does europa have a subsurface ocean& evaluation of the ...europa's mean density and moment...

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 104, NO. El0, PAGES 24,015-24,055, OCTOBER 25, 1999

Does Europa have a subsurface ocean? Evaluation of the geological evidence

R. T. Pappalardo, • M. J. S. Belton, 2 H. H. Breneman, 3 M. H. Carr, 4 C. R. Chapman, 5 G. C. Collins, • T. Denk, 6 S. Fagents, 7 P. E. Geissler, s B. Giese, 6 R. Greeley, 7 R. Greenberg, 8 J. W. Head, 1 P. Helfenstein, 9 G. Hoppa, s S. D. Kadel, 7 K. P. Klaasen, 3 J. E. Klemaszewski, 7 K. Magee, 3'•ø A. S. McEwen, s J. M. Moore, TM W. B. Moore, •2 G. Neukum, 6 C. B. Phillips s L. M Prockter, • G. Schubert, •2 D. A. Senske, 3'•ø R. J. Sullivan, 9 B. R. Tufts, s E. P. Turtle, s R. Wagner, 6 and K. K. Williams 7

Abstract. It has been proposed that Jupiter's satellite Europa currently possesses a global subsurface ocean of liquid water. Galileo gravity data verify that the satellite is differentiated into an outer H20 layer about 100 km thick but cannot determine the current physical state of this layer (liquid or solid). Here we summarize the geological evidence regarding an extant subsurface ocean, concentrating on Galileo imaging data. We describe and assess nine pertinent lines of geological evidence: impact ...... •"'•'•':"• '•'-':^'" .......... '---:- •: ........... " ..... •_A__,_ A, ........... 111UI}JIIUIU[I•D• l%•11tl•UllZt• %.•IJUVUIk•I, IIIL• I•tLUI•D• }Y till- •t}Y lZtl t Ui:ilIU3• iJlI•U3• 11U•U5•

surface frosts, topography, and global tectonics. An internal ocean would be a simple and comprehensive explanation for a broad range of observations; however, we cannot rule out the possibility that all of the surface morphologies could be due to processes in warm, soft ice with only localized or partial melting. Two different models of impact flux imply very different surface ages for Europa; the model favored here indicates an average age of-50 Myr. Searches for evidence of current geological activity on Europa, such as plumes or surface changes, have yielded negative results to date. The current existence of a global subsurface ocean, while attractive in explaining the observations, remains inconclusive. Future geophysical measurements are essential to determine conclusively whether or not there is a liquid water ocean within Europa today.

1. Introduction and Relevance

It is possible that Jupiter's satellite Europa currently possesses a global subsurface ocean. Voyager-era observations (Figure 1)and modeling offered tantalizing clues regarding the geology and state of Europa (see Lucchitta and Soderblotn [1982] and Malin and Pieri [1986] for reviews). Recent Galileo data complement these data and add credence to the ocean hypothesis. As described below (section 2.1), Galileo gravity measurements indicate that Europa is differentiated. The satellite can be considered to be a rocky object slightly smaller than Earth's moon but with an outer lay_er of H20 some 100 km

1Department of Geological Sciences, Brown University, Providence, Rhode Island.

2National Optical Astronomy Observatory, Tucson, Arizona. 3jet Propulsion Laboratory, Pasadena, California. 4U.S. Geological Survey, Menlo Park, California. 5Southwest Research Institute, Boulder, Colorado. 6DLR-Institut far Planetenerkundung, Berlin, Germany. 7Geology Department, Arizona State University, Tempe. 8Lunar and Planetary Laboratory, University of Arizona, Tucson. 9Center for Radiophysics and Space Research, Cornell University,

Ithaca, New York. løSterling Software, Pasadena, California. •lNASA Ames Research Center, Moffett Field, California. 12Department of Earth and Space Sciences, University of California

Los Angeles.

Copyright 1999 by the American Geophysical Union.

Paper number 1998JE000628. 0148-0227/99/1998JE000628509.00

thick [Anderson et al., 1998b]. Some thermal models (section 2.2) indicate that a significant portion of the H20 layer could be liquid today [Cassen et al., 1979; Squyres et al., 1983; Ross and Schubert, 1987; Ojakangas and Stevenson, 1989b]. If so, the volume of water would exceed that of Earth's oceans.

Galileo imaging observations reveal a low crater density, suggesting that the Europan surface is relatively young and the satellite may be geologically active today [Chapman et al., 1998; Bierhaus et al., 1998; Zahnle et al., 1998] (see section 3). The geological observations imply that warm, mobile material, and perhaps liquid water, lay at shallow depths within the subsurface at the time of its recent geological deformation (see section 4). Europa's elliptical orbit means that substantial tides generated by Jupiter might drive current-day geological activity and cause significant surface disruption. These characteristics make Europa in many respects unique in the Solar System and a fascinating topic for comparative planetology.

It has been considered for many years that a liquid water ocean within an icy satellite, and specifically Europa, might possibly support life [Consolmagno, 1975; Pellegrino and Stoff, 1983; Reynolds et al., 1983, 1987; Squyres, 1989; Chyba and McDonald, 1995; Lunine and Lorenz, 1997]. This prospect is made even more intriguing by suggestions that life on Earth may have arisen at hydrothermal systems on the ocean floor [e.g., Baross and Hoffman, 1985] and that photo- synthesis might be able to proceed in the absence of sunlight, powered by dim near-infrared light of hydrothermal vents [Van Dover et al., 1996]. In examining criteria for the origin and evolution of life as we know and understand it, at least three

24,015

24,016 PAPPALARDO ET AL.: DOES EUROPA HAVE AN OCEAN?

Figure 1. Voyager 2 mosaic of Europa at a resolution of --2 km/pxl. A dearth of impact structures, low topographic relief, and the existence of enigmatic landforins (including ridges seen along the terminator and dark wedge-shaped bands in the antijovian region, center left) hint that the icy surface is young, tectonically deformed, and perhaps still geologically active. In combination with thermal modeling, these observations led to suggestions that Europa may have a subsurface ocean. Photomosaic produced by the U.S. Geological Survey, Flagstaff, Arizona.

PAPPALARDO ET AL.: DOES EUROPA HAVE AN OCEAN? 24,017

conditions must be satisfied: the presence of organic compounds; sufficient heat to support a biomass; and water, probably required to be in the liquid form. A survey of the Solar Sys- tem shows that Europa is one of the very few objects which may meet these conditions. Of course, fundamental issues are whether there is sufficient heat to maintain water in liquid form, and if organic evolution ever actually occurred. Moreover, it is not clear that Europa has experienced enough internal heating to produce a significant biota [Jakosky and Shock, 1998]. But the possibility of a subsurface ocean makes Europa a high pri- ority for exobiological exploration.

The Galileo Solid State Imaging (SSI)experiment has contributed substantially to our knowledge of Europa's surface geology. While the SSI experiment is not capable of directly detecting liquid water beneath an icy shell, Galileo images do show a variety of features that we interpret as suggestive of near-surface liquid water, ductile ice, or water-ice slurries at the time that the features formed. Depending on whether the satellite's surface is old or young, these features may be indicative of current or past subsurface conditions.

The purpose of this paper is to review and critically evaluate the geological and geophysical evidence relevant to the existence of an ocean within Europa currently or in the geologically recent past. In section 2 we summarize our knowledge about Europa's interior as revealed by Galileo gravity data and its expected physical state (ice or water) as predicted by thermal models. In section 3 we discuss the age of Europa based on its observed crater density, finding that our interpretations of Galileo images support a young, probably currently active Europa. Section 4 investigates nine lines of geological evidence, each of which has been noted in existing literature as suggesting a Europan ocean. The relevant Galileo SSI observations are reviewed, along with pertinent models derived from the imaging data; each line of evidence is evaluated critically, with discussion of whether the Galileo observations are indica-

tive of liquid water, or whether warm ductile ice or a localized ice-liquid slurry could account for the observations. In section 5 our SSI search for current activity is reviewed. In section 6 we summarize the geological evidence and discuss implications for Europa's interior. In section 7 we review data from other Galileo instruments pertinent to the ocean issue. Section 8 considers how the ocean hypothesis can be tested with the up- coming Europa Orbiter spacecraft and potential future lander missions. Finally, in section 9 we present major outstanding questions relevant to a Europan ocean.

2. Europa's Interior

2.1 Gravity Measurements and Internal Structure

Prior to the Galileo mission there were two competing models for the internal structure of Europa [Schubert et al., 1986]. In one model, Europa consisted of an anhydrous rocky core with the density of Io or the Moon surrounded by a layer of H20 (ice and possibly liquid water) that could be more than 100 km thick. In the other model, most of the water in Europa was retained in a hydrated silicate interior surrounded by a thin ice layer. Constraints on these pre-Galileo models included the known mean density of Europa (about 3040 kg m '3) and knowledge that Europa's surface is largely water ice. The possibility that Europa might have differentiated a metallic core did not re- ceive discussion. Debate centered around the extent to which

Europa's interior was dehydrated and, in the fully dehydrated model, whether the outer water layer was completely frozen or had a liquid layer beneath an outermost layer of ice. Today, as a consequence of the Galileo measurements of Europa's gravitational field, we know that the model of Europa with a thin ice shell above a largely hydrated silicate interior is no longer tenable.

Europa's gravitational field has been determined on the basis of radio Doppler data from four encounters of the Galileo spacecraft with Europa (E4, E6, E11, and El2) [Anderson et al., 1997, 1998b]. An inversion of these data together with ground-based astrometric data on the positions of the four Gali- lean satellites and optical navigation data from the Voyager missions to Jupiter gives values for all the spherical harmonic gravitational coefficients of degree 2 including J2 = 435.5 + 8.2 and C22 = 131.0 + 2.5 (J2 and C22 in units of 10 -6) [Anderson et al., 1998b]. It is assumed that the source of Europa's spherical harmonic degree 2 gravitational field is an equilibrium ellipsoidal distortion of the satellite caused by its spin and the tidal forces it experiences as it revolves around Jupiter in effec- tively synchronous rotation with its orbital period. With this assumption, the value of C22 can be used to infer Europa's axial moment of inertia C, normalized to MR 2 (M is the mass of Europa and R is its radius), as C/MR 2 = 0.346 _+ 0.005 [Ander- son et al., 1998b]. This value of C/MR 2 is substantially less than 0.4, the value of C/MR 2 for a uniform density sphere, and it necessitates a concentration of mass toward the center of Eu-

ropa. Preliminary determinations of Europa's global shape from limb profiles of Galileo images show that Europa is indis- tinguishable from an equilibrium figure (P. C. Thomas, personal communication, 1998).

Anderson et al. [1998b] have explored the implications of Europa's mean density and moment of inertia for the structure of the interior in terms of simple two-layer and three-layer models, concluding that Europa must have a three-layer structure with an Fe or Fe-FeS core at its center, an anhydrous rocky mantle surrounding the metallic core, and an H20 layer around the rock. The thickness of Europa's outer H20 layer must lie in the range of about 80 to 170 km, with smaller layers corresponding to larger metallic cores and smaller mantle densities. An outer layer thickness of about 100 km is most plausible by analogy to the inferred differentiated structure of Io [Anderson et al., 1996]. Thus the gravity data suggest that Eu- ropa has an outer layer of H20 about 100 km thick surrounding an anhydrous silicate mantle overlying a metallic core [Ander- son et al., 1998b]. Debate still centers on the physical state of the thick H20 layer, i.e., whether there is a liquid water ocean beneath an outer ice shell (Figure 2).

2.2. Thermal Models for Europa

Thermal models of Europa have focused on the question of the existence of a liquid water ocean under Europa's outermost ice shell today [Schubert et al., 1986]. Accretional and radiogenic heat sources suffice to dehydrate Europa early in its evolution leaving the satellite covered with a layer of liquid water 100 km or more thick, so it is most probable that Europa had a water ocean in the ancient past. Early models [Lewis, 1971; Consolmagno and Lewis, 1976; Fanale et al., 1977] considered only the conductive cooling and freezing of the outer layer of water with time, predicting a current configuration of liquid water beneath an ice shell. However, Reynolds and Cassen [1979]


Figure 2. Illustration of the possible interior structure of Europa based on Galileo data. (Left) Galileo gravity data indicate that Europa has an outer layer of H20 about 100 km thick surrounding an anhydrous silicate mantle, overlying a metallic core. (Right) The physical state of the H20 layer (liquid or solid) is uncertain; it is plausible that Europa has a liquid water ocean beneath an outer ice shell. Illustration courtesy Pam Engebret- son, after a design by Eric M. DeJong and Zareh Gorjian.

showed that the outer ice shell would become unstable to

convection with sufficient thickening, thereby promoting heat transfer through the ice and the cooling and solidification of the underlying liquid water. Their models resulted in complete freezing of the outer layer of water in --108 years, a small fraction of geological history [Reynolds and Cassen, 1979; Cas- sen et al., 1982]. The predicted freezing of Europa's ocean by efficient subsolidus convection in the ice cover accounted only for radiogenic heating in the silicate interior. Cassen et al. [1979] included the heat produced by tidal dissipation in Eu- ropa's outer ice shell and found that this heat source could off- set the subsolidus convective cooling of the ice and prevent complete solidification of the water ocean. A steady state could be achieved in which tidal dissipative heating in an ice shell above a liquid water ocean would be carried upward by convection in the ice; the balance between the dissipative heat source and the convective cooling would leave the ice layer with a constant thickness. Cassen et al. [1980] later revised their estimate of tidal heating downward, again opening the question of whether the water layer on Europa could freeze completely over geological time.

The competition between the tendency of tidal heating to maintain a liquid water ocean and that of subsolidus ice convection to freeze the ocean has now been analyzed for nearly two decades without achieving a definitive conclusion [Squyres et al., 1983; Ross and Schubert, 1987; Ojakangas and Stevenson, 1989b; Fattale et al., 1990; Deschamps and Sotin, 1998; McKinnon, 1999]. The major uncertainty in the modeling is the rheology of ice [Durham et al., 1997; Goldsby and Kohlstedt, 1997a, b], both in its control of convection and dissipation. The phenomena of dissipative heating and convective cooling involve nonlinear feedback mechanisms associated with the dependence of rheology on temperature and the dependence of temperature on the heating and cooling mecha-

nisms. The amount of tidal heating in the ice depends on the magnitude of tidal deformation, which might vary through geological time [Greenberg, 1982; Ojakangas and Stevenson, 1986], as well as the rheology of ice at tidal periods. Notably, tidal dissipation is expected to be greatest where the ice is warmest and most deformable, nominally at the base of the ice layer [Ojakangas and Stevenson, 1989b]. The temperature profile and degree of tidal deformation in the ice shell depend on the satellite's internal structure, in particular the existence of a global liquid ocean beneath the ice layer and the ice thickness. Overall, these models predict that Europa's current orbital eccentricity may be able to produce a heat flow of about 25 to 50 mW m '2 which may be able to maintain a conductive ice shell about 10 to 30 km thick [Squyres et al., 1983; Thomas and Schubert, 1986; Ross and Schubert, 1987; Ojakangas and Stevenson, 1989b; W. McKinnon, personal communication, 1999].

Other properties of the ice are also important and highly uncertain. The thermal conductivity of the ice depends on temperature and physical state of the ice (density and distribution of cracks, for example). A thermally insulating layer at the surface of Europa would promote stabilization of a liquid water ocean [Squyres et al., 1983; Ross and Schubert, 1987]. The occurrence of minor constituents in the ice and ocean such as salts

[Kargel, 1991; McCord et al., 1998, 1999] and ammonia [Deschamps and Sotin, 1998] would affect the rheology of the ice and the freezing temperature of an ocean.

Tidal heating on major faults or within convective upwel- lings in Europa's ice shell may be important locally [Steven- son, 1996; McKinnon, 1999], and tidal heating due to forced circulations in a thin liquid water ocean sandwiched between the rock interior and the overlying ice may prevent complete solidification of an ocean [Yoder and Sjogren, 1996]. Tidal heating is too dependent on too many unknown or poorly known


properties of Europa's ice for this type of modeling to settle the debate on the current existence of a liquid water ocean within Europa.

3. The Age of Europa's Surface

In the sections that follow, evidence is presented that Eu- ropa has, or had in the past, a rigid outer icy shell overlying mobile material, which may be warm ice or liquid water. Was this configuration a transitory phase in the evolution of the satellite, for which evidence was left frozen into the surface at

some early stage in its history, or is the configuration an indication of the current, or very recent, state of the satellite? The answer to this question is central to determining not only how the satellite has evolved but how we should conduct its

future exploration. A young age is suggested by the fact that Europa is more sparsely cratered than almost all other solid objects observed so far in the Solar System (Io and Earth being exceptions). Lucchitta and Soderblom [1982] suggested on the basis of rather sparse Voyager data and estimates of cratering rates by Shoemaker and Wolfe [1982] that the Europan surface has a crater retention age of - 108 years. However, our knowledge of cratering rates in the outer Solar System is uncertain and the absolute ages have a corresponding uncertainty [Zahnle et al., 1998]. In the following sections we discuss two different models that have been proposed for estimating cratering rates, and hence ages, in the Jovian system. This discussion is followed by a summary of the crater densities in representative areas viewed by Galileo and what they might imply for the age and geological evolution of the surface.

Cratering rates, and hence absolute ages, in the Jovian system have been estimated in two very different ways. In Model I [Neukum, 1997; Neukum et al., 1998], cratering rates are estimated through extrapolation from lunar cratering rates along with the assumption that Gilgamesh, the youngest impact basin on Ganymede, is the same age as the youngest impact basin on the Moon. This assumption gives ages for the Europan surface of the order 10 9 years. Model II [Shoemaker and Wolfe, 1982; Shoemaker, 1996; Zahnle et al., 1998] derives cratering rates from the observed distribution of comets and asteroids

within the Solar System. This model results in ages less than 107 years for some parts of the Europan surface.

3.1. Model I

Neukum [1997] and Neukum et al. [1998] suggest that the size-frequency distribution of craters on the Galilean satellites is similar to the size-frequency distributions of craters on the terrestrial planets [Neukum et al., 1975; Neukum and Ivanov, 1994; G. Neukum et al., manuscript in preparation, 1999] and to those on the asteroids Gaspra and Ida [Chapman et al., 1996]. Specifically, the slope of the cumulative size-frequency distribution curve between a few kilometers diameter and a few

tens of kilometers diameter is lower than at smaller and larger diameters. These changes in cumulative slope are readily described in terms of a relative crater size-frequency distribution diagram, or R-plot [Crater Analysis Techniques Working Group, 1979] on which a cumulative slope of-2 is represented by a horizontal line, a slope of -3 is represented by a line slop- ing down to the fight, and a slope of-1 is shown by a line slop- ing down to the left. On an R-plot, the Ganymede and Callisto distributions are U-shaped below about 10 km diameter, have a minimum at a few kilometers diameter, and show an inverted U-

shape above 10 km. From (empirical) intersatellite crater scaling between Callisto, Ganymede, and Europa, G. Neukum et al. (manuscript in preparation, 1999) then shift the lunar distribution curve laterally to fit crater size-frequency distributions on Europa.

From the apparent similarity in the size-frequency distributions, Neukum [1997] and Neukum et al. [1998] contend that the same population of objects, thought to be derived mostly from the asteroid belt, has dominated the cratering in both the Jovian system and inner Solar System. They further argue that the cratering rate must have had a similar history in the outer and inner Solar System, having been very high and declining rapidly prior to about 3.5 billion years ago and low and essentially constant ever since. Moreover, they reason that the youngest basin on Ganymede (Gilgamesh) should have approximately the same 3.8 Gyr age as the youngest basin on the Moon (Orientale).

The cratering rate on the Moon can be approximated by the following expression:

N(1) = 5.44 x 10 -•4 [exp(6.93t) - 1] + 8.38x10 '4t (1)

where N(1) is the number of 1-km-diameter craters and t is the age in billions of years [Neukum and Ivanov, 1994]. N(1) for a 3.8 Gyr surface on the Moon is 1.81x10 -2. For Gilgamesh, assumed to be 3.8 Gyr old, N(1) is 3.49 x 10 -3. The cratering rate on Europa is approximately twice that on Ganymede [Shoe- maker and Wolfe, 1982], so that Model I yields the following expression for Europa:

NE(1 ) = 2.05x10 -•4 [exp(6.93t)- 1] + 3.16x10 '4t (2) Various parts of the Europan surface have extrapolated Ne(1) values ranging from 2 x 10 '4 km -2 to 8 x 10 '4 km -2 [Neukum, 1997]. It should be noted that these Ne(1) values are extrapolated from actual counts with the assumptions that the production population and size distribution of impactors at the Gali- lean satellites are the same as for the Moon. If we further as-

sume that all these craters are primaries, the resulting ages range from 0.7 to 2.8 Gyr.

Thus, in estimating an old age for Europa's surface, Neukum [1997] and Neukum et al. [1998] make two fundamental assumptions: 1) Gilgamesh is 3.8 Gyr old, and 2) the cratering flux in the Jovian system has decayed over time in a manner identical to the lunar decay. We note that assumption (1) is uncertain, and assumption (2) would not be directly relevant if Eu- ropa's surface is young and dominated by more recent cometary impacts.

3.2. Model II

A very different chronology is obtained from cratering rates derived from what is known about the current distribution of objects within the Solar System. Cratering rates within the inner Solar System have been estimated from the observed distribution of impacting objects and checked against the numbers of craters on the Earth and the Moon [Shoemaker et al., 1990; Weissman, 1990; Rabinowitz et al., 1994; Shoemaker et al., 1994; Grieve and Shoemaker, 1994; McEwen et al., 1997]. The estimated rate of production of craters >20 km diameter on the Earth based on astronomical observations is consistent

with the terrestrial geological record for the last 120 Myr, giving credence to the techniques used for estimating the numbers of undiscovered objects, impact probabilities, and the sizes of craters produced.


Shoemaker [1996] and Zahnle et al. [1998] both find that main belt asteroids contribute negligibly to crateting in the Jovian system. Zahnle et al. point out that a million kilometer-sized asteroids must be ejected outward from the asteroid for just one to hit Ganymede (or Europa) and form a single 20 km crater, and there are simply not enough asteroids in the asteroid belt to cause significant cratering of the Galilean satellites at this low yield. Therefore they challenge the assumption of Model I that the same population of objects cause the cratering in the inner and outer Solar System. Shoemaker [1996] and Zahnle et al. [1998] both conclude that Jupiter family comets, both active and extinct, cause most of the cratering in the Jo- vian system, with small additional contributions from long- period comets, Halley-type comets, and Trojan asteroids. Jupi- ter family comets have generally low, prograde orbits and are now generally thought to originate in the Kuiper belt [Duncan et al., 1988]. Their masses and numbers are difficult to estimate, the active ones because their coma masks their true size, and inactive ones because many have large perihelion distances (>2 AU) so are difficult to detect. Zahnle et al. review in detail the various means for estimating the numbers of active and inactive comets at Jupiter and the implied cratering rates. Their reported estimate of the crateting rate on Europa is 1.0 x 10 -•3 km '2 yr -• for craters 10 km in diameter or larger, and more recent dynamical modeling by this group reduces the estimated impact rate by about factor of two, to 4.6 x 10 '• km '2 yr '• (K. Zahnle, personal communication, 1999). These values can be compared to the estimate of 1.7 x l0 -13 km -2 yr '• by Shoemaker [1996]. Zahnle et al. conclude that the uncertainty in the crateting rate is about a factor of 5, mainly (a factor of 3) due to the uncertainty in converting comet magnitudes to masses. Because of observational limitations, estimates cannot be

made of production rates for smaller size craters. Although Zahnle et al. estimate a factor of 5 uncertainty for

the current cratering rate, they do not explicitly discuss the uncertainty involved in extrapolating backward in time to estimate the age of satellite surfaces, instead assuming a constant rate. Astronomical surveys and observations of Earth and the Moon have shown that the current crateting rate is within a factor of 3 of the average cratering rate over the past -3.2 Gyr [Shoemaker et al., 1990; McEwen et al., 1997]. However, Zahnle et al. [1998] argue convincingly that the populations leading to impacts on Earth and the Moon are largely separate from those affecting Jupiter. Few observational constraints are currently available on the long-term temporal variability of Jupiter-farrdly comets, but theoretical models provide some in- sight. According to Levison and Duncan [1997], the median dynamical lifetime of massless particles as they evolve from Neptune-encountering orbits in the Kuiper belt into Jupiter- family comets is 4.5 x 107 years. As we will see, this median lifetime is comparable to the hypothesized average age of Eu- ropa's surface, so we can be reassured that the current cratering rate is typical over this timescale. (The extrapolation issue be- comes a greater concern for estimating the age of Ganymede.)

The current cratering rate of Zahnle et al. is much higher than that estimated by Neukum [1997] and Neukum et al. [1998]. Figure 3 combines crater counts on a small portion of the antijovian wedges area (observation C3WEDGES01; 15S, 195W), imaged at -420 meters per pixel (m/pxl) during orbit C3, and an area of mottled terrain (observation E4MACSTR02; 5N, 330W) imaged at ~30 rn/pxl during orbit E4 [Neukum, 1997]. If we extrapolate from the observational limit of 3 km up to 10 km, we obtain a cumulative crater frequency of 2.5 x

10 '6 km '2 for craters larger than 10 km in diameter. For this crater density, the revised Zahnle et al. cratering rate of 4.6 x 10 '•4 results in an age of 54 Myr (or 11 to 270 Myr, considering the factor of 5 uncertainty), as compared with a Model I age of 2.5 Gyr. This crater density may be typical of the Europan surface in general, although only a few areas have been yet counted over a large enough area to enable statistically significant results in the 1 - 10 km size range. We have several areas with counts for craters <1 km in diameter,

but the Shoemaker-Zahnle model makes no predictions of cratering rates for craters below 10 km in diameter, there being no observational data for cometary objects at Jupiter's distance that make craters of this size.

We can alternatively consider only those Europan craters with ctiameter >20 km, formed by impactors >1 km in size, for which the observational data are most reliable. Zahnle et al.

[1998] reported a formation rate of one such crater each 1.4 Myr; they have since revised this rate to one such crater each 3.2 Myr (K. Zahnle, personal communication, 1999). Galileo and Voyager imaging have revealed seven craters >20 in diameter [Turtle et al., 1999], over the roughly one-third to one-half of the surface imaged well enough to recognize such craters. The existence of ~14 to 21 craters with diameters >20 km would

imply a surface age of ~56 Myr (or ~11 to 280 Myr, considering the factor of 5 uncertainty). This age is in excellent agreement with that derived above from analysis of smaller craters.

3.3. Europa's Surface Age: Discussion

3.3.1. Evaluation of the crater age models. The two age models are irreconcilable. According to the premise of Model I, Gilgamesh, with a crater density of 10-km-diameter craters of 3.1x10 '5 km '2 [Neukum et al., 1998], is 3.8 Gyr old. In contrast, according to Model II, the 10 km diameter cratering rate is 2.4 x 10 '•4 km '2 yr '•, which gives the age of Gilgamesh as approximately 1 Gyr [Zahnle et al., 1998; K. Zahnle, personal communication, 1999]. Model II implies that Ganymede may have been active in recent Solar System history or it could be very old, considering the factor of 5 uncertainty of this model. Some support for a young Ganymede surface comes from the discovery that the satellite has a magnetic field [Kiv- elson et al., 1996], suggesting that it has undergone significant tidal heating at some time within the past 1 Gyr [Showman et al., 1997]. Lending further support to Model II, there have been four observed close passes of three comets to Jupiter in the past 150 years [Zahnle et al., 1998]: Brooks 2 (1886, 2 Rj), Gehrels 3 (1970, 3 Rj), Shoemaker-Levy/9 (1992, 1.3 Rj), and Shoemaker-Levy/9 (1994, 0.5 Rj). According to Model II, a comet the size of Shoemaker-Levy/9 (1.5 to 1.8 km diameter) should hit Jupiter each 200 to 300 years. But according to Model I, an asteroid would impact Jupiter every several tens of thousands of years and a comet would impact even less often, so that the chances of observing such an event would be extremely small. Finally, the argument that asteroids contribute negligibly to cratering in the Jovian system today is persuasive. Model I requires that asteroids were a significant impactor population in the Jovian system during late heavy bombardment, while comets had to be almost nonexistent.

Therefore, most of the authors of this article favor Model II and implied ages for the Europan surface of ~50 Myr with an er- ror of about a factor of 5. We find the arguments of Zahnle et al. [1998] persuasive, whereas the assumptions of Model I


10-2

o

O-a

10-4

1o 0

10 -6 10 -2

0 EUROP8: C3ES NEDGES 01

EUROPS: E4ESMACSTRO• AREA4

, I |, I ! l ill

I lO 8

Crater I]iameter D(Km)

Figure 3. Cumulative crater size-frequency plots for two sites on Europa. Measurements were made across portions of Galileo target regions C3WEDGES01 (image resolution -420 m/pxl) and E4MACSTR02 (image resolution -30 m/pxl). Crater data are fit with the lunar production size-frequency distribution adjusted to Europa impact conditions [after Neukum, 1997].

(i.e., the assumed age of Gilgamesh, the application of lunar cratering chronology to the Jovian system, and the signifi- cance given to crater size-frequency shapes) seem less plausible and inconsistent with current-day astronomical observations of comets. Perhaps the only way to reconcile Model I would be to discover that we are currently in the midst of a Jupiter-family comet "shower," but there is no evidence for this, and it is inconsistent with dynamical modeling.

3.3.2. Age constraints from sputtering. An independent method for constraining the age of Europa's surface comes from estimates of ice sputtering (section 4.7.2). Based on Galileo Energetic Particle Detector (EPD)measurements, Ip et al. [1998] estimate that H20 is lost from Europa at a rate of 10- 20 cm/Myr. While there is plentiful evidence for mass wasting on Europa in Galileo's highest resolution (-6- 35 m/pxl) imaging [Sullivan et al., 1999a], the preservation of small-scale topography on the satellite indicates that sputter erosion is probably not the dominant process shaping the

landscape. The high rate of sputter-induced erosion modeled by Ip et al. seems reasonable only if Europa's surface is young overall. From high-resolution images (e.g., Figure 12a), we estimate that Europa's stratigraphically oldest units (the ridged plains) display topography on vertical scales of-10 m. If sputtering rates have been essentially constant over time, then a surface age of _< 108 years is implied for Europa. Europa's nominal Model II crater age of -50 Myr would imply sputter erosion of-5 - 10 m of material, a scale comparable to but generally less than the smallest scale topography identifiable in Galileo imaging. If the Model I derived age of-1 - 3 Gyr were correct, then -100 - 600 m of surface ice should have been lost

through sputtering, denuding Europa of most of its topography, contrary to observations.

3.3.3. Europa's youngest terrains. Although ap- proximate crater densities can be calculated that are applicable to widespread regions of Europa, typically yielding -50 Myr ages, we are especially interested in the ages of the youngest

24,022 PAPPAI•RDO ET AL.: DOES EUROPA HAVE AN OCEAN?

terrains on Europa, as inferred from stratigraphic relationships. In particular, the regions of chaos appear relatively young, and these are also where the morphology suggests that the rigid ice shell may be thin and could have been disrupted by processes originating just beneath the surface (see section 4.3.). These terrains are not sufficiently extensive to have been superimposed by larger craters.

Chapman et al. [1998] and Bierhaus et al. [1998] have studied the smaller crater populations in Conamara Chaos (see Figure 1 l c) in the hopes of dating these smaller, apparently more youthful units. In places, there are abundant small craters on the surface of the chaos, but they can be definitively identified as secondary craters derived from the distant, fresh, and large crater Pwyll (see section 4.1); the small craters are numerous in visible patches of Pwyll rays and they fall off dramatically with distance from the rays. Well away from the rays, the density of craters a few hundred meters in diameter is surprisingly low, about a factor of 30 less than a simple extrapolation to smaller diameters (with a cumulative slope of -3) from the density of multikilometer craters on typical Europan plains. This density is an upper limit to the density of small primary impact craters if Pwyll or other recent large impacts have contributed some secondary craters far from the visible rays. We note it is possible that small craters might be poorly retained or recognized on the hummocky material of the chaos matrix.

Using Model II and assuming the extrapolation is valid, the far-from-ray crater density suggests crater retention ages on the floor of the chaos less than 1 Myr, an age overall consistent with the expected -3 Myr frequency of Pwyll-sized impacts on Europa. However, as stated above, we do not know the size distribution of smaller comets so the extrapolation may not be correct. Still, it is likely that (1) the stratigraphically young Conamara Chaos is younger in absolute age than typical surfaces on Europa (by about an order of magnitude), and/or (2) there is a prominent lack of small cometary impactors (compared with, for example, the asteroidal size distribution adopted in Model I).

Galileo has surveyed only a very small fraction of Europa's surface at adequate resolutions to measure spatial densities of subkilometer craters. In addition, we know that there are many other mottled terrains at low to middle latitudes on Europa which at low resolution look similar to Conamara Chaos. It is

plausible that there is a wider distribution of chaos terrain ages and that there are many younger than those imaged so far; therefore it is possible that terrains exist with Model II ages significantly younger than a million years.

The young ages derived by the preferred Shoemaker-Zahnle model imply that the processes described subsequently in this article are probably active today because 10 ? years represents such a small fraction of the satellite's history. Thus, if the deformation in Conamara Chaos is due to warm ice or water at

shallow depths, as seems certain (see section 4.5), there may be places on Europa where there is warm ice or water at shallow depths today.

The youthful age of Europa presents an enigma, in that geological relationships in most areas observed suggest that chaos and lenticulae are forming at the expense of the more typical Europan surface of criss-crossing ridges and bands; evidence for ridges and bands cutting across chaos and lenticulae is more sparse. If the surface is in quasi-equilibrium we should be seeing both occurring with comparable resurfacing rates. Con- ceivably the ridge-forming and chaos-forming processes oper-

ate episodically, and chaos-forming processes have been most recently in the ascendancy. This and other possible evolutionary scenarios are explored in section 6.2.

4. The Geological Evidence

Various geological lines of evidence have been used to argue for or against the existence of a Europan ocean, some suggested based on Voyager data (see Pappalardo et al. [1996] for a summary). Some lines of evidence are amplified, some are diminished, and others are newly suggested by Galileo imaging data. Here we provide an overview of nine lines of geological evidence, and their possible implications for the presence of an ocean concurrent with Europa's surface deformation (i.e., very recently and probably today). The arguments summarized here are primarily derived from the large body of work produced to date by the Galileo Imaging Team.

In investigating the implications for subsurface liquid, it is important to evaluate whether Europa's surface features necessi- tate liquid water, or whether "warm ice" or a localized "water-ice slurry" can explain the observed geological deformation. The viscosity of ice is highly temperature dependent [Durham et al., 1997]. Extrapolation of known ice flow parameters to the --100 K average surface temperature of Europa implies a viscosity of -.-10 2• Pa s for the "cold" and rigid near-surface ice [Goldsby and Kohlstedt, 1997a, b]. "Warm" ice, near (but below) its melting temperature, can flow readily in response to geological stresses. At Europan conditions, ice with homolo- gous temperature ~0.8- 1.OT m (where T m is the melting temperature) has a viscosity of only -10 •3 - 10 •5 Pa s [Goldsby and Kohlstedt, 1997a, b; Pappalardo et al., 1998a; McKinnon, 1999]. Moreover, a range of planetologically relevant salts can depress the freezing temperature of water-ice [Zent and Fa- nale, 1986]; therefore, if salts are present within Europa's icy subsurface, some amount of briny liquid may exist at temperatures below 273 K. We refer to such possible configurations of ice crystals and liquid as a "water-ice slurry" or a "partial melt." It is unlikely that a liquid-rich slurry with interconnected pore spaces could exist globally within Europa or persist for long periods of time, as the melt fraction would be expected to rapidly settle out, leaving the ice behind. But such a slurry could exist locally and transiently in response to heating, and might account for some observed Europan landforms. Moreover, in- terstitial briny fluids might significantly lower the viscosity of otherwise solid-state ice. The viscosity of a water-ice slurry is expected to be less than that of warm ice and greater than essentially inviscid water, but rheological studies have not yet been performed on candidate icy partial melts.

4.1. Impact morphologies

Few unambiguous craters were observed in Voyager images of Europa [Lucchitta and Soderblom, 1982; Malin and Pieri, 1986]. Galileo imaging reveals that Europa's rare large (>25 km in diameter) impact craters show morphologies that are unusual compared to those on other similarly sized icy satellites (Figure 4). The impact craters Pwyll and Mannann'an, each --26 km in diameter, are reminiscent of similar-sized complex craters on other icy satellites; thus these craters are believed to have formed entirely within a solid target that exhibited brittle behavior on timescales of the impact event [Moore et al., 1998]. However, these Europan craters have unusually low depth-to-diameter ratios. In fact, stereo-derived topography across Pwyll (Figure 5), the most recent of Europa's large im-


,' .•. : .... ., ...... . ..:..,•--- .:.. :•'•. .•:'• -":' :.:•:*:.•:*•;•L ' ........ ', ,..•

•, :..: . ..•:•;$T; ..... ;:.,..•:• .......... , ..,•. .,.•!• .. . ..... :•;,•

:-:• --

,.•.

.....:?

.• .... c•-. . ..

:.,:.%,:,:.....-


pact features, shows that its interior is at a level similar to its exterior [Giese et al., in preparation]. This is suggestive of isostatic relaxation of large scale crater topography in a weak viscoelastic lithosphere [Passey and Shoemaker, 1982; Tho- mas and Schubert, 1986; Hillgren and Melosh, 1989]. Most relaxation should occur relatively soon after crater formation, when topographically induced stresses are highest and resulting Maxwell times are shortest, and for Ganymede this timescale is modeled as <106 years [Hillgren and Melosh, 1989]. Photoclinometric profiles across Govannan crater (Figure 4c), 10 km in diameter, yield a large diameter-to-depth ratio of 30, implying relaxation of this crater as well [Moore et al., 1998].

Based on Voyager data, it was suggested that the absence of large craters on Europa might be due to relaxation of topography in a low viscosity lithosphere maintained above a liquid layer [Squyres et al., 1983; Thomas and Schubert, 1986; Shoe- maker, 1996]. If this hypothesis were true, then Europa's surface might be more ancient than counts of large craters would suggest. The small number of craters observed by Gali- leo makes conclusions regarding systematic trends in relaxation uncertain, but relaxation does seem to have subdued the

topography of the satellite's larger craters. However, with the exception of the multiringed structures discussed next, we do not observe subcircular patches of missing or destroyed ridged plains that might indicate a population of former large craters that have completely relaxed away, nor is there a surplus of small secondary craters with unidentified source craters. In- stead, the most important process in erasing the satellite's large craters appears to have been resurfacing through formation of Europa's abundant bands, ridges, and mottled terrain (chaos and lenticulae).

Voyager images of Europa showed that Tyre might be a mul- titinged structure [Lucchitta and Soderblom, 1982], reminiscent of larger multiring systems surrounding some palimpsests on Ganymede and Callisto [e.g., McKinnon and Melosh, 1980]. On those Galilean satellites, palimpsests and multiringed structures probably formed where relatively large ancient impacts "punched" through the brittle lithosphere into subsurface low- viscosity material [McKinnon and Melosh, 1980; Thomas and Squyres, 1990]. While some Voyager-based models attribute such low-relief impact scars to viscous relaxation [e.g., Passey and Shoemaker, 1982], palimpsests imaged at high resolution by Galileo show little evidence of long-wavelength relaxation and are interpreted as primary topography (K. B. Jones et al., manuscript in preparation, 1999), along the lines of those Voyager-based models that emphasize the role of fluid-rich ejecta [e.g., Greeley et al., 1982; Croft, 1983]. From Galileo imaging, Europa's known multiring structures, Tyre and Callanish, are inferred to be impact scars [Moore et al., 1998]. This conclusion is based on the presence of circular pits that we interpret to be secondary craters surrounding both features (Figure 4). The lack of regional relief across them and the morphological facies within are analogous to the characteristics of Ganymede's palimpsests (K. B. Jones et al., manuscript in preparation, 1999). Like similar structures on Ganymede and Callisto, the multi-ringed morphologies of Callanish and Tyre are indicative of impacts that have penetrated into a low- viscosity sublayer.

Modeling has been performed to constrain the conditions necessary for ring formation around impacts into ice layers overlying material that is "fluid" on the timescale of the crater formation [Turtle et al., 1998b; Moore et al., 1998]. The results suggest that penetration of an ice shell -10 to 15 km

thick is consistent with the formation of Callanish and Tyre. An ice shell just -3 to 6 km thick would produce multiringed structures from impacts the size of Pwyll or Mannann'an, implying that the ice shell was >6 km thick when such sized impacts formed. Pwyll-sized impacts through an ice shell thinner than 3 km should produce a series of concentric rings along with radial structures [McKinnon and Melosh, 1980], a type of structure which has not been observed on Europa. This argues that an ice shell -6 to 15 km thick (with possible local variations) could account for the morphologies of Europa's large- impact features. This modeling does not require that the low- viscosity sublayer be liquid water but that it have a Maxwell time •-- 2rl/E (where r/ is viscosity and E is Young's modulus) comparable to or less than the timescale of crater collapse, i.e., a few minutes. Though the viscoelastic properties of icy materials are very poorly known at the high strain rates of impact events, if E- 10 7 to 10 •ø is assumed, r/< 10 9 to 10 •2 Pa s is required of the low-viscosity sublayer. Warm solid-state ice (l•,•10 •3 to 10 TM Pa s) ordinarily does not meet this criterion. However, ice will experience much lower viscosities in response to the very high stresses of an impact's transient cav- ity, likely sufficient to permit ring formation (W. McKinnon, personal communication, 1999).

Overall, Europa's large impact features suggest formation in a-6 to 15 km thick lithosphere, overlying a low-viscosity material. Significant crater relaxation implies a weak lithosphere, potentially because the lithosphere has been warm. Modeling of multiringed structure formation suggests that the impacts which created them penetrated to material of relatively low viscosity. This material may be liquid water but instead could be warm solid-state ice which experienced high stresses within the transient crater.

4.2. Lenticulae and Solid-State Convection

Galileo images show that among the stratigraphically most recent features on Europa's surface are circular to elliptical pits, domes, and dark spots -7 to 15 km in diameter, spaced -5 to 20 km apart (Figure 6) [Carr et al., 1998; Greeley et al., 1998b; Pappalardo et al., 1998a]. These features, collectively termed "lenticulae," generally modify and disrupt the preexisting ridged plains, which are widespread and consist of subparallel ridges and grooves that overlap in successive generations. Lower resolution images show that lenticulae, along with regions of chaos (see section 4.5.), appear to comprise the previously defined "mottled terrain" of Europa, which occurs over a large portion of the satellite's imaged surface [Lucchitta and Soderblom, 1982]. Therefore we infer that these pits, domes, and spots are widespread across Europa's mottled terrain.

The similarity in size and spacing of pits, domes, and spots and the gradation in morphology among them suggest that they are genetically related. The features have all altered the original topography through upward or downward deformation; most have also disrupted and/or resurfaced the preexisting plains (Figure 6). The features have been interpreted to represent a sequence in which the surface was upwarped by diapiric intrusion beneath a relatively thin rigid surface layer, along with localized disruption of the surface and/or extrusion of buoyant diapiric material [Head et al., 1997; Pappalardo et al., 1998a].

As discussed in section 4.3, one alternative model for some

lenticulae is that they indicate the rise and extrusion of viscous cryovolcanic materials [Fagents et al., 1998]. Extrusion of icy


.:

....

. ß

..:

diapirs is another type of cryomagmatic extrusion, meaning that diapirism may account for the range of intrusive and extrusive lenticula morphologies observed. If lenticulae instead formed as dike-fed laccoliths, then their dimensions are predicted to vary significantly as a function of the driving pressure and intrusion depth [Pollard and Johnson, 1973], and they might be expected to be aligned along the trends of feeder dikes; instead, lenticulae show generally consistent size and relatively random distribution in any given region, more consistent with diapirism [Spaun et al., 1999]. An additional model for lenticula formation might include melting and surface disruption [Greenberg et al., 1998b], perhaps above a narrow thermal plume that rises from the silicate mantle to locally melt a floating ice shell. It seems extremely unlikely that a narrow plume could rise and maintain its width through ---100 km of liquid water while being sufficiently warm and positionally stable to locally melt the ice shell.

Diapirism can be triggered by compositional layering, with density inversions leading to gravitational instability. Such may have occurred on Triton to create that satellite's "canta- loupe terrain" [Schenk and Jackson, 1993]. Calculated growth times for plausible density configurations in a cold ice shell on Europa are-•10 Myr to 1 Gyr, long compared to the nominal -•10 Myr age of the surface [Moreau et al., 1997; J. Moreau, personal communication, 1998]. Tidal heating could significantly lower the viscosity of Europan ice and speed diapirism by inducing thermal convection. Thermally driven diapiris are the manifestation of solid-state convection, which may occur within any icy body if a sufficient thermal gradient exists, especially if the temperature at the base of the ice layer is near the solidus [Schubert et al., 1986; McKinnon, 1998].

Some Voyager-based models for the internal structure of Eu- ropa suggest that tidal heating might be sufficient to maintain a liquid water ocean beneath a -•10 to 30 km thick conductive ice shell [Cassen et al., 1982; Squyres et al., 1983; Ross and Schubert, 1987; Ojakangas and Stevenson, 1989a]. Reynolds and Cassen [1979] determined that a liquid water layer should drive solid-state convection within the outer ice shell if the

shell ever cooled to 30 km thickness. Pappalardo et al. [1998a] and McKinnon [1999] use more realistic estimates of

Figure 6. Galileo views of lenticulae. (a) Regional view of an area dense with lenticulae, illustrating their common -• 10 km diameters and morphological variety as pits, domes, and spots (15ESREGMAP01; 230 m/pxl). Enlarged views of lenticulae (from observation E6ESDRKLIN01; 180 m/pxl): (b) a dome showing evidence for localized uplift of the preexisting ridged plains; (c) pits showing surface disruption and/or extrusion to create hummocky material; (d) lenticula with an annu- lus of smooth, relatively low albedo material, suggesting thermal alteration of the surface or localized extrusion of melt; (e) lenticula with convex margins that stand slightly above the surrounding terrain, suggesting extrusion of viscous material. (f) High-resolution view of a lenticula illustrating in situ disruption of preexisting ridged plains and replacement by hummocky material, in a manner similar to the formation of larger chaos regions (11ESMORPHY01; 33 m/pxl). These features may represent the surface expression of diapirs that have risen buoyantly through Europa's ice shell [Pappalardo et al., 1998a]. Illumination is from the right in all images; north is toward the top in all images except Figure 6f, which is south toward the top.

PAPPALARDO ET AL.' DOES EUROPA HAVE AN OCEAN? 24,027

ice viscosity and recognize that the effective viscosity of Eu- ropa's ice shell is controlled by the tidal kneading of Europa, necessitating modeling of its stress- and strain-dependent theology. In this way, they estimate that Europa's icy shell has a viscosity as low as -10 •3 Pa s at its base and could convect when the shell is -10 - 25 km thick, potentially before the shell's steady state conductive equilibrium thickness is reached. In turn, convection may increase tidal dissipation within the shell, increasing the vigor of convection and perhaps allowing for localized runaway convective overturn [McKinnon, 1999].

Solid-state convection would suggest that Europa's temperature profile is adiabatic beneath a thin conducting lid (see section 4.4). If ice convection occurs above a liquid water layer, then the temperature of the convecting ice is expected to be -260 K [McKinnon, 1999; C. Sotin, personal communication, 1998]. This temperature is above the eutectic point of a variety of brines [Zent and Fanale, 1986]. In turn, a melt component within Europa's ice shell could substantially affect the viscosity and temperature profile of the convecting ice matrix.

Rathbun et al. [1998] have modeled the diapiric rise of warm ice plumes through Europan ice. Assuming Newtonian theology, they show that diapirs of initial temperature -273 K, and with initial dimension appropriate to produce Europa's lenticulae (radii -2 to 5 km), can rise through 5 40 km of ice at -260 K before they cool to ambient temperature and their rise is halted. Because the total thickness of Europa's H20 layer is likely -100 km (see section 2.1), and because diapirs are expected to initiate at the warm base of Europa's ice shell, a diapir rise distance of <_ 40 km argues for the presence of a liquid water ocean 60 km or more deep during lenticula formation. Rathbun et al. determine that diapirs will rise to the surface in •< 10 • years, significantly less time than the estimated average age of Europa's surface.

Solid-state convection is an efficient heat loss mechanism.

Cassen et al. [1982] predicted that convective heat loss would freeze a Europan ocean in -108 years, unless sufficient heat is continually supplied to the convecting ice, as through tidal heating (see section 2.2). Because tidal dissipation is strongly dependent on ice temperature [Ojakangas and Stevenson, 1989b], tidal heating may be enhanced in an ice shell warmed by convection, perhaps allowing an ocean to be maintained [McKinnon, 1999].

Models of Europan convection are relatively new. Further work is required to determine whether a subsurface ocean is necessary to drive convection within the satellite, and whether an ocean can be maintained below a convecting ice shell. For example, scenarios might be envisioned in which solid-state convection does not require a liquid layer but is driven by a steep thermal gradient through a thick ice layer.

4.3. Cryovolcanic Features

Voyager images show little evidence for volcanic units that have flooded widespread portions of Europa's surface. Higher resolution Galileo images show the presence of numerous features of possible "cryovolcanic" origin, suggestive of the existence of some amount of liquid water in the subsurface (Figure 7). The surface of liquid water exposed to a vacuum environment would initially be vigorously disrupted as a result of nucleation and growth of bubbles of H20 vapor [Cassen et al., 1979; Allison and Clifford, 1987]. This evaporation will ex- tract latent heat from the water, causing rapid cooling and ice

formation. The turbulent disruption would delay the formation of a coherent ice carapace, but a layer -0.5 m thick will sup- press further evaporation and permit flow propagation [Cassen et al., 1979; Allison and Clifford, 1987; Wilson et al., 1997]. Ice-liquid mixtures or carapace development may cause flow morphologies to differ significantly from simple liquid floods.

Patches of relatively dark, smooth material occupying topographic lows may represent effusions of low-viscosity fluid onto Europa's surface. Figures 7a and b shows two such examples: a smooth deposit some 3 km in diameter, inferred to have a thickness of <50 m and a volume of <0.5 km 3 [Head et al., 1998c], and a portion of a larger dark patch -23 km across, within which embayment relationships are inferred. These morphologies suggest the extrusion of water-rich melt onto the surface.

Effusion of more viscous material is suggested by certain dome-like lenticulae that exhibit well-defined margins and surface textures distinct from the surrounding terrain (e.g., Fig- ure 6e). These morphologies indicate viscosities that preclude effusions of liquid water or warm ice, if they have extruded from central vents [Fagents et al., 1998]. More appropriate candidate extrusive materials might include slurries containing compounds such as ammonia [Kargel, 1991], ice-water slurries, or liquids injected within an ice carapace [Fagents et al., 1998].

Explosive cryovolcanism and mantling by "cryoclastic" material is a possible origin for dark, reddish brown, diffuse- margin albedo features associated with triple bands, lenticulae, and some cracks (Figure 7c) [Greeley et al., 1998a; Head et al., 1998c; Fagents et al., 1999]. Explanations for these features that do not require liquid water include changes in ice grain size [Dalton and Clark, 1999] or formation of lag deposits in response to subsurface heat sources (dikes or ice diapirs) [Fagents et al., 1999].

The greater density of water with respect to ice precludes buoyant rise of liquid water through a pure water-ice crust. However, the driving force required to erupt material onto the Europan surface could be provided in several ways (Figure 8).

1. The possible presence of volatile compounds (such as CO2, CO, CH4, SO 2 [Crawford and Stevenson, 1988]) dissolved within the liquid water might promote explosive cryovolcanism (Figure 8a). Release of confining pressure on water beneath or within an ice layer would permit nucleation of bubbles of the volatiles, lowering the bulk density of the fluid and producing a positive buoyancy. Further bubble growth by decom- pression and coalescence would ultimately disrupt a bubbly liquid into a rapidly expanding gas-droplet spray, which would erupt to form a gas-dominated explosive plume. The separation of the gas and water phases implies that large-volume water flows would be difficult to achieve by this mechanism. Craw- ford and Stevenson [1988] suggested that this mechanism might operate in water-filled cracks propagating upward from an ocean through a thin ice shell.

2. Buoyancy-driven ascent is possible if contaminants in the ice and/or water significantly modify the density of one phase with respect to the other [e.g., Wilson and Head, 1998]. Alternatively, lithospheric density could be increased by formation of clathrates [Lunine and Stevenson, 1985] or the presence of non-ice material (-4 to 9% by volume of silicate particles would produce a density contrast of 10 to 100 kg m-3).

3. Pressurization of isolated liquid reservoirs within the ice shell could arise from regional or tidal stresses, or from ice crystallization-induced volume changes in the reservoir. These


...

..

: ß

ß . . ..--.•.: R.:.. ß .•.:::.::...•,: . .:,•:•,:.....::.:....,q• ...:;:::'}:.::•.' .' .. :.:

:-;•,• L.?.!:.-;:.;;?'"'•:'•*'""•'ø•'•--'-'••••• ""'• ?"'•-'• ........... ""*'"•'•"•••• • •- *••:--•-• -..........-. -.....•:•: ' ':'"'""•- -;':•.;:!½;;½:i !!!!:'"'"'"'"'•;;;

•**..•i•::..?.;•....'-•'..<...';•i::,X::,;,:x;:i•:;:;:.: .::::j!!, :-: .......... '."'• .... .**:*::.::'.'.--t ::::•;-::!;',;:**:::,::.::-...:..-.;•.:::..::::•:;•, "27':' :'"•-":"

.!;• .:-:.::.:.:?}• ............ : ......... ;;;•:., ......................... . ............... *¾...:.'•C""'",,,::***--..---..-.'**•':-'*•'""Sig:,:i•.?•';.- "- - '••••....-..•a: ß :. ..;•::.::.:r,.-- (:.• ,::..., ............................. a.:,,:,;; ,;s.s. :.,½ .........

:. ............. :-:L.• ..... •:•.;• •-::..-.::•;.. -•:.. - a•½- .;;:,.:..;:..:...;:..:,.• ,:•:!•:,•.-•:--.• s•: -::..:::• .....

'" .....•.½'?:½?..;.;:-::;;.;.: .....:.½:'"".-'-•:½:.•-;.•'... ,•-•:,::,:::•.-'"'•'-.<•,;•-•,•;,?;• .....................

. ..................... ,,..-:::,,½: ........... ' ..................... •s•:•.*;•:'*-.'-..;•8-.?$-.::'-.½•sw" ""• --" ' ':':•!.

.,• ::.:::. .............. ß ...<-•s?x::?-..?.s::½•: ½•..•;• .½--'•.• .... •:•:•;;:,:-'i:•-'.-.;•: ..................... ß ...... :-::-

i..•:•f•::-• ...................................... ::•!;E'"'"'"":':"'" ""•:•:;"-"4•;•!,....,?*'•--*--'-.--.;•..... • ' •f/i•i:ii• 'tF"•'•:::.::?a•:.......•;:.i:•$½$:•,::;;•;•4:.-..JB½:•:X;,;.::,,S--;>-a :**" '*•'•:• --.., -.•;•i:• "** '•--- - -**- --'•5-.-**-a:;½:•aa•-;.,;•?:•j•i•:•'• ............. :';'•;--'-'-•...:........• ,•:::*-*," -':•:::•: ;•:•:*S':•:' :-•-".-.--':-'-'•:•-..j;.'•::::•-:' '-•.::- .½' r•::•::•,;•i:•:•!.i•!i•*:D•;•""*:•*:•*>:*•i•*•j•i:?•:•1•*;•;•;X?•;•*:•;}•?•,•,•`;•jjc :'""• ............ *-,;•:•.½:'½•:•.•;.'.::.;/::*•.•.5•::•;• ...... •:::-:.•:½

cryomagma chambers might represent the residual water pockets of a solidifying global ocean [Wilson et al., 1997], localized melts within the ice shell, or the consequence of heating at the interface between a thick ice layer and the silicate interior (Figure 8b). Pressure-driven ascent of a cryomagma could occur through fractures between the zone of overpressurization and the surface.

4. Initiation of fractures and their subsequent cyclic opening and closing due to diurnal tidal stresses might be another mechanism for pumping water/ice mixtures to the surface [Greenberg et al., 1998a].

Models for cryovolcanism on Europa are severely limited by poor compo3itional constraints. Uncertainty exists regarding the presence, amounts, and composition of volatiles and contaminants in the ice and presumed liquids. This precludes more definitive determination of the origins of the landforms in question, and their implications for Europa's interior structure and the presence of an ocean. Based on our current knowledge, we conclude that none of the inferred cryovolcanic feature types requires a global ocean. In fact, other configurations are pref- erable for some features. For example, if buoyancy mechanisms cannot be invoked to explain liquid effusions, then the requirement for pressure-driven eruption could be met by having discrete water reservoirs; it would be difficult or impossible to achieve local pressurization of a global water layer [Wilson et al., 1997]. Cryoclastic volcanism requires that a volatile- bearing liquid be depressurized by the opening of a fracture. Detailed modeling of fracture and dike propagation in ice are necessary to understand whether these processes would be more likely to occur in the presence of an ocean or above pressurized cryomagma chambers.

The question of Europa's ocean remains unresolved on the basis of the potential cryovolcanic features discussed here. These features are significant in that they offer good evidence for the presence of some amount of subsurface heating and melting in Europa's recent past, and perhaps currently. However subsurface liquid might be contained within isolated pockets, so cryovolcanic features have no certain bearing on the presence or absence of a global subsurface ocean.

4.4. Pull-Apart Bands

4.4.1. Morphology and origin of pull-apart bands. Opposing sides of relatively dark wedge-shaped and gray bands on Europa can be closed and thereby reconstructed

Figure 7. Examples of candidate cryovolcanic deposits on Europa.. (a) A circular smooth deposit probably emplaced as a cryovolcanic eruption of low-viscosity material, perhaps liquid water (E4ESDR•AT02; 20 m/pxl; high-incidence illumination from the right). (b) Outer margin of a dark patch --23 km across; arrows mark places where dark material appears to em- bay older ridged plains (14ESDRKSPT01; 23 m/pxl; solar illumination from the right). (c) Example of a triple band (a ridge flanked by diffuse dark material) imaged at high resolution; a possible model suggests the dark material was entrained in a ballistically emplaced cryoclastic deposit (14ESTRPBND01; 80 m/pxl; solar illumination from the lower right). Note the different scales of each image; north is toward the top in each.


silicate

b

Figure 8. Schematic illustration of two mechanisms for erupting negatively buoyant water through an ice lithosphere. (a) A volatile-bearing water body can experience a decrease in pressure through the opening of a fracture (A) or at the tip of an upward-propagating crack (B). The consequent volatile exsolu- tion and expansion may result in eruption of a gas-driven spray of water droplets. (b) An isolated reservoir of water within or at the base of an ice lithosphere may develop an excess pressure (ziP) due to volume change from ice crystallization (C) or regional stresses (D). The resulting stresses around the reservoir could cause fractures to open and may be able to drive water to the surface to create a cryovolcanic eruption.

with few gaps, restoring structures that were apparently displaced as the bands opened along fractures [Schenk and McKin- non, 1989; Pappalardo and Sullivan, 1996; Sullivan et al., 1998]. Collectively, such features have been termed "pull- aparts" [Greeley et al., 1998a]. Reconstructions of pull-apart bands imply that Europa's surface layer has behaved in a brittle manner, separating and translating atop a low-viscosity subsurface material, with the region of separation being infilled with relatively dark, mobile, endogenic material. Thus, pull- apart bands offer compelling evidence for warm, mobile material in the shallow Europan subsurface at the time of their formation.

The morphological relationships of pull-apart bands are most clearly seen in Galileo images within the region southwest of Europa's antijove point (Figure 9), where small-scale plates separate darker material and can be locally reconstructed through plate translation and minor rotation of small subplates [Belton et al., 1996; Tuffs et al., 1997; Sullivan et al., 1998]. The smallest recognizable plates in the region are approximately 4 km across, suggesting a similar or lesser depth to the mobile material. Fracture mechanics and crevasse theory suggest that Europan cracks probably can penetrate through a brittle lithosphere this thick, to reach a mobile sublayer [Golom- bek and Banerdt, 1990; Leith and McKinnon, 1996].

Analogies have been made between pull-apart formation on Europa and the formation of leads in terrestrial sea ice [Pap- palardo and Coon, 1996; Greeley et al., 1998b]. However, analysis of Galileo images suggests that their formation may be more analogous to terrestrial spreading centers. In terrestrial plate tectonics, "spreading" occurs where the lithosphere has been rifted through, and new lithospheric material is intruded and extruded as the opposing original lithospheric plates are pulled apart. Regional-scale Galileo images of pull-aparts show an internal structure of ridges and troughs trending subparallel to each other and the boundaries of the band (Figure 9), and an overall bilateral symmetry suggests that a spreading analog may be appropriate [Sullivan et al., 1998].

High-resolution (-15 - 55 m/pxl) images of pull-aparts were acquired during Galileo orbit El2 (Figure 9c), including stereo transects. Examination of several pull-apart bands at high resolution shows that they contain similar morphological elements, though not each pull-apart band contains each element [Prockter et al., 1999a]. Their margins are generally sharp, and some show a rounded bright rim (a bounding ridge). A narrow central trough is common and is remarkably linear and uniform in width along the length of each band. To either side of this trough a hummocky textured zone commonly occurs, and outside of the hummocky zone are subparallel ridges and troughs 300 - 400 m in width.

By analogy to terrestrial spreading centers in oceanic crust [e.g., Macdonald, 1982], the axial trough may be the site of plate separation, with the flanking hummocky material emplaced symmetrically on either side of the band axis, analogous to the terrestrial neovolcanic zone which is formed through cycles of volcanic and tectonic activity. Moreover, the subparallel ridges and troughs may be analogous to the abyssal hills observed along terrestrial midocean ridges, formed by normal faulting in response to extension and spreading. Stereo imaging of the dark pull-apart band of Figures 9b and 9c indicates that the entire band stands topographically higher than the surrounding ridged plains, consistent with emplacement of relatively buoyant material in a fashion broadly analogous to emplacement of terrestrial oceanic crust.

In summary, along with the analogous processes of complete lithospheric separation and creation of new lithospheric material, the units and characteristics of Europan pull-apart bands are similar to those in terrestrial oceanic-spreading environments. This suggests that a terrestrial spreading and rifting model is an appropriate analog for Europan pull-apart bands [Sullivan et al., 1998; Prockter et al., 1999a].


4

10

'••...• lOO

25 •

o

0 2 4 6 8

differential stress for failure (MPa)

4.4.2. Models of lithospheric structure. It has been speculated that the subsurface mobile material beneath pull-apart zones might be liquid water or brine, or warm ductile ice [Schenk and McKinnon, 1989; Golombek and Banerdt, 1990; Pappalardo and Sullivan, 1996]. Strength envelopes can be constructed to constrain the strength of Europa's ice shell and evaluate whether ductile subsurface ice could serve to

decouple a cold upper lithosphere from the lower portion of Europa's ice shell. To update the ductile strength curves potentially applicable to Europa's ductile lower lithosphere [Golombek and Banerdt, 1990], we use the "superplastic" flow law for ice recently derived by Goldsby and Kohlstedt [ 1997a, b], which describes strain rate as a function of temperature, differential stress, and grain size as it varies in each of three creep regimes. Figure 10 displays the ductile strength of Europa's ice shell for a variety of possible conductive thermal gradients as labeled, assuming a surface temperature of 100 K, ice grain size of 1 mm, and imposed strain rate of 2 x 10 '1ø s '1 (typical of Europa's tidally strained shell [Ojakangas and Stevenson,

10 1989b]). An upper limit to the strength of Europa's brittle lithosphere

assumes it is composed of polycrystalline ice with tensile Figure 10. Model strength profiles through Europa's ice strength 2 MPa, with a Griffith failure criterion providing a lithosphere. Ductile strength curves (solid) are based on the good approximation of strength as a function of lithostatic rheology of pure ice [Goldsby and Kohlstedt, 1997a, b] assum- overburden [Golombek and Banerdt, 1990]. As an opposite ing a strain rate of 2 x 10 '•ø s '1 [Ojakangas and Stevenson, endmember, the brittle lithosphere may be pervasively 1989b], for linear thermal gradients as labeled (K km'l). The

fractured with its strength controlled by the frictional failure of strength of Europa's brittle lithosphere is bounded by assumptions of a Griffith criterion, appropriate to annealed polycrys- ice, with essentially no cohesion in the most extreme case talline ice (dashed), and a cohesionless frictionally controlled [Beeman et al., 1988]. On Figure 10, these two brittle strength ice lithosphere (dotted). Lithospheric strength at the brit- end-members are shown as dashed and dotted, respectively. tie/ductile transition depth for a given thermal gradient is rep- Various methods have been applied to estimate the thick- resented by the intersection of the appropriate ductile and brit- ness of Europa's brittle lithosphere, suggesting a brittle/ductile tle curves. The open circle at the lower extent of each ductile transition depth-1 - 2 km (Table 1). This range seems reason- curve represents the ice melting depth for that thermal gradient. able locally on Europa, considering that Europa shows even Dot-dash line represents the very low strength (0.1 MPa) of ice at-260 K, as may be appropriate to a convecting (essentially isothermal) lower lithosphere or "asthenosphere." If a convecting asthenosphere exists, then melting of water-ice does not occur at the shallow depths implied by the open circles. For high thermal gradients, an order of magnitude change in strength occurs over a small depth interval from the brittle/ductile transition depth to the nearly strengthless lower portion of the ice shell.

more evidence for surface deformation than does Ganymede, which is estimated to have had a brittle lithosphere -2 km thick in grooved terrain at the time of its geological deformation [e.g., Collins et al., 1998]. In a conductive ice lithosphere, this would imply a thermal gradient-50- 100 K km '1, essentially independent of assumptions regarding the degree of fracturing of the brittle lithosphere (Figure 10). The corresponding maxirm•m qtrength of the lithosphere (at the

Table 1. Galileo-Based Estimates for the Thickness of Different Portions of Europa's Ice Ice Shell Portion Thickness, km Estimation Technique

Effective elastic lithosphere 0.4 Ridge-induced flexure Effective elastic lithosphere 0.1 -0.5 Dome-induced flexure Elastic lithosphere 1 - 2 Flexure correction factor

Tensile lithosphere < 1 Fracture mechanics

Local rigid portion of shell < 3 Dimension of chaos plates

Local shell thickness 2 Buoyant blocks Local shell thickness 0.2 - 3 Buoyant blocks

Convecting sublayer > 2 - 8 Initiation of convection and spacing of lenticulae

Convecting sublayer < 40 Diapir rise time

Whole ice shell > 5 - 20 Initiation of convection Whole ice shell > 12 - 25 Initiation of convection

Whole ice shell > 6- 15 Lar[[e impact morphologies *As adjusted by McKinnon [1999].

Reference

Tufts et al. [ 1997] Williams and Greeley [1998] this work

Moore et al. [1998]

Carr et al. [1998]

Carr et al. [1998] Williams and Greeley [1998]

Pappalardo et al. [1998a]

Rathbun et al. [1998]

Pappalardo et al. [ 1998a]* McKinnon [1999] Turtle et al. [1998b]


brittle/ductile transition depth) is -1 - 5 MPa. At these high thermal gradients the transition from maximum lithospheric strength to negligible (-0.2 MPa) strength in the ductile lower lithosphere occurs within a depth interval of just -0.5 - 1 km. We note that geological indicators of very a thin brittle lithosphere on Europa (Table 1) could be misleading if heat is distributed nonuniformly within the satellite's icy shell. For example, tidal heating may be preferentially concentrated along fractures [Stevenson, 1996] or in warm diapirs [McKinnon, 1999].

If we assume thermal gradients of-50 - 100 K km -• through a purely conductive ice lithosphere, it is expected that melting would occur at a depth h < 3.5 km. For Europa's average surface temperature of about 100 K [Ojakansas and Stevenson, 1989b], and an ice thermal conductivity k = 2.7 W m -• K -• (appropriate to an average lithospheric temperature of about 180 K [Hobbs, 1974; Ojakansas and Stevenson, 1989b]), the implied heat flow q = kAT/h > 0.13 W m '2. If regions of thin lithosphere (Table 1)are representative of Europa as a whole, this would imply -4 times the global, time-average heat flow predicted by steady-state models of Europa's current tidal heating [Ojakan- gas and Stevenson, 1989b].

An alternative scenario is that a conductive lithosphere exists above a convecting lower ice lithosphere (or convecting ice "asthenosphere"). Below the conductive lithosphere, Eu- ropa's asthenosphere may be nearly isothermal. The temperature of convecting ice is predicted to be -260 K [McKinnon, 1999; C. Sotin, personal communication, 1998], with a corresponding strength of-0.1 MPa and effective viscosity of -10 TM Pa (assuming a grain size of 1 mm and strain rate of 2 x 10 -•ø s '•) [Pappalardo et al., 1998b]. For the case of a < 2 km brittle lid, the implied heat flow is > 0.22 W m -2. If regions of thin lithosphere (Table 1) are representative of Europa as a whole, this model would be even more challenging than the conductive scenario to models of Europa's predicted current average heat flow (a conundrum similarly recognized by Ruiz and Tejero [1999] and Ruiz [1999]). However, preliminary modeling of convection with a Europan ice shell predicts that upwelling convective plumes may be significantly warmer than neighbor- ing downwelling regions [McKinnon and Gurnis, 1999], implying that Europa's heat may be very nonuniformly distributed in a convecting ice shell.

In a convecting ice asthenosphere, the transition from conductive to convective heat loss would occur along an upper thermal boundary layer, which would also correspond to a rheological transition. The sharp decrease in strength from the brittle-ductile transition to the depth of a convecting ice asthenosphere is a prime candidate for the sharp decollement implied by Europa's pull-apart bands. A convecting ice asthenosphere could in turn, but does not necessarily, overlie a deeper liquid water ocean. Hence we concur with previous workers [Schenk and McKinnon, 1989; Golombek and Banerdt, 1990; Pap- palardo and Sullivan, 1996] that pull-apart bands do not provide direct evidence for a liquid water ocean, but may have opened above warm ductile ice.

4.5. Chaos

Images obtained during Galileo's sixth orbit first revealed that portions of the relatively dark, mottled terrain of Europa consist of chaotic terrain [Carr et al., 1998]. Chaos regions are typically composed of polygonal blocks of preexisting ridged plains that have translated and rotated with respect to one an-

other within a matrix of hummocky material. Chaos matrix material can be either low-lying or high-standing relative to the surrounding plains [Collins et al., 1999; Chuang et al., 1999]. In Conamara Chaos (Figure 11), -60% of the preexisting terrain has been replaced by and/or converted into low- lying matrix matehal, and many of the surviving blocks can be reconstructed in jigsaw-puzzle fashion [Spaun et al., 1998a]. Galileo imaging has revealed that chaos is widespread across Europa and, along with smaller, subcircular lenticulae, com- prises the satellite's widespread regions of mottled terrain. Nearly everywhere it has been observed to date, chaos appears to be stratigraphically young: chaos material consistently crosscuts the ridged plains and its constituent bands and ridges, while only some individual troughs and rare individual ridges crosscut chaos materials [Carr et al., 1998; Greeley et al., 1998b; Pappalardo et al., 1998a; Spaun et al., 1998a; Prockter et al., 1999c].

Chaos regions have been interpreted as areas of amplified heat flow and perhaps complete local melting of the Europan surface [Carr et al., 1998; Williams and Greeley, 1998; Green- berg et al., 1998b]. In a melting model, blocks are analogous to buoyant icebergs, which have translated by plowing through a thin ice rind formed atop freezing liquid water. Some chaos blocks appear to be tilted (Figure 1 lb), and this could be consistent with isostatic adjustment of blocks immersed within a low-viscosity material such as water or a water-ice slurry. With the assumption that chaos blocks have been buoyant within a liquid, their heights above matrix material have been used derive a local ice shell thickness of-0.2 - 3 km in Conamara

Chaos [Carr et al., 1998; Williams and Greeley, 1998]. To account for Europa's widespread chaos, a melting model requires heat output much greater than predicted from existing models of tidal and radiogenic heating [McKinnon, 1997]. Assuming an initial ice shell just 2 km thick, the latent heat of fusion necessary to melt 60% of the ice volume of Conamara Chaos requires -1021 J of heat. Models of Europa's tidal heating predict a current heat flow of a few tens of milliwatts per square meter, so the relatively small Conamara region (just 0.024% of Europa's total surface area) would require the entire predicted heat output of Europa concentrated there for several hundred years to over- come the latent heat of fusion of ice; additional heat would first be required to raise the icy lithosphere to the melting temperature. If the necessary heat energy were gained over a longer time, the fraction of Europa's total heat budget necessary to cause local melting would be proportionally reduced, but heat loss by conduction would then become significant, again causing difficulty for the melting model. It should be reiterated that significant uncertainties exist in the tidal heating predictions, as discussed in section 2.2.

At the other end-member, it is possible that solid-state ice has risen diapirically toward the surface to form chaos, disrupt- ing the relatively cold and rigid upper lithosphere in a manner similar to that proposed for lenticulae [Pappalardo et al., 1998a; McKinnon, 1999; C. Sotin, personal communication, 1998] and translating fragmented slabs of colder lithospheric material. Chaos regions appear to have formed contemporane- ously with the similarly youthful lenticulae, and examples of lenticulae surround some large areas of chaos; moreover, many lenticulae contain regions of chaos-like material or "micro- chaos" [Head et al., 1997; Pappalardo et al., 1998a; Spaun et al., 1999], suggesting that the origins of lenticulae and larger chaos regions may be related. If lenticulae have formed by


thermal convection, chaos areas may represent regions of more intense diapiric upwelling, perhaps related to local variations in the scale or intensity of endogenic heating or in the character of the rigid near-surface layer. Because tidal heating is expected to become intensified within warm convecting ice, a modest amount of convective overturn might lead to runaway convection, perhaps giving rise to chaos regions [McKinnon, 1999].

To test these endmember models, a portion of Conamara Chaos was imaged at 10 m/pxl during Galileo orbit El2 (Figure 1 l c). These high-resolution images reveal that the matrix has a rough and hummocky texture. The matrix material shows no obviously lobate boundaries as might be expected if it was emplaced as solid-state ice, nor are smooth patches seen as might be expected if the matrix material was emplaced as a liquid (compare with the much smoother materials seen at similar resolution in Figure 7). Small-scale blocks are visible within the chaos, with a lower block size limit of -2 km before preexisting ridged terrain is no longer recognizable on them. This suggests that the texture of the smallest blocks might be destroyed by significant mass wasting and/or by tilting of the small blocks, once dissected to this critical size. Some candidate small block remnants are traceable within the matrix, im-

plying that the matrix in part consists of preexisting lithospheric blocks which were disrupted in situ.

The fact that small-scale blocks have been mobilized and

possibly tilted within the chaos challenges the notion that the matrix was emplaced as solid-state ice because the timescale of block movement in warm ductile ice is expected to be longer than the timescale of thermal diffusion causing blocks to cool in place [Collins et al., 1999]. On the other hand, the absence of very smooth materials challenges the model in which the matrix material was emplaced as liquid water. It is expected that liquid water emplaced in a vacuum would be turbulent, perhaps quickly forming an ice rind, perhaps accounting for the rough texture of chaos matrix. The existence of very smooth materials elsewhere on Europa (section 4.3 and Figure 7a) demonstrates that they do occur and are probably the signature of very low viscosity extrusion. However, a chaos melt zone plausibly could be significantly thicker and emplaced in a different style than a cryomagmatic melt deposit, and might develop a different surface character as it cools.

It is possible that the chaos matrix was emplaced as a slurry of partially melted icy material [Head and Pappalardo, 1999]. Impurities such as salts within the ice lithosphere would depress its melting temperature, so heated icy materials on Eu- ropa could consist of a mixture of briny liquid and solid-state components. Heating and partial melting of the near-surface could occur above warm ice diapirs, especially if low melting point hydrated salts exist within the lithosphere, as suggested

Figure 11. Conamara Chaos seen at three scales of imaging. (a) At -180 m/pxl, blocks of preexisting ridged plains material are inferred to have translated and rotated within a low-lying hummocky matrix (E6ESDRKLIN01). Stratigraphic relationships imply that the chaos is young relative to the surrounding plains. (b) At-54 m/pxl, structures can be identified on individual high-standing plates, and some plates (arrows) appear to have tilted (E6ESBRTPLN01). (c)At -10 m/pxl, high-standing blocks and massifs as small as-1 km across (arrows) are identified, surrounded by hummocky matrix material (12ESCHAOS_01). The rough texture of the matrix is distinct from the extremely smooth texture of some plains patches imaged at similar resolution and scale which are thought to indicate liquid extrusion (cf. Figures 7a and 7b). The upper right portion of the 10 m/pxl image (including the plate of preexisting ridged plains) has been brightened by ray material from the crater Pwyll,-1000 km to the south-southeast. This bright material shows an excess of craters <_ 100 m in size, inferred to be secondaries (see section 3.3).


by Galileo near-infrared mapping spectrometer (NIMS) multispectral data [McCord et al., 1998, 1999] (see section 7.1).

Chaos suggests the presence of warm icy material at shallow depths, potentially in a molten or partially molten state. However, chaos as direct evidence for an ocean is uncertain, as

chaos could be the manifestation of localized partial melting triggered by solid-state diapirism. Work remains in understanding the process of ridged plains destruction, and in quantifying the viscosity, emplacement style, and melt fraction of chaos materials.

4.6. Ridges

Long, narrow ridges less than a few kilometers wide and up to several hundred kilometers long were observed near the terminator on the best 2 km/pxl Voyager images of Europa [Smith et al., 1979; Lucchitta and Soderblom, 1982]. Few ridge details could be distinguished at this resolution, but Malin and Pieri [1986] estimated maximum relief as not exceeding a few hundred meters and recognized ridges as the most abundant landform on the surface.

Much higher resolution Galileo images, many obtained under near-terminator illumination conditions, have led to a new

appreciation of the importance of ridges as a Europan landform, and have revealed their true geomorphic complexity (Figure 12). Singular ridges seen at lower resolutions are revealed at higher resolution to be doublet ridges with central troughs. These doublet ridges predominate, and true singular ridges are rare. Triple bands, named primarily for their alternating albedo appearance in low-incidence angle, lower resolution Voyager images, have been revealed as complex, multicomponent ridges flanked by darker units. The darker materials might be produced by mantling of surrounding terrain by darker materials, or in situ thermal alteration [Greeley et al. 1998a; Head et al., 1998a; Fagents et al., 1999]. While most ridges are linear to curvilinear in planform, some ridges show an enigmatic repeating cycloidal planform (see section 4.9.2 for discussion).

Perhaps most important, high-resolution Galileo images show that the surface of Europa in most places is completely covered with intricately overlapping doublet ridges. If ridges are the most ubiquitous landform of Europa, what can their morphology and inferred origin tell us about the theology and state of subsurface materials? This question can be addressed by considering how proposed ridge models account for observed ridge morphologies, and the implications each model carries for subsurface theology and the presence and distribution of liquid water.

Several ridge formation models have been proposed on the basis of ridge morphologies observed in high resolution Gali- leo images (Figure 13). None of the models completely ac- counts for all the ridge characteristics observed in the images. Study of the ridge origin problem is in its preliminary stages; current models are rapidly evolving and need testing by laboratory or numerical simulations and with additional high-resolution observations of Europa. Five major ridge formation proposals exist to date: (1) ridges represent linear debris piles deposited at the surface by explosive volcanism along fractures; (2) ridges represent linear debris piles created by lateral stresses periodically opening and closing fractures, squeezing material to the surface as a slurry of crushed ice and boiling water; (3) ridges represent preexisting surface materials that have been upwarped and intruded by tabular diapiric "walls"; (4)

ridges have been upwarped or even further deformed by lateral compressive stresses which deform material at shallow depths; (5) ridges represent preexisting surt•tce materials upwarped by addition of melt into shallow cracks. We concisely review each model in turn.

4.6.1. Volcanism model. Kadel et al. [1998] propose that doublet ridges represent volcanic debris deposited ballistically by gas-driven fissure eruptions. In this scenario, water vapor with other exsolving gasses erupt onto the surface, de- positing debris that piles up on either side of the fissure to form a doublet ridge with a central trough. This idea is related to earlier ideas proposed to explain the distribution of darker materials along triple band margins, including the discontinu- ous deposits along Rhadamanthys Linea [Belton et al., 1996; Fagents et al., 1997, 1999; Greeley et al., 1998a]. The outstanding consistency of ridge cross-sectional profile with ridge length would require corresponding consistency of eruptive power and discharge volumes along fissure vents that would extend in some cases for hundreds of kilometers.

4.6.2. Tidal squeezing model. Greenberg et al. [1998a] propose that ridges are linear debris piles that accumu- late along preexisting fractures that open and close slightly in response to diurnal stresses, pumping icy debris to the surface with each cycle. This idea is a closely related outgrowth of work exploring nonsynchronous rotation and diumal stresses to explain global fracture pattems [Geissler et al., 1997; Geissler et al., 1998a; Greenberg et al., 1998a; Hoppa et al., 1999]. These authors hypothesize a liquid water ocean underlying a thin icy shell and propose that diumal global stresses drive a one-way, ratchet-like transfer of materials up through fractures to build a ridge, while at the same time surrounding terrain adjusts isostatically to the increased line-load. Tuffs et al. [1998] and Greenberg et al. [1998a], adapting a pre-Galileo suggestion of Pappalardo and Coon [1996], cite cracks and apparent downward flexure of terrain alongside ridges as evidence that the increased load provided by a ridge depresses the icy shell to cause flexure and cracking. These effects are most completely displayed in large triple bands, where it is proposed that darker areas on either side of the central ridge complex have been downwarped below a water table and affected by flooding. Incomplete extrusion can leave ice jammed in a crack, such that it exerts lateral pressure on the walls, ratchet- ing the crack open over many tidal cycles. These ideas, rooted in a global tectonic perspective, require additional development to reconcile an organized global stress field as the cause for the myriad of ridge orientations and cross-cutting relationships seen in the highest resolution images. Further work is also required to develop a physical model for a narrow, planar conduit structure capable of systematically transferring water upward several kilometers through a freezing transition to the surface, while flexing open and closed on the rapid timescale of Europa's 3.55 day orbital period.

4.6.3. Diapirism model. Head et al. [1998a, 1999] propose that doublet ridges form as a response to cracking and consequent linear diapirism (see also Pappalardo et al. [1998b]). This idea is an extension to ridge morphology of diapiric models proposed for domes, pits, and spots (see section 4.2). In this scenario, slabs or "walls" of diapiric material rise buoyantly and passively in response to cracking of the brittle lithosphere above. As the buoyant icy material intrudes into the cracks, the preexisting surface is envisioned as warp- ing upward, undergoing faulting and mass wasting. In the diapiric model, prominent cracks trending parallel to the ridge


, ß "•* •'•" ... • ::-':" • .... ""%.: ..-' .-' "---•.-. '- .-.'.• .-' ..... "----' ..... ' "-.: ..... "".• ..-. ":.." '-•' , ': '.- ':",".:.•:'•:.•.X....'..--.' ß ....

.... '•' •':. :-.- "•:.-'. :-': ':.. • -' '..... :'. , .:". '• :'" ...•:•' '" '• .... .':' ' •-"", ' '::-:.. '%.. •'.: ""•:-•.,.•:t ...... :" ..... "?: •:'. '"'•,

.... :'" •'•" --.. •':' '"' :::::::::::::::::::::::'"?--::': .... ' "'- * ' •':-'•-•*:•:': '"': -'-::: ...... :•:'• ':::" '•':" ..... :'t? .' "" :':-:::' '"•'•"-. ':": "::'•::•-:--•'.7%•:::•:*:. '•:'•:•"?::';:•:•:':'"' •t '•'"- :. .•--'-:•.-.. :-.:, •:•. ::.:-v..:.-:.----:...-. • • •-:• •::.:.:.: ..... ..::.. . ..... ,•.:..•.. ........ :..•.:..... .... c%e:•td•:. ...... .. ..,::... .... •: • :::•:. ..... .... • ..... :: .•::::..... •:..:.-..•:•.....•: •. :..... ,.•. •

.•..•.? ..- •..•:.. •-::...... •...?::...&•. • .•.."•.-•'--: , :•:•..•:• ...... . ...... ----:-:-:-.::-•- :...:.-•;:•:•....•. ................. --•:.• :,•:.:.::•,•:.•::::•:.•.. :•. ....... . .......... .•.•.•:•: ....... •. :•.::: .•. .... -•::::::•:•.:... • .... .:•q....:.•.•:•..• .. •- :.•:.•...•.•.•;.- ß •..•:.?- .:•..::..•.•.' .. -./....... ,.. '::.....::•...• :. .. '??•:•::•.:.:.::.• ...... ----:- . ....... .....:.?'""'•'•? :•.:•::;•::..:.:::$•:•.•;•.:• .......... ß .... :.. :•:•...'. ::.•:.. :....• ..•:...;:--:•::--:•'•:.• ...... •'.::• ß ....... ::• .... . :'."--•----•...:•.. •..:.•......: . -.:.-•:;:•:-;• ..... ;::•-.• •:...-•:-:: .... '..•......-.. •.. '•:.... .... . ? •,--.•::.:::• ..•.•-.• .::.:;:•:•"- .•- .... ---:---•:•:•:•:•:. •:- •..•- •::: •:•?•;..•:?.,•.:... •' :-.-:. ......... ::..:.::.'-.-'.-.-• '• .- . • ..

...... ...... ....... ...:::.•:• .... • ....... -----•?•?: •:,. •----•;•-•:.. ....... . .....:•:..• •:. . ..•.. ............. .....•. .• ....:-•::::.::...• ::.::• ..: ..•/•: :'•":•:-:--•': ' •:•. ":• ' ::'• -::•'::• f ..... :: ':•' :'. "': ' ....... ":• '::-.'• :' :::' -'"'"' •':' ':': ."• :::- •"•:'" ':•:•:• '•'"* -"" .... ' ......... ' .... •:i .- ':"'.' "' ." .......... ' '• ..........

...? :: ?--'.? ....... .... ... ...... ? ...... .... .... .......................... . ........................... ....................... ............ ................ .......................... .. ......... ::•.• ""'•'•'•:•;•:•.• '•:..•'E::•Z• '•'•"•:•'• :'.? ................. • • ................ _...::• •-: ........... '" •':':"•:• :• •'-•"• • ..•..•?. ......................... ./•.•:'•:..•. -?•?;•;• .•- .•....•..•.•• ...................... .•. •:::•. ,:::..• ..• .•..:•-•:• ..":::.;:.:..?:::.:::::?'• ..-•: :::....• ,..• --;:.:•:•-•--.• :..•.•.:.•:.• .... ::'..' -.-:..•:•:•:•:.•. •:.•.:

.......................................................... .•..,--.-:•:•?.:.•,:.::•.:.•, ...............

..:•.:e•:'•-.-?•:•?•;. ß ......... ..:. •.• •-:•':•'.• .. .-:•;•;•:::•...,::'? ,:•:.;•:•:.:.•;•;.•..•:::.•.:.•::: .•..::::...:•::•-•:-. .. •,:•::•:.•-.:•:;;•;•.•:z•.-::;•:•:•::•:.•.•:•:•:•;•::•:•;;•:• ...... •::::.. :..:•:.•:: ?.•::.. :..::..: ...... . ........... .::.::.:.:•... :.::--:--•.'•;•::..•.:•.•:•..::::---z•,-•-•...•. ......... •:.•.?:.?:• ::•:.-•:•:•::?/•?•.,. ,•::,:,.•:•: ........................ •: •.::. :•......-•::•:•....• .:.:•'•':.?• . :.:....:,•:•;:•;•,:5:• :•;::•.•:. '-.?•::•::.-.:.•....•:::,,• ........ ...• •:....•::• •:.•.•'• .•..•-'•-':::•-•:-... •..•:• .... :,-•-.•:/-•,•:.?:::.•.,: .:::.:....: ................. •:•,•:•, •:•::•.•:•:::•:"•. ....... •.,•-• •..•::•: ...-.•:. . ..... ,..•:•:-: ...:--•--,•:•::'. ............... •:•:•z':•;:.:::•:::,:•'•:•/•:•.•.; ............. . ...... •'"•:.....•.•?::•"- .... ::.:..• '•:-•.•-•••.•-"'-•--•-•;• .... :•. :•:•:.•:• '-• .• .• ............. ß ......... ..:..•,..• ...e/w•::.:..:•::.•::,:::...::•-,-:•:•?.:.:.:/.'•':'•::.•:•:• .... ':-..-.-• -/•:•::....:•.•?;:::•,, ...... ., .•..•.: ...•-•--.-.-.-••• ........... ,,•.....•?.. .... ::•..:: ...... : ......... .:• :.;•..-.::•:..... :.:,.•'.•.• ........,.•-..•-•.•,':',•'•-•:•.•.•:,.•?•:•.•:•.:.•-......••••,••••.::.• ......... ..•...•-..•::•:..;•?•::'• ....... '.:•..•.•.,. ........ : ............................. •:•:.;:y•-.,•.•:• •:;•,•.,,•,•' . •.::::....•:•••?'?•?.:... •.,.•••;•.:..•,•,•.:•.•,: :...•-•:• •.-•:.•,.

% -.-:•. ,:•;•?•:•:,•'-•-:-•/;:• "•:':' ====================== .................. •.•.:;•:-?•.._•;;•,.•.;•.•:•::•:.•?,•c;•:•,:•::::::.•::.•::: ............ .-• -,•:•:'•?'•'• ...... .::::•..'.•:•;:;::,.'•?:?•:.•?'•'•;:•?::':, .:.-.• ••-•.•,:••:••:.••• ............ . •:.•..•: ....... ..•;•:•:• •-:•::•*•.•:•:•::•:• .... :...• :.• .............. .... ;•:-::• -..:..•. ..,?--' ................................. ••-•'•.:::•:•:, ........... .• ................ ,.:• ......... ...:....:...?:;•: :.• ........ . .... .•z•:•'.:•:•'•'-':'-• ::•::::::•:$;•:•'• :.. ...... • ..:::::....:..:. ...........

. •.... :,.

...•:•:•.•:.,,:•"-"-"•:-":-•:.--•.•;:•:•.• •"':"•'..-••:':::::'.... •..,•.• .... ::.; -.--e•:.:• ;•::.•:• ......... .•..• :.• ..• ß .:: :•:•::-:.•.•,•?:•.•..: .... •:.;•.-..•:7:::..•- ..•:•. ............ •.•.

..................... '•:•'•::•* •,.:•:•:•,:•: .•.:•.:.., ..'•'•:•"':•:"•::'7 ............ , .......... .:.:•;•;•:•::'.•:•:.•?-:-. ................... • •.• ..... . .... ;'• ....• •,•::::::?:.•? :?:.•.. ....... :•::.:.•-,• ......... ..:•...•,•:•. ..... ,z,..• :•?:;•;:,:•??/•:? :?.•.•-....•::.:,•'

-.-.. -.. • ................... .... :•:.•:•.: ............. .•?:•.:•.:...::;:.:•:.:. .............. ...

ß .• " .-.-::.- •'"•'•:; • .... . . .. •.• :::•. .•::.•::.:::... '-•:•:•'"". ........ .:'.• .... •:•::.•:: .............

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

ß e•:•.... ..... - ... ::•.•:• ......... ....,.:. :,•. :..: .... ß .................. • • .......... :::..:.'.../..:::.::• ..•::..::..:::.•:•.:.::•;:.:.•:.•.:•:•:•:•:•..:::..•...::: ß ,..-... :.-?-• • :•?: ....... ..::.'::.-.:.:• . -•.. . :. :..z:•..•, .•:•%?;•z.•.•:•:•.•..:•:•...:.:..•:•:•. :..:::..::....:-.• ...... .•. ....... • .... - ........ • •..•.:•.,.?•.;•:..:....• .:•..•.;..•:•::.•.•:.•....::.•.•:•:;•:•:•:.•.::..:....•:;.•...•...•:•:•::•.:•:•:•.:•:•.•.•.•::•:•:•::•:•.• .............. -.-.-•-•.,•-•

" .• "":::•?..:• • .' •;•,,• • •:•. p .... •:.:', "•:..r•?• ::-• •::::•:•.; •. '• •'"'•'-'•"•

.... •?,•:...,. '.•;•:::.•;•::,..• :•"•'•' .•,. ::•. .::::, ....... • ..:,•:• ........ = ................... •;:.• .,.:-:.• ......... .•.:•:..:•,• •.: ..•..•-.. •:.,.. :;•:..:.'•. ':.':•-: "•:: ":-:'• •;:: ..... ":•. "•.. •. ':.•::•:•.• ".. ' .... '"'. :': ........... •:.'".,"....•"•;::'.•:•: :-•:•..'"• ...... :•" •:-.: .-'•"z?•

...: -• .•.. :.??•:•: .::.•-:•:..•.,•:.. •. ?, .,:•:' '.- .• .• .•••:•..•.•••••••::•.•,•::..•.:-:•.:;.•::•::.•.??•....-•?. •.. . •. ;•:'...•::.•:.-" :•..,• ...... ::• -.•:::::•.'-% •;•:•,,•,::•. •::•:.$z•;:::•.:•..:..•:•:.......•.::;:•::..:•::...•...•...•;•::•:.:. ,•:•.".• .•:•::•:;-:•:.•.:-:.'.....•-•;•:.•:'•.•..•::...:•.•:•:..:',•:•::::/:::.'•:•::•-:•:•-•:.•

Figure 12. High-resolution views of Europan ridges. (a) View of doublet ridges crossing ridged plains, seen at high solar incidence angle, emphasizing shadows and topographic shading. Where younger doublet ridges cross older ones (e.g., at arrow), it is uncertain whether the preexisting terrain has been upwarped. The oldest features in the region are small-scale low-relief ridges that comprise the background plains; the largest ridges in the region have vertical relief-200 m. Younger chaos material crosscuts the ridged plains, as in the upper right and upper left corners of this image (11ESMORPHY01; 33 m/pxl). (b) View of a prominent doublet ridge that transects a gray band in the antijovian "wedges" region (12ESWEDGE_02; 26 m/pxl). The image was obtained at relatively low solar incidence angle, which emphasizes albedo variations. Dark material flanks the ridge and lies along the bottom of its central trough (arrow); dark and bright materials also delineate fine-scale topography within the surrounding bands. North is toward the top; note the different scales of these images.

in surrounding terrain would be the brittle surface expression of extrusion is that preexisting terrain features should be identifi- rim synclines, formed by withdrawal of diapiric material from able on ridge flanks. This would suggest that ridge flanks prin- the broader source region that feeds them, with consequent cipally represent preexisting terrain that has been upwarped by depression of the surface above. The key observation that intrusion from below, with some additional mass-wasted debris would distinguish this and other models from those involving superimposed. Sullivan et al. [1997] point out that mass- linear debris piles created either by volcanism or tidally driven wasted debris on the lower flanks of most ridges, in draping out


a) volcanism b) tidal squeezing c) diapirism d) compression e) wedging

L__.j ice '•-'• warm ice • water •.•• rubble Figure 13. Schematic illustration of the five ridge formation models described in the text: (a) volcanism model [after Kadel et al., 1998], (b) tidal squeezing model [after Greenberg et al., 1998a], (c) diapirism model [after Head et al., 1998a], (d) compression model [after Sullivan et al., 1997], and (e) incremental wedging model [after Turtle et al., 1998a]. These models have different implications for the presence and distribution of liquid water in Europa's subsurface.

onto preexisting terrain, can disguise or mimic in subdued form the appearance of preexisting surface features buried under- neath.

4.6.4. Compression model. Sullivan et al. [1997] propose a model in which ridges form when opposing plates are stressed compressively. Compressive stress would deform warm subsurface material more than the cold near-surface ice.

Upwarp of ridge flanks may occur as deformed icy material is compressed and forced upward along preexisting fractures, pushing apart the relatively cold near-surface plates. An origin of ridges in a compressive stress regime would provide a balance for the extension implied by Europa's pull-apart bands (section 4.4). Apart from different stress sources, details of the Head et al. [1998a, 1999] and Sullivan et al. [1997] models differ in their explanations of cracks and depressions observed ad- jacent to ridge flanks. Recent high resolution images reiterate the fact that mass-wasting has affected ridge flank morphology pervasively, complicating efforts to reconstruct pristine morphologies and determine how much ridge relief is attributable to upwarping of preexisting terrain, extrusion, tidal flexing, or other processes.

4.6.5. Incremental wedging model. Turtle et al. [1998a] have investigated a model in which ridges form in response to addition of melt within a shallow vertical crack which is constricted at its bottom. Injection of melt into a shallow crack in Europa's ice lithosphere would cause outward and upward plastic deformation of the near-surface to build a ridge. Through finite-element analysis, they find that model ridge height corresponds to the amount of material intruded into the shallow crack. This model is somewhat analogous to the volcanic ridge model or the tidal squeezing model in that a periodically recharged shallow water source is implied along the length of a ridge; furthermore, injection of material into the crack might be controlled by tidal squeezing. The model incor- porates elements of the diapirism model, in that uplift of the ridge flanks builds the ridge. Currently it is unclear whether water-filled cracks would necessarily have the geometry required of the model, i.e., whether either the bottom of a crack would be constricted or the amount of intruded material would

decrease sharply at depth.

4.6.6. Discussion. The five ridge formation proposals carry different implications for the presence of liquid water or warm ductile ice at the time of ridge formation (Figure 13). The first ridge model, involving accumulation of explosively erupted materials along surface fractures, and the fifth ridge model, involving addition of liquid into shallow cracks, require liquid water at or very near the surface, although most likely on a temporary basis subject to periodic recharge. The second ridge model, proposing that ridges represent debris piles squeezed into place by periodic tidal stresses, requires a liquid water ocean at most several kilometers under an icy shell. The third and fourth ridge models, involving ridge formation through upwarping of preexisting terrain into ridge flanks in response to diapirism or compression, require that subsurface materials deform ductily in response to the stresses involved in moving this material several kilometers or more to the surface. These solid-state models do not directly require the presence of subsurface liquid during ridge formation, but they do require sufficient heat at depths of several kilometers for ductile deformation of icy materials to occur.

Understanding ridge formation is a fundamental outstanding problem in Europan geology, and one important to the presence and distribution of liquid water in the satellite's subsurface. It must be emphasized that all model-dependent require- ments for the presence and location of liquid water, ice-water mixtures, and warm deformable ice apply only during periods of ridge formation and modification, and might not reflect the current state of near-subsurface materials in any of the imaged areas. Overall, it must be concluded that the implications of ridges for the presence and location of subsurface liquid water remain uncertain.

4.7. Surface frosts

Based on Voyager imaging, it was suggested that Europa may be "painted" through continuous deposition of frost onto its surface, perhaps a result of venting from a liquid interior region [Cassen et al., 1979; $quyres et al., 1983]. There were four Voyager-based lines of evidence which might lend support to the hypothesis that surface frost is regularly deposited on Europa. First, much controversy over possible geyser-like


•. :.... .. :.:..•; .; ..

"•"•"•"' •'•'•'•"'"•'• ........ ':s;i•'"'" '": .... ' ........ "':':,..-' !•.:'•"-,;';'-•...".-'.?--:: ......... "•::'" '•'•"'•:iil i'.....:. i '"• ;•i•11!5 ":•'•'•"' -?: :?'" ...... :'"'•' ..

:. .: •.; ..........

............ . ..... ; •--.,. -.. . ................. . . .... .:.:. :,..'* ;:-.,:.:.; -.•.- .... . ;,.•:,....... .:....•.:':'. .... .. ..... ...,:-. •--.-, . . .. ...... .... :; •.-

,.. ::;:......;?. .;..•;. -:-: ..•. ß .

ß : ..... ..-?. ........ ......., ..,.:,.:,.¾ ..... ...

..

'-{ .:: ,-:-- ,. . . '(;,•"::::..:... ,:(:•:•**.. •... :'-:;'-5 --; .• .... .. . -• ,.,:.,•, ...... :• %.::..... ,• ......... ..':',..: .. -• ß . . , ..... ;. -....½<:•... ß :..:;½ ':.... '... :.,::::;-;..:..., .... ..... . '... '•':;•;}.:.:.. :•:•.•½":;,.. ].';•'bZZ: ':?*;• ..... **•' ...... "*"-.: .......... ...::d..':,'.':...'.:,.;-:•.-:::; ....... .:;.:..:.;.::;:..7%'..';..','' ,.. .'.'•::b."'E'"':•:'-*.•*'.:.. -"-•:':::**:'::' ---•;:-..:: ...... :'-.:: ..::'-': •a'•:•4*::•;.;•:;.::: '•;. '•!• ::..%. :.:½•.%:'/:: .. •' •;•.:. • .... ;.;.•;::;-.;j;;.:.• .-...:? :. --...;..: :•.'-';::...,;•;s-.• .:.--.•.: ... ..•' :-:. :. ., . ..;•-•.:. ?.:. :.... ß ...... . ......... :... . . / •{:$:•. ::'**..;•;::s**•:**:, *-,:•,,•'$.•. ..-,:,-,.• ........ ":,::;•-•-:" .,;• ;•-s½.•:•½•:•:'T'::?; ..,.;,.;'.*-..--. ;:,•.;;;;.... ."%;•;•.•:..: .: ..:..: ........... :-•

•?:.. :?. '%;;;;•:•.;.,•;::;•-, '%.."-?:.. ........... ;.;;;./1'•;;"....' ............ :..:..,?-;•.;.::.• :-..•-,:-'.¾:•. .... - :. ...... .,•.:.•.: ... ;,::,.....•.. ........ ' ..:?a,:;..{,, ¾.' ?'-::::.... % *-:.'Q .... . ..... :,:-;--.r,.-;" ;:-,..: .•" ...:.•....,... ........ . ........... -• ........ : ..... . .:?' :,. ,Y " ':'. •'* .:,•;;':;'":}, .... " "::•*:• ;::':• :'".}*: q;:;.:': ......... : .... '• ?':.::.'}a,.,:,, :::.X:.;.';...•,. , ' ..... .:'• ...... ..•;::.;.;•;:"?;;•':;:,'a•.;;:L:.:..;.;:,;:-:.:*w,:•;:..•;½:-X:•':.c -"•-;

.'%. '":-?;j;:.•::;**;:.:::., .... ?:,:'•'.. ß ,.•'?'. ..... .•,.... .,.....,.. ........ :•..::•:•:•::.•::•;;;'•:.::::•:• . .... '½". ' •*:':•$•:' ' ........................ ½; ..•::.:**.::,.,•**.<::':;;%•:•:•.•:':;'. ..... .E '•½•'---- -:'-•,:•'*E*:•i;;½•**?'•-•:• '. .•*•.:;-•,, .:: *?¾p'•);'*a*- '•?•;;::½:'•' ...... '::':5..'•;{:..';'.;..-..,,' '•"'•½• ......... ;:,•;'**'"' '...

'*:*; :...- 'e*a-z•**?:?½?½]•.:'.-...:• ........ -.:*-*:</ •:• ....•a.......•-•'•.:-:.:;}'-..:%::'-•:•;:;• .: .•;., ..... % ................. .....:S•-• •;.`.•.•.•.•:.•..::.::•:•::•.•:..•.•.•..••••.•;`;;•:•.•.......•:.:..•:•;•.,•;:;•.;•::..*... ?:' .., •: •::.4•:'• • '"'½•;.":..•'::;:;-" .,.;.. ":%•:½&:•:;•;:.•;;-•.-?.:•;•;'%;:•<• %.' ':"•;•:,•';;;•::½:;•,.•;'•:?::/•'; '•;:•:.'•':::q'" ;.; •.;•.•*•* ;!:•'•:::•.•;,5.:;$•½•' . ............... :. ,:• •;.•,•::::;•;•-'7%..;...::;•?:' :.? ......a.'-.- . ............ ,;.::;:•:;•:•;':•;•;.:.•2;.•½::•?•.:•½•:s-; : ?::; ;•:.-:.::::•'";•;' 'E'a•W'TS.•.•:.;• -:;½*::'•½':'•':•..: :•;•"•'•'•...,.•';;;;:•"":.• .... •?"•:•..:.•-•:.::-•..•;•;;;:::::;;:; ............. •;;;½":":'•:;•:..

ß • '.• ß .•-.• •.;.;•;; .... •'••• .•.,x-•"•*' ....... •,•'•-•:.•;e½:•-.'•'•:c• :;•,:. •:'s•??•½ •e;•.?.,:;•.r •' ....•',•',•: ' .... . ':.

ß '* ................. ' ......... •..'•; "a

. ß ..... :•::.½ .•:' .............. .;•e•.•.:•"•"• ......... •,. :.•.s•.,•,•,•..:..:•;:•:....,. .......... •:,::.;• ................. "•' .S; .......................................... ' .......

Figure 14. Bright and dark materials seen at the highest Galileo resolution. This oblique-looking view is the highest resolution image of Galileo's orbital tour, with-6 m/pxl horizontal scale. Topography is the chief control on the albedo patterns. Bright material generally correlates with higher topography, and dark material occupies topographic lows. Segregation of surface materials into bright (icy) and dark (non-ice) patches is suggestive of sublimation-driven thermal segregation, which acts on very short timescales and can dominate over sputtering in Europa's equatorial region [Spencer, 1987]. Warmer temperatures there and down- slope movement of non-ice materials may initially act to concentrate dark materials in topographic lows. There is no direct evidence seen for venting of bright frosts from ridges or cracks. North is to the right, and the scene is illuminated from the east-northeast (lower right). Galileo observation 12ESMOTTLE01.

venting on Europa has centered on the identification by Cook et al. [1982, 1983] of a putative plume in one Voyager 2 image; however, this identification is highly uncertain, as is discussed in section 5.2 (see Figure 17). Second, the rate of magnetospheric sulfur implantation on Europa's trailing hemisphere has been inferred from ultraviolet absorption depths and seems to be far less than predicted by modeled particle implantation rates; masking of sulfur by water frost might account for this apparent discrepancy, if water frost is deposited onto the surface at a rate of -4 cm Myr 4 [Squyres et al., 1983; Eviatar et al., 1985]. Third, Europa exhibits unusual photometric behavior, specifically a very narrow opposition surge, and the standard Hapke [1986] photometric model applied to Voyager and groundbased observations of the satellite's phase curve suggests that Europa has extremely high surface porosity (> 95%) [Domingue et al., 1991]. Fourth, crosscutting relationships inferred from Voyager images imply that some dark reddish bands are relatively youthful,

suggesting that Europa's surface features might brighten with age, and frost deposition or comminution of surficial material into smaller grains offer plausible mechanisms for progressive brightening and disappearance of initially dark bands [Pappalardoand Sullivan, 1996]. Galileo observations shed light on each of these Voyager-based hypotheses, with respect to the distribution and sources, means of redistribution, photometric properties, and possible evolution of ice "frosts" on Europa's surface. In this way, the Galileo data address the overarching hypothesis that frosts might be vented from a liquid water interior.

4.7.1. Presence and distribution of bright and dark materials. High-resolution (6- 50 m/pxl), low solar incidence angle images of Europa were obtained by Galileo for the first time during orbit El2. At a scale of tens to hundreds of meters the surface shows a strong bimodal distribution of albedo (Figure 14). Dark material generally occupies topographic lows, and bright material of particularly uniform al-


bedo is preferentially located on topographic highs and upper slopes. This bimodal albedo pattern is suggestive of sublimation-driven thermal segregation, which can act on a timescale of just decades, with dark materials representing lag deposits of non-ice material and bright material being deposits of relatively pure ice or frost [Spencer, 1987; Pappalardo et al., 1998d]. No distinct correlation is observed between the distribution of bright material and features such as ridges or fractures that might be candidate source vents. In fact, the distribution of dark and bright materials seen in Galileo imaging suggests that venting is a more likely origin for some dark deposits on Europa (specifically, those flanking some ridges to form triple bands [Fagents et al., 1999] than for bright units. If some dark materials are vented and consist of dark material entrained in a

mixture of water and other volatiles, it is possible that H20 molecules could travel farther than the dark material and might be deposited as diffuse patches of frost. This model is speculative, as no unequivocal plumes have been observed at Europa (see section 5.3).

4.7.2. Sputtering. Europa is deeply immersed within the charged particle environment of Jupiter's magnetosphere. Based on Galileo Energetic Particle Detector measurements, sputtering by surface impact of magnetospheric particles is expected to redistribute H20 molecules across Europa at a rate of 20 - 80 cngMyr and cause loss of H20 at a rate of 5 - 20 cngMyr [Ip et al., 1998]. On relatively cold surfaces, notably throughout Europa's polar regions, these rates imply that sputtering should dominate over sublimation as the chief means of H20 redistribution on Europa [Spencer, 1987]. Ip et al. find their Galileo-based sputtering rates leave essentially unchanged the arguments of Eviatar et al. [1985], who infer current deposition of H20 onto Europa's surface based on inferred sulfur column abundances and modeled sulfur implantation rates. However, as pointed out by McEwen [1986a], the estimates of Eviatar et al. are dependent on a number of uncertain parameters, including the spatial and temporal variability of the Europan plasma environment.

Noll et al. [1995] have compared Europa's ultraviolet spectrum to laboratory spectra of sulfur ions implanted in ice, and SO,• grown on H,•O ice [Sack et al., 1992]. They find that the spectrum of SO,• on H,•O provides a much better match to Europa than does the spectrum of sulfur implanted in ice. Therefore they call into question the basic assumption of Eviatar et al. that Europa's trailing hemisphere absorption band is directly related to implantation of magnetospheric sulfur ions. The frost deposition argument would be invalidated if SO• on Europa is found to he nrincin_allv ondng•.nic (with it• distribution only modified by magnetospheric effects), or if implantation of sulfur ions in water ice creates SO,_, but does so inefficiently. Mapping of Europa's ultraviolet absorber by the Galileo Ultra-Violet Spectrometer instrument indeed suggests an endogenic source for Europa's SO,_ [Hendrix, 1998]. There- fore there seems to be no compelling observational requirement for venting of water frost from Europa's interior.

4.7.3. Europa's opposition surge. During orbit G7, Galileo images captured Europa's opposition surge with three- color imaging of the same location at phase angles of 0 ø and 5 ø and a resolution of 400 m/pxl. Helfenstein et al. [1998] find that the opposition surge of bright icy plains is 1.5 times stronger than inferred from pre-Galileo studies, and that low- albedo materials have an even more intense opposition surge than bright materials. This indicates that shadow hiding provides a significant contribution to Europa's opposition ef-

fect, in combination with the narrower contribution of coher-

ent backscatter. Recent laboratory data imply that Europa's narrow coherent backscatter opposition effect might be produced by scatterers that permit long optical transport lengths for scattered light [Helfenstein et al., 1998]. Therefore, in contrast to porosity estimates which model the narrow coherent backscatter component in terms of the Hapke [1986] shadow- hiding model [Dorningue et al., 1991; Dorningue and Verbiscer, 1997], regolith cover on Europa may not have unusually high porosity. Analysis of recently acquired high-phase images will allow the contributions to Europa's opposition effect from coherent backscatter and shadow hiding to be further distinguished, permitting a quantitative measure of surface porosity for different Europan materials.

4.7.4. Fading of bands. Crosscutting relationships inferred from Galileo images (e.g., Figure 9) confirm that initially dark bands have faded (i.e., have lost contrast) with age [Belton et al., 1996; Geissler et al., 1998a, c; Greeley et al., 1998b]. Redistribution of ice through sputtering is a viable means for brightening Europa's bands. Specifically, fading and disappearance of some kilometer-scale bands toward Europa's poles are consistent with the expected dominance of sputter redistribution of H20 in the cold polar regions [Pappalardo et al., 1998d]. Bands in Europa's equatorial regions might brighten with age more slowly if a brightening trigger (e.g., radiation damage [Johnson, 1995] or chemical alteration [Geissler et al., 1998a, c]) ultimately allows sputter-deposition to dominate over thermal segregation. Because sputtering should promote frost deposition and surface brightening essentially independent of topography or terrain type, this H20 redistribution model does not readily explain why Europan plains units (specifically the stratigraphically old infrared-dark plains of Geissler et al. [1998a, c]) apparently have not brightened over time. Chemical alteration of dark band material, perhaps in response to charged partical bombardment, provides an alternative mechanism to explain band brightening with age [Geissler et al., 1998c]. The chemistry involved, however, remains speculative. In either of these models of band brightening, venting of water from a liquid ocean is neither required nor implied.

4.7.5. Surface frosts: Discussion. Galileo observa-

tions lend no definite support to Voyager-based arguments that water frost is continuously vented on Europa from a liquid interior. Of the four lines which have been suggested as supporting water venting (geyser-like deposition of bright frost, masking of implanted sulfur, extreme surface porosity, and brightening ,-,r surface features over time), we ,,,•, positive evidence only for brightening of initially dark surface features over time. This process can be plausibly explained by charged particle ir- radiation of the surface, redistributing H20 molecules and/or causing chemical changes of endogenic surface materials. Though not disproven, the other three Voyager-based lines of evidence for water venting are not required or supported by Galileo observations.

4.8. Topography

Near-terminator Voyager images indicated that Europa has extremely subdued topographic relief [Malin and Pieri, 1986]. Analysis of Galileo stereo images demonstrates that the broad- scale topography of Europan craters and pits can reach -750 m [Giese et al., 1999; B. Giese et al., manuscript in preparation, 1999]. The more typical local topography of Europa's ridges,


bands, and chaos is remarkably subdued, with relief of only -200- 300 m, based on photoclinometric profiles (R. Kirk, personal communication, 1998), digital terrain modeling from stereo images [Giese et al., 1999; B. Giese et al., manuscript in preparation, 1999], and near-terminator shadow measurements. Europa's short-wavelength topography has apparently formed in such a way that most features have not grown taller than this height. Stratigraphic relationships suggest that older Europan features display generally less relief than younger ones (Figure 12) [Sullivan et al., 1999b]. Moreover, the character of surface features appears to have changed through time, with sets of low-relief closely spaced ridges and troughs generally being the oldest identifiable units on Europa [Head et al., 1998b]. In- ability to build or retain topography is suggestive of a warm weak lithosphere, consistent with a high near-surface thermal gradient. Europa's subdued topography and its evolution could be accounted for by any of three alternative models, with topographic elevation limited by (1) the brittle strength of near- surface materials, (2) buoyancy, or (3) viscoelastic relaxation. An example of more extreme topography was found in early Galileo images of Europa, suggesting the presence of isolated topographic knobs >1 km in height [Belton et al. 1996] and challenging the notion of a thin, weak ice shell. Here we investigate the implications of Europa's topography for the character and evolution of the satellite's lithosphere, and we remark on the extent to which Europa's topography bears on the question of a subsurface ocean.

4.8.1. Strength-limited model. Europa's topography might be self-limiting due to the strength of near-surface materials. Pappalardo and Coon [1996] speculated that Europa's ridges might progressively load the brittle lithosphere as they grow, causing flexure, then cracking and subsidence once a critical height is achieved. In this model, a brittle near-surface layer lies atop a relatively mobile substrate, which might be liquid water, ductile ice, or a transient ice-liquid slurry. In high- resolution Galileo images it is observed that parallel cracks comrnonly occur alongside Europan ridges [Greenberg et al., 1998a; Head et al., 1998a]. From the distance to such cracks, the effective elastic lithospheric thickness is found to be very thin (-400 m) [Tuffs et al., 1997]. In terrestrial flexural analyses the effective elastic thickness is found to be -20 - 40% of

the actual thickness of the seismogenic (brittle) lithosphere, due to significant weakening by faults [Stein et al., 1988]; this scaling factor applied to Europa would suggest a brittle lithosphere -1 - 2 km thick. This adjusted thickness would be in ac- cord with brittle thickness estimates based on fracture mechan-

ics [Moore et al., 1998] and models of buoyant chaos blocks [Carret al., 1998; Williams and Greeley, 1998] (see Table 1). The thickness of the brittle lithosphere does not necessarily constrain the thickness of the underlying mobile layer, so Eu- ropa's total ice shell thickness could be significantly greater. In this model of Europa's topography the apparent increase in topographic relic/ through time suggests that Europa's lithosphere strengthened over time, for example, through progressive cooling and thickening.

4.8.2. Buoyancy-limited model. Another way that the height of Europan topography could be self-limiting is if buoyancy plays a significant role. Europa's ridges and bands may have formed by rise of warm diapiric material to its level of neutral buoyancy (section 4.6.3). Topography would be relatively uniform across the satellite if the composition and source depth of warm, buoyant material were everywhere similar. This is consistent with diapirs rising from a thermally

controlled interface, such as the brittle/ductile transition or the

base of a floating ice shell. The apparent increase in Europa's topography over time could be understood if older topography formed with lesser elevation, and the relative buoyancy of diapiric material evolved and increased over time, for example, if the source depth of Europa's diapiric material increased over time through cooling and thickening of ti•e lithosphere. The hummocky matrix material of chaos regions is found to be either low-lying or high-standing relative to the surrounding plains [Collins et al., 1999; Chuang et al., 1999]. In a diapiric model, more buoyant material (warmer or of different composition) would be implied by higher-standing chaos materials.

4.8.3. Relaxation model. Finally, Europa's topography may have relaxed over time. From Voyager images, isostatic relaxation has been invoked to explain reduction of crater topography on Europa and other icy satellites [e.g., Thomas and Schubert, 1986; Hillgren and Melosh, 1989]. Though not yet explicitly modeled for Europa's ridge topography, the satellite's presumed high thermal gradient and warm interior may have promoted relaxation of ridges and other topography over time. This would imply that the limiting height of Europa's topography is controlled by the viscoelastic properties of the lithosphere, and the competing rates of building and subsidence of topography. In this model, the fact that more recent topography displays greater elevation might suggest that the lithosphere is stronger (cooler and/or thicker) today than in the past. An alternative interpretation is that Europa's topographic relief is in a steady state, with higher topography relaxing at a rate similar to that at which new topography is formed. Quantitative analysis is necessary to determine whether relaxation of Europa's fine-scale topography is a viable hypothesis.

4.8.4. Isolated knobs. During Galileo's first orbit of Jupiter, Europa was imaged at 1.6 km/pxl, surpassing the best Voyager resolution [Belton et al., 1996]. One SSI image reveals two features with brightness characteristics suggesting they are topographic knobs (Figure 15). Subsequent high-resolution imaging reveals that one of them, Cilix, is instead a flat- floored impact crater. To investigate the topography of the other unnamed massif (33.55øN, 169.47øW), the Galileo image and two Voyager 2 images (FDS 20651.55, 2.0 krn/pxl; and FDS 20649.10, 2.3 krn/pxl) were mapped to a common ortho- graphic projection and analyzed for stereo content. The unnamed massif is approximately 13 km wide and appears to stand 2+_0.5 km above its surroundings, the tallest known feature on Europa.

If the massif is an extrusive construct, its high topography could locally challenge the concept of a thin ice shell floating on water. If the massif loads Europa's lithosphere, a flexurally induced topographic moat should surround the massif, unless the Europan ice shell at this location is very thick. Figure 16 illustrates the amount of flexure expected at the edge of the massif if it loads an elastic ice shell resting atop an inviscid substrate (either warm ice or water). The flexure was determined using solutions from Brotchie [1971] and Comer et al. [1985] for a cylindrical load an elastic lithosphere. The massif is modeled as two cylinders: one of radius 6.5 km and height 1.5 km, and atop it another of radius 4.5 km and height 0.5 km. Solutions are evaluated for a range in possible values of Young's modulus E from 10 7 to 10 lø Pa. The plotted deflections correspond to those calculated at the edge of the massif (the larger deflection along the load axis would be hidden by the load itself).


50 k- ß .:,½-:•-:/;...:::,..•,•.:..;;i:q•:•: ========================================= ': :.:•½:;½;;...::;;::•:..:........,:.•,•;.

Figure 15. High-standing massif on Europa (arrow), estimated as 2_+0.5 km tall. Cilix crater is at lower left' local albedo patterns give the false impression that Cilix might also be a dome. Note the abundant triple bands criss-crossing the image. Galileo observation GIGSGLOBAL01, frame s0349875113. Resolution is 1.6 km/pxl, and north is toward the upper right.

,

These results predict that if the massif loads a floating ice shell <20 km thick, a moat >100 m deep should result, unless the Europan ice shell is strong (E > 108 Pa). Though the broad flexural wavelength of the lithosphere may preclude observation of a minor load-induced deflection, a deep moat should be identifiable in Galileo high resolution images. Of course, the massif may have formed in another manner that does not cause it to load Europa's shell. For example, perhaps the massif is an unusually large tilted block analogous to those seen in chaos regions [Carr et al., 1998; Spaun et al., 1998a]. Unfortunately, a spacecraft sating during Galileo orbit El8 precluded Galileo high-resolution imaging of this feature to test these hypotheses.

4.8.5. Topography: Discussion. Modeling of the strength and buoyancy of candidate surface and subsurface materials should be able to distinguish among the three principal means by which Europa's topography might be self-limiting and whether these processes might have worked in combination. Each of the proposed models is suggestive of a warm, weak lithosphere and mobile material at relatively shallow depth; however, none of them seems to require the presence of a subsurface ocean. The apparent increase in Europa's relief through time could mean that the lithosphere has progressively cooled, thickened, and strengthened over time; a steady state model is also plausible, involving relaxation of older topography over time. We find that Europa's subdued topography is


5OO

4OO

300

2OO

lOO

o lO 20 30 40 50 60

shell thickness (km) Figure 16. Downward flexural depression at the edge of an ice load equivalent to the unnamed massif, as a function of thickness and elastic modulus of an ice shell. Curves are labeled with the log•o of Young's modulus E in units of Pa. Polycrystalline ice has E-10 •ø Pa, while a fractured ice lithosphere is expected to have a significantly smaller modulus. Greater deflection is predicted for a thinner ice shell and smaller values of Young's modulus (a more fractured shell).

not a direct indication that Europa possesses a subsurface ocean but is suggestive of weak (warm) lithospheric materials. If an unusually large massif observed on Europa loads the icy lithosphere, then a topographic moat should surround it if Europa's ice shell is thin.

4.9. Global Tectonics

4.9.1. Nonsynchronous rotation. If Europa's rigid near-surface is decoupled from the rocky interior, then the surface may rotate at a rate faster than synchronous, presenting an ever-changing face toward Jupiter. A satellite in an eccentric orbit experiences a torque that tends to increase its spin rate to faster than synchronous, unless it possesses a permanent mass asymmetry large enough to counter the tidal torque [Goldreich, 1966; Greenberg and Weidenschilling, 1984; Oja- kangas and Stevenson, 1989a]. Europa has a forced eccentricity of 1% (due to its three-body resonance with Io and Gany- mede), so its orbital velocity varies along its path around Jupi- ter. At perijove, where tidal forces are greatest, Europa's tidal bulge lags behind the Jupiter-facing direction, creating a torque that tends to accelerate the satellite's rotation to slightly faster than synchronous. The rocky interior of Europa is expected to maintain a permanent mass asymmetry sufficient to counter this torque and therefore should be synchronously locked, but Europa's icy shell may be able to rotate nonsynchronously.

Mechanisms and timescales of nonsynchronous rotation of Europa have been considered by Greenberg and Weidenschill- ing [1984] and Ojakangas and Stevenson [1989a]. Greenberg and Weidenschilling [1984] envision a satellite locked in synchronous rotation with its tidal bulge frozen-in at the torque minimum, such that the average torque over an orbital period is zero. The bulge will attempt to creep toward hydrostatic equi-

librium (toward the potential minimum), with the result that the satellite's surface will rotate nonsynchronously. For viscosities appropriate to warm ice, nonsynchronous rotation would proceed on the viscous relaxation timescale of the bulge, -10 8 tO 10 9 years. Because Europa's rocky mantle is probably synchronously locked, decoupling of the icy outer shell from the rocky interior is necessary if the surface is to rotate nonsynchronously. Decoupling is easily achieved if a global ocean exists. In a warm ice endmember case, viscous resistance may significantly retard nonsynchronous rotation; however, it has not yet been demonstrated whether effective decoupling could be achieved for realistic (nonlinear) ice rheologies.

Ojakangas and Stevenson [1989a] consider that nonsynchronous rotation could result from thermal adjustment of a floating ice shell. Predicted longitudinal variations in ice thickness (due to variations in tidal strain rate) will drive the ice shell out of hydrostatic equilibrium. In attempting to move back toward hydrostatic equilibrium, the shell will rotate nonsynchronously on the thermal diffusion timescale of-•10 7 to 10 8 years. In this model, liquid water beneath the ice shell as- sures decoupling. Nonsynchronous rotation is not guaranteed in the floating' ice shell model, however, as it is possible for the ice shell to maintain a mass asymmetry sufficient to prevent or halt rotation.

The observational evidence suggests that nonsynchronous rotation probably has occurred on Europa, and may be ongoing today. Based on Voyager data, Helfenstein and Parmentier [1985], McEwen [1986b], and Leith and McKinnon [1996] concluded that Europa's global-scale dark lineaments might have resulted from the stresses produced by shell reorientation due to nonsynchronous rotation. Galileo color images (at 1.6 km/pxl) of the northern high-latitude region of Europa's trail-


a b

240 ø

d

210 ø

o

30 ø

Plate 1. Global-scale color imaging provides evidence for nonsynchronous rotation of Et•ropa's icy shell. Crosscutting relationships indicate that Europa's youngest bands are generally dark; older bands are brighter, notably in the near infrared (Galileo observation GIGSGLOBAL01; 1.6 km/pxl). (a) In this mosaic of a Gali- leo three-filter composite (green, 756 nm, and 968 nm), younger bands are darker and redder, and older bands are brighter and whiter. (b) Geissler et al. [1998a] use color classification to isolate the oldest bands, found to trend generally NE-SW in this region. (c) Intermediate-age bands generally have more E-W trends. (d) The youngest features in the scene are "greenish" in color and trend NW-SE. Consistent with the predictions of nonsynchronous rotation of Europa's ice shell, the color-based stratigraphy suggests a clockwise rotation of bands, and hence stresses, in this region over time. The orientations here are consistent with -•60 ø of nonsynchronous rotation between tbrmation of the earliest and latest structures [Greenberg et al., 1998a].

ing hemisphere support and strengthen this argument. The improved long-wavelength sensitivity of Galileo SSI makes it possible to discern at near-infrared wavelengths older lineaments which were largely invisible to Voyager [Clark et al., 1998; Geissler et al., 1998a, c]. Based on their color characteristics, Geissler et al. [1998a] mapped lineaments of various ages, inferring that the orientations of these features have progressively rotated in a manner consistent with a migration of stress patterns as predicted by nonsynchronous rotation (Plate 1).

Regional-scale and high-resolution images reveal that Eu- ropa's plains units are everywhere overprinted by a network of

intersecting ridge-sets with orientations that vary with stratigraphic age, as shown by their crosscutting relationships [Kraft and Greeley, 1997; Spaun et al., 1998b; Sullivan et al., 1999b]. While correlation remains to be demonstrated conclusively, these ridge orientations may be related to continually shifti.t•g patterns of stresses over time, possibly due to nonsynchronous rotation.

Nonsynchronous rotation predicts regions of isotropic extension to the west of Europa's subjovian and antijovian points as the surface is stretched over its tidal bulges [Helfenstein and Parmentier, 1985; Greenberg et al., 1998a]. Europa shows a distinct zone of dark pull-apart bands close to the predicted re-


gion of rifting west of the antijovian point [Belton et al., 1996]. As discussed in section 4.4, extension within this zone has been inferred from both Voyager and Galileo image analyses [Schenk and McKinnon, 1989; Sullivan et al., 1998]. The Jupiter-facing hemisphere has not yet been imaged at the resolution required to distinguish pull-apart bands. Zones of compressional stress are also predicted, centered 90 ø in longitude from the tensional zones, and concentrations of mottled terrain at these locations might be an indirect manifestation of biaxial horizontal compression on Europa [Pappalardo et al., 1998c].

Both theory and observation are consistent with nonsynchronous rotation of Europa's icy shell during its recorded geological history. The relative youth of the surface argues that nonsynchronous rotation may be ongoing today. Because the surface should shift eastward over time, the positions of features seen in Galileo observations might be displaced from their locations in Voyager images if the rotation were sufficiently rapid. The lack of perceptible change in the positions of surface features in the 17-year interval between the two sets of spacecraft observations allows a lower limit of 104 years to be placed on Europa's current nonsynchronous rotation period [Hoppa et al., 1997]. Nonsynchronous rotation is readily achieved if the upper icy crust is decoupled from the rocky mantle by a liquid water layer. It remains to be determined whether sufficient decoupling can be achieved by warm ice alone, in the absence of a subsurface ocean.

4.9.2. Diurnal stressing. While Europa's global pattern of lineaments as known from Voyager display a good fit to the stress predictions of nonsynchronous rotation [Helfenstein and Parrnentier, 1985; McEwen, 1986b; Leith and McKinnon, 1996], Galileo's higher resolution imaging reveals the formation sequence of Europa's lineaments and shows small-scale lineaments that were previously unknown. The orientation of Europa's most recent structures (e.g., Plate ld) cannot be accounted for by nonsynchronous rotation stresses alone, but indicate the addition of a "diurnal" stress component produced by tidal flexing as the satellite orbits Jupiter [Greenberg et al., 1998a; Geissler et al., 1998a].

Recent tectonic studies provide support for the hypothesis that diurnal tidal stressing is important in influencing Europa's geology. High-resolution Galileo imaging shows that strike- slip displacement of surface features is common. Hoppa et al. [1999] show that there is a tendency for strike-slip motion to be left-lateral in Europa's northern hemisphere and right-lateral in the southern hemisphere. This could be explained if strike- slip motion is driven by diurnal tidal stressing, causing faults to open and then to horizontally shear, with an opposite sense predicted to predominate in each hemisphere. Observations suggest a match to the predicted preferred sense of strike-slip motion in each hemisphere, for example, plausibly explaining motion along the right-lateral fault Astypalaea Linea [Hoppa et al., 1999; Tuffs, 1996].

Hoppa and Tuffs [1999] have recently proposed an elegant model for the formation of Europa's enigmatic cycloidal ridges. The model suggests that diurnal stressing creates arcuate cracks in response to the locally changing direction and magnitude of diurnal stress through the Europan day. One arcuate crack seg- ment forms during a single Europan day, with the cusps mark- ing an abrupt change in propagation direction following a brief period of crack inactivity when stress magnitude drops below a critical value for cracking. The visible cycloidal ridge (or other cycloid-bounded structure) forms at some time after cracking, exploiting the cycloidal cracks. This scenario implies that cy-

cloidal cracks form very rapidly, on the order of weeks. How- ever, as discussed in section 5.1 below, no visible surface changes have yet been detected in Galileo images of Europa. New cracks may not have been observed in the limited regions observed at high resolution, they may be below Galileo resolution, or perhaps they no longer form today.

The orbital fluctuation of Europa's tidal bulge is predicted to be -30 m if Europa has an ice shell tens of kilometers thick above a subsurface ocean, and only -1 m if the interior is solid [e.g., Edwards et al., 1997]. Tensional stresses produced by diurnal flexing approach 0.1 MPa if the satellite has a floating ice shell [Hoppa et al., 1999; Hoppa and Tufts, 1999] and become much less significant if Europa's interior is solid ice. While polycrystalline ice has a strength of-2 MPa, Europa's lithosphere may be significantly weaker if frictionally controlled (Figure 10; see section 4.4), potentially allowing shallow cracking to shallow depth in response to diurnal stresses. The apparent importance of diurnal stressing in influencing Europa's deformation argues for a floating ice shell at the time its structural features formed.

The patterns of Europa's lineaments suggest that nonsynchronous rotation and diurnal stressing have operated in tandem to influence the surface geology. The inferred youth of the surface and its lineaments (see section 2) argues that deformation due to nonsynchronous rotation and diurnal flexing have occurred recently, and the rapid deformation predicted by diurnal stressing models begs the question of whether and how frequently they form today. Both nonsynchronous rotation and diurnal stressing are greatly facilitated by a liquid interior, arguing for a global subsurface ocean today. However the effects of realistic warm ice rheologies have not yet been fully quantified.

5. The Search for Current Activity on Europa

As discussed in section 3, Europa's surface is probably very young relative to the age of the Solar System; thus it is likely geologically active today. Furthermore, models for the tidal heating and tectonic stresses suggest that ample energy and forces are available to drive continuing activity. This activity could take several forms, including (1)tectonic activity, forming new lineaments or widening existing features; (2) solid-state convection and extrusion, forming or changing the appearance of lenticulae or chaos terrains; (3) venting of water, producing plumes and bright frost or (if non-ice material is entrained) diffuse dark mantling deposits; and (4) mass wasting of surficial materials, changing the distribution of bright and dark materials. An ongoing process is exogenic modification by sputtering erosion and redeposition, which may act to mask some of the evidence for recent geological activity.

Although we believe that the cratering record implies a young Europa, the discovery of current activity would dramatically confirm that interpretation. For example, if evidence for venting of water could be found, this would provide a "smoking gun" for current activity and would confn-m the hypothesis that at least pockets of liquid exist today below Europa's surface. However, such a discovery still would not prove the existence of a global ocean. If no surface changes are detected, their absence can nevertheless put an upper limit on rates of ongoing cryovolcanic resurfacing and tectonic activity.

There are several ways in which imaging can be used to search for current activity, similar to techniques used on Io.


First, we can look for changes over time in the color and albedo patterns or topography of the surface. Second, we can search for surface units with unusual photometric scattering properties. Third, we can search for volcanic plumes by trying to detect the scattering of fine particles above the bright limb, or in stereo views of the surface. Fourth, images of Europa in eclipse (inside Jupiter's shadow) may reveal diffuse glows associated with escaping gases. Each of these techniques and the findings to date are discussed below.

5.1. Search for Surface Changes

One way to detect potential ongoing surface changes is by comparing images from Voyager (1979) to those from Galileo (1996 to 1999). It is more likely that a detectable change would have occurred over a time span of 17 to 20 years than over the few years of the Galileo tour. Phillips et al. [1998] and C. B. Phillips et al. (The search for geologic activity on Europa, submitted to J. Geophys. Res., 1999; hereinafter referred to as submitted paper) have compared Voyager 2 images of Europa (Figure 1) to Galileo images obtained at similar resolution and phase angle during orbit El4. Direct comparisons are limited by the -2 krn/pxl resolution and limited areal coverage of Voyager 2 imaging, and by the somewhat different phase angles of the data sets. No certain changes have been detected. If Eu,'opa's deformation occurs in a steady state, this negative resu!•t limits Europa's current resurfacing rate to <1 km2/yr and its surface age to >30 Myr (C. B. Phillips et al., submitted paper, 1999). This is in agreement with the -50 Myr average surface age implied by impact crater abundances (section 3).

Another possible set of Europa comparisons is between images taken on sequential Galileo orbits which cover the same area of the surface. Changes which could be found by these comparisons would have much less time to take place, but the scale of changes which can be detected in Galileo-Galileo comparisons is much smaller because of the increased resolution of the Galileo images and the improved geometric fidelity of the SSI camera. However, our preliminary examinations show no apparent changes. Perhaps the best comparisons will be between images from Galileo and the future Europa Orbiter. Eu- ropa has a photometric function that depends on wavelength and the illumination, emission, and phase angles [McEwen, 1986a; Domingue and Hapke, 1992; Phillips et al., 1997; Clark et al., 1998; Helfenstein et al., 1998]. These complex photometric properties complicate change detection but may also provide an additional means of searching for recent activity, as next discussed.

5.2. Unusual Photometric Scattering

Another approach to determining if volcanic and/or tectonic emplacement of new surface materials has been occurring on Europa in recent geological history is to search for exposed deposits of frost or ice that appear to be anomalously fresh in comparison to other Europan icy surfaces. Photometric measurements should detect such differences if the light scattering behavior of freshly deposited ice is different from icy regolith that has been exposed on Europa's surface for long periods of time. Investigations of Voyager and Galileo imaging data have revealed two different types of photometric anomalies that may be interpreted to represent relatively fresh exposures of icy material: surface deposits of anomalously transparent grains, and

surface deposits of relative large ice grains that have evidently not been exposed to the space weathering environment long enough to evolve to a fine-grained, mature regolith.

The best Voyager evidence for the presence of anomalously fresh frost on Europa's surface is from the controversial, 143 ø crescent "plume" image of Cook et al. [1982, 1983] (Figure 17). In addition to showing an off-limb feature that was interpreted as a volcanic event [Cook et al., 1982, 1983], the image shows a conspicuous bright spot on the surface at-34 ø, 337 ø. Helfenstein and Cook [1984] measured the photometric contrast of the bright feature relative to surrounding terrains and compared it with that measured of the same geographic region viewed at 13 ø phase. They found that, relative to the surrounding features, the brightness of the anomalous spot increased as the phase angle increased from 13 ø to 143 ø phase by more than seven standard deviations above the average surface change. Although Helfenstein and Cook [1984] interpreted this brightness change to be due to active emplacement of surface materials, a more conservative interpretation would be that the feature represents relatively transparent frost that was deposited on Eu- ropa's surface in recent geological history [cf. Verbiscer et al., 1990; Verbiscer and Veverka, 1990; Verbiscer and Helfenstein, 1998]. McEwen [1986a] suspected that this bright region was not actually anomalous compared to other bright regions on Europa which had not been seen at high phase angles by Voy- ager. Indeed, in low-phase global-scale Galileo images, this area appears similar to other bright plains regions. This hypothesis will be tested with high-phase images of Europa's plains recently obtained by Galileo.

Evidence for exposures of relatively large-grained ice comes from Galileo SSI multispectral data [Geissler et al., 1998c; Clark et al., 1998] and from low-phase photometry images obtained during orbit G7 which reveal variations in the opposition effects of different terrains [Helfenstein et al., 1998]. Both data sets suggest that coarse-grained ice is exposed on the summits and wall escarpments of prominent, stratigraphically young ridges. High-resolution images show that these materials are likely exposed by masswasting of debris into topographically low-lying regions [Sullivan et al., 1997, 1999a, b; Head et al., 1998a], and these exposures occur on young ridges with comparatively large topographic expression. Strati- graphically older ridges have more subdued topography and are interpreted to be mantled by Europan regolith [Helfenstein et al., 1998]. This suggests that the most prominent ridges were eraplaced recently enough in Europa's geological history that mass wasting and regolith maturation on exposed surfaces has not proceeded far enough to homogenize surface materials. While the timescale for development of regolith on horizontal Europan surfaces may be as short as 10 4 years [Veverka et al., 1986], Helfenstein et al. note that significantly more time may be required to build a regolith cover on the observed topographic slopes. In order to derive a true timescale for maturation of Europa's surface scattering properties, first a realistic understanding is needed of how rapidly the optical properties of fresh Europan ices are altered by the space environment.

Agenor Linear is a rare bright band on Europa's surface and has unusual photometric properties in being less backscattering than other terrain, suggesting that it might have been active relatively recently [Geissler et al., 1998b]. However, high-resolution observations during Galileo orbit El7 show that it is crosscut by small-scale fractures and mottled terrain,


.... - .... bright area .... ,:':' ......... :.-::.,:.:--- ....... ... ....... reseau

..

Figure 17. Unprocessed Voyager 2 clear filter image (FDS 20767.37) at a resolution of -44 km/pxl and phase angle of 143 ø, argued by Cook et al. [1982, 1983] to show an active plume along Europa's bright limb. Stretched inset image shows detail of the bright limb and putative plume, and an unusually bright area on the surface. The "plume" feature, with a signal level of just 5 DN, is not observed in subsequent images. Its location in the corner of the Voyager vidicon image, where noise and distortion are most severe, suggests that it is a camera artifact.

significantly reducing the likelihood that its unusual photometric properties are a sign of current activity [Prockter et al., 1999b].

5.3. Search for Plumes

Galileo images have shown no definitive sign of plumes above Europa's bright limb. Our ability to search for plumes has been limited because plume-like scattering is best detected at high phase angles. A few high-phase (179 ø) images of Eu- ropa were acquired during orbit C10, but these were at very low spatial resolution (--73 km/pxl), so any small plumes could have been missed. Images of Europa's bright limb were obtained during orbit El9 at high resolution and high phase angle, but over a very limited portion of the globe (about 20 ø of arc). We note that a tantalizing dim "haze" seems to exist above Europa's limb in one high-phase El9 Galileo image (s484888253); however, our analysis shows that this is a false double-exposure rather than a real phenomenon, probably resulting from incomplete closing of the camera shutter blades in combination with an imaging mode that does not include a pre- exposure erasure of the CCD array. Stereo observations offer

another way to detect active venting, and this is the way Tri- ton's plumes were discovered [Smith et al., 1989]. No such evidence has been seen from stereo views of Europa, but a systematic search has not yet been completed.

5.4. Search for Airglow

In order to search for airglow that might be associated with current activity, images of Europa's trailing hemisphere were acquired at a resolution of 21 km/pxl while the satellite was in Jupiter's shadow during orbit G7. A diffuse glow was observed in the region within 30 ø of the subjovian point. While this phenomenon could possibly be due to geological activity, we note that other explanations are viable. First, this region is also the subsolar region just prior to eclipse, so solar heating provides an alternative means for concentrating atmospheric gasses. Second, the regions near 0 ø and 180 ø longitude are places where charged particles might be precipitating into Eu- ropa's ionosphere along magnetic field lines connected to the flux-tube footprints on Jupiter, possibly producing an enhanced flux of charged particles on the Jupiter-facing side of the satellite (M. McGrath, personal communication, 1998).


Third, and potentially most relevant, light refracted through Jupiter's atmosphere might provide enough illumination to make Europa visible in eclipse. Thorough modeling of alternative scenarios is necessary before the hypothesis of geological activity as a cause can be properly evaluated. Higher resolution eclipse observations are planned for future Galileo orbits.

6. Discussion

6.1. Summary: Evaluation of the Evidence

No single piece of geological evidence discussed in this paper is itself wholly convincing of a current subsurface Europan ocean. In fact, we find that some perceived lines of evidence (watery flows, surface frosts, and topography) have no certain relationship to the existence or nonexistence of an ocean. Viewed as a whole, we conclude that there is compelling evidence for a warm, low-viscosity material at shallow depths during the history of Europa's pervasive deformation, and warm subsurface matehal (liquid water, warm ice, and/or transient partially melted ice slumes) probably played a key role in control- ling much of Europa's surface geology. Based on the low crater density and inferred youthful crater age of the satellite, we conclude that Europa's surface is probably geologically very young. This implies that the Europa's deformation is geologically recent, and that its subsurface has been warm in the recent past. However, no definitive evidence of ongoing activity (e.g., surface changes) is yet revealed in images of the satellite.

The nine lines of geological evidence regarding a Europan ocean are summarized as follows. Relaxation of large craters suggests a warm weak lithosphere; multiringed impact structures indicate penetration to a low viscosity layer, suggesting liquid water or warm low-viscosity ice at a depth of -6 - 15 km. A model-dependent indication of a water ocean comes from the suggestion that Europa's lenticulae are produced by thermal convection; if warm ice diapirs initiate at an ice-water interface, modeling of their rise height limits the ice shell to <40 km thick, which is less than total thickness of Europa's outer H20 layer as inferred from Doppler gravity data. Extru- sion of liquid water is implied by some very smooth (probably cryovolcanic) deposits, indicating localized water reservoirs but providing no certain evidence of a global ocean. Pull-apart bands could have formed by lithospheric separation atop a ductile ice asthenosphere, suggesting a steep thermal gradient but not requiring a liquid water ocean. The presence of small mobile ice blocks (some of which may have tilted) imaged in chaos regions argues that chaos matrix material was quite mobile at the time of its emplacement, suggesting era- placement as an ice-liquid slurry, but this is not necessarily indicative of a global ocean. Some ridge formation models suggest liquid water (perhaps an ocean) at depth, but solid-state deformation models are also viable. There is no definitive

evidence for active venting of liquid water from Europa's interior; instead, sputtering may be the dominant method of H20 redistribution. Subdued topography is indicative of warm lithospheric material but has no certain bearing on subsurface liquid. Europa's global-scale tectonic patterns indicate that nonsynchronous rotation and diurnal flexing have been the primary stress influences on the satellite, most readily accounted for if Europa's icy shell is decoupled from the rocky interior by liquid water; however, the effects of realistic ice rheologies have not yet been investigated, and it is uncertain that these processes are ongoing today.

A liquid water ocean within Europa would be a simple and comprehensive explanation for the range of geological landforms and geophysical phenomena inferred from Galileo image data. However, we cannot determine conclusively that a liquid water layer necessarily exists or has existed in the recent past within Europa. In accounting for most of the geological evidence discussed here, liquid water is not strictly required: warm low-viscosity ice and/or transient ice-liquid slurries can allow for the observed deformation. In fact, a ductile ice layer is expected to exist between the cold upper ice lithosphere and any liquid water below (Figure 10), making it difficult for any subsurface liquid water to directly affect the surface geology. A subsurface liquid water ocean remains a viable, but unproven, hypothesis.

Determining the depth to Europa's low-viscosity (plausibly liquid) layer is challenging. Table 1 summarizes the results of many different methods for deriving ice thickness on Europa. Each method makes its own assumptions and each senses a different part of the Europan ice shell; not all of the methods are necessarily self-consistent. The results can be largely recon- ciled if Europa has an effective elastic lithosphere ~0.1 to 0.5 km thick; an actual brittle-elastic lithosphere ~1 to 2 km thick; a ductile convecting asthenosphere 4 to >20 km thick; and a whole ice shell 6 to 40 km thick, plausibly above liquid water. If chaos blocks have been buoyant, then the rigid ice shell may have been locally ~0.5 - 3 km thick. The thickness of a floating Europan ice shell is predicted to vary globally due to latitu- dinal and longitudinal variations in insolation and tidal heating [Ojakangas and Stevenson, 1989b], as well as regionally and locally in response to local thermal anomalies and geological activity.

6.2. How Could a Young Europa Change Its Geological Style?

The existence of a subsurface Europan, and scenarios for sta- bility and evolution of an ocean over geological time, are issues intricately tied to the satellite's surface age and style of geological activity. Mottled terrain units (lenticulae and chaos) have been imaged during each close encounter of Galileo with Europa. It is most commonly found that these are among Europa's youngest materials, crosscutting older bands and ridged plains, while ridged plains materials are commonly inferred to be Europa's stratigraphically oldest units (Figures 11 a and 12a) [Carr et al., 1998; Greeley et al., 1998b, Head et al., 1998b; Pappalardo et al., 1998a; Spaun et al., 1998a, b; Sulli- van et al., 1999b; Prockter et al., 1999c]. Individual troughs commonly crosscut the mottled terrain units, but there are only rare examples of mottled terrain units being converted back into ridged plains. It appears that Europa may have changed its geological style over time, from ridged plains formation to mottled terrain formation. The mechanism for such a change is uncertain. Perhaps mottled terrain units formed by diapirism triggered as an ice shell cooled and thickened to the point where it reached a critical thickness and solid-state convection

initiated [Pappalardo et al., 1998a; McKinnon, 1999]; in this case, a change in geological style could represent a general thickening of Europa's ice shell over time until this critical shell thickness is reached locally.

If Europa's surface is an average of just -50 Myr old, how could a fundamental change in geological style have occurred during only the last -1% of the satellite's history, and what are the implications for the maintenance and current existence of a


ancient time • today past

Figure 18. Schematic representation of Europa's geological activity as a function of time, for four possible evolutionary scenarios. Black represents periods of mottled terrain formation and white represents periods of ridged plains formation. ("Ancient past" demarks an ancient but indefinite time in the past.) (a) Steady-state model: ridged plains and mottled terrain have each formed throughout Europa's history. (b) Special time model: Europa recently went through a change in its geological style producing mottled terrain most recently; ridged plains formation might extend back to ancient times or could have been triggered by a relatively recent global heating event. (c) Episodic resurfacing model: Europa's history is cyclical, expe- riencing periods of ridged plains formation interchanged with periods of mottled terrain formation (perhaps interspersed with periods of inactivity). (d) Old after all: Europa's activity ceased long ago, and mottled terrain formation was the last gasp of its geological activity.

subsurface ocean? (Interestingly, an analogous planetary problem exists for Venus, which has a relatively young (-500 Myr old) surface based on its crater abundance and shows evidence for widespread geological activity prior to this time [Basilevsky et al., 1997].) For Europa, four possible evolutionary scenarios can be envisioned: (1) the satellite is in a steady state and resurfaces itself in a patchy style; (2) the satellite •s now at a very special time in its history; (3) the satellite's global resurfacing activity is episodic or sporadic; (4) the satellite's surface is instead very old. These four possible evolutionary scenarios are illustrated schematically in Figure 18.

6.2.1. Steady state model. If Europa's surface is very young, then our current view of its surface is likely representative of what Europa has been like for some significant geological time rather than being indicative of some unique stage through which the satellite is passing. The current surface and subsurface structure should represent a quasi-equilibrium configuration rather than some step in an evolutionary sequence. So perhaps change in geological style on Europa is not global, but patcry. (A patchy resurfacing model would be broadly analogous to a similar model proposed for Venus [Phillips et al., 1992].) If Europa is in a quasi steady state, any given patch of the surface might go through a change from ridged plains formation through lenticula and chaos formation. Older regions of mottled terrain might become unrecognizable with age, perhaps brightening and fading over time. If mottled terrain formation is linked to global compressional stresses, for example, then regions of mottled terrain formation may sweep across the satellite as nonsynchronous rotation proceeds, fol-

lowed by episodes of ridged plains formation [Pappalardo et al., 1998c].

This scenario implies that mottled terrain should be converted back into ridged plains, but areas showing such transi- tions are rare. Limited Galileo coverage means it is possible that we have not observed the right places on the satellite, but this explanation seems less likely as the mission nears a close. Perhaps the conversion back to ridged plains occurs rapidly and completely, so that all evidence of preexisting terrain is lost. The continuation of some Europan lineaments for very long distances (up to thousands of kilometers) across the •ellite makes this explanation, and indeed any invoking patchy resurfacing, seem unlikely.

If a subsurface ocean has existed within Europa in the geologically recent past, a steady state scenario would imply that it has been relatively stable over long periods of geological time and should still exist today. Differences in the resurfacing style or degree of activity across Europa might be linked to random or systematic variations in stress or heat flow.

6.2.2. A special time. Though seemingly improbable, it is not impossible that we have observed Europa during a special time in its history. If the Laplace resonance is ancient and stable, it is possible that Europa's tidal heating caused the satellite to cool from its initial warm history much more slowly than Ganymede or Callisto [Cassen et al., 1982]. Perhaps a slow monotonic cooling allowed an internal ocean to progressively freeze and the satellite's lithosphere to progressively thicken, promoting a very recent critical change in geological style. (For Venus, it has been suggested that monotonic cooling and lithospheric thickening might have caused a critical change in convective style to explain its apparently rapid decrease in resurfacing rate -500 Myr ago [e.g., Schubert et al., 1997].) This seems an unlikely explanation for Europa because of the necessary coincidence in timing. Moreover, a slow cooling of the satellite might be expected to bring about a progressive change in the satellite's geological style, leading to differently aged surface units which might be reflected in crater counts, but such is not evident from our preliminary examinations of small craters on Europa.

Perhaps we are seeing a unique time in Europa's history because the satellite's near-surface and deep interior recently cooled from a unique transient heating event. Malhotra [ 1991 ] and Showman and Malhotra [1997] have investigated possible scenarios for the three-body tidal evolution of the Galilean satellites, from the assumption that the satellites' orbital configuration evolved into the Laplace resonance over time. Although the authors are principally concerned with the tidal evolution of Ganymede, they find that the orbital eccentricity of Europa can increase during passage through several candidate Laplace- like tidal resonances. Ganymede's large free eccentricity suggests that passage through such a resonance may have occurred relatively recently, likely <1 Gyr ago [Showman et al., 1997]; this model is consistent with the current warm interior of Ga-

nymede implied by its magnetic field [Kivelson et al., 1996]. In this way, it is possible we are seeing Europa at a unique time in its history. Perhaps Europa's visible surface deformation occurred over a very short period of time when its orbital eccentricity was increased during passage through tidal resonance prior to entry into the stable Laplace resonance. In this scenario, tidal heating would have been greatly amplified, and the satellite's H20 layer partially or wholly melted. Today the satellite would be cooling off, with any internal ocean in the process of freezing or perhaps wholly frozen.


6.2.3. Episodic resurfacing model. Another model is one in which Europa's history is cyclical, or sporadically episodic. (For Venus, plausible episodic resurfacing models have been invoked to explain the planet's apparent global scale resurfacing--500 Myr ago [e.g., Turcotte, 1993].) Per- haps Europa experiences an era of ridged plains formation, followed by an era of chaos formation, then another period of ridged plains formation, and so on. This may reflect changes in the satellite's orbital eccentricity and associated intensity of tidal heating. Greenberg [1982, 1987] considers a model in which Io, Europa, and Ganymede formed in orbits deep within the Laplace resonance, evolving outward from resonance as tidal dissipation in Io increased. The outward evolution would decrease dissipation within Io, moving the satellites back toward deep resonance. Variations in the coupled thermal and orbital evolution of Io may drive variations in Europa's orbital eccentricity and tidal heating [Ojakangas and Stevenson, 1986; Fischer and Spohn, 1990]. This cycle could continue on a period of--108 years, with activity currently predicted to be on the wane [Greenberg, 1982; Ojakangas and Stevenson, 1986]. The tidal evolution model of Showman and Malhotra [1997] of-

fers the possibility of sporadic change in orbital eccentricity and corresponding tidal heating, if the Galilean satellites have evolved through multiple Laplace-like resonances.

A period of increased tidal heating in the relatively recent past could help to reconcile the high heat flow suggested by some Europan surface features with a relatively young Europan surface. While dynamical modeling offers a possible means of driving cyclical or sporadic tidal heating of Europa, it is not clear whether the satellite's orbital eccentricity would change sufficiently to affect the degree of internal melting or the surface geological activity [cf. Showman et al., 1997]. Aside from tidal variations, other means might be envisioned to in- duce episodic change in Europa's geological activity or resurfacing style. We note that in any episodic model, the large heat of fusion of ice, combined with the tendency for heating t o trigger convection rather than melting of ice, make it difficult to remelt Europa's interior if it ever froze [Cassen et al., 1982]. The close Galileo flybys of Io in late 1999 will help us to distinguish among Io interior models and better measure its heat flow, which in turn will place constraints on orbital evolution models (A. S. McEwen et al., manuscript in preparation, 1999).

6.2.4. Old after all. From the standpoint of thermal evolution, a simple explanation for a change in Europa's geological style is one in which the lithosphere and interior cooled monotonically after the satellite's formation, but owing to tidal heating, Europa remained geologically very active until just after a period of early heavy cratering that affected Gany- mede and Callisto [Cassen et al., 1982]. In this scenario, Europa's lithosphere thickened and its primordial liquid water ocean froze solid or retreated to depth long ago. This scenario would be consistent with the assumptions of Neukum [1997] and Neukum et al. [1998], who argue that Europa's visible surface is --1 - 3 Gyr old. However, as discussed in detail in section 3, such an old Europa is inconsistent with our current understanding of the population of impactors in the Jovian system.

Improved observational constraints on Jupiter family comets are of first-order importance to the question of Europa's surface age, and therefore to the evolutionary history of the satellite and its possible subsurface ocean. But to test this model further, it is essential to develop independent measures of Eu- ropa's surface age. For example, it may be possible to con-

strain the surface age by quantifying the rate of talus buildup and regolith development. Refined models of surface sputtering, erosion, and deposition could be used to derive the erosion rate of Pwyll's rays and other surface materials, constraining the maximum age of the satellite's landforms. If the process of brightening of Europa's bands can be better understood, then the brightening rate and band age might be constrained. Im- proved geophysical modeling can better constrain the tidal heating rate and the cooling rate of the satellite, notably through refining our understanding of the role of convection. Finally, continued mapping of the surface will provide infor- mation on the orientations, distribution, and relative ages of Europa's lineaments and material units, constraining the rate and degree of Europa's nonsynchronous rotation and testing the hypothesis of a near-global change in style of Europa's deformation.

7. Additional Galileo Tests for an Ocean

Galileo instruments other than SSI have obtained measure-

ments relevant to the ocean hypothesis, and these results are reviewed briefly here.

7.1. NIMS Multi-Spectral Data

Data from the Galileo NIMS experiment shows that water-ice absorptions at 1.45 and 2.0 i. an are notably asymmetrical in shape and shifted slightly toward shorter wavelengths. McCordet al. [1998, 1999] interpret these spectra as those of hydrated salts. A good fit to the data is provided by heavily hydrated magnesium sulfate salts, specifically MgSO4øXH2 ¸, where X_>6. Presence of the asymmetric water bands shows good correlation to the lower albedo, red endogenic features observed in SSI images (triple band flanks, pull-apart bands, lenticulae, and chaos) [Granahan et al., 1997; McCord et al., 1998]. Hydrated MgSO 4 salts are predicted to be a significant component of a Europan ocean, based on evolutionary modeling that assumes an initial carbonaceous chondrite composition [Kargel, 1991]. While such salts are not themselves colored, a colored liquid that includes hydrated magnesium salts has been produced in experiments in which primitive meteor- ites are leached in water [Fanale et al., 1998].

Hydrated salts are common in terrestrial evaporite deposits, and McCord et al. [1998, 1999] infer that the observed spectra may be the signature of salt-rich brine extruded onto Europa's surface. Presumably solid-state convection can also serve to transport salt-rich material from a brine-rich sublayer toward the surface. Comparison of NIMS spectra of Europa to laboratory salt spectra reveals some discrepancies, and grain size effects could play an important role [Dalton and Clark, 1999]. NIMS data provide a probe of the composition of Europa's endogenic materials and potentially the composition of a subsurface ocean. Additional Galileo and laboratory spectral data will help to refine the nature of Europa's surface impurities, their distribution, and their relative abundances.

7.2. PPR Search for Thermal Anomalies

The volcanic activity of Io was first detected from infrared observations of hot spots [Witteborn et al., 1979]. From Galileo and other observations it is clear that Europa does not share Io's energetic and high-temperature hot spots, but low- temperature anomalies remain a possibility. In 1981, ground-


based observations measured a possible but unconfirmed thermal outburst on Europa at 4.8 prn [Tittemore and Sinton, 1989]. Observations of Europa with Galileo's Photopolar- imeter-Radiometer (PPR) instrument [Orton et al., 1996] can provide a test for currently active hot spots by searching for nighttime temperature anomalies. If Europa's nighttime surface temperature is observed to be anywhere greater than about 130 K, endogenic activity can be confidently inferred even if the albedo and thermophysical properties of the surface are unknown [Spencer et al., 1999]. Results to date have indicated that no part of the mapped area (18% of Europa's surface) shows a nighttime temperature that would require endogenic activity [Spencer et al., 1999].

PPR data can also indicate the thermal inertia of surface ma-

terials, which might be related to endogenic activity. For example, regions of anomalously low thermal inertia might indicate recently deposited materials. Anomalous high postsunset equatorial temperatures have been observed by PPR across 0 to 60øW longitude on Europa. No obvious geological explanation exists for preferred equatorial occurrence of fine (low thermal inertia) ice grains, and unexpectedly high endogenic heat flow (-1 W m -2) is a speculative but possible explanation for this thermal anomaly [Spencer et al., 1999].

7.3. Magnetometer Indications of an Induced Field

If Europa possesses a conductive (salty) subsurface ocean, then the time-varying Jovian magnetic field can drive electric currents within the satellite, producing an induced magnetic field [Colburn and Reynolds, 1985; Kargel and Consolmagno, 1996]. The signature of such a field is suspected based on data from the Galileo magnetometer experiment. Khurana et al. [1998] and Kivelson et al. [1999] show that the data from Gali- leo's close encounters with Europa are suggestive of an induced magnetic field, or possibly a weak, highly tilted intrinsic dipole field. An induced field would be strong evidence for an extant subsurface ocean. Assuming a conductivity similar to terrestrial oceans, a water layer >10 km thick is implied [Khurana et al., 1998]. The intrinsic and induced field models predict very different signatures for encounters which occur south of Jupiter's magnetic equator, and a Galileo encounter in January 2000 could provide a definitive test of the induced field hypothesis.

Interestingly, data from the C3 and C9 encounters with Cal- listo show a very good fit to the model of an induced dipole [Khurana et al., 1998; Kivelson et al., 1999]. However, Cal- listo displays little surface expression of endogenic activity [Klemaszewski et al., 1998] and is inferred the have experienced only a modest degree of internal differentiation [Ander- son et al., 1998a]. Thus it is not clear how a subsurface ocean would have formed on Callisto and how it could have been

maintained to the current day. It seems possible for an ocean to have formed near the minimal melting point pressure of ice, at 150 km depth within Callisto, with ammonia or salts serving as antifreeze that preserved an ocean to the present day [Ste- venson, 1998]. Better understanding of the cause and implications of Callisto's magnetic signature will have important implications for the state of Europa's interior.

8. Future Missions to Europa

ocean currently exists within the satellite [Johnson et al., 1999]. Possible follow-on missions are being investigated to land on the satellite's surface [Chyba et al., 1999]. There are a variety of geophysical, geomorphological, and geochemical experiments which could be performed from Europa orbit or from the surface to provide more definitive tests of the existence and nature of a subsurface ocean.

8.1. Potential Experiments From Orbit

Geophysical measurements will be fundamental to detecting an ocean during a Europa Orbiter mission. As Europa orbits Jupiter, its tidal bulge rises and lowers as Europa moves through perijove and apojove; simultaneously, the position of the tidal bulge librates due to the ellipticity of the orbit. These tidal variations are reflected in the satellite's gravitational field and can be detected in the satellite's shape, as quantified by the gravitational and tidal Love numbers (k 2 and h2, respectively). In combination, measurement of both tidal Love numbers can

distinguish among various ice and mantle rheologies and would constrain the depth to an ocean [Edwards et al., 1997; Castillo et al., 1998]. This can be achieved through precise Doppler tracking of an orbiting spacecraft to determine the time- varying gravitational field, in combination with precise laser altimetry to measure the time-varying shape. For an ice shell tens of kilometers thick, Europa's tidal bulge would flex by at most-30 m; for a solid ice shell, the tidal bulge fluctuation would be only -1 m [Edwards et al., 1997]. Thus geophysical measurements should be able to distinguish between solid-state ice and liquid water, answering the question of whether an ocean exists below Europa's surface today. An orbiting laser altimeter would have the added benefit of providing detailed topographic profiles across Europa's surface features.

A camera on board a spacecraft in a near-polar low-altitude orbit of Europa would be able to image the entire globe at regional-scale resolution and could image selected features at very high resolution, addressing aspects of all nine lines of geological evidence discussed above, in greater detail than is possible from the limited Galileo imaging coverage. Notably, such imaging would reveal the global distribution of surface features, permitting a global stratigraphy to be assembled. This would permit correlation of geological units and their relative ages across the surface, allowing interpretation of the spatial and temporal evolution of the satellite's surface and interior.

Multispectral capability, notably high resolution coverage in the near-infrared portion of the spectrum, would expand upon Galileo NIMS results in addressing the composition of Europa's surface features and could, by inference, address the composition of a subsurface ocean. Furthermore, a spectrometer sensitive to organic materials (notably at-3.4 prn, where many complex organic molecules have a characteristic absorption [e.g., Moore and Donn, 1982]) would facilitate the search for organic matedhals on Europa's surface, ingredients that would be vital for life. Surface composition might also be inferred with an orbiting mass spectrometer, which could detect species sputtered from the surface [Johnson et al., 1998].

A radar sounder has the potential to directly detect subsurface liquid within Europa [Squyres, 1989; Chyba et al., 1998]. At sufficiently long wavelengths, cold clean ice is relatively transparent to radar. In the terrestrial Antarctic, ice penetrating

NASA has recently announced its intent to send an space- radar has detected the water interface of Lake Vostok through -4 craft into orbit about Europa in the next decade, with the prin- km of overlying ice [Kapitsa et al., 1996]. Chyba et al. [1998] cipal goal of determining definitively whether a subsurface model the two-way attenuation within Europa's ice shell for a


variety of ice temperature profiles and impurity concentrations, for a radar wavelength of 6 m. For an ice lithosphere with a conductive temperature profile, they find that an ice/ocean interface can be detected to -5 - 10 km depth if the ice is relatively clean (-1 - 10% lunar-like impurities). If the ice shell is convecting with a warm (-250 K) adiabatic temperature profile, the detection depth is reduced to -3 - 5 km for relatively clean ice; detection is precluded for a warm convecting lithosphere with impurity levels -50%. Thus the prospect of detecting a -10 km deep ice/water interface is reasonable if Europa's ice is relatively cold and clean; however, warm convecting ice may preclude radar detection of an ice-ocean interface. Even if a radar signal does not detect an ocean interface, a radar system might detect melt zones within the shallower Europan lithosphere, if such zones occur beneath the satellite's youngest ridges, lenticulae, or chaos.

An orbiting thermal instrument has the potential to detect the infrared signature of ice or water flows recently erupted onto Europa's surface. If the local surface albedo is known and a region is observed at several different times of day to allow com- putation of the local radiative balance, an extruded water flow can be detected for decades to centuries (J. Van Cleve, Detecting Europan thermal palimpsests, submitted to Icarus, 1999; hereinafter referred to as submitted paper). If nighttime surface temperatures exceed 130 K, then endogenic activity can be confidently inferred even if the albedo and thermophysical properties of the surface are unknown [Spencer et al., 1999]. J. Van Cleve (submitted paper, 1999) demonstrates that the thermal radiation from a warm ice eruption _> 200 m in diameter ex- lauded at Europa's equator can be detected for -30 years and a similarly sized eruption of a thick liquid water flow can be detected for -90 years; a freezing water deposit at 60 ø latitude can be detected thermally for -400 years. Confirmation of current activity would imply that Europa's surface is indeed young and would increase the likelihood that a subsurface ocean has

persisted to the current day. The Galileo magnetometer suggestion of an induced mag-

netic field at Europa, plausibly produced by induction in a salty subsurface ocean, could be definitively tested with an orbiting magnetometer. Induced and intrinsic field signatures could be distinguished by monitoring Europa's magnetic field as the radial component of the external Jovian field reverses sign on the timescale of half the satellite's 11.1 hour synodic period about Jupiter.

8.2. Potential Surface Experiments

Following an orbiter mission, one or more surface landers or penetrators might be sent to Europa, as outlined by Chyba et al. [1999]. A science goal of primary importance would be direct compositional sampling of near-surface materials. This would be indicative of the chemistry of Europa's icy shell and potentially of its subsurface ocean. Testing for organic materials would be of fundamental importance, and it would be vital to sample materials which are shallowly buried and so protected from the intense Jovian radiation environment [Varnes and Jakosky, 1999]. Direct sampling might allow Europa's surface materials to be radiometrically dated [Kargel, 1992].

Lander instruments could determine the internal structure of

Europa, notably the depth to a subsurface ocean. If Europa is geologically active, then abundant tectonic activity probably occurs today; tidal flexing may produce detectable seismic activity even on a geologically inactive world. If seismic detec-

tors are positioned on the surface, the satellite's subsurface structure could be mapped by monitoring the types and arrival times of seismic waves. The time-varying Jovian field can serve as an ideal source for passive electromagnetic investigation of Europa's subsurface. Electromagnetic sounding, such as the magnetotelluric method of simultaneous measurement of electric and magnetic fields, would allow the local conductivity structure of Europa to be determined, quantifying the local depth to an ice/water interface [Grimm, 1999]. A measure of endogenic heat flow is important to understanding Europa's degree of current geological activity and its internal heat sources, and placement of temperature sensors within the subsurface may permit measurement of local heat flow. Very high resolution imaging would be important in providing the geological context for surface measurements and to understanding the small-scale characteristics of the local terrain.

Ultimately, it may be possible for a "hydrobot" craft to melt its own path through Europa's ice shell, in order to explore a subsurface ocean in situ [Trowell et al., 1996]. The heat energy required could be supplied by a nuclear source, and a surface sta- tion would relay data back to Earth. Perhaps once it reached liquid, a craft could detach itself to investigate Europa's physical and chemical oceanography and to search for the signature of indigenous biology, exploring as an autonomous underwater vehicle.

9. Major Outstanding Questions

Many questions about Europa have been answered with Gali- leo data, but many remain, as outlined throughout this work. Here are summarized five principal outstanding questions relevant to the current existence of a Europan ocean.

1. What is the distribution of ice and liquid water within Eu- ropa today? Recently acquired Galileo data along with continued analysis and geophysical modeling will help to address this fundamental issue. Future geophysical measurements are necessary to answer this question more conclusively.

2. What is the age of the surface, and is Europa currently active? Improved determination of impactor fluxes, analysis of crater size-frequency distributions, understanding of processes that age surface materials, and searches for current activity will help to constrain Europa's age and its degree of ongoing activity. Innovative techniques to more directly measure Europa's surface age would be of great value.

3. Have there been changes in Europa's heat flow and geological style through time? Geological mapping, morphological analyses, and geophysical modeling are vital to this issue, along with more extensive imaging coverage.

4. What impurities exist on and within Europa's H20-dominated outer layer? A combination of multispectral, theoretical, and laboratory investigations can address the nature and geological effects of Europan non-water-ice components, and their possible relevance to exobiology.

5. Have there been variations in Europa's rotational and orbital characteristics through time? Geological mapping and theoretical modeling are necessary to better constrain the degree and rate of Europa's nonsynchronous rotation and tidal evolution. Improved understanding of the geological histories of all four Galilean satellites is needed to address the issue of

tidal evolution.


Acknowledgments. This paper is dedicated to the memory of C. Sagan, who advocated critical evaluation of any hypothesis independent of its inherent appeal. We thank John Spencer, Ray Reynolds, and William McKinnon for constructive reviews, and Kevin Zahnle for providing his group's most recent cratering rate estimates. The concept for this synthesis began with a presentation by R.T.P. at the Autonomous Underwater Vehicles Laboratory of MIT. This work is supported by NASA's Galileo Project.

References

Allison, M. L., and S. M. Clifford, Ice-covered water volcanism on Ganymede, J. Geophys. Res., 92, 7865-7876, 1987.

Anderson, J. D., W. L. Sjogren, and G. Schubert, Galileo gravity results and the internal structure of Io, Science, 272, 709-712, 1996.

Anderson, J. D., E. L. Lau, W. L. Sjogren, G. Schubert, and W. B. Moore, Europa's differentiated internal structure: Inferences from two Galileo encounters, Science, 276, 1236-1239, 1997.

Anderson, J. D., G. Schubert, R. A. Jacobsen, E. L. Lau, W. B. Moore, and W. L. Sjogren, Distribution of silicates and ices in Europa and Callisto (abstract), Eos Trans. AGU, 79, Spring Meet. Suppl., S199, 1998a.

Anderson, J. D., G. Schubert, R. A. Jacobsen, E. L. Lau, W. B. Moore, and W. L. Sjogren, Europa's differentiated internal structure: Infer- ences from four Galileo encounters, Science, 281, 2019-2022, 1998b.

Baross, J. A., and S. E. Hoffman, Submarine hydrothermal vents and associated gradient environments as sites for the origin and evolution of life, Origin Life Evol. Bioaphere, 15, 327-345, 1985.

Basilevsky, A. T., J. W. Head, G. G. Schaber, and R. G. Strom, The resurfacing history of Venus, in Venus II.' Geology, Geophysics, At- mosphere, and Solar Wind Environment, edited by S. W. Bougher et al., pp. 1047-1084, Univ. of Ariz. Press, Tucson, 1997.

Beeman, M., W. B. Durham, and S. H. Kirby, Friction of ice, J. Geophys. Res., 93, 7625-7633, 1988.

Belton, M. J. S., et al., Galileo's first images of Jupiter and the Galilean satellites, Science, 274, 377-385, 1996.

Bierhaus, B., et al., Secondary cratering on Europa: A chronology of the •11 impact event and the Conamara region (abstract), Eos Trans. AGU, 79, Spring Meet. Suppl., S198, 1998.

Brotchie, J. F., Flexure of a liquid-filled spherical shell in a radial gravity field, Mod. Geol., 3, 15-23, 1971.

Carr, M. H., et al., Evidence for a subsurface ocean on Europa, Nature, 391,363-365, 1998.

Cassen, P., R. T. Reynolds, and S. J. Peale, Is there liquid water on Europa?, Geophys. Res. Lett., 6, 731-734, 1979.

Cassen, P., S. J. Peale, and R. T. Reynolds, Tidal dissipation in Europa: A correction, Geophys. Res. Lett., 7, 987-988, 1980.

Cassen, P.M., S. J. Peale, and R. T. Reynolds, Structure and thermal evolution of the Galilean satellites, in Satellites of Jupiter edited by D. Morrison, pp. 93-128, Univ. of Ariz. Press, Tucson, 1982.

Castillo, J., A. Mocquet, and C. Sotin, Gravitational effects of internal oceans in the icy satellites of Jupiter, in The Jovian System After Galileo: The Saturnian System Before Cassini-Huygens, pp. 45-46, Lab. de G6ophys. et Plantt., Nantes, France, 1998.

Chapman, C. R., E. V. Ryan, W. J. Merline, G. Neukum, R. Wagner, P. C. Thomas, J. Ververka, and R. J. Sullivan, Cratering on Ida, Icarus, 120, 77-86, 1996.

Chapman, C. R., W. J. Merline, B. Bierhaus, S. Brooks, and the Galileo Imaging Team, Cratering in the Jovian system: Intersatellite comparisons, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1927, 1998.

Chuang, F. C., S. D. Kadel, J. E. Klemaszewski, R. Greeley, and the Galileo SSI Team, Europa: Geomorphologic mapping and stereo analysis of chaos regions southeast of the Tyre multi-ringed impact structure, (abstract), Eos Trans. AGU, 79, Fall Meet. Suppl., F540- F541, 1999.

Chyba, C. F. and G. D. McDonald, The origin of life in the solar system: Current issues, Annu. Rev. Earth Planet. Sci., 23, 215-250, 1995.

Chyba, C. F., S. J. Ostro, and B.C. Edwards, Radar detectability of a subsurface ocean on Europa, Icarus, 134, 292-302, 1998.

Chyba, C. F., et al., Europa and Titan: Preliminary recommendations of the Campaign Science Working Group on Prebiotic Chemistry in the Outer Solar System, Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract i537, 1999.

Clark, B. E., P. Helfenstein, J. Veverka, M. Bell, P. E. Geissler, C. B. Phillips, A. S. McEwen, R. Greeley, R. Sullivan, and T. Denk,

Multispectral terrain analysis of Europa from Galileo's first images, Icarus, 135, 95-106, 1998.

Colbum, D. S., and R. T. Reynolds, Electrolytic currents in Europa, Icarus, 63, 39-44, 1985.

Collins, G. C., J. W. Head, and R. T. Pappalardo, The role of extensional instability in creating Ganymede grooved terrain: Insights from Gali- leo high-resolution stereo imaging, Geophys. Res. Lett., 25, 233-236, 1998.

Collins, G. C., J. W. Head, R. T. Pappalardo, and N. A. Spaun, Evaluating models for the formation of chaotic terrain on Europa, Lunar Planet. Sct. Conf, [CD-ROM], XXX, abstract 1434, 1999.

Comer, R. P., S.C. Solomon, and J. W. Head, Mars: Thickness of the lithosphere from the tectonic response to volcanic loads, Rev. Geo- phys., 23, 61-92, 1985.

Consolmagno, G. J., Thermal history models of icy satellites, Master's thesis, Mass. Inst. of Technol., Cambridge, 1975.

Consolmagno, G. J. and J. S. Lewis, Structural and thermal models of icy Galilean satellites, in Jupiter, edited by T. Gehrels, pp. 1035-1051, Univ. of Ariz. Press, Tucson, 1976.

Cook, A. F., E. M. Shoemaker, L. A. Soderblom, K. F. Mullins, and R. Fiedler, Volcanism in ice on Europa, Bull. Am. Astron. $oc., 14, 736- 737, 1982.

Cook, A. F., E. M. Shoemaker, L. A. Soderblom, K. F. Mullins, and R. Fiedler, Volcanism in ice on Europa, Report on the Planetary Geol- ogy Program 1982, NASA Tech. Memo., TM-85127, 415-416, 1983.

Crater Analysis Techniques Working Group, Standard techniques for presentation and analysis of crater size-frequency data, Icarus, 37, 467-474, 1979.

Crawford, G. D., and D. J. Stevenson, Gas-driven water volcanism and the resurfacing of Europa, Icarus, 73, 66-79, 1988.

Croft, S. K., A proposed origin for palimpsests and anomalous pit craters on Ganymede and Callisto, Proc. Lunar Planet. $ci. Conf. 14th, Part I, J. Geophys. Res., 88, suppl., B71-B89, 1983.

Dalton, J. B. and R. N. Clark, Observational constraints on Europa's surface composition from Galileo NIMS data, Lunar Planet. $ci. Conf., [CD-ROM], XXX, abstract 2064, 1999.

Deschamps, F., and C. Sotin, Thermal convection in the outer ice I shell of icy satellites, in The Jovian System After Galileo: The Saturnian System Before Cassini-Huygens, p. 51, Lab. de G6ophys. et Planttol., Nantes, France, 1998.

Domingue, D. L., and B. Hapke, Disk-resolved photometric analysis of Europan terrains, Icarus, 99, 70-81, 1992.

Domingue, D., and A. Verbiscer, Re-analysis of the solar phase curves of the icy Galilean satellites, Icarus, 128, 49-74, 1997.

Domingue, D. L., B. W. Hapke, G. W. Lockwood, and D. T. Thompson, Europa's phase curve: Implications for surface structure, Icarus, 90, 30-42, 1991.

Duncan, M., T.Quinn, and S. Tremaine, The origin of short-period comets, Astrophys. J., 328, L69-L73, 1988.

Durham, W. B., S. H. Kirby, and L. A. Stem, Creep of water ices at planetary conditions: A compilation, J. Geophys. Res., 102, 16,293- 16,302, 1997.

Edwards, B. C., et al., The Europa Ocean Discovery Mission, in Instru- ments, Methods, and Missions for the Investigation of Extraterrestrial Microorganisms, edited by R. B. Hoover, Proc. $PIE lnt. $oc. Opt. Erg., 3111, 249-261, 1997.

Eviatar, A., A. Bar-Nun, and M. Podolak, Europan surface phenomena, Icarus, 61, 185-191, 1985.

Fagents, S. A., R. Greeley, and the Galileo SSI Team, Explosive venting on Europa? A preliminary assessment, Lunar Planet. $ci. Conf., XXVIII, 345-346, 1997.

Fagents, S. A., S. D. Kadel, R. Greeley, R. L. Kirk, and the Galileo SSI team, Styles of cryovolcanism on Europa: Summary of evidence from the Galileo nominal mission, Lunar Planet. Sci. Conf., ,[CD ROM], XXIX, abstract 1721, 1998.

Fagents, S. A., R. Greeley, R. J. Sullivan, R. T. •Pappalardo, and L. M. Prockter, Cryomagmatic mechanisms for the ftrmation pf Rhadamanthys Linea, triple band• margins, 'and other• low albedo features on Europa, Icarus, in press,, 1999.

Fanale, F. P., T. V. Johnson; and D. L. lMatson, Io's surface and the his•- ries of the Galilean satellites, in,Planetary:Satellites, edited by 'J. Bums, pp. 379-405, Univ. of Ariz. Press, Tucs•on, 1977.

Fanale, F. P., J. R. Salvail, D. L. Matsoh, and R. H. Brown, The effect of volume phase changes, mass transport, sunlight penetration, and den- siftcation on the thermal regime of iicy regoliths, Icarus, 88, 193-204, 1990.


Fanale, F. P., Y.-H. Li, E. Decarlo, N. Domergue-Schmidt, S. K. Sharma, K. Horton, J. C. Granahan, and the Galileo NIMS Team, Laboratory simulation of the chemical evolution of Europa's aqueous phase, Lu- nar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1248, 1998.

Fischer, H.-J., and T. Spohn, Thermal-orbital histories of viscoelastic models of Io (J1), Icarus, 83, 39-65, 1990.

Giese, B., R. Wagner, and G. Neukum, The local topography of Europa: Stereo analysis of Galileo SSI images and implications for geology (abstract), Geophys. Res. Abstr., 1, 742, 1999.

Geissler, P. E., et al., Galileo multispectral imaging of Europa: Evidence for non-synchronous rotation, Lunar Planet. Sci. Conf., XXVIII, 401- 402, 1997.

Geissler, P. E., et al., Evidence for non-synchronous rotation of Europa, Nature, 391,368-370, 1998a.

Geissler, P. E., et al., A recently active lineament on Europa?, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1904, 1998b.

Geissler, P. E., et al., Evolution of lineaments on Europa: Clues from Galileo multispectral imaging observations, Icarus, 135, 107-126, 1998c.

Goldreich, P., Final spin states of planets and satellites, Astraphys. J., 71, 1-7, 1966.

Goldsby, D. L., and D. L. Kohlstedt, Grain boundary sliding in fine- grained ice I, Scr. Mater., 37, 1399-1406, 1997a.

Goldsby, D. L., and D. L. Kohlstedt, Flow of Ice I by dislocation, grain boundary sliding, and diffusion processes (abstract), Lunar Planet. Sci. Conf., XXVIII, 429-430, 1997b.

Golombek, M.P., and W. B. Banerdt, Constraints on the subsurface structure of Europa, Icarus, 83, 441-452, 1990.

Granahan, J. C., et al., A Galileo multi-instrument study of Europa's color heterogeneities, Bull. Am. Astron. Sac., 29, 982, 1997.

Greeley, R., J. H. Fink, D. E. Gault, and J. E. Guest, Experimental simulation of impact cratering on icy satellites, in Satellites of Jupiter, edited by D. Morrison, pp. 340-378, Univ. of Ariz. Press, Tucson, 1982.

Greeley, R., M. Coon, R. Sullivan, P. E. Geissler, J. W. Head III, R. T. Pappalardo, and J. M. Moore, Terrestrial sea ice processes: Consid- erations for Europa, Icarus, 135, 25-40, 1998a.

Greeley, R., et al., Europa: Initial Galileo geological observations, Icarus, 135, 4-24, 1998b.

Greenberg, R., Orbital evolution of the Galilean satellites, in Satellites of Jupiter, edited by D. Morrison, pp. 65-92, Univ. of Ariz. Press, Tucson, 1982.

Greenberg, R., Galilean satellites: Evolutionary paths in deep resonance, Icarus, 70, 334-347, 1987.

Greenberg, R., and S. T. Weidenschilling, How fast do the Galilean satellites spin?, Icarus, 58, 186-196, 1984.

Greenberg, R., P. E. Geissler, G. Hoppa, B. R. Tufts, D. D. Durda, R. Pappalardo, J. W. Head, R. Greeley, R. Sullivan, and M. H. Carr, Tectonic processes on Europa: Tidal stresses, mechanical response, and visible features, Icarus, 135, 64-78, 1998a.

Greenberg, R., G. Hoppa, R. Tufts, and P. Geissler, Chaos regions: Widespread melt-through to the surface of Europa?, Bull. Am. Astron. Sac., 30, 1086, 1998b.

Grieve, R. A. F., and E. M. Shoemaker, The record of past impacts on Earth, in Hazards Due to Comets and Asteroids, edited by T. Gehrels, pp. 417-462, Univ. of Ariz. Press, Tucson, 1994.

Grimm, R. E., Natural electromagnetic exploration for water on Mars and Europa, Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 1442, 1999.

Hapke, B., Bidirectional reflectance spectroscopy, 4, The extinction co- efficient and the opposition effect, Icarus, 67, 264-280, 1986.

Head, J. W., and R. T. Pappalardo, Brine mobilization during lithospheric heating on Europa: Implications for formation of chaos terrain, lenticulae texture, and color variations, Lunar Planet. Sci. Conf., [CD- ROM], XXX, abstract 1664, 1999.

Head, J. W., R. T. Pappalardo, R. Greeley, R. Sullivan, and the Galileo Imaging Team, Origin of ridges and bands on Europa: Morphologic characteristics and evidence for linear diapirism from Galileo data, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1414, 1998a.

Head, J. W., N. D. Sherman, R. T. Pappalardo, R. Greeley, R. Sullivan, D. A. Senske, A. McEwen, and the Galileo Imaging Team, Geologic history of the FA region of Europa: Implications for ridge formation, cryovolcanism, and chaos formation, Lunar Planet. Sci. Conf., [CD- ROM], XXIX, abstract 1412, 1998b.

Head, J. W., N. D. Sherman, R. T. Pappalardo, C. Thomas, R. Greeley, and the Galileo SSI team, Cryovolcanism on Europa: Evidence for the emplacement of flows and related deposits in the EA region (5N,

305W) and interpreted eruption conditions, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1491, 1998c.

Head, J. W., R. T. Pappalardo, R. Sullivan, and the Galileo SSI team, Europa: Morphological characteristics of ridges and triple bands from Galileo data (E4 and E6) and assessment of linear diapirism model, J. Geophys. Res., this issue, 1999.

Helfenstein, P., and A. F. Cook, Active venting on Europa? Analysis of a transient bright surface feature, (abstract), Lunar Planet. Sci. Conf. XV, 531-532, 1984.

Helfenstein, P., and E. M. Parmentier, Patterns of fracture and tidal stresses due to nonsynchronous rotation: Implications for fracturing on Europa, Icarus, 61,175-184, 1985.

Helfenstein, P., et al., Galileo obervations of Europa's opposition effect, Icarus, 135, 41-63, 1998.

Hendrix, A. R., Galileo UVS observations of Europa: Variations in ultraviolet absorber with surface features, (abstract), Eos Trans. AGU, 79, Fall Meet. Suppl., F535, 1998.

Hillgren, V. J., and H. J. Melosh, Crater relaxation on Ganymede: Impli- cations for ice rheology, Geophys. Res. Lett., 16, 1339-1342, 1989.

Hobbs, P.V., Ice Physics, 837 pp., Oxford at the Clarendon Press, London, 1974.

Hoppa, G. V., and B. R. Tufts, Formation of cycloidal features on Eu- ropa, Lunar Planet. Sci. Conf, [CD-ROM], XXX, abstract 1599, 1999.

Hoppa, G., R. Greenberg, P. Geissler, J. Plassmann, and B. R. Tufts, Ro- tation of Europa: Constraints from terminator positions, (abstract), Lunar Planet. Sci. Conf, XXVIII, 597, 1997.

Hoppa, G., B. R. Tufts, R. Greenberg, and P. E. Geissler, Strike-slip faults on Europa: Global shear patterns driven by tidal stress, Icarus, in press, 1999.

Ip, W.-H., D. J. Williams, R. W. McEntire, and B. H. Mauk, Ion sputtering and surface erosion at Europa, Geophys. Res. Lett., 25, 829-832, 1998.

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

Johnson, R. E., Polar frost formation on Ganymede, Icarus, 62, 344-347, 1995.

Johnson, R. E., R. M. Killen, J. H. Waite Jr., and W. S. Lewis, Europa's surface composition and sputter-produced ionosphere, Geophys. Res. Lett., 25, 3257-3260, 1998.

Johnson, T. V., C. Chyba, K. P. Klaasen, and R. J. Terrile, Outer Plan- ets/Solar Probe Project: Europa orbiter, Lunar Planet. Sci. Conf., [CD- ROM], XXX, abstract 1423, 1999.

Kadel, S. D., S. A. Fagents, R. Greeley, and the Galileo SSI Team, Trough-bounding ridge pairs on Europa - Considerations for an endogenic model of formation, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1078, 1998.

Kapitsa, A. P., J. K. Ridley, G. Q. Robin, M. J. Siegert, and I. A. Zotikov, A large deep freshwater lake beneath the ice of central East Antarc- tica, Nature, 381,684-686, 1996.

Kargel, J. S., Brine volcanism and the interior structures of asteroids and icy satellites, Icarus, 94, 368-390, 1991.

Kargel, J. S., Ammonia-water volcanism on icy satellites: Phase relations at 1 atmosphere, Icarus, 100, 556-574, 1992.

Kargel, J. S. and G. J. Consolmagna, Magnetic fields and the detectability of brine oceans in Jupiter's icy satellites, (abstract), Lunar Planet. Sci. Conf , XXVII, 643-644, 1996.

Khurana, K. K., M. G. Kivelson, D. J. Stevenson, G. Schubert, C. T. Rus- sel, R. J. Walker, and C. Polanskey, Induced magnetic fields as evidence for subsurface oceans in Europa and Callisto, Nature, 395, 777-780, 1998.

Kivelson, L. A., K. K. Khurana, C. T. Russel, R. J. Walker, J. Warnecke, F. V.Coriniti, C. Polanskey, D. J. Southwood, and G. Schubert, Dis- covery of Ganymede's magnetic field by the Galileo spacecraft, Nature, 384, 537-541, 1996.

Kivelson, M. G., K. K. Khurana, D. J. Stevenson, L. Bennett, S. Joy, C. T. Russel, R. J. Walker, and C. Polanskey, Europa and Callisto: Induced or intrinsic fields in a periodically varying plasma environment, J. Geophys. Res., 104, 4609-4625, 1999.

Klemaszewski, J. E., et al., Galileo at Callisto: Overview of nominal mission results, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1866, 1998.

Kraft, M.D., and R. Greeley, Ridge orientations on Europa: Observa- tions from Galileo images, Geol. Sac. Am. Abstr. Programs, 29, A- 406, 1997.

Leith, A. C., and W. B. McKinnon, Is there evidence for polar wander on Europa?, Icarus, 120, 387-398, 1996.


Levison, H. F., and M. J. Duncan, From the Kuiper Belt to Jupiter-family Ojakangas, G. W., and D. J. Stevenson, Thermal state of an ice shell on comets: The spatial distribution of ecliptic comets, Icarus, 117, 13-32, Europa, Icarus, 81,220-241, 1989b. 1997. Orton, G. S., J. R. Spencer, L. D. Travis, T. Z. Martin, and L. K. Tamp-

Lewis, J. S., Satellites of the outer planets: Their physical and chemical pari, Galileo photopolarimeter-radiometer observations of Jupiter and nature, Icarus, 15, 174-185, 1971. the Galilean satellites, Science, 274, 389-392, 1996.

Lucchitta, B. K., and L. A. Soderblom, Geology of Europa, in Satellites of Pappalardo, R. T., and M.D. Coon, A sea ice analog for the surface of Jupiter, pp. 521-555, edited by D. Morrison, Univ. of Ariz. Press, Tucson, 1982.

Lunine, J. L., and R. D. Lorenz, Light and heat in cracks on Europa: Im- plications for prebiotic synthesis, (abstract), Lunar Planet. Sci. Conf., XXVIII, 855-856, 1997.

Lunine, J. L., and D. J. Stevenson, Thermodynamics of clathrate hydrate at low and high pressures with application to the outer Solar System, Astrophys. J. Suppl., 58, 493-531, 1985.

Macdonald, K. C., Mid-ocean ridges: Fine scale tectonic, volcanic and hydrothermal processes within the plate boundary zone, Annu. Rev. Earth Planet. Sci., 10, 155-190, 1982.

Malhotra, R., Tidal origin of the Laplace resonance and the resurfacing of Ganymede, Icarus, 94, 399-412, 1991.

Malin, M. C., and D.C. Pieri, Europa, in Satellites, edited by J. A. Bums and M. S. Matthews, pp. 689-717, Univ. of Ariz. Press, Tucson, 1986.

McCord, T. B., et al., Salts on Europa's surface detected by Galileo's Near Infrared Mapping Spectrometer, Science, 280, 1242-1245, 1998.

McCord, T. B., et al., Hydrated salt minerals on Europa's surface from the Galileo near-infrared mapping spectrometer (NIMS) investigation, J. Geophys. Res., 104, 11,827-11,851, 1999.

McEwen, A. S., Exogenic and endogenic albedo and color patterns on Europa, J. Geophys. Res., 91, 8077-8097, 1986a.

McEwen, A. S., Tidal reorientation and the fracturing of Jupiter's moon Europa, Nature, 321, 49-51, 1986b.

McEwen, A. S., J. M. Moore, and E. M. Shoemaker, The Phanerozoic impact cratering rate: Evidence from the farside of the Moon, J. Geophys. Res., 102, 9231-9242, 1997.

McKinnon, W. B., Sighting the seas of Europa, Nature, 386, 765-766, 1997.

McKinnon, W. B., Geodynamics of icy satellites, in Solar System Ices, edited by B. Schmitt et al., pp. 525-550, Kluwer Academic, Norwell, Mass., 1998.

McKinnon, W. B., Convective instability in Europa's floating ice shell, Geophys. Res. Lett., 26, 951-954, 1999.

McKinnon, W. B., and M. Gumis, On initiation of convection in Europa's floating ice shell (and the existence of the ocean below), Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 2058, 1999.

McKinnon, W. B., and H. J. Melosh, Evolution of planetary lithospheres: Evidence from multiring basins on Ganymede and Callisto, Icarus, 44, 454-471, 1980.

Moore, J. M., et al., Large impact features on Europa: Results of the Galileo nominal mission, Icarus, 135, 127-145, 1998.

Moore, M. H., and B. Donn, The infrared spectrum of a laboratory-syn- thesized residue: Implications for the 3.4 micron interstellar absorption feature, Astrophys. J., 257, L47-L50, 1982•.

Moreau, J., R. Greeley, R. Pappalardo, J. W. Head III, and the Galileo Solid State Imaging Team, A Rayleigh-Taylor instability model for domes on Europa, (abstract), Eos Trans. AGU, 46, Fall Meet. Suppl., F415, 1997.

Europa, (abstract), Lunar Planet. Sci. Conf., XXVII, 997-998, 1996. Pappalardo, R. T., and R. J. Sullivan, Evidence for separation across a

gray band on Europa, Icarus, 123, 557-567, 1996. Pappalardo, R. T., J. W. Head, R. Greeley, and the Galileo SSI Team, A

Europan ocean? The (circumstantial) geological evidence, in Europa Ocean Conference, Capistrano Conf. 5, pp. 59-60, San Juan Capis- trano Res. Inst., San Juan Capistrano, Calif., 1996.

Pappalardo, R. T., et al., Geological evidence for solid-state convection in Europa's ice shell, Nature, 391,365-368, 1998a.

Pappalardo, R. T., J. W. Head, N. D. Sherman, R. Greeley, R. J. Sullivan, and the Galileo Imaging Team, Classification of Europan ridges and troughs and a possible genetic sequence, in Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1859, 1998b.

Pappalardo, R. T., et al., Distribution of mottled terrain on Europa: A possible link to nonsynchronous rotation stresses, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1732, 1998c.

Pappalardo, R. T., J. Spencer, P. E. Geissler, L. M. Prockter, G. C. Collins, J. W. Head, P. Helfenstein, and the Galileo Imaging Team, What controls the albedo patterns on Europa?, (abstract) Eos Trans. AGU, 79, Spring Meet. Suppl., S198-S199, 1998d.

Passey, Q. R., and G. Shoemaker, Craters and basins on Ganymede and Callisto: Morphological indicators of crustal evolution, in Satellites of Jupiter, edited by D. Morrison, pp. 379-434, Univ. of Ariz. Press, Tucson, 1982.

Pellegrino, C. R., and J. A. Stoff, Darwin's Universe: Origins and Crises in the History of Life, Van Nostrand Reinhold, New York, 1983.

Phillips, C. B., A. S. McEwen, P. E. Geissler, R. Greeley, R. Sullivan, R. Pappalardo, and the Galileo SSI Team, New mosaics of Europa and mapping of endogenic units, (abstract), Lunar Planet. Sci. Conf., XXVIII, 1103-1104, 1997.

Phillips, C. B., A. S. McEwen, G. V. Hoppa, P. E. Geissler, R. T. Pap- palardo, D. P. Simonelli, R. Greeley, and the Galileo SSI Team, The search for active resurfacing on Europa and Io, Bull. Am. Astron. Soc., 30, 1085, 1998..

Phillips, R. J., R. F. Raubertas, R. E. Arvidson, I. C. Sarkar, R. R. Herrick, N. Izenberg, and R. E. Grimm, Impact craters and Venus resurfacing history, J. Geophys. Res., 97, 15,923-15,948, 1992.

Pollard, D. D., and A.M. Johnson, Mechanics of growth of some lacco- lithic intrusions in the Henry Mountains, Utah, II, Bending and failure of overburden layers and sill formation, Tectonophysics, 18, 311-354, 1973.

Prockter, L. M., R. T. Pappalardo, R. Sullivan, J. W. Head, J. G. Patel, B. Giese, R. Wagner, G. Neukum, and R. Greeley, Morphology and evolution of Europan bands: Investigation of a seafloor spreading analog, Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 1900, 1999a.

Prockter, L. M., J. W. Head, R. T. Pappalardo, G. C. Collinq, J. E. Kle- maszewski, P. E. Geissler, and the Galileo SSI Team, Geological mapping of central Agenor Linea, Europa (212 ø- 226ø), Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 1299, 1999b.

Neukum, G., Bombardment history of the Jovian system, in The Three Prockter, L. M., A. Antman, R. T. Pappalardo, J. W. Head, and G. C. Galileos: The Man, the Spacecraft, the Telescope, edited by C. Bar- bieri et al., pp. 201-212, Kluwer Acad., Norwell, Mass., 1997.

Neukum, G., and B. A. Ivanov, Crater size distributions and impact probabilities on Earth from lunar, terrestrial-planet, and asteroid cratering data, in Hazards Due to Comets and Asteroids, edited by T. Gehrels, pp. 359-416, Univ. of Ariz. Press, Tucson, 1994.

Ne•tkum, G., B. Konig, and J. Arkani-Hamad, A study of lunar impact crater size distributions, Moon, 12,201-229, 1975.

Neukum, G., R. Wagner, U. Wolf, B. A. Ivanov, J. W. Head, R. T. Pap- palardo, J. E. Klemaszewski, R. Greeley, and M. J. S. Belton, Crater- ing chronology in the Jovian system and derivation of absolute ages, Lunar Planet. Sci. Conf., [CD-ROM], XX/X, abstract 1742, 1998.

Collins, Europa: Stratigraphy and geological history of the anti-Jovian region from Galileo El4 solid-state imaging data, J. Geophys. Res., 104, 16,531-16,540, 1999c.

Rabinowitz, D. L., E. Bowell, E. M. Shoemaker, and K. Muinonen, The population of Earth-crossing asteroids, in Hazards Due to Comets and Asteroids, edited by T. Gehrels, pp. 285-312, Univ. of Ariz. Press, Tucson, 1994.

Rathbun, J. A., G. S. Musser Jr., and S. W. Squyres, Ice diapirs on Eu- ropa: Implications for liquid water, Geophys. Res. Lett., 25, 4157- 4160, 1998.

Reynolds, R. T., and P.M. Cassen, On the internal structure of the major satellites of the outer planets, Geophys. Res. Lett., 6, 121-124, 1979.

Noll, K. S., H. A. Weaver, and A.M. Gonnella, The albedo spectrum of Reynolds, R. T., S. W. Squyres, D. S. Colbum, and C. P. McKay, On the Europa from 2200 • to 3300 •, J. Geophys. Res., 100, 19,057- habitability of Europa, lcarus, 56, 246-254, 1983. 19,059, 1995. Reynolds, R. T., C. P. McKay, and J. F. Kasting, Europa, tidally heated

Ojakangas, G. W., and D. J. Stevenson, Episodic volcanism of tidally oceans, and habitable zones around giant planets, Adv. Space Res., 7, heated satellites with application to Io, Icarus, 66, 341-358, 1986. 5125-5132, 1987.

Ojakangas, G. W., and D. J. Stevenson, Polar wander of an ice shell on Ross, M., and G. Schubert, Tidal heating in an internal ocean model of Europa, Icarus, 81,242-270, 1989a. Europa, Nature, 325, 133-134, 1987.


Ruiz, J., Onset of convection, heat flow and thickness of Europa's ice shell, Earth Moon Planets, in press, 1999.

Ruiz, J., and R. Tejero, Heat flow and brittle-ductile transition in the ice shell of Europa, Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 1031, 1999.

Sack, N.J., R. E. Johnson, J. W. Boring, and R. A. Baragiola, The effect of magnetospheric ion bombardment on the reflectance of Europa's surface, Icarus, 100, 534-540, 1992.

Schenk, P., and M.P. A. Jackson, Diapirism on Triton: A record of crustal layering and instability, Geology, 21,299-302, 1993.

Schenk, P.M., and W. B. McKinnon, Fault offsets and lateral crustal movement on Europa: Evidence for a mobile ice shell, Icarus, 79, 75- 100, 1989.

Schubert, G., T. Spohn, and R. T. Reynolds, Thermal histories, composi- tions, and internal structures of the moons of the Solar System, in Sat- ellites, edited by J. A. Bums and M. S. Matthews, pp. 224-292, Univ. of Ariz. Press, Tucson, 1986.

Schubert, G., V. S. Solomatov, P. J. Tackley, and D. L. Turcotte, Mantle convection and the thermal evolution of Venus, in Venus II--Geology, Geophysics, Atmosphere, and Solar Wind Environment, edited by S. W. Bougher, et al., pp. 1245-1287, Univ. of Ariz. Press, Tucson, 1997.

Shoemaker, E. M, The age of Europa's surface, in Europa Ocean Con- ference, Capistrano Conf. 5, pp. 65-66, San Juan Capistrano Res. Inst., San Juan Capistrano, Calif., 1996.

Shoemaker, E. M., and R. F. Wolfe, Cratering time scales for the Gali- lean satellites, in Satellites of Jupiter, edited by D. Morrison, pp. 277- 339, Univ. of Ariz. Press, Tucson, 1982.

Shoemaker, E. M., R. F. Wolfe, and C.S. Shoemaker, Asteroid and comet flux in the neighborhood of Earth, in Global Catastrophes in Earth History, edited by V. L. Sharpton and P. D. Ward, Spec. Pap. Geol. Soc. Am., 247, 155-170, 1990.

Shoemaker, E. M., P. R. Weissman, and C. S. Shoemaker, The flux of periodic comets near Earth, in Hazards Due to Comets and Asteroids, edited by T. Gehrels, pp. 313-335, Univ. of Ariz. Press, Tucson, 1994.

Showman, A. P., and R. Malhotra, Tidal evolution into the Laplace resonance and the resurfacing of Ganymede, Icarus, 127, 93-111, 1997.

Showman, A. P., D. J. Stevenson, and R. Malhotra, Coupled orbital and thermal evolution of Ganymede, Icarus, 129, 367-383, 1997.

Smith, B. A., and the Voyager Imaging Team, The Galilean satellites of Jupiter: Voyager 2 imaging science results, Science, 206, 951-972, 1979.

Smith, B. A., and the Voyager Imaging Team, Voyager 2 at Neptune: Imaging science results, Science, 246, 1422-1449, 1989.

Spaun, N. A., J. W. Head, G. C. Collins, L. M. Prockter, and R. T. Pap- palardo, Conamara Chaos Region, Europa: Reconstruction of mobile polygonal ice blocks, Geophys. Res. Lett., 25, 4277-4280, 1998a.

Spaun, N. A., J. W. Head, R. T. Pappalardo, and the Galileo Imaging Team, Geologic history, surface morphology and deformation sequence in an area near Conamara Chaos, Europa, Lunar Planet. Sci. Conf., [CD-ROM], XXIX, abstract 1899, 1998b.

Spaun, N. A., L. M. Prockter, R. T. Pappalardo, G. C. Collins, A. Antman, R. Greeley, and the Galileo SSI Team, Spatial distribution of lenticulae and chaos on Europa, in Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 1847, 1999.

Spencer, J. R., Thermal segregation of water ice on the Galilean satellites, Icarus, 69, 297-313, 1987.

Spencer, J. R., L. K. Tamppari, T. Z. Martin, and L. D. Travis, Tem- peratures on Europa from Galileo PPR: Nighttime thermal anomalies, Science, 284, 1514-1516, 1999.

Squyres, S. W., Prospects for the existence and detectability of an ocean on Europa, Adv. Space Res., 9, 279-285, 1989.

Squyres, S. W., R. T. Reynolds, P. Cassen, and S. J. Peale, Liquid water and active resurfacing on Europa, Nature, 301,225-226, 1983.

Stein, R. S., G. C. P. King, and J. B. Rundle, The growth of geological structures by repeated earthquakes, 2, Field examples of continental dip-slip faults, J. Geophys. Res., 93, 13,319-13,331, 1988.

Stevenson, D., Heterogeneous tidal deformation and geysers on Europa, in Europa Ocean Conference, Capistrano Conf. 5, pp. 69-70, San Juan Capistrano Res. Inst., San Juan Capistrano, Calif., 1996.

Stevenson, D. J., An ocean within Callisto?, (abstract), Eos Trans. AGU, 79, Fall Meet. Suppl., F535, 1998.

Sullivan, R., et al., Ridge formation on Europa: Examples from Galileo high resolution images, Geol. Soc. Am. Abstr. Programs, 29, A-312, 1997.

Sullivan, R., et al., Episodic plate separation and fracture infill on the surface of Europa, Nature, 391, 371-372, 1998.

Sullivan, R., J. Moore, and R. Pappalardo, Mass-wasting and slope evolution on Europa, Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 1747, 1999a.

Sullivan, R., R. Greeley, J. Klemaszewski, J. Moreau, J. W. Head III, R. Pappalardo, J. Moore, and B. R. Tufts, High resolution Galileo geological mapping of ridged plains on Europa, Lunar Planet. Sci. Conf., [CD-ROM], XXX, abstract 1925, 1999b.

Thomas, P. J. and G. Schubert, Crater relaxation as a probe of Europa's interior, Proc. Lunar Planet. Sci. Conf. 16 th, Part 2, J. Geophys. Res., 91, suppl., D453-D459, 1986.

Thomas, P. J., and S. W. Squyres, Formation of crater palimpsests on Ganymede, J. Geophys. Res., 95, 19,161-19,174, 1990.

Tittemore, W. C., and W. M. Sinton, Near-infrared photometry of the Galilean satellites, Icarus, 77, 82-97, 1989.

Trowell, S., J. Wild, J. Horvath, J. Jones, E. Johnson, and J. Cutts, Through the Europan ice: Advanced lander mission options, in Europa Ocean Conference, Capistrano Conf 5, p. 76, San Juan Cap- istrano Res. Inst., San Juan Capistrano, Calif., 1996.

Tufts, B. R., A San Andreas-sized strike-slip fault on Europa, (abstract), Lunar Planet. Sci. Conf , XXVII, 1343-1344, 1996.

Tufts, B. R., R. Greenberg, P. Geissler, G. Hoppa, R. Pappalardo, R. Sullivan, and the Galileo Imaging Team, Crustal displacement features on Europa, Geol. Soc. Am. Abstr. Programs, 29, A-312, 1997.

Tufts, B. R., R. Greenberg, P. Geissler, G. Hoppa, R. T. Pappalardo, and the Galileo Imaging Team, Europan lithospheric displacement features and their interpretation, Lunar Planet. Sci. Conf, [CD-ROM], XXIX, abstract 1694, 1998.

Turcotte, D. L., An episodic hypothesis for venusian tectonics, J. Geo- phys. Res., 98, 17,061-17,068, 1993.

Turtle, E. P., H. J. Melosh, and C. B. Phillips, Tectonic modeling of the formation of Europan Ridges, (abstract), Eos Trans. AGU, 79, suppl., F541, 1998a.

Turtle, E. P., C. B. Phillips, A. S. McEwen, J. M. Moore, R. Greeley, and the Galileo SSI Team, Numerical modeling of multi-ring impact craters on Europa: Implications for subsurface structure, Lunar Planet. Sci. Conf, [CD-ROM], XXIX, abstract 1325, 1998b.

Turtle, E. P., C. B. Phillips, G. C. Collins, A. S. McEwen, J. M. Moore, R. T. Pappalardo, P.M. Schenk, and the Galileo SSI Team, Europan impact crater diameters and inferred transient crater dimensions and excavation depths, Lunar Planet. Sci. Conf, [CD-ROM], XXX, abstract 1882, 1999.

Van Dover, C. L., G. T. Reynolds, A.D. Chave, and J. A. Tyson, Light at deep-sea hydrothermal vents, Geophys. Res. Lett., 23, 2049-2052, 1996.

Vames, E. S. and B. M. Jakosky, Lifetime of organic molecules at the surface of Europa, Lunar Planet. Sci. Conf, [CD-ROM], XXX, abstract 1082, 1999.

Verbiscer, A. and P. Helfenstein, Reflectance spectroscopy of icy satellite surfaces, in Solar System Ices, edited by B. Schmitt et al., pp. 157- 197, Kluwer Acad., Norwell, Mass., 1998.

Verbiscer, A., and J. Veverka, Scattering properties of natural snow and frost: Comparison with icy satellite photometry, Icarus, 88, 418-428, 1990.

Verbiscer, A., P. Helfenstein, and J. Veverka, Backscattering frost on icy satellite surfaces, Nature, 347, 162-164, 1990.

Veverka, J., P. Thomas, T. V. Johnson, D. Matson, and K. Housen, The physical characteristics of satellite surfaces, in Satellites, edited by Bums, J. A. and M. S. Mathews, pp. 342-402, Univ. of Ariz. Press, Tucson, 1986.

Weissman, P. R., The cometary impactor flux at the Earth, Geol. Soc. Am. Spec. Pap., 247, 171-180, 1990.

Williams, K. K., and R. Greeley, Estimates of ice thickness in the Cona- mara Chaos region of Europa, Geophys. Res. Lett., 25, 4273-4276, 1998.

Wilson, L., and J. W. Head, Europa cryovolcanism: Ascent and eruption of magma and its role in resurfacing, Lunar Planet. Sci. Conf, [CD- ROM], XXIX, abstract 1138, 1998.

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

Wittebom, F. C., J. C. Bregman, and J. B. Pollack, Io: An intense brightening near 5 micrometers, Science, 203, 643-646, 1979.

Yoder, C. F., and W. L. Sjogren, Tides on Europa, in Europa Ocean


Coro':erence, Capistrano Conf. 5, pp. 89-90, San Juan Capistrano Res. Inst., San Juan Capistrano, Calif., 1996.

Zahnle, K., L. Dones, and H. F. Levison, Cratering rates on the Galilean satellites, Icarus, 136, 202-222, 1998.

Zent, A. P., and F. P. Fanale, Possible Martian brines: Equilibrium and kinetic considerations, Proc. Lunar Planet. Sci. Conf, 16 th, Part 2, in J. Geophys. Res., 91, suppl., D439-D445, 1986.

M. J. S. Belton, National Optical Astronomy Observatory, 950 Cherry Street, Tucson, AZ 85719.

H. H. Breneman, K. P. Klaasen, K. Magee, and D. A. Senske, Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109.

M. H. Carr, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025ø

C. R. Chapman, Southwest Research Institute, 1050 Walnut Street, Boulder, CO 80302.

G. C. Collins, J. W. Head, R. T. Pappalardo, and L. M. Prockter, Department of Geological Sciences, Box 1846, Brown University, Provi- dence, RI 02912-1846. ([email protected])

T. Denk, B. Giese, G. Neukum, and R. Wagner, DLR-Institut ftir Planetenerkundung, Rudower Chaussee 5, 12489 Berlin, Germany.

S. Fagents, R. Greeley, S. D. Kadel, J. E. Klemaszewski, and K. K. Williams, Geology Department, Arizona State University, Tempe, AZ 85287-1404.

P. E. Geissler, R. Greenberg, G. Hoppa, A. S. McEwen, C. B. Phillips, B. R. Tufts, and E. P. Turtle, Lunar and Planetary Laboratory, University of Arizona, Tucson, AZ 85721.

P. Helfenstein and R. J. Sullivan, Center for Radiophysics and Space Research, Cornell University, Ithaca, NY 14853.

J. M. Moore, NASA-Ames Research Center, Moffett Field, CA 94035.

W. B. Moore and G. Schubert, Department of Earth and Space Sciences, University of California Los Angeles, CA 90095.

(Received October 8, 1998; revised May 4, 1999; accepted May 6, 1999.)

does europa have a subsurface ocean& evaluation of the ...europa's mean density and moment...

Documents