Map: Research / Habitability / Jupiter's Galilean Moons	< Back		Next >

2.2 Habitability of Jupiter’s Galilean Moons (Schubert, Moore, Nimmo, Veradi)

2.2.1 Tidal Forces and the Implications for Oceans within the Icy Satellites

	Figure 2.2. Cross sections of Europa. (Image courtesy of NASA image exchange, artist’s conception by Pam Engebretson of Mountain View, CA, from a design by Eric M. DeJong and Zareh Gorjian of JPL, Pasadena, CA).

There is strong evidence for the existence of a subsurface liquid water ocean on Europa and evidence for the existence of oceans buried in the deep interiors of Callisto and Ganymede. A liquid water ocean, especially relatively near the surface of Europa, is a possible habitat for life. It is important therefore to understand how liquid water oceans can exist on Jupiter’s icy moons in spite of the tendency for the satellites to cool and the oceans to freeze. The answer lies in a source of heat that offsets the tendency to freeze. It is possible that this heat source is simply the radiogenic heat supplied by the rocky material in the moons (Spohn and Schubert, 2003), but especially in the case of Europa, the crucial heat production might originate in the tidal flexing of the satellite by Jupiter. We are therefore motivated to study the role of tides in establishing and maintaining oceans of liquid water beneath the surfaces of the icy Jovian satellites.

From a more general perspective, it is evident that the large satellites of the giant planets have experienced significant orbital, rotational, and thermal evolution due to tides. Tidal dissipation makes Io the most volcanically active body in the solar system, but at the same time, Io clearly demonstrates that too much tidal heating can dessicate a body and render it inhospitable to life. In order to identify the boundaries of a tidally supported habitable zone we must study the orbital, rotational, and thermal evolution of systems of bodies around giant planets. The need to study giant planet moons as a system, even if interest lies mainly in the icy bodies, is clear from the coupling of Io, Europa, and Ganymede in the Laplace resonance.

Schubert, Moore, Varadi and colleagues are developing a numerical model that will calculate the coupled orbital, rotational and thermal evolution of satellite systems of giant planets and will apply the model to the Jupiter, Io, Europa, and Ganymede system. The orbital and rotational evolution of these satellites is obtained by following the dynamical equations forward in time using both symplectic and traditional integrators. We take a primitive-variables approach that does not arbitrarily restrict the orbits or rotations. Thermal evolution models using parameterizations also lead to differential equations that may also be integrated by these methods.

The key to coupling the orbital dynamical and thermal models is development of a dynamical approach to the tidal and rotational deformation of the satellites that represents the deformation equations within each body as a system of ordinary differential equations in time (Hanyk et al.,1996). The entire coupled system can then be advanced forward in time using the same integrator. Novel heat transport parameterizations must be developed to adequately describe the magmatic and convective transport of heat in Io; similar parameterizations are necessary for the rocky mantles and icy shells of the other Galilean satellites. The result will be quantitative estimates for the times and places that sufficient heat and liquid water combine to yield potential habitats for life in the icy Galilean moons.

The model includes the tidal interactions among the bodies and the tidal dissipation within them. The model accounts for the feedback between tidal heating and heat transport in the bodies through the temperature-dependence of viscosity; this is a fundamental aspect of the dynamics of close satellites. Since tidal heating might be the only heat source capable of sustaining geologic activity on the moons for billions of years, understanding this process will be critical for evaluating their habitability. Tidal dissipation is also the mechanism driving orbital evolution and therefore it connects the interior dynamics to the observable orbital state, allowing precise astrometric measurements to constrain the thermal state of the interior of a satellite using only telescopic observations. Since we will not be able to visit any extrasolar satellite systems directly, the development of this theory will be vital to interpreting observations of such systems in terms of astrobiological potential.

For example, by incorporating the tidal models of Schubert and colleagues into ephemeris computations, these workers will be able to tell how much the orbits of the Galilean satellites have changed over recent decades, for different internal satellite structures. These predictions will be compared to past observations of the mutual events of the Galilean satellites. Our model will also predict the physical librations of the satellites, i.e., departure from rotation around a principal axis of the moments of inertia tensor. These predictions can be tested against observations of the librations, perhaps by a future orbiter (for example, the planned Jupiter Icy Moon Orbiter) to constrain the interior structures of the satellites.

From results already obtained on the physical librations of the Galilean satellites, it has become apparent that improvements in our understanding of the dynamics of synchronous rotation are necessary. Straightforward numerical simulations of rigid satellites reveal that physical librations can have very large amplitudes, comparable to those of spin axis precession. We expect that tides would damp such motions, but it is not obvious how. Nevertheless, the heating rate implied by this dissipation is potentially significant, and could explain the excess heat flow observed from Io relative to that expected from Keplerian tides alone. Europa may also be subject to this excess tidal heating, and additionally may experience differential rotation between its ice shell and rocky interior. Measuring the amplitude of such librations (from an orbiting spacecraft) would provide an excellent constraint on the thickness of the ice when compared with the predictions of the theory being developed as part of this project.

2.2.2 Estimating the thickness of Europa’s icy crust

The thickness of the icy crust of a Galilean satellite has major implications for its thermal history, habitability, and suitability for future missions. On Ganymede and Callisto, the crust is thought to be O(100 km) thick; on Europa, the crust is perhaps O(10 km) thick or smaller, but there is considerable uncertainty in this estimate (Pappalardo et al., 1999). Some inferences of crustal thickness from interpretations of the surface geology have the crust only a few kilometers thick (Geissler et al.,1998; Hoppa et al.,1999; Greenberg et al.,2000) As discussed more thoroughly in the next section, the thickness of Europa’s crust is key to the transport of nutrients from the surface to an underlying liquid water ocean. So much depends on how close Europa’s ocean is to the surface, that it is essential to explore all avenues at our disposal for estimating the thickness of Europa’s icy crust.

One approach to estimating the thickness of Europa’s crust is to construct theoretical models of its internal structure as part of a thermal history investigation. This has been done by Hussmann et al. (2002) and Spohn and Schubert (2003). These models predict ice crusts that are a few tens of kilometers thick. The coupled orbital dynamical, rotational, and thermal evolution models discussed in the previous section contain more physics and will provide improved theoretical estimates of ice crust thickness.

Another way of estimating the crustal thickness is to measure its rigidity, or effective elastic thickness. Ice retains its elastic strength only at relatively cold temperatures; since the temperature gradient within the ice crust depends on its total thickness, the measured elastic thickness can be used to infer the crustal thickness. The elastic thickness measured is the lowest since the deformation occurred.

On the Galilean satellites, stereo topography and flexural analysis can be used to derive the elastic thickness (Nimmo et al., 2002). This has been done for topographic profiles across two rifts on Ganymede, with the result that both profiles give an effective elastic thickness of about 1 km, implying a total crustal thickness of about 3 km at the time of loading (Nimmo et al., 2002). The crustal thickness at present is much larger; however, these calculations indicate that Ganymede once had an ice crust only a few km thick, probably due to an episode of tidal heating.

We will extend this approach to Europa. In a preliminary analysis of several topographic profiles across Europa we find an elastic thickness of 6 km, suggesting that the present-day crustal thickness is at least 15 km. This is larger than some estimates, but it agrees with evidence from impact crater studies (Schenk, 2002) and predictions from thermal history models (Hussmann et al., 2002; Spohn and Schubert, 2003). Possible links between the surface and the internal ocean through such thick ice will be studied in the project described in the next section.

2.2.3 Transport of nutrients through an icy crust – feeding an isolated world

The most likely abode for life on Europa, its ocean, is covered by a continuous ice shell about 10 km thick (the large uncertainty in ice thickness was discussed in the previous section). The ice shell has two effects. First, it greatly impedes the transfer of nutrients and energy (e.g., sunlight) from the near surface to the interior. Second, it makes the detection of subsurface life much more difficult. Accordingly, an extremely important question is the extent to which material from the surface can be transported to the interior, and vice versa.

Nutrients could be produced by either ⁴⁰K decay within the ice shell (Chyba and Hand, 2001) or the collision of particles accelerated by Jupiter's magnetic field with the surface (Carlson et al., 1999). Europa is located at the outer edge of Io's plasma torus. The co-rotating plasma from Jupiter's magnetosphere continually overtakes Europa and bombards it with plasma with energies ranging from a 100 eV to several MeV. The particles that have a significant effect on the surface chemistry and morphology have energies exceeding tens of keV. It is estimated that upwards of 2x10⁷ ions (mostly protons and various charge states of oxygen and sulfur atoms) and 2x10⁸ electrons impact each cm² of Europa's surface every second (Cooper et al., 2001). The energy flux contained in these charged particles is an 8x10¹⁰ keV/cm2s^-1,with the electrons dominating this flux. Whereas ions cause the most amount of sputtering of surface constituents, the electron stopping depth is the highest (exceeding 0.6 mm for typical energy electrons) and they are the leading cause of chemical alterations.

Below the depth of 1 mm, radiolysis is caused mainly by the secondary electrons and the electron bremsstrahlung photon products (Paranicas et al., 2002) and the radiation dose rate is significant in the upper 10 cm of ice. It is estimated (Cooper et al., 2000, ) that 20 mm of water ice is sputtered away every million years from the bombardment of Europa's surface. The chemical products formed from the radiolysis of water and CO₂ ice by charged particles include H, OH, H₂O₂, O₃ and many CHO compounds like CH₃OH, H₂CO, and CH₂CO etc (Delitsky and Lane, 1998). The yield factors for these chemical products are not very well known but yields of 0.01 for O₂ molecules and 0.2- 0.4 for H₂O₂ molecules for each 100 eV of deposited energy have been reported (Brown et al., 1982, Moore and Hudson, 2000). Carlson et al. (1999) using infrared spectroscopy from the Galileo spacecraft estimate that hydrogen peroxide (H2O2) abundance is 0.13% relative to water ice on the surface of Europa.

Whether these nutrients could sustain an oceanic biosphere depends on the rate at which they are produced, and the rate at which they are transported to the ocean from the ice shell (Gaidos et al., 1999). Because, photosynthesis is severely inhibited by the thick ice cover and the primary energy from geothermal and chemical weathering processes would be quite limited, Gaidos et al. (1999) argue that most metabolic pathways that power the life cycle on Earth would be denied to organisms on Europa. Therefore, many authors (Chyba, 2000; Chyba and Phillips, 2001; Cooper et al., 2001; Chyba and Hand, 2001) have considered the potential of the radiolysis produced oxidants to power life in an oceanic ecosystem. The rate at which the oxidants produced at the surface are transferred to the liquid ocean depends on the primary yield from radiolysis, erosion of surface from sputtering and impact gardening, the ultraviolet processing of the surface and the oxidants and the overturn rates of the surface material by endogenic geological processes. The radiation not only creates oxidants, it also destroys them and through sputtering can eject a large fraction of the radiolysis products from the surface altogether.

Chyba and Phillips (2001) have analyzed the competition between particle sputtering and impact gardening in creating, destroying and preserving (through regolith burial) oxidants on the surface of Europa and suggest that if the regolith is well mixed and communicating with the deeper layers, as much as 2.5x10²⁵ molecules of H₂O₂ would be produced per square centimeter of the surface over a time period of 10 Myr. Depending on the models of ice thickness and subsurface geology a wide variety of scenarios can be supposed for the delivery of these oxidants to the liquid ocean. Using fairly conservative estimates of oxidant creation and transfer to the ocean, Chyba and Phillips (2001) suggest that the ocean would be able to support ~ 10²³– 10²⁴ prokaryotic-analog cells in the oceanic biomass. However, if the upper layers of ice could be constantly replenished with material exchange from the interior so that a maximum transfer of the oxidants occurs to the subsurface ocean, the biomass estimate would increase by a factor of 10³ and the level of oxygen in the Europan ocean could be comparable to that in the Earth's ocean (Cooper et al., 2001).

The solid ice shell of Europa consists of two parts, a near-surface cold region in which deformation occurs by brittle processes, and a deeper warmer region in which ductile deformation predominates. The latter region may experience solid-state convection (see below) and is likely to be the area in which heat is generated by tidal deformation of the ice shell.
We will investigate possible transport mechanisms in both these regions, and focus in particular on the likelihood of melting the ice. Melting is attractive because it forms an efficient transport mechanism, and there is observational evidence for it, such as the surface chaos regions.
Melting within the brittle near-surface layer is difficult to achieve, but one potential mechanism is shear heating on strike-slip faults (Nimmo and Gaidos, 2001).

This model suffers some drawbacks, notably its inability to calculate the thickness of the brittle layer self-consistently. We intend to remedy this deficiency in order to better establish the conditions under which near-surface melting can occur. A further consequence of the shear-heating model is that it may produce linear diapirs, which can potentially advect material towards the surface. It is well known that Europan ridges are darker than the surrounding material (Fagents et al., 2000); we will investigate whether this observation is consistent with diapiric activity.

Transport of material within the convective zone of the Europan ice crust is estimated to occur on timescales of about 10³ yr to about 1 Myr. If nutrients can be transported through the brittle lid, such timescales are probably sufficient to sustain a modest oceanic biosphere (Chyba and Phillips, 2001).
Previous convection models of Europa have generally assumed Newtonian behavior (e.g., Pappalardo et al., 1998), but the ice may actually be behaving in the non-Newtonian regime. An important consequence of non-Newtonian behavior is episodicity, which means that previously calculated transport timescales could be incorrect. Furthermore, convection might lead to discrete diapirs (e.g., Nimmo and Manga, 2002) rather than the steady currents previously envisaged (e.g., Pappalardo et al., 1998). We will carry out a suite of convective models incorporating non-Newtonian behavior and tidal heating, and examine the consequences for material transport and melt generation. While we intend to focus on Europa initially, the equations are general so that we can also investigate convective processes on Ganymede, Callisto and the Saturnian icy satellites.

A key feature of the convection models will be the incorporation of a realistic composite ice rheology (Goldsby and Kohlstedt, 2001), including the different creep deformation mechanisms (dislocation, superplastic, and diffusional) that come into play at different stresses and grain sizes. The former two of these are non-Newtonian, which can lead to deformation that is episodic and localized (e.g., Larsen and Yuen, 1997), as mentioned above. The models will also include a simple parameterization of brittle failure of the lid using a pseudoplastic yield stress , which can also lead to highly episodic behavior, as well as mobile-lid-like features (Tackley, 2000). This modeling can be performed using an existing code, Stag3D, which has previously been used for modeling silicate convection in 2-D and 3-D and already includes the necessary rheological capabilities. Failure of the brittle layer would be especially interesting in this context as it would permit advective transport of materials from/to the surface.

The convection models will be coupled to a tidal dissipation calculation, in order to self-consistently treat the feedback between viscosity variations caused by convection, and tidal dissipation (Sotin et al., 2002). Tidal flexing is a long-wavelength phenomenon so it couples mostly to long-wavelength viscosity variations, and a global treatment is necessary to calculate this correctly. The main interest in such heating is that it is another possible melt-generating (and hence nutrient-transporting) mechanism.

< Previous Section	Next Section >