Magnetic Signatures of the Interaction Between Europa and Jupiter's Magnetosphere During the Juno Flyby

Based on a hybrid model of Europa's magnetospheric interaction, we provide context for the magnetic field perturbations observed by the Juno spacecraft during its only close flyby of the moon in September 2022. By systematically varying the incident flow conditions and the density profile of Europa's atmosphere, we demonstrate that the observed, large‐scale signatures of magnetic field draping are consistent with a dawn‐dusk asymmetry in the moon's neutral envelope. During the flyby, such an asymmetry would have enhanced the magnetic perturbations in Europa's anti‐Jovian hemisphere, explaining why the spacecraft already detected strong field line draping while still several moon radii away. Conversely, a reduced neutral density in the sub‐Jovian hemisphere can explain why the perturbations in the flow‐aligned field component remained nearly constant as Juno approached Europa. While a dawn‐dusk asymmetry in Europa's atmosphere has been predicted by theoretical work, our results provide the first in situ hints of its presence.

10.1029/2023GL106810 2 of 10 position (defined by the Jovian Local Time) and its distance to the center of Jupiter's magnetospheric plasma sheet (defined by the System III longitude).Therefore, any such plasma interaction model inherently needs to make assumptions on potential asymmetries present in Europa's neutral gas envelope.For instance, the magnetohydrodynamic (MHD) model of Blöcker et al. (2016) treated the moon's global O 2 atmosphere as spherically symmetric, and this symmetry is broken only by localized plumes of water vapor at various positions across the surface.The MHD models of Rubin et al. (2015), Jia et al. (2018), and Harris et al. (2021Harris et al. ( , 2022) ) as well as studies with the AIKEF hybrid model (e.g., Addison et al., 2021;Arnold et al., 2019) assumed a ram-wake asymmetry to be present in Europa's O 2 envelope, that is, the neutral density in these models peaks at the moon's ramside apex and is reduced in the wakeside hemisphere.This assumption is based on the notion that surface sputtering by energetic ions (a major contributor to Europa's atmosphere) is largely concentrated around the moon's ramside apex (Cassidy et al., 2013).The presence of enhanced ion sputtering rates near the ramside apex was recently put into question by Addison et al. (2021Addison et al. ( , 2022) ) who calculated the trajectories of the incident energetic ions in a realistically draped electromagnetic environment.In their recent MHD modeling study, Cervantes and Saur (2022) again treated Europa's O 2 atmosphere as spherically symmetric.These authors also included a "bulge" of enhanced H 2 O density around the moon's dayside (or ramside) apex, consistent with observations by the Hubble Space Telescope (Roth, 2021).Oza et al. (2019) developed a Monte Carlo model of O 2 dynamics in Europa's atmosphere, taking into account the moon's motion around Jupiter along its 85-hr orbit and the associated, tidally locked rotation around its polar axis.Under the assumption that O 2 is mainly generated near Europa's dayside apex, Oza et al. (2019) demonstrated the presence of a dawn-dusk asymmetry in the neutral gas density: an atmospheric "bulge" is formed in Europa's dusk hemisphere, while a corresponding reduction in column density was identified in the dawn hemisphere.Depending on surface location, the model of Oza et al. (2019) suggests O 2 column densities in the range of  (7.4 − 8.1) ⋅ 10 17 m −2 .The dawn-dusk asymmetry of Europa's atmosphere proposed by these authors has not yet been included in any model of the moon's plasma interaction.
On 29 September 2022, the Juno spacecraft carried out its only close flyby of Europa, reaching a closest approach altitude of 354.5 km (0.23R E ) at 09:36:29 UTC. Figure 1 displays the trajectory of this flyby in the Cartesian EPhiO coordinate system.The origin of this system coincides with the center of Europa.Its (+X) axis is aligned with the direction of corotation, and the (+Y) axis points toward Jupiter.The (+Z) axis completes the right-handed system, pointing northward.Juno approached Europa from its anti-Jovian side while traveling northward and toward upstream.However, the spacecraft remained downstream of the moon during most of the flyby.The encounter occurred at a Jovian Local Time of 18:40, that is, the direction of the incident magnetospheric plasma was nearly aligned with the Sun-Europa line (see Figure 1b in Liuzzo et al. (2015)).In other words, Europa's dusk terminator was located in the anti-Jovian half space (Y < 0), approximately 90° in longitude eastward of the subsolar point.For this reason, Europa's location during the Juno flyby facilitates the search for any dawn-dusk asymmetries in the plasma interaction region which, in turn, may provide hints of similar hemispheric dichotomies in the moon's neutral envelope.
In this letter, we analyze the time series recorded by Juno's magnetometer (Connerney et al., 2017) during the Europa flyby.By applying the AIKEF hybrid model (Müller et al., 2011), we demonstrate that a dawn-dusk asymmetry in the moon's atmosphere provides a possible explanation for the magnetic perturbations observed along the Juno trajectory.

Application of the AIKEF Hybrid Model to Europa
The AIKEF model treats ions as individual macroparticles, whereas electrons form a massless, charge-neutralizing fluid.The model has an extensive history of applications to Europa's plasma interaction (Addison et al., 2021(Addison et al., , 2022;;Addison, Liuzzo, & Simon, 2023;Arnold et al., 2019;Arnold, Liuzzo, & Simon, 2020;Arnold, Simon, & Liuzzo, 2020;Breer et al., 2019;Haynes et al., 2023).To study Europa's environment during the Juno flyby, we carried out over 60 model runs, systematically varying the incident magnetospheric flow conditions and the structure of the moon's atmosphere.This letter discusses results from six of these setups (see Table 1) which were found to be most instructive for the interpretation of Juno magnetometer observations.These six runs assume the ion population of the impinging thermal plasma to consist of a singly charged species with mass m 0 = 18.5 amu (analogous to, e.g., Addison et al., 2021Addison et al., , 2022) ) and number density n 0 = 100 cm −3 .This value is (approximately) the average between the densities obtained from the empirical models of Bagenal and Delamere (2011) are computed for magnetospheric ions (index s = i) and electrons (s = e) separately, using temperatures T s from Kivelson et al. (2004).The magnitude of the Alfvén velocity   0 =  0 √  0  0  0 as well as the Alfvénic (M A ), sonic (M S ), and magnetosonic (M MS ) Mach numbers are also given.The bulk velocity of the impinging flow is set to = 100 km∕s (setups #3-#6), which is within the range of velocity magnitudes deduced from Galileo observations (Bagenal & Dols, 2020).In setups #1, #4, and #5, the upstream flow travels along the (+X) direction, while setups #2, #3, and #6 take into account the small radial flow component observed near Europa's orbital distance (Bagenal et al., 2016): in these two configurations, the flow vector   0 = 0(cos  sin  0) is inclined toward Jupiter by ϕ = 15°.This tilt may be caused by radial plasma transport in the Jovian magnetosphere and the angle is within the range covered by the error bars in Figure 5 of Bagenal et al. (2016).The uniform background magnetic field   0 is the same in all six setups and was obtained by interpolating the magnetospheric field observed before entering and after exiting Europa's interaction region (see Section 3) to the point of Juno's closest approach.Taking into account Equation 2from Addison et al. (2021), the vector   0 also determines the moon's induced magnetic moment   ind at the time of the flyby.We use a cuboid-shaped domain with extensions of −8R A critical input parameter for our model is the shape of Europa's atmosphere, which is partially ionized by electron impacts to form the moon's ionosphere (see, e.g., Arnold et al. (2019) for details).Unlike preceding studies with AIKEF (e.g., Addison et al., 2021Addison et al., , 2022)), setups #1-#3 and #6 no longer include an enhancement in the neutral density near Europa's ramside apex and an associated decrease in density around the wakeside apex.Instead, we describe the neutral profile n n with the expression where  = √  2 +  2 +  2 is the distance to the moon's center and H is the atmospheric scale height.The parameter ψ denotes the angle between the vector  = (   ) and a fixed radial unit vector   , pointing from the center of Europa toward a point on the moon's surface.The profile defined by Equation 1 is axially symmetric around the direction of   , producing a bulge in number density at ψ = 0° and a minimum for ψ ≥ 170°.The exponent λ in Equation 1 is used to define the "steepness" of the decrease in atmospheric density with growing angular distance ψ from the bulge.A slightly modified form of Equation 1 was used by Liuzzo et al. (2015) to emulate asymmetries in Callisto's neutral envelope.
In our setup #6, the location of the atmospheric bulge is defined by  = (0, −1, 0) , that is, it coincides with the moon's anti-Jovian apex.The Juno flyby occurred around 18:40 Jovian Local Time; that is, Europa's duskside apex was displaced by about 10° in longitude away from the X = 0 plane and toward the solar apex.However, Figure 4 of Oza et al. (2019) shows that their atmospheric density maximum does not occur precisely along the semi-meridian at dusk, but is slightly displaced toward the nightside by 10° − 15° in longitude.These two longitudinal displacements (one toward the dayside, the other toward the nightside apex) approximately compensate each other.For our study of the Juno flyby, we therefore center the atmospheric bulge around Europa's anti-Jovian apex.Addison et al. (2021) demonstrated that a non-zero B 0,X component of the magnetospheric background field breaks the symmetry of energetic ion precipitation onto the moon's surface between its northern and southern hemispheres.The "Case (1)" scenario studied by these authors includes a negative B 0,X component, and it was revealed that this tilt of   0 toward upstream leads to slightly elevated influx of magnetospheric ions onto Europa's southern hemisphere.Conversely, the positive B 0,X component observed during the Juno flyby (see Table 1) corresponds to enhanced magnetospheric ion precipitation onto Europa's northern hemisphere.For this reason, we also explored several configurations with the atmospheric density bulge located along the moon's anti-Jovian semi-meridian, but slightly displaced toward the north: in our setups #1-#3, the location of the atmospheric density peak is defined by  = (0, − cos 20 • , + sin 20 • ) , that is, it is rotated northward by 20° within the X = 0 plane.We emphasize that Addison et al. (2021) did not provide a conversion of their modeled ion influx patterns into actual atmospheric density profiles.In this sense, scenarios #1-#3 assume that the north-south asymmetry seen by Addison et al. (2021) in ion precipitation patterns likewise maps into the resulting atmospheric density profile.However, this assumption is currently heuristic in nature and does not follow from rigorous quantitative modeling of Europa's neutral envelope.For all setups discussed here, the values of the atmospheric scale height and the "steepness parameter" have been set to H = 100 km (see, e.g., Addison et al., 2021Addison et al., , 2022;;Haynes et al., 2023) and λ = 5, respectively.Neither parameter was found to have significant influence on the conclusions drawn.
In addition, we have carried out simulations to demonstrate that the atmospheric profiles used in earlier modeling studies are not suitable to explain the magnetic signatures observed by Juno.Setup #4 includes a ram-wake asymmetry in the neutral density, using Equation 1 from Arnold et al. (2019) and adopting their value of A = 10 for the "asymmetry parameter": the density at Europa's ramside apex then exceeds that at the wakeside apex by a factor of A + 1 = 11.In setup #5, Europa's neutral envelope is treated as spherically symmetric (analogous to, e.g., Blöcker et al., 2016), that is, A is set to zero.In all model setups, Europa's atmosphere is assumed to consist of O 2 only, that is, we include neither transient plumes of water vapor (Arnold et al., 2019;Jia et al., 2018) nor a bulge of H 2 O molecules around the subsolar or dayside apex (Roth, 2021).Each of these additional atmospheric components was found to generate only very localized perturbations to the magnetic field on length scales much smaller than the extension of the signatures observed by Juno (e.g., Cervantes & Saur, 2022;Haynes et al., 2023).The surface density n 0 in all six setups is chosen such that the column density n 0 H is of the same order (≈10 18 m −2 ) as derived from remote observations (see, e.g., Addison et al., 2021 for details) and proposed by Oza et al. (2019).

Juno Magnetometer Observations and Model Results
In Figure 2 we display Juno magnetic field observations from the Europa flyby as well as the synthetic time series from the six AIKEF runs.We also show (in plot (b)) the magnetic signature obtained by taking the sum of   0 and the induced field from Europa's subsurface ocean.This setup assumes the moon's environment to be devoid of any plasma currents.Juno's closest approach to Europa approximately coincided with the spacecraft's passage through the Z = 0 plane as it traveled from southern to northern latitudes.The geometry of Europa's plasma interaction during this flyby is somewhat tricky to capture, since   0 possesses three non-vanishing components.This implies that, for example, none of the three planes of the EPhiO system represents a symmetry plane between the moon's northern (−) and southern (+) Alfvén wing characteristics   ± =  0 ±  0 (Neubauer, 1980).Besides, while Europa's induced dipole moment is still contained within the Z = 0 plane, it is inclined by 33° against the Y axis.As shown in Figure 2, the observed B X component exhibits a broad enhancement, commencing in Europa's southern hemisphere around 09:34 UTC and reaching up to Z = +0.6RE into the northern hemisphere (until 09:39 UTC).The occurrence of positive perturbations δB X (where δB i = B i − B 0,i for i = X, Y, Z) on both sides of Europa's equatorial plane is not unexpected: due to the tilt of   0 toward downstream (B 0,X > 0, see Table 1), the Alfvén characteristics are rotated counter-clockwise around the (+Y) axis, allowing the southern wing tube (where δB X > 0) to penetrate into the northern half space (e.g., Simon and Motschmann (2009)).Likewise, the northern wing is rotated away from Juno's trajectory, causing the spacecraft to entirely miss the region of δB X < 0 associated with   − .Europa's induced dipole alone also generates perturbations δB X > 0 above the Z = 0 plane (red in Figure 2b).In setups that include the atmospheric bulge in Europa's anti-Jovian hemisphere (#1-#3 and #6), this B X enhancement is further amplified by the plasma interaction.
While the B X perturbations seen by Juno (black) remained positive throughout the encounter, the strength δB X of this enhancement changed non-monotonically with time, achieving a broad local minimum in Europa's southern hemisphere around 09:35 UTC.The subsequent B X enhancement is "interrupted" by several smaller dips, but these are way less prominent than the feature observed around 09:35.The large-scale shape of the modeled B X in setups #1-#3 ("enhancement-dip-enhancement") is similar to the observed signature.However, the locations of the modeled, large-scale B X features do not precisely coincide with observations.In these three runs, the central dip is displaced slightly toward upstream: around 09:35:30, AIKEF output still displays the dip while observations show that Juno had already entered the outbound enhancement.These deviations may stem from uncertainties in the atmosphere parameters.The smaller dips imposed on the outbound enhancement are likely caused by fine structures in Europa's ionosphere which the model is not designed to resolve.In setup #1 (which uses ϕ = 0°), the onset of the modeled B X increase occurs about 90 s earlier than observed, corresponding to a distance of 1.3R E traveled along Juno's trajectory.However, setups #2 and #3 almost precisely match the observed location and width of the B X enhancement.The slight tilt of   0 toward Jupiter in the latter two setups rotates the Alfvén characteristics into the Jupiter-facing half space.Juno approached Europa from the anti-Jovian side; this implies that in setups #2 and #3 the spacecraft would enter the region of draped field lines later than in setup #1.
Setup #6 (see Figure 2b) includes less "fine-tuning" than configurations #2-#3: the atmospheric bulge coincides precisely with Europa's anti-Jovian apex.Nevertheless, the model still matches both the locations and the magnitudes of the two spikes associated with the asymmetric B X signature.However, setup #6 overestimates the reduction near the inbound edge of the B X feature, with the modeled δB X clearly turning negative.This run still emphasizes the "robustness" of our conclusions, illustrating that the notion of a dawn-dusk asymmetry in Europa's atmosphere during the Juno flyby can qualitatively explain the shape of the observed B X perturbations over a broad range of upstream and atmospheric parameters.
Setups #4 and #5 do not include an atmospheric bulge in the anti-Jovian hemisphere.The magnetic signatures obtained from these runs reveal significant qualitative differences to Juno observations.First, the onset of the inbound B X enhancement occurs earlier than detected.Analogous to our discussion of setups #1-#3, this displacement may again be reduced by including a small component of   0 along (+Y).Second, and more importantly, runs #4 and #5 both reveal a region of negative B X in the outbound segment of the flyby, corresponding to a passage through the center of Europa's northern Alfvén wing tube in the Z > 0 half space.Additional simulations suggested that this B X < 0 feature cannot be eliminated by tilting the upstream flow vector within the observational bounds from Bagenal et al. (2016).The only way we found to prevent Juno's trajectory from intersecting the B X < 0 region is a localized reduction of the atmospheric column density near Europa's sub-Jovian apex by several orders of magnitude, as realized by Equation 1.Using such a low column density (≈2.5 ⋅ 10 13 m −2 ) uniformly across Europa's entire neutral envelope would be inconsistent with magnetometer observations: in this case, the modeled B X would be very similar to the "superposition case" (red) that does not include plasma effects.Hence, the observed B X perturbations support the notion of Europa's atmosphere at the time of the flyby having been more dense in the anti-Jovian than in the sub-Jovian hemisphere.Such an asymmetric configuration can explain why Juno encountered a strong increase in B X while traveling through the anti-Jovian half space and still several R E downstream of the moon: the locally enhanced atmospheric density causes the draping to be stronger for Y < 0. Besides, this atmosphere model can explain why the B X enhancement observed close to Europa for Y > 0 is comparable in strength to the inbound feature seen farther downstream: the reduced atmospheric density in the Y > 0 half space locally weakens the plasma interaction; that is, despite the proximity to Europa the draping signature is not amplified.
The shape of the δB Y < 0 perturbation observed shortly before closest approach can already be matched by the superposition of   0 and Europa's induced field (Figure 2b).The four setups including a dawn-dusk asymmetry in the atmosphere (#1-#3 and #6) largely reproduce the width and magnitude of the observed B Y signature (Figure 2).These runs predict a broad enhancement δB Y > 0 in the outbound segment, centered around the isolated B Y > 0 spike seen by Juno at 09:37 UTC.The magnitude of the modeled feature matches the peak value of the observed signature (B Y ≈ −57 nT), while its width is overestimated.Therefore, it remains elusive whether this spike is related to Europa's plasma interaction or is magnetospheric in origin.The time series from runs #4 and #5 overestimate the strength and, especially, the width of the observed perturbations δB Y .Particularly in the outbound region, an atmosphere with high column densities generates δB Y > 0 signatures that are still discernible around 09:40 UTC, while the observed B Y already returns to the Jovian background field at 09:37 UTC.This again suggests that, during the Juno flyby, the "strength" of Europa's plasma interaction (e.g., Simon et al., 2021) in the sub-Jovian hemisphere may have been locally weakened.
The observed B Z component reveals a broad depletion feature (i.e., δB Z > 0) around closest approach, commencing at 09:34 UTC.Such a depression region has already been identified in earlier studies of Europa's interaction (e.g., Arnold, Liuzzo, & Simon, 2020): the elevated plasma pressure associated with pick-up ions partially pushes the magnetospheric field lines out of the tail region.Analogous to B Y , the width and magnitude of the B Z feature can be reproduced only by runs that include a dawn-dusk asymmetry of Europa's neutral envelope.None of our modeled time series reproduce the double-spike signature observed in B Z shortly before closest approach.This may indicate that the representation of the ionosphere in AIKEF does not include sufficient complexity to reproduce such fine structures.
To provide some context for the qualitative differences in the modeled B X signatures, Figure 3 displays two-dimensional profiles of this component in the three planes of the EPhiO system.The left column illustrates results from setup #2 (which does not produce a δB X < 0 segment after Juno's closest approach), while the right column shows output from run #4 (which suggests δB X < 0 in the outbound region).As can be seen from the top and middle rows, Europa's Alfvén wings are rotated toward Jupiter around the Z axis and northward around the Y axis.The bottom row in Figure 3 reveals the reason for the morphological differences in B X along the Juno trajectory.In run #2, the core of the northern Alfvén wing (δB X < 0, blue) is shifted slightly toward the Y < 0 half space where the atmospheric density is larger.This happens despite the inclusion of a small upstream flow component toward Jupiter.The "ray" of δB X > 0 in the northern, Jupiter-facing half space (red in panel 3(e)) corresponds to the outer region of the   − wing: in planes perpendicular to   ± , the magnetic perturbations associated with an Alfvén wing can be represented by a two-dimensional dipole (Neubauer, 1980); that is, the field lines need to close outside of the core region.Therefore, the field lines inside the northern wing tube are draped (δB X < 0), but in the periphery of the wing they appear "anti-draped" (δB X > 0).
In setup #2, the "thinning" of the northern wing tube for Y > 0 caused Juno to first travel through the center of the southern wing tube (where the field is draped, δB X > 0), and then graze the outer regions of the northern wing where the field is "anti-draped."For this reason, the observed δB X in the north has the same sign as in the south.In setup #4, the northern wing tube (characterized by δB X < 0, blue) extends slightly farther toward Jupiter than in setup #2 (panels 3(f) versus (e)).Therefore, in run #4 Juno would have encountered the center of the northern wing as it traveled through the Z > 0 half space.However, this picture is incompatible with observations.

Concluding Remarks
Results from the AIKEF model suggest that reproducing the large-scale shape and magnitude of the magnetic perturbations observed during Juno's Europa flyby requires the inclusion of a dawn-dusk asymmetry in the moon's neutral envelope.Within our model, neither a spherically symmetric atmosphere nor the notion of a ram-wake asymmetry can reproduce the overall shape and strength of the signatures seen by Juno.However, this interpretation of the magnetic perturbations is possibly not unique, as we have explored only a small corner of the parameter space spanned by the upstream flow parameters and the neutral density profile.Various fine structures in the observed magnetic field are not reproduced by any of the analyzed model setups.In addition, Juno was still several R E downstream when traveling through Europa's anti-Jovian hemisphere; that is, a potential atmospheric bulge around dusk was not directly sampled.Magnetometer and particle data from additional close flybys in Europa's dusk and dawn hemispheres are needed to further substantiate the presence of such an atmospheric asymmetry.

Figure 1 .
Figure 1.Trajectory of the Juno spacecraft during its close flyby of Europa on 29 September 2022.The panels display the projections of the trajectory onto the (a) Z = 0 and (b) Y = 0 planes of the EPhiO system.The blue circle represents the position of Juno's closest approach at 09:36:29 UTC.The green asterisks along the trajectory are 3 min apart, starting at 09:30:00 UTC.During the dashed portion of the trajectory in panel (b), Juno was located "behind" Europa in the Y > 0 half space.The blue arrow denotes Juno's direction of travel.The orange arrow in panel (a) represents the direction of the incident solar radiation.

Figure 2 .
Figure 2. Magnetic field  = ( ,  ,  ) near Europa in EPhiO coordinates: model results versus Juno data.The figure displays time series of the observed magnetic field components (black) as well as output from AIKEF runs (a) #1-#3 and (b) #4-#6.Plot (b) also shows the magnetic field obtained from a mere superposition of   0 and Roth et al. (2014)for Europa's system III longitude during the flyby (λ III = 136°).At the time of this writing, ambient plasma moments from the Juno flyby were not yet available in the peer-reviewed literature.