Galileo Magnetometer Measurements: A Stronger Case for a Subsurface Ocean at Europa

See allHide authors and affiliations

Science  25 Aug 2000:
Vol. 289, Issue 5483, pp. 1340-1343
DOI: 10.1126/science.289.5483.1340


On 3 January 2000, the Galileo spacecraft passed close to Europa when it was located far south of Jupiter's magnetic equator in a region where the radial component of the magnetospheric magnetic field points inward toward Jupiter. This pass with a previously unexamined orientation of the external forcing field distinguished between an induced and a permanent magnetic dipole moment model of Europa's internal field. The Galileo magnetometer measured changes in the magnetic field predicted if a current-carrying outer shell, such as a planet-scale liquid ocean, is present beneath the icy surface. The evidence that Europa's field varies temporally strengthens the argument that a liquid ocean exists beneath the present-day surface.

Europa, one of the icy moons of Jupiter, may have a layer of liquid water beneath its surface. Features of Europa's tortured surface revealed by Galileo's imaging system may have formed as its surface ice stretched, broke, and rearranged itself while floating on a watery subsurface sea (1–4). But even if water shaped the surface we see today, it may have frozen hundreds of thousands of years ago. Those searching for possible abodes of life elsewhere in the solar system would like to know whether water exists beneath the surface at the present epoch.

In the initial report (5), the magnetic perturbations measured on Galileo's first pass by Europa were characterized as the signature of an internal dipole moment. The Galileo magnetometer team subsequently reported (6, 7) that many features of their data on close passes by Europa in 1996 and 1998 can be modeled as the electromagnetic response to Jupiter's time-varying magnetospheric magnetic field if a layer of electrically conducting material exists near Europa's surface. Although the dominant, southward-oriented magnetic field imposed by Jupiter's magnetosphere at Europa's position is about constant, the projection of the magnetospheric field into Europa's equatorial plane varies with the synodic period (Jupiter's rotation period, 11.2 hours as seen from Europa) and has a mean value close to zero. Such a time-varying magnetic field, referred to as the primary field, can drive inductive currents through an electrical conductor. Inductive currents, in turn, generate a secondary field, whose source can be represented as a time-varying magnetic dipole moment lying in Europa's equatorial plane with an orientation approximately antiparallel to the instantaneous orientation of the primary field (8, 9). If one models Europa as an idealized, highly conducting sphere (conductivity ≫ 1 S m−1) of radius 1 RE (radius of Europa = 1560 km) and if the primary field (Bx (t),By (t), Bz ) is uniform over the scale of Europa, then the components (10) of the induced magnetic moment (M) are −1/2(Bx (t),By (t),0) in magnetic field units, implying that at the surface of Europa, the radial components of the induced field and the time-varying primary field cancel. Table 1 gives values of this idealized induced dipole moment at the time of closest approach (labeled “Ind”) and other key parameters of all Europa passes for which the magnetometer acquired data (E4, E11, E12, E14, E15, E17, E19, and E26). We shall focus hereafter only on passes that came within 2000 km of Europa's surface (E4, E12, E14, E19, and E26) for which the signatures of internal sources can have amplitudes large enough to be clearly detected (11).

Table 1

Idealized induced dipole moment at closest approach and key parameters of Europa passes. Fitted int, fitted interval; Mag lat and SIII long, System III latitude and longitude; CA, closest approach above Europa's surface; Europa lat and E-long, Europacentric latitude and longitude measured eastward; Bf, best-fit (or measured) dipole model; Ind, induced dipole model.

View this table:

The idealized induction model is consistent with the orientation and the magnitude of the equatorial dipole moment observed on the close Europa passes on Galileo's fourth and fourteenth orbits around Jupiter (referred to as E4 and E14; data plotted in Figs. 1 and 2).

Although measurements from the E4 and E14 passes were consistent with perturbations arising from inductive currents, one could not exclude the possibility that a permanent dipole moment tilted toward they axis was the source of the observed magnetic perturbations. Both passes occurred in Jupiter's northern magnetic hemisphere at times when the magnetospheric field at Europa pointed away from Jupiter and therefore the induced dipole moment pointed toward Jupiter. However, an induced equatorial dipole moment changes orientation and amplitude over a synodic period in a predictable manner. Consequently, a flyby in the southern magnetospheric hemisphere in the opposite phase of the driving signal would test the validity of the model by determining whether Europa's inferred magnetic moment is or is not variable. The inset in Fig. 3shows the variation of the x and y components of Jupiter's magnetospheric field at Europa over a synodic period (12). Also marked are the values of these components at closest approach to Europa for the passes at altitudes <2000 km. The E26 encounter (3 January 2000) (see data plotted in Fig. 4) was designed to occur when Europa was in the southern magnetospheric hemisphere and the driving field was almost 180° out of phase with its value on the E4 and E14 passes. In this situation, Europa's equatorial dipole moment would point away from Jupiter if it were induced but would remain pointing toward Jupiter if it were quasi-static.

The best-fit dipole moment (13) and the idealized induced dipole moment track each other relatively closely and represent the changes on the scale of Europa's diameter. (Table 1 provides the components of these dipole moments for all passes.) In contrast, the signature that would have been observed on E26 if the dipole moment had been that fitted to the E4 data (moment pointing toward Jupiter) is at variance with the measurements. This reveals unambiguously that the orientation of the equatorial dipole moment changed.

Although the modeled signature of an inductive dipole moment tracks the data (Figs. 1, 2, and 4) fairly well, there are discrepancies. By understanding how plasma currents contribute to the field perturbations, one can identify the components that reveal most about the internal sources of the magnetic perturbations. The effect of currents coupling Europa's environment to Jupiter's ionosphere, referred to as Alfvén wing currents, are well understood (14, 15). These currents roughly align with the background field north and south of Europa, and they close through Europa or its ionosphere and through currents produced by ionization of neutral species in the surrounding plasma. The Alfvén wing currents bend the background field in the direction of the flow and the closure currents affect the field magnitude near Europa, increasing it upstream and decreasing it downstream. Bending produces a negative change in Bx north of Europa and a positive change south of Europa. The field may also bulge around Europa producing perturbations in By , which are generally smaller than those in Bx . Away from Europa's equator (Table 1), the Alfvén wing perturbation may contaminate the signature of internal sources in the xand y components of the field. It may also produce fluctuations of the field components comparable in magnitude to the perturbations estimated for an idealized induced response over distances shorter than 1 RE [e.g., E26, Fig. 4; E12, Web fig. 1 (16); and E19, Web fig. 2 (16)]. However, on near equatorial passes (E4 and E14), Alfvén wing currents contribute little toBx or By but do affectBz . Fortunately, near the equator, an induced dipole moment (which lies in the x-y plane) contributes only to the Bx and By components. Thus, the close correspondence between the model and the data for these components on the low-latitude passes is meaningful. Away from the equator (passes E19 and E26), all components (but particularly Bx ) may be affected by Alfvén wing currents. For example, Bx is modified by the negative (positive) contribution of the Alfvén wing bendback on the E19 pass north of the equator (E26 pass south of the equator). This accounts for part of the discrepancy between the data and the induced field model [Fig. 4 and Web fig. 2 (16)]. Pass E12 occurred close to the center of the jovian current sheet, where exceptionally strong pickup produced unexpectedly large perturbations over many RE (the field near closest approach was double the background field in magnitude). The large error of fit to a dipole moment (Fig. 3) reveals that currents external to Europa dominate these measurements.

Figure 1

Magnetometer data in the EphiO (10) coordinate system (three components and the field magnitude in nT) from pass E4 on 19 December 1996 plotted against spacecraft event time in UT for a range of ∼4 RE from Europa's center. The solid curves are data points. The slowly changing thin curves, obtained from a polynomial fit to the field components measured before and after the encounter, represent the background field without Europa perturbations. The dotted curves are predictions from the induced dipole model without plasma effects. The heavy solid curves are the least squares fits to a tilted dipole moment, referred to in the text as the measured magnetic moment.

Figure 2

The E14 pass on 29 March 1998 showing data within 5 RE of Europa's center. Curves are as in Fig. 1 with addition of the dashed curves, which represent the values along this orbit from the fitted E4 dipole moment.

Figure 3

The radial (positive toward Jupiter) component of the equatorial dipole moment fitted to the observed data plotted against predictions for a 1 RE highly conducting sphere (the induced model). The heavy line represents the model. The narrow line is a least squares fit to the data. Error bars give the error of fit for the best-fit dipole. Estimates of the contributions of plasma currents have not been removed. The critical test occurred on E26. If the dipole moment fitted to E4 and E14 had been permanent, the point would have fallen at the approximate position of the triangle in the upper left-hand quadrant. (Inset) Variation over Europa's synodic period of theBx and By components of Jupiter's magnetospheric field at Europa's location. Dots show the values observed at closest approach to Europa on the relatively close encounters. Time increases clockwise.

Figure 4

The E26 pass on 3 January 2000: data within 5 RE from Europa's center. Curves are as in Fig. 2. The dashed curves computed from the dipole moment fit to the E4 pass are in antiphase to both the dipole fitted to the E26 data (heavy solid curves) and to the inductive response model (dotted curves).

Because My is the largest component of the induced (time-varying) magnetic moment and the one least modified by plasma effects for the low latitudes passes, we focus on its measured values. The measured My and theMy from the induced field model vary in phase with each other if the dipole moments are produced by induction. Figure 3 shows that the modeled and measured moments track one another within experimental uncertainty. The critical aspect of Fig. 3 is that if the dipole moment were permanent, the measured My on E26 would have been the same as it was on E4 and E14 within the uncertainty of the data fits; it would have appeared in the upper left quadrant instead of the lower left quadrant, as predicted by the inductive model.

Europa's response to the changing magnetic field at its orbit (inset to Fig. 3) has been modeled as that of a perfectly conducting sphere responding to a uniform time-varying primary field. Alfvén wing perturbations are also time varying but their nonuniform structure contributes only to internal spherical harmonics of higher than dipole order (17). Nonetheless, there is harmonic aliasing in any fit to data acquired along a spacecraft trajectory rather than everywhere on a closed surface. Alfvén wing currents do, therefore, contribute to the best-fit dipole moment.

A spherical shell conductor with 1 RE outer radius would produce nearly the same signature as a sphere of the same size for sufficiently high conductivity (7). For example (18, 19), the amplitude of the response of a spherical shell of ∼7.5-km depth below Europa's surface and a conductivity of 2.75 S m−1 for Earth's oceans (20) would be 90% that of a highly conducting sphere with a phase lag relative to the primary field of 25°. The amplitudes of the induced magnetic moment for E4, E14, and E26 would be reduced by the phase lag (21). If the ocean were thicker, but still close to the surface, or if the conductivity were higher, the phase lag would decrease, and the amplitude of the response would increase.

On passes E4, E12, E14, E19, and E26, currents induced by the time-varying primary field in a Europan ionosphere may contribute to the magnetic perturbation. Elsewhere (18), we have calculated the signature of induced currents flowing above the surface, close to Galileo's trajectory, in a conducting shell of larger than 1 RE radius. Based on the measured ionospheric density and scale height (22) and measured (23) or modeled (24) atmospheric properties, we find that ionospheric currents fail to account for the observations by an order of magnitude.

The case for a subsurface electrical conductor on a planetary-wide scale passed the test of the E26 flyby with flying colors. Although the electrically conducting layer need not be salty water, water is the most probable medium on Europa. Geological evidence has been interpreted as consistent with surface effects of subsurface liquid water, but the defining features could have been formed in the distant past. The magnetometer result makes it likely that liquid water persists in the present epoch.

  • * To whom correspondence should be addressed. E-mail: mkivelson{at}


View Abstract

Navigate This Article