Europlanet Science Congress 2020
Virtual meeting
21 September – 9 October 2020
Europlanet Science Congress 2020
Virtual meeting
21 September – 9 October 2020

Oral presentations and abstracts


The exploration of the outer solar system by Galileo at Jupiter, Cassini-Huygens at Saturn, and New Horizons at Pluto-Charon, has revealed that several icy worlds harbor a subsurface salty ocean underneath their cold icy surface. By flying through the icy-vapor plume erupting from Enceladus' south pole, Cassini proceeded for the first time to the analysis of fresh materials coming from an extraterrestrial ocean, revealing its astrobiological potentials. Even if there is no direct evidence yet, similar oceanic habitats might also be present within Europa, Ganymede and Titan, which will be characterized by future missions currently under development for the exploration of icy Galilean moons (JUICE, Europa Clipper) and of Saturn’s moon Titan (Dragonfly).
Understanding these icy ocean worlds and their connections with smaller icy moons and rings requires input from a variety of scientific disciplines: planetary geology and geophysics, atmospheric physics, life sciences, magnetospheric environment, space weathering, as well as supporting laboratory studies, numerical simulations, preparatory studies for future missions and technology developments in instrumentation and engineering. We welcome abstracts that span this full breadth of disciplines required for the characterization and future exploration of icy worlds and ring system.

Convener: Gabriel Tobie | Co-conveners: Carly Howett, Alice Lucchetti, Frank Postberg, Federico Tosi

Session assets

Session summary

Chairperson: Gabriel Tobie and Carly Howett
Nicolas Ligier, Christoper Paranicas, John Carter, François Poulet, Wendy Calvin, Tom Andre Nordheim, and Colin Snodgrass

Icy bodies are the most numerous and diverse bodies in the Solar System, but only a few have been visited or will be visited by a space probe. The Galilean satellites are one of those, especially Ganymede which is the primary target of the future L-class mission JUICE (launch scheduled in 2022) in ESA’s Cosmic Vision programme. Ganymede's surface visually exhibits an important geological diversity, with young bright areas to older dark terrains. This diversity also expresses itself through the moon’s surface composition, which was studied extensively in situ by the NIMS instrument of the Galileo mission (NASA) in the late 90s; like the majority of giant planet satellites, Ganymede's surface is dominated by H2O-ice and some non-icy components, very likely hydrated salts based on the distorted shape of the spectral signatures (McCord et al., 2001). However, this binary composition has been obtained with a spectral sampling (~25 µm) not allowing to detect specific absorptions for the non-icy materials, thus not allowing their identification. Hence, many questions about Ganymede’s surface composition remain unanswered while important technical advances have been made since.

In preparation of the JUICE mission, and specifically of the near-infrared imaging spectrometer MAJIS of the JUICE mission, a ground-based campaign was performed using an instrument with a much finer, i.e. better, spectral sampling: SINFONI (SINgle Faint Object Near-IR Investigation). SINFONI is installed on the UT4 of the Very Large Telescope (VLT hereafter) at the European Southern Observatory (ESO hereafter) in Chile. It combines one adaptive optics module and an integral field spectrometer operating in the near-infrared covering from the beginning of the J-band (~1.1 µm) to the end of the K-band (~2.45 µm). Here we present the results derived from the analysis and the modeling of four observations acquired at different dates, from October 2012 to March 2015, all covering the 1.45 – 2.45 µm wavelength range with a spectral resolution about 0.5 µm and a spatial sampling of 12.5 x 12.5 mas2. These results were recently published in Icarus (Ligier et al., 2019).


The first result we obtained concerns the physical properties of Ganymede’s surface. Indeed, the data reduction process highlights that the Lambertian model is not sufficient to remove the photometric effects due to observations geometry. Instead, the Oren-Nayar model (Oren & Nayar, 1994), which generalizes the Lambertian law for rough surfaces, produces excellent results where no inclination residuals are observed up to inclination angles around 65°. The quality of the photometric correction is thus used as proxy to infer Ganymede’s surface roughness: from 16° ± 6° to 21° ± 6° depending on the observations.

Then, concerning the surface composition of the moon, our modeling confirms that it is dominated by H2O-ice, predominantly the crystalline form. The abundance maps of the ice show two main patterns: (1) a latitudinal gradient in terms of abundance, with large polar caps, and (2) a latitudinal gradient in terms of grain size, where the smaller grains (>50 µm) are located at the highest latitudes, showing a sharp transition around ±35°N coinciding with the transition between open and closed field lines of Ganymede’s own magnetic field (figure 1a). Ice sublimation explains this distribution, redistributing H2O-ice from the “hot” equator to the colder poles, very likely redeposited as finer grains.

Apart from the ice, another major compound is required to fit Ganymede’s spectra: a darkening agent. Similarly to previous studies, this darkening agent could not be identified, but we were able to provide new constraints on it. First of all, this unknown material cannot be organic matter since its reflectance level, about 0.25, is more or less five times higher than that of organic matter. Instead, the reflectance level suggests a silicate-type material, as already mentioned in a previous study about Callisto (Calvin & Clark, 1991). Its abundance map shows the highest abundances (up to 0.85) at equatorial latitudes in the trailing hemisphere (figure 1b). The well-known surface sputtering engendered by Jupiter’s magnetosphere is the simplest process to explain such distribution. Even with the magnetic field of the moon, it is possible for corotating singly-charged ions to become neutralized near the moon and continue to the surface as neutrals, sputtering mostly around the trailing apex.

Last but not least, our study highlights the necessity of secondary species, i.e. >10% overall, to better fit the measurements: sulfuric acid hydrate and salts, likely sulfates and chlorinated. While the sulfuric acid hydrate is, like H2O-ice, mostly located at high latitudes (figure 1c), the abundance map of the salts shows a heterogeneous distribution which seems neither related to the Jovian magnetospheric bombardment nor craters (figure 1d). These species are mostly detected on bright grooved terrains surrounding darker areas. Endogenous processes, such as freezing of upwelling fluids going through the moon’s ice shell, may explain this heterogeneous distribution.

How to cite: Ligier, N., Paranicas, C., Carter, J., Poulet, F., Calvin, W., Nordheim, T. A., and Snodgrass, C.: Pending the JUICE/ESA mission … surface composition and properties of Ganymede from near-infrared ground-based measurements, Europlanet Science Congress 2020, online, 21 Sep–9 Oct 2020, EPSC2020-655,, 2020.

Samuel Howell


While the thickness of Europa’s ice shell is underconstrained by the current knowledge of the body, it is not unconstrained. That is, while there are many possible ice shell thicknesses, not all are equally plausible. In this study, I survey the current knowledge of Europa and quantify distributions in the parameters and processes that control ice shell thickness. I then create a Monte Carlo simulation of 107 plausible Europa’s with ice shells in thermodynamic equilibrium, and condition the result into a probability distribution for ice shell thickness. I predict a best estimate thickness of 22.0 km, but  with great uncertainty. These results may inform future planning and inferences from NASA’s Europa Clipper mission and ESA’s JUICE mission later this decade [1].



Within its predominantly water-ice shell, Jupiter’s moon Europa likely harbors a global saltwater ocean, placing it among an emerging class of celestial objects known as Ocean Worlds. The ice shell is thought to consist of a brittle upper lithosphere, where heat transfer occurs by thermal conduction, and a ductile interior asthenosphere that may be convecting in solid-state. The icy surface of this shell exhibits one of the youngest average ages in the solar system (~40 – 90 Myr) [2], requiring the recent or current geologic resurfacing.

The geologic processes within the ice shell responsible for the young surface may convey material between the surface and interior ocean, critically influencing chemical disequilibria within the watery interior and the habitability of Europa’s ocean [3], [4]. However, the thickness of the ice shell and therefore the potential for geologic material exchange is unknown, and estimates span three order of magnitude.

Following discoveries at Europa by the Galileo mission, a summary of ice shell thickness predictions was made by Billings and Kattenhorn [5], providing a name to the “Great Thickness Debate.” That study catalogued dozens of previous publications, collating estimates for icy thicknesses from a few hundred meters to many tens of kilometers.

In order to estimate a probability distribution for Europa’s ice shell thickness using the Monte Carlo method, several parameters must be estimated. Therefore, I first identify current best estimate (CBE) values and construct distributions that capture the uncertainty in those values for several key parameters. In each instance, I determine both the range of values for the parameter of interest and the form of the probability distribution for that parameter.



Assuming an ice shell in thermodynamic equilibrium, the heat flux out of the icy lithosphere radiated to space is equal to the heat flux into the base of the lithosphere from the silicate interior plus the internal heat generated by tidal dissipation within a ductile convecting asthenosphere, integrated over the depth of the asthenosphere.

For each of these terms, I identify the underlaying uncertainties associated with their calculation. For Europa as a whole, I include uncertainties in total H2O layer and iron core thickness. In the silicate interior, I consider uncertainties in radiogenic heating associated with the material Europa accreted from, as well as the potential for silicate tidal dissipation based on mechanical constraints. For the ice shell, I consider uncertainties in convective and melting temperatures, grain size, empirical diffusion creep constants, non-ice composition, porosity, mechanical properties, and tidal response.

The thickness of the conductive and convective layer are then sampled according to the governing equations for heat transfer and tidal heat dissipation (Figure 1).


Results and Discussion

The CBE ice shell thickness for Europa is 22.0 km (Figure 2), though the spread in potential thicknesses is great. Only ~1% of possible configurations are thinner than 15 km, and the median thickness is ~40 km. Further, ice shells are primarily dominated by heat transfer through thermal conduction (Figure 3), with thermal convection being relegated most commonly to the bottom 1/3rd or less of the ice shell.

I will address several surprising results in the underlaying distributions, and their affect on the overall solution. Additionally, there are many key sensitivities built in to this model that may be retired through future laboratory analysis and spacecraft observation. In parallel, some unresolvable issues will persist until the ice shell is explored in situ.




Figure 1. Differential and cumulative probability distributions for Europa’s ice shell thickness. The steep rise is attributed to the inverse relationship between minimum ice shell thickness and total heat flux, and the tail to the right is controlled by silicate heat flux and H2O thickness.



Figure 2. Probability heat map showing conductive and convective thicknesses. Warmer colors denote higher probability. White dashed lines show lines of constant ice shell thickness. Note, for example, that the red line showing a constant thickness of 15 km falls outside the majority of solutions, while the line showing 20 km thickness passes through regions of high probability.



This work was performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.



[1]  S. M. Howell and R. T. Pappalardo, “NASA’s Europa Clipper—a mission to a potentially habitable ocean world,” Nat. Commun., vol. 11, no. 1, Art. no. 1, Mar. 2020, doi: 10.1038/s41467-020-15160-9.

[2]  E. B. Bierhaus, K. Zahnle, C. R. Chapman, R. T. Pappalardo, W. R. McKinnon, and K. K. Khurana, “Europa’s crater distributions and surface ages,” Europa, pp. 161–180, 2009.

[3]  K. P. Hand, C. F. Chyba, J. C. Priscu, R. W. Carlson, and K. H. Nealson, “Astrobiology and the Potential for Life on Europa,” in Europa, R. T. Pappalardo, W. B. McKinnon, and K. Khurana, Eds. Tucson: University of Arizona Press, 2009, p. 589.

[4]  S. M. Howell and R. T. Pappalardo, “Can Earth-like plate tectonics occur in ocean world ice shells?,” Icarus, 2019, doi: 10.1016/j.icarus.2019.01.011.

[5]  S. E. Billings and S. A. Kattenhorn, “The great thickness debate: Ice shell thickness models for Europa and comparisons with estimates based on flexure at ridges,” Icarus, vol. 177, no. 2, pp. 397–412, Oct. 2005, doi: 10.1016/j.icarus.2005.03.013.


How to cite: Howell, S.: The Likely Thickness of Europa’s Icy Shell, Europlanet Science Congress 2020, online, 21 Sep–9 Oct 2020, EPSC2020-173,, 2020.

Olivier Witasse and the JUICE Teams

JUICE - JUpiter ICy moons Explorer - is the first large mission in the ESA Cosmic Vision 2015-2025 programme. The mission was selected in May 2012, and entered in implementation phase “D” in March 2019. Due to launch in May 2022 and to arrive at Jupiter in October 2029, it will spend at least three ½ years making detailed observations of Jupiter and three of its largest moons, Ganymede, Callisto and Europa.  The status of the project and the main milestones in 2020-2021 are presented.

Science Objectives

The focus of JUICE is to characterise the conditions that might have led to the emergence of habitable environments among the Jovian icy satellites, with special emphasis on the three worlds, Ganymede, Europa, and Callisto, likely hosting internal oceans. Ganymede, the largest moon in the Solar System, is identified as a high-priority target because it provides a unique and natural laboratory for analysis of the nature, evolution and potential habitability of icy worlds and waterworlds in general, but also because of the role it plays within the system of Galilean satellites, and its special magnetic and plasma interactions with the surrounding Jovian environment. The mission also focuses on characterising the diversity of coupling processes and exchanges in the Jupiter system that are responsible for the changes in surface and space environments at Ganymede, Europa and Callisto, from short-term to geological time scales. Focused studies of Jupiter’s atmosphere and magnetosphere, and their interaction with the Galilean satellites will further enhance our understanding of the evolution and dynamics of the Jovian system. The overarching theme for JUICE is the emergence of habitable worlds around gas giants. At Ganymede, the mission will characterize the ocean layer, provide topographical, geological and compositional mapping of the surface, study the physical properties of the icy crust, characterize the internal mass distribution, investigate the exosphere, study the intrinsic magnetic field and its interactions with the Jovian magnetosphere. At Europa, the focus will be on the surface composition, understanding the formation of surface features and subsurface sounding of the icy crust over recently active regions. Callisto will be explored as a witness of the early solar system, trying to elucidate the mystery of its internal structure. JUICE will also perform a multidisciplinary investigation of the Jupiter system as an archetype for gas giants. The Jovian atmosphere will be studied from the cloud top to the thermosphere. Concerning Jupiter’s magnetosphere, investigations of the three dimensional properties of the magnetodisc and of the coupling processes within the magnetosphere, ionosphere and thermosphere will be carried out. JUICE will study the moons’ interactions with the magnetosphere, gravitational coupling and long-term tidal evolution of the Galilean satellites.

The payload

The JUICE payload consists of 10 state-of-the-art instruments plus one experiment that uses the spacecraft telecommunication system with ground-based instruments. This payload is capable of addressing all of the mission's science goals, from in situ measurements of the plasma environment, to remote observations of the surface and interior of the three icy moons, Ganymede, Europa and Callisto, and of Jupiter’s atmosphere. A remote sensing package includes imaging (JANUS) and spectral-imaging capabilities from the ultraviolet to the sub-millimetre wavelengths (MAJIS, UVS, SWI). A geophysical package consists of a laser altimeter (GALA) and a radar sounder (RIME) for exploring the surface and subsurface of the moons, and a radio science experiment (3GM) to probe the atmospheres of Jupiter and its satellites and to perform measurements of the gravity fields. An in situ package comprises a powerful suite to study plasma and neutral gas environments (PEP) with remote sensing capabilities of energetic neutrals, a magnetometer (J-MAG) and a radio and plasma wave instrument (RPWI), including electric fields sensors and a Langmuir probe. An experiment (PRIDE) using ground-based Very Long Baseline Interferometry (VLBI) will support precise determination of the spacecraft state vector with the focus at improving the ephemeris of the Jovian system.

The mission profile

The mission is due to launch from Kourou with an Ariane 5 ECA. The baseline launch is 20 May 2022, with two backup launch slots in September 2022 and August 2023. The baseline interplanetary transfer sequence relies on gravity assist with Venus, the Earth and Mars. The Jupiter orbit insertion will be performed in October 2029. A Ganymede swing-by is performed just before the capture manoeuvre. The tour of the Jupiter system, as currently designed, starts with a series of three Ganymede swing-bys. The spacecraft is transferred to Callisto to initiate the Europa science phase, one year after the Jupiter insertion. This phase is composed of two fly-bys, separated by 15 days, with closest approach at 400 km altitude. The next phase is a 200-day period characterized by an excursion at moderate inclinations, in order to investigate regions of the Jupiter environment away from the equatorial plane. A series of resonant transfers with Callisto raise the inclination with respect to Jupiter’s equator to a maximum value of about 28 deg. The spacecraft is then transferred from Callisto to Ganymede with a series of Callisto and Ganymede flybys, followed by a gravitational capture with the moon. The orbital phase around Ganymede is split into a first elliptical subphase, a circular orbit at 5000 km altitude followed by another elliptical sub phase, and then a circular phase at 500 km altitude. The duration of the Ganymede phase is about nine months, the end of mission being planned in September 2033. The spacecraft will eventually impact the surface.

The key project milestones in 2020-2021

  •  February 2020: First instrument delivered
  •  May 2020: Interface test between the spacecraft engineering model and the mission operations centre
  •  Second quarter of 2020: Instrument deliveries and integration
  •  January 2021: Start of the environmental tests
  •  First half of 2021: Spacecraft tests
  •  November 2021: Flight acceptance review


How to cite: Witasse, O. and the JUICE Teams: JUICE (Jupiter Icy Moon Explorer): A European mission to explore the emergence of habitable worlds around gas giants, Europlanet Science Congress 2020, online, 21 Sep–9 Oct 2020, EPSC2020-76,, 2020.

Audrey Vorburger and Peter Wurz

It has long been postulated that Europa might have a sub-surface ocean covered by an icy crust. First clues for the existence of such a sub-surface ocean were obtained by the Galileo magnetometer in the late '90s [1]. If such an ocean indeed exists, it might sporadically erupt in plumes. Indeed, in 2014, [2] reported on increased Lyman-α and oxygen OI130.4 nm emissions, which the authors interpreted as transient water vapor measurements resembling water plumes. In this presentation we analyze different plume models to determine which model would result in an observation as the one presented by [2], and analyze what implications this might have for the upcoming measurements of the Neutral and Ion Mass spectrometer (NIM) onboard JUICE.

1. Plume observations
[2] characterized the source of the increased Lyman-α and oxygen OI130.4 nm emissions through modeling. According to their results, the emissions are best described by two individual water vapor plumes that are located at 90°W/55°S and 90°W/75°S, respectively, both of which exhibit a radial expansion of ~200 km and a latitudinal expansion of ~270 km, and the surface densities of which amount to 1.3x1015 m-3 and 2.2x1015 m-3, respectively.

2. Plume model
The source of the observed plume might either be a liquid or a solid (icy) reservoir that evaporates or sublimates as it comes into contact with space. Figure 1 shows different scenarios which could lead to the localized release of H2O particles resulting in a plume-like structure. In the first scenario, a surface, or near-surface, liquid reservoir is exposed to near-vacuum conditions upon which water directly evaporates into space. In the second scenario, a crack in the ice shell all the way to the bottom leads to the exposure of oceanic water to space, resulting in the formation of an oceanic plume. If the flow is chocked (by the conduit's geometry), the plume may become a supersonic jet. In the third scenario a rising diapir results in the warming of local surface ice (indicated by the shaded region in Figure 1), which sublimates into space. The water temperature was set to 280 K in scenarios one and two, whereas the ice temperature was set to 250 K in scenario three, and the reservoir areas were set to ~1'000 m2 and ~20'000 m2, respectively. 

Figure 1: The three analyzed scenarios: Scenario one shows evaporation of a surface, or near-surface, liquid, scenario two shows evaporation of oceanic water (in form of a jet), and scenario three shows the sublimation of surface ice heated by diapirs.

3. Monte-Carlo Model
To simulate Europa's plumes, we use a 3D Monte Carlo model originally developed to model Mercury's exosphere [3]. In this model particles are created ab initio, travel on collision-less trajectories, and are removed as they are either ionized, fragmented, or lost either to space or by freezing out on the surface. The grids are ~25x25x25 km3 in size, thus almost a factor 10 higher in resolution than the [2] measurements. For comparison with the [2] observations we also merged our model results into 200x200x200 km3 bins.

4. Model Results
Figure 2 shows our model results for the three different scenarios presented above. For each scenario, we present the [2] measurement on the left, the reduced resolution result in the middle, and the high resolution result on the right. All measurements are normalized to one and span six orders of magnitude. 

Figure 2: Model results for the three analyzed scenarios. The top row shows the surface liquid scenario, the middle row shows the oceanic water jet scenario, and the bottom row shows the diapir scenario. The [2] measurement is shown on the left, the reduced resolution model results are shown in the middle, and the high resolution model results are shown on the right.

5. Discussion
Whereas the observed ~200 km scale height is met by all three modeled scenarios, only scenario number two, the oceanic water jet scenario, results in a narrow enough plume structure. Both scenarios number one and three are too broad to be in good agreement with the [2] observations. It thus seems that for the high radial scale height but the low latitudinal expansion a narrowing factor needs to be present, as for example a conduit-like geometry provides. The presence of a nozzle does not necessarily require that the liquid stems from the ocean, though. If a throat exists close to Europa's surface, it is also possible that a water inclusion close to the surface results in a jet-like geometry.

6. NIM Plume Observations
NIM is a highly sensitive neutral gas and ion mass spectrometer designed to measure the exospheres of the Europa, Ganymede, and Callisto. The detection limit is at 10-16 mbar for a 5 second integration time, which translates to a particle density of ~1 cm-3. NIM's mass resolution is M/ΔM > 1100 in the mass range 1-1000 amu. In addition to the modeled 3D plume density profiles, we will also present modeled mass spectra for the individual scenarios, and discuss their implications for positive plume identification possibilities.

7. Conclusion
The origin of the observed Europa water vapor plumes is of high scientific interest because the plume's chemical composition directly represents the reservoir's chemical composition. If the plume thus indeed originates in the ocean expected to lie underneath Europa's surface ice layer, analysis of its chemical omposition with a mass spectrometer would offer us direct information about the chemical composition of the water ocean itself. Such information would in turn teach us more about Europa's habitability, and about the possibility of Europa harboring life.

[1] Khurana, K. K., et al.: Induced magnetic fields as evidence for subsurface oceans in Europa and Callisto, Nature, Vol. 395, pp. 777-780, 1998.

[2] Roth, L., et al.: Transient water vapor at Europa's south pole, Science, Vol. 343, pp. 171-174, 2014.

[3] Wurz, P., and Lammer, H.: Monte-Carlo simulation of Mercury's exosphere, Icarus, Vol. 164, pp. 1-13, 2003.

How to cite: Vorburger, A. and Wurz, P.: Monte-Carlo Model of Europa's Water Vapor Plumes, Europlanet Science Congress 2020, online, 21 Sep–9 Oct 2020, EPSC2020-78,, 2020.

Ines Belgacem, Frédéric Schmidt, and Grégory Jonniaux


We carried out a regional study of Ganymede and Callisto’s photometry by estimating the parameters of the Hapke photometric models on a handful of regions of interest. We found that the photometry was diverse across the surface and highlighted areas with a distinct behavior that could indicate localized activity on Ganymede


Jupiter’s icy moons are at the center of future space exploration missions such as ESA’s JUpiter ICy moons Explorer [1] and NASA’s Europa Clipper [2]. Ganymede, in particular, will be the primary target of the JUICE mission. Knowledge of the surface of Ganymede and Callisto is paramount to best plan the mission, help with navigation [3] and understand the its geology. In this study, we focus on the photometry of the surface and how we can describe it using the Hapke photometric model [4].

1. Dataset

We use images from the Voyager probes dataset taken with the Imaging Science System (ISS) [5] and from the New Horizons spacecraft taken with the LOng Range Reconnaissance Orbiter (LORRI) [6] with a ground resolution between 10 km and 30 km at the subspacecraft point. All Voyager images are limited to the clear filter.

2. Method

We need two elements to carry out this study: the reflectance and the geometry of observation. The first can be obtained after radiometric calibration of the images. The second necessitates accurate projections of each pixel.

2.1 Correction of metadata

We simulated images with SurRender [7] and compared those simulations to the real images to correct for spacecraft pointing and moon attitude. Additional distortion and distance corrections were needed and implemented on the Voyager images [8].

2.2 Model and Bayesian inversion

For this study we are considering the Hapke model detailed in Hapke, 1993 [4]. Six parameters are estimated: b, c, ω, θ, h and B0.

We have developed an inversion tool using a Bayesian approach based on previous work done on Mars [9, 10]. No a priori knowledge of the parameters were inferred except for their physical domain of variation. We also include in the model uncertainties on the absolute level  reflectance (radiometric uncertainties) to correct for potential bias. This work is detailed in our previous study of Europa [11].

3. Results

3.1. Ganymede's regional photometry

We realized a regional photometric study of 15 areas of Ganymede with a very limited dataset of 16 images matching our criteria (see section 1) for which we corrected the metadata (spacecraft position and orientation) and radiometric calibration discrepancies.

We found that most of our areas are consistent with a global backscattering behavior of the surface with two notable exceptions - ROIs #2 and #4 (see fig. 1). They are situated in the polar latitudes, respectively in the north and in the south hemispheres. They are both very bright and rough and exhibit a strong and narrow forward scattering. This could be due to them being particularly favorable areas for ice redistribution at the poles or signs of possibly localized activity.