structures, and aspects of evolution.Icarus76,437–464.

Nash, D. B., and F. P. Fanale 1977. Io’s surface composition based on reflecspectra of sulfur/salt mixtures and proton-irradiation experiments.Icarus31,40–80.

Neidhardt, F. C., J. L. Ingraham, and M. Schaechter 1990.Physiology of theBacterial Cell: A Molecular Approach.Sinauer, Sunderland, MA.

Neukum, G., R. Wagner, U. Wolf, B. A. Ivanov, J. W. Head III, R. T. PappalarJ. E. Klemaszewski, R. Greeley, M. J. S. Belton, and the Galileo SSI T1998. Cratering chronology in the jovian system and derivation of absoages.Lunar Planet. Sci.29, abstract 1742.

Noll, K. S., H. A. Weaver, and A. M. Gonnella 1995. The albedo spectrof Europa from 200 angstroms to 3300 angstroms.J. Geophys. Res.100,19,057–19,059.

Ojakangas, G. W., and D. J. Stevenson 1989. Thermal state of an ice shEuropa.Icarus81, 220–241.

Orlando, T. M., M. T. Sieger, G. Teeter, N. G. Petrik, T. B. McCord, and L. vKeulen 1999. The role lof electronic excitations in the aging and processinmaterials relevant to the Galilean satellites.Eos, Trans. Am. Geophys. Unio80, F600.

Orphan, V. J., L. T. Taylor, S. K. Goffredi, and E. F. DeLong 1998. Charac

zation of microbial assemblages associated with a high temperature offsoil reservoir with molecular and culture-based methods. InAbstracts of the

OCEAN 263

rgyal

oneter.

er-the

pa’ses-

ik,tes

ll.

.atew

calice.

con-

sults

ion

alnor

de

ance

o,amute

m

ll on

ng of

ri-

American Society for Microbiology 98th General Meeting, Atlanta, GA, abstr.N-101, p. 382.

Pappalardo, R. T., and J. W. Head 1999. Europa: Role of the ductile layer.LunarPlanet. Sci.30, abstract 1967.

Pappalardo, R. T., and R. J. Sullivan 1996. Evidence for separation acrossband on Europa.Icarus123, 557–567.

Pappalardo, R. T., and 10 colleagues 1998. Geological evidence for solidconvection in Europa’s ice shell.Nature391, 365–368.

Pappalardo, R. T., and 31 colleagues 1999. Does Europa have a subsocean? Evaluation of the geologic evidence.J. Geophys. Res.104, 24,015–24,056.

Pierazzo, E., and C. F. Chyba 2000. Impact delivery of organics to Europa.LunarPlanet. Sci.31, abstract 1656. [CD-ROM]

Pledger, R. J., and J. A. Baross 1991. Preliminary description and nutriticharacterization of a chemoorganotrophic archaeobacterium growing atperatures of up to 110◦C isolated from a submarine hydrothermal vent enronment.J. Gen. Microbiol.137, 203–211.

Pledger, R. J., B. C. Crump, and J. A. Baross 1994. A barophilic respby two hyperthermophilic, hydrothermal vent Archaea: An upward shifthe optimal temperature and acceleration of growth rate at supra-optemperatures by elevated pressure.FEMS Microbiol. Ecol.14, 233–242.

Pollack, J. B. 1985. Formation of the giant planets and their satellite-ringtems: An overview. InProtostars and Planets II(D. C. Black and M. S.Matthews, Eds.), pp. 791–831, Univ. of Arizona Press, Tucson.

Pontius, D. H., and T. W. Hill 1982. Departure from corotation of the Io plastorus: Local plasma production.Geophys. Res. Lett.9, 1321–1324.

Prentice, A. J. R. 1996. Origin and bulk composition of the Galilean sateland the primitive atmosphere of Jupiter: A pre-Galileo analysis.Earth MoonPlanets73, 237–258.

Prieto, O., M. Gutierrez-Florez, J. S. Kargel, and E. Quilez 2000. Theranalysis of some hydrated salts: Implications for Europa.Lunar Planet. Sci.31, abstract 1728. [CD-ROM]

Prieto, O., J. S. Kargel, E. Quilez, and M. P. Martinez 1999. Reflectance strum of MgSO4 · 12H2O+ ice: Compositional implications on the surfaceEuropa jovian satellite.Lunar Planet. Sci.30, abstract 1762.

Prinn, R. G., and B. Fegley Jr. 1989. Solar nebula chemistry: Origin of planesatellite, and cometary volatiles. InOrigin and Evolution of Planetary andSatellite Atmospheres(S. K. Atreya, J. B. Pollack, and M. S. Matthews, Edspp. 78–136, Univ. of Arizona Press, Tucson.

Priscu, J. C., C. H. Fritsen, E. E. Adams, S. J. Giovannoni, H. W. Paerl, CMcKay, P. T. Doran, D. A. Gordon, B. D. Lanoil, and J. L. Pinckney 199Perennial Antarctic lake ice: An oasis for life in a polar desert.Science280,2095–2098.

Priscu, J. C., and 10 colleagues 1999. Geomicrobiology of subglacial ice aLake Vostok, Antarctica.Science286, 2141–2144.

Prockter, L. M., and R. T. Pappalardo 2000. Evidence for regional-scale fon Europa.Lunar Planet. Sci.31, abstract 1182. [CD-ROM]

Prockter, L. M., A. M. Antman, R. T. Pappalardo, J. W. Head, and G. C. Col1999a. Europa: Stratigraphy and geological history of the anti-jovian refrom Galileo E14 solid-state imaging data.J. Geophys. Res.104, 16,531–16,540.

Prockter, L. M., R. Pappalardo, R. Sullivan, J. Head, J. Patel, B. GieseWagner, G. Neukum, and R. Greeley 1999b. Morphology and evolutioEuropan bands: Investigation of a seafloor spreading analog.Lunar Planet.Sci.30, 1900.

Prockter, L. M., R. T. Pappalardo, and J. W. Head, III 2000. Strike-slip duplexon Jupiter’s icy moon Europa.Jour. Geophys. Res.105, 9483–9488.

Psenner, R., and B. Sattler 1998. Life at the freezing point.Science280, 2073–2074.

horeRansford, G. A., A. A. Finnerty, and K. D. Collerson 1981. Europa’s petrologicalthermal history.Nature289, 21–24.

r

i

it

dn

e

a.

g

.c

See

n

x

.

s

t

),

ly

J

a

ts

te

rus.

unds

eous

trano

nge

es of

ts on.

cent

rdoonal

, R.lae

eralK–

ures

ctive

ents.

llites.

.

hell.

. Inna

H.998.

shipan.

lter-r dry

264 KARGEL

Rathbun, J. A., G. S. Musser Jr., and S. W. Squyres 1998. Ice diapirs on EuImplications for liquid water,Geophys. Res. Lett.25, 4157–4160.

Resing, J. A., and F. J. Sansone 1999. The chemistry of lava–seawateractions: The generation of acidity.Geochim. Cosmochim. Acta63, 2183–2198.

Reynolds, R. T., S. W. Squyres, D. S. Colburn, and C. P. McKay 1983. Onhabilitability of Europa.Icarus56, 246–254.

Richardson, S. M. 1978. Vein formation in the CI carbonaceous chondrMeteoritics31, 141–159.

Roberts, H. H., and P. Aharon 1994. Hydrocarbon-derived carbonate builof the northern Gulf of Mexico slope: A review of submersible investigatioGeo-Marine Lett.14, 135–148.

Roberts, H. H., and R. S. Carney 1997. Evidence of episodic fluid, gas,sediment venting on the northern Gulf of Mexico continental slope.Econ.Geol.92, 863–879.

Ross, M. N., and G. Schubert 1987. Tidal heating in an internal ocean modEuropa.Nature325, 133–134.

Rutherford, M. J., and J. D. Devine 1996. Pre-eruption pressure-temperconditions and volatiles in the 1991 dacitic magma of Mount PinatuboFire and Mud, Eruptions and Lahars of Mount Pinatubo, Philipppines(C. G.Newhall and R. S. Punongbayan, Eds.), pp. 751-766. Univ. of WashinPress, Seattle.

Sassen, R., S. Joye, S. T. Sweet, D. A. DeFreitas, A. V. Milkov, and IMacDonald 1999b. Thermogenic gas hydrates and hydrocarbon gases inplex chemosynthetic communities, Gulf of Mexico continental slope.OrganicGeochem.30, 485–497.

Sassen, R., I. R. MacDonald, N. L. Guinasso, S. Joye, S., A. G. Requejo,Sweet, J. Alcala-Herrera, D. A. DeFreitas, and D. R. Schink 1998. Bactmethane oxidation in sea-floor gas hydrate. Significance to life in extrenvironments.Geology26, 851–854.

Sassen, R., C. McCabe, J. R. Kyle, and E. W. Chinn 1989. Deposition of magpyrrhotite during alteration of crude oil and reduction of sulfate.OrganicGeochem.14, 381–392.

Sassen, R., H. H. Roberts, P. Aharon, J. Larkin, E. W. Chinn, and R. Carney 1Chemosynthetic bacterial mats at cold hydrocarbon seeps, Gulf of Mecontinental slope.Organic Geochem.20, 77–89.

Sassen, R., S. T. Sweet, A. V. Milkov, D. A. DeFreitas, G. G. Salata, and EMcDade 1999a. Geology and geochemistry of gas hydrates, central GuMexico continental slope.Trans. Gulf Coast Assoc. Geol. Soc.49, 462–468.

Schenk, P. M., and W. B. McKinnon 1989. Fault offsets and lateral crumovement on Europa: Evidence for a mobile ice shell.Icarus79, 75–100.

Schenk, P. M., and C. K. Seyfert 1980. Fault offsets and proposed plate mofor Europa.Eos Trans. Am. Geophys. Union61, 286. [Abstract]

Schidlowski, M.,et al. 1985. Carbon isotope geochemistry and 14C agesmicrobial mats from the Gavish Sabkha and the Solar Lake. InHypersalineEcosystems: The Gavish Sabkha(G. M. Friedman and W. E. Krumbein, Eds.pp. 381–401. Springer-Verlag, New York.

Schleper, C., G. P¨uhler, B. Kuhlmorgen, and W. Zillig 1995. Life at extremelow pH. Nature375, 741–742.

Schrenk, M. O., K. J. Edwards, R. M. Goodman, R. J. Hamers, andBanfield 1998. Distribution ofThiobacillus ferrooxidans and Leptospirillumferrooxidans: Implications for generation of acid mine drainage.Science279,1519–1522.

Schut, F., R. A. Prins, and J. C. Gottschal 1997. Oligotrophy and pelagic mbacteria: Facts and fiction.Aquat. Microb. Ecol.12, 177–202.

Scott, H. P., Q. Williams, and F. J. Ryerson 2000. Experimental constrainthe chemical evolution of icy satellites.Lunar Planet. Sci.31, abstract 2002.[CD-ROM]

Seyfried, W. E. Jr., and D. R. Janecky 1985. Heavy metal and sulfur transporing subcritical and supercritical hydrothermal alteration of basalt: Influenc

fluid pressure and basalt composition and crystallinity.Geochim. Cosmochim.Acta49, 2545–2560.

ET AL.

opa:

nter-

the

es.

upss.

and

l of

tureIn

ton

R.om-

. T.rialme

etic

993.ico

C.lf of

tal

ions

of

. F.

rine

on

dur-of

Shemansky, D. E. 1987. The ratio of oxygen to sulfur in the Io plasma toJ. Geophys. Res.92, 6141–6146.

Shock, E. L. 1990. Geochemical constraints on the origin of organic compoin hydrothermal systems.Origins Life Evol. Biosphere20, 331–367.

Shock, E. L., and M. D. Schulte 1990. Amino-acid synthesis in carbonacmeteorites by aqueous alteration of polycyclic aromatic hydrocarbons.Nature343, 728–731.

Shoemaker, E. M. 1996. The age of Europa’s surface. InEuropa Ocean Confer-ence, Capistrano Conference, No. 5, Nov. 12–14, 1996, San Juan CapisResearch Institute, San Juan Capistrano, California, pp. 65–66. [Abstract]

Siegert, M. J., R. Kwok, C. Mayer, and B. Hubbard 2000. Water exchabetween the subglacial Lake Vostok and the overlying ice sheet.Nature403,643–646.

Sloan, E. D. 1998.Clathrate Hydrates of Natural Gases, 2nd ed. Dekker, NewYork.

Smith, B. A., and 21 colleagues 1979. The Jupiter system through the eyVoyager 1.Science204, 951–972.

Sonnenfeld, P. 1984.Brines and Evaporites. Academic Press, New York.

Sotin, C. J., J. S. Kargel, and O. Grasset 1999. High-pressure experimenhydrates and the fate of a deep ocean within Europa.Eos Trans. Am. GeophysUnion80, F37.

Spaun, N. A., and J. W. Head 2000. Modeling Europa’s crustal structure: ReGalileo results and implications for an ocean.J. Geophys. Res.in press.

Spaun, N. A., J. W. Head, G. C. Collins, L. M. Prockter, and R. T. Pappala1998. Conamara Chaos region, Europa: Reconstruction of mobile polygice blocks.Geophys. Res. Lett.25, 4277–4280.

Spaun, N., L. Prockter, R. Pappalardo, J. Head, G. Collins, A. AntmanGreeley, and the Galileo SSI Team 1999. Spatial distribution of lenticuand chaos on Europa.Lunar Planet. Sci.30, abstract 1847.

Spencer, R. J., N. Moller, and J. H. Weare 1990. The prediction of minsolubilities in natural waters: A cehmical equilibrium model for the Na–Ca–Mg–Cl–SO4–H2O system at temperatures below 25◦C. Geochim. Cos-mochim. Acta54, 575–590.

Spencer, J. R., L. K. Tamppari, T. Z. Martin, and L. D. Travis 1999. Temperaton Europa from Galileo PPR: Nighttime thermal anomalies.Science284,1514–1516.

Squyres, S. W., R. T. Reynolds, and P. M. Casse 1983. Liquid water and aresurfacing on Europa.Nature301, 225–226.

Stetter, K. O. 1999. Extremophiles and their adaptation to hot environmFEBS Lett.452, 22–25.

Stevenson, D. J. 1982. Volcanism and igneous processes in small icy sateNature298, 142–144.

Stevenson, D. J. 1998. An ocean within Callisto?Eos Trans. Am. GeophysUnion79, F535. [Abstract]

Stevenson, D. J. 2000. Limits on the variation of thickness of Europa’s ice sLunar Planet. Sci. Conf.31, abstract 1506. [CD-ROM]

Stevenson, D. J., A. W. Harris, and J. I. Lunine 1986. Origins of satellitesSatellites(J. A. Burns and M. S. Matthews, Eds.), pp. 39–88. Univ. of ArizoPress, Tucson.

Sullivan, R., R. Greeley, K. Homan, J. Klemaszewski, M. J. S. Belton, M.Carr, C. R. Chapman, R. Tufts, J. W. Head III, and R. Pappalardo 1Episodic plate separation and fracture infill on the surface of Europa.Nature391, 371–373.

Takami, H., K. Kobata, T. Nagahama, H. Kobayashi, A. Inoue, and K. Horiko1999. Biodiversity in deep-sea sites located near the south part of JaExtremophiles3, 97–102.

Tannenbuam, E., and I. R. Kaplan 1985. Role of minerals in the thermal aation of organic matter—I: Generation of gases and condensates undeconditions.Geochim. Cosmochim. Acta49, 2589–2604.

Thomas, C., and L. Wilson 2000. Photoclinometric analysis of resurfaced regionson Europa.Lunar Planet. Sci.31, abstract 1328. [CD-ROM]

’S

g

a

s9

a

n

ifc

h

yaao

h

,

str

m

f

n

in

por

layer

trano

ofthe

ilityto

trano

lean

–

ion

de

e

the

ele-

pa:

ed

EUROPA

Tomeoka, K., and P. R. Buseck 1985. Indicators of aqueous alteration incarbonaceous chondrites: Microtextures of a layered mineral containinS, O, and Ni.Geochim. Cosmochim. Acta49, 2149–2163.

Torii, T., N. Yamagata, S. Nakaya, S. Murata, T. Hashimoto, O. Matsuband H. Sakai 1975. Geochemical aspects of the McMurdo saline lakesspecial emphasis on the distribution of nutrient matters. InGechemical andGeophysical Studies of Dry Valleys, Victoria Land in Antarctica, Memoirthe National Inst. of Polar Res.(T. Torii, Ed.), Spec. Issue No. 4. pp. 5–2Tokyo.

Trickett, Y. L., I. Baker, and P. M. S. Pradhan 2000. The effects of sulfuricon the mechanical properties of ice single crystals.J. Glaciology, submittedfor publication.

Tufts, B. R. 1996. A San Andreas-sized strike-slip fault on Europa.Lunar Planet.Sci.27, 1343–1344.

Tufts, B. R., R. Greenberg, G. Hoppa, P. Geissler 2000. Lithospheric dilatioEuropa.Icarus146, 75–97.

Ventosa, A., D. R. Arahal, and B. E. Volcani 1999. Studies on the microbof the Dead Sea—50 years later. InMicrobiology and Biogeochemistry oHypersaline Environments(A. Oren, Ed.), pp. 139–147. CRC Press, BoRaton, FL.

Ventosa, A., J. J. Nieto, and A. Oren 1998. Biology of moderately halopaerobic bacteria.Microbiol. Mol. Biol. Rev.62, 504–544.

Von Damm, K. L. 1995. Controls on the chemistry and temporal variabilitseafloor hydrothermal fluids. InSeafloor Hydrothermal Systems: PhysicChemical, Biological, and Geological Interactions, Geophysical Monogr(S. E. Humphris, R. A. Zierenberg, L. S. Mullineaux, and R. E. ThomsEds.), Vol. 91, pp. 222–247. Amer. Geophys. Union, Washington, DC.

Vreeland, R. H., and J. H. Huval 1991. Phenotypic characterization of halopbacteria from ground water sources in the United States. InGeneral andApplied Aspects of Halophilic Microorganisms(F. Rodriguez-Valera, Ed.)pp. 53–60. Plenum Press, New York.

Wallmann, K., E. Suess, G. H. Westbrook, G. Winckler, and M. B. Cita 19Salty brines on the Mediterranean sea floor.Nature387, 31–32.

Wang, C. L., P. C. Michels, S. C. Dawson, S. Kitisakkul, J. A. BaroJ. D. Keasling, and D. S. Clark 1997. Cadmium removal by a new sof Pseudomonas aeriginosain aerobic culture.Appl. Environ. Microbiology63, 4075–4078.

Wei-Dongyan, Liu-Zhenmin, Deng-Xiaolin, and Xu-Shaokang 1998.Carbon-ates and Evaporites. 13, 86–89.

Williams, K. K., and R. Greeley 1998. Estimates of ice thickness in the ConaChaos region of Europa.Geophys. Res. Lett.25, 4273–4276.

Wilson, J. K., and N. M. Schneider 1994. Io’s fast sodium: Implicationsmolecular and atomic atmospheric escape.Icarus111, 31–44.

Wilson, L., and J. Head 1999. Cryomagmatism: Processes of generatio

cent and eruption and application to Europa.Lunar Planet. Sci.30, abstract1689.

OCEAN 265

CMFe,

ya,with

of.

cid

on

ota

a

ilic

ofl,phn,

ilic

97.

s,ain

ara

or

, as-

Wilson, L., J. W. Head, and R. T. Pappalardo 1997. Eruption of lava flowsEuropa: Theory and application to Thrace Macula.J. Geophys. Res.102,9263–9272.

Wilson, R. E. 1921. Some new methods for the determination of the vapressure of salt-hydrates.J. Am. Chem. Soc.43, 704–709.

Winebrenner, D. P., and J. Sylvester 1996. Sounding the thickness of an icein the presence of multiple reflections and scattering. InEuropa Ocean Confer-ence, Capistrano Conference, No. 5, Nov. 12–14, 1996, San Juan CapisResearch Institute, San Juan Capistrano, California, pp. 87–88. [Abstract]

Winkler, H. G. F. 1979.Petrogenesis of Metamorphic Rocks, 5th ed. Springer-Verlag, New York.

Woodside, J. M., M. K. Ivanov, A. F. Limonov, and Shipboard Scientiststhe ANAXIPROBE Expeditions 1998. Shallow gas and gas hydrates inAnaximander Mountains region, eastern Mediterranean Sea. InGas Hydrates:Relevance to World Margin Stability and Climate Change(J.-P. Henriet andJ. Mienert, Eds.), pp. 177–193. Geological Society of London, London.

Yayanos, A. A. 1995. Microbiology to 10,500 meters in the deep sea.Annu. Rev.Microbiol. 49, 777–805.

Yen, A. S., B. Burray, G. R. Rossman, and F. J. Grunthaner 1999. Stabof hydroxylated minerals on Mars: A study on the effects of exposureultraviolet radiation.J. Geophys. Res.104, 27,031–27,041.

Yoder, C. F., and W. L. Sjogren 1996. Tides on Europa. InEuropa Ocean Confer-ence, Capistrano Conference, No. 5, Nov. 12–14, 1996. San Juan CapisResearch Institute, San Juan Capistrano, California, pp. 89–90. [Abstract]

Zahnle, K., L. Dones, and H. F. Levison 1998. Cratering rates on the Galisatellites.Icarus136, 202–222.

Zolensky, M., and H. Y. McSween Jr. 1988. Aqueous alteration. InMeteoritesand the Early Solar System(J. F. Kerridge and M. S. Matthews, Eds.), pp. 114143. Univ. of Arizona Press, Tucson.

Zolensky, M. E., W. L. Bourcier, and J. L. Gooding 1989. Aqueous alteraton the hydrous asteroids: Results of EQ3/6 computer simulations.Icarus78,411–425.

Zolotov, M. Yu., and B. Fegley 1998a. Volcanic production of sulfur monoxi(SO) on Io.Icarus132, 431–434.

Zolotov, M. Yu., and B. Fegley 1998b. Volcanic origin of disulfur monoxid(S2O) on Io.Icarus133, 293–297.

Zolotoov, M. Yu., and B. Fegley 1999. Oxidation state of volcanic gases andinterior of Io. Icarus141, 40–52.

Zolotov, M. Yu., and E. L. Shock 2000a. Mass balance constraints on themental composition of the ocean on Europa.Lunar Planet. Sci.31, abstract1580. [CD-ROM]

Zolotov, M. Yu., and E. L. Shock 2000b. Freezing of oceanic water on EuroTheoretical modeling.Lunar Planet. Sci.31, abstract 1726. [CD-ROM]

Zolotov, M. Y., and E. L. Shock 2000c. Thermodynamic stability of hydrat

salts on the surface of Europa.Lunar Planet. Sci.31, abstract 1843. [CD-ROM]