Subsurface Properties of Lucus Planum, Mars, As Seen by MARSIS

Subsurface properties of Lucus Planum, Mars, as seen by MARSIS

Roberto Orosei1, Angelo Pio Rossi2, Federico Cantini3, Graziella Caprarelli4,5, Lynn Carter6, Irene Papiano7 1INAF (IT), 2JacobsUni (DE), 3EPFL (CH),4UniSA (AU), 5IRSPS (IT), 6NASA (US), 7Liceo Righi BO (IT) Why Lucus Planum

• Located in key area: Dichotomy, Tharsis, Elysium, Medusa Fossae Formation (MFF), 2 (3) landing sites nearby

• MFF partially radar-transparent

• Lucus Planum variably radar transparent (challenging with SHARAD)

• Nature of subsurface material?

2 Geology

Gale

Gusev

Htu AHtu Hesperian Amazonian-Hesperian Source: Tanaka et al. (2014) 3 transitional unit transitional unit Geology

Htu AHtu Source: Tanaka et al. (2014) 4 Lucus Planum

Htu AHtu Source: Tanaka et al. (2014) 5 Data and Methods

• MARSIS radargrams and simulations

• Thickness estimation

• Subsurface radar reﬂectors

• Estimation of Lucus Planum thickness and ROIs

• Interpolation of MOLA topography

• Estimation of real depth of subsurface reﬂector

• Derivation of dielectric permittivity

• Loss tangent

6 Data

• MARSIS, SHARAD respectively on board MEX, MRO

• Low-frequency synthetic aperture radar sounders

• MARSIS ➔ deep penetration, free- space range resolution of approximately 150 m, a footprint size of 10-20 km in the across-track Image courtesy NASA/JPL-Caltech. direction ranges and 5-10 km in the along-track direction. MARSIS • SHARAD ➔ shallow penetration, but f 1 f 2 f 3 f 4 10x spatial resolution 1.8 Mhz 3.0 Mhz 4.0 Mhz 5 Mhz

7 MARSIS coverage

8 Simulations

• Simulations of surface clutter from MOLA MEGDR grid (1/128 px/ deg)

• Features visible in the radargram and not in the simulations are most likely real subsurface reﬂectors

• Real reﬂectors mapped

MARSIS - Orbit 4011 Frequency band 1 9 Subsurface reﬂectors

• Position, strength of subsurface echoes extracted manually across Lucus Planum ➔ Echo time delay to apparent depth

• Strongest subsurface echoes (weak internal attenuation, strong subsurface reﬂectivity, or both) in deposits located NW of Apollinaris Patera

• No detected subsurface echoes in the central section of Lucus Planum

• Subsurface reﬂections common in the E and NW sectors, up to 2 km (assuming dielectric permittivity = 3)

10 MARSIS radagrams

MARSIS visualisation tool —> See Poster Cantini et al. #12545, PS9.5

11 MARSIS - Orbit 13522 Frequency band 1 MARSIS radagrams

MARSIS - Orbit 12319 Frequency band 1 MARSIS - Orbit 12319 Frequency band 2 12 MARSIS radagrams

MARSIS - Orbit 7013 Frequency band 1 MARSIS - Orbit 7013 Frequency band 2

50 km 13 MARSIS radagrams

MARSIS - Orbit 7013 Frequency band 1 MARSIS - Orbit 7013 Frequency band 2

50 km 14 Regions of Interest

B • 3 sub-regions selected C within the broader Lucus Planum terrain A • Slightly different response and properties

• Surface morphology, erosional stage, age is also slightly different

15 R. Orosei et al.: Radar sounding over Lucus Planum, Mars 3

The depth of reﬂectors can be estimated from the round- trip timeSubsurface delay between surface and subsurface reﬂectors echo through the following relation:

c ⌧ z = (1) 2p"

5 where z is depth, c the speed of light in vacuo, ⌧ the round- ztrip = depth, time delay between surface and subsurface echo, and " the real part of the relative dielectric permittivity of the Lucus cPlanum = light speed material. in The values of the apparent depths shown in vacuum Fig. 2 have been computed neglecting the effect of ", and thus 10 τrefer = TWT to the distance covered by an electromagnetic wave in εfree = real space part of during the the same round-trip time. As such, appar- relativeent depths dielectric overestimate the thickness of Lucus Planum by a permittivityfactor comprised Lucus between p3 and 3, depending on the nature Planum deposits of the material through which the wave propagates (see e.g., Figure 3. MARSIS coverage over a shaded relief map of Lucus 15 Ulaby et al., 1986, Appendix E). Planum. Ground tracks are plotted as black lines. An estimate the dielectric permittivityassuming for the different propagation re- in vacuum gions of Lucus Planum would provide some insight on their nature and a more precise evaluation of their thickness. Fol- 16 lowing the approach ﬁrst presented in Picardi et al. (2005) 20 and used also in Watters et al. (2007), we estimate the left side of Eq. 1 by interpolating the topography beneath Lucus Planum from that of the surrounding area.

G 2 P = P R 2 (2) s t · 8⇡H ·| s| ✓ ◆

Figure 4. Apparent depth of subsurface echoes detected by MAR- SIS, presumably originating at the base of Lucus Planum. The real depth is obtained dividing the apparent depth by the square root of the relative dielectric permittivity of the medium.

2 2 G 2 Pss = Pt 1 Rs · 8⇡(H + z) · | | · both, are found within the deposits located NW of Apolli- ✓ ◆ 2 ⇣ ⌘ naris Patera, while no subsurface echoes could be detected in 25 Rss exp( 4⇡f tan⌧) (3) | | · the central section of Lucus Planum, in spite of several high- SNR observations. Subsurface reflections are common in the Eastern and Northwestern sectors, in some cases to depths of 35 more than 2000 m assuming a dielectric permittivity of about 3 (Watters et al., 2007; Carter et al., 2009). Because subsurface echoes were clustered in specific ar- The positions and strengths of subsurface echoes were ex- eas, Lucus Planum has been subdivided in three Regions of tracted manually from radargrams and mapped across Lu- Interest (ROI), as shown in the map below (Fig. 5) 40 cus Planum, converting echo time delay to apparent depth To estimate the dielectric permittivity, the topography be- (Fig. 4). The strongest subsurface echoes, resulting from neath Lucus Planum has been extrapolated from that of the 30 weak internal attenuation, strong subsurface reflectivity, or surrounding terrains. If this extrapolation is sufficiently ac- Interpolation

1 2 Interpolation of basal topography with MOLA MEGDR

2 interpolated base vs. MOLA surface 17 Apparent vs interpolated

A C

18 equation

April 15, 2016 equation Dielectric permittivityR. Orosei et al.: Radar sounding over Lucus Planum, Mars 3 1 Introduction The depth of reflectors can be estimated from the round- • In a medium characterized by permittivity ε, the speed of propagation of an EM wave is: trip time delay between surface and subsurface echo through v April=thec/ followingp✏ 15, 2016 relation: (1) • In such medium, the relationship between echo time delay � and depth z is thus: 2 Gc ⌧ 2 Ps = Pt z = Rs (2) (1) · 8⇡H2p!" ·| | 1 Introduction B • The ratio between the extrapolated thickness and the 5 where z is depth, c the speed of light in vacuo, ⌧ the round- A apparent thickness provides an estimate of 2 trip timep delay✏ between surface and subsurfaceC echo,(1) and " G 2 2 • Pss = Pt the real part of1 the relativeRs dielectric permittivity of the Lucus The relative dielectric permittivity· 8⇡(H estimated+ z) for! ROIs· A | | · and C is comprised between 5 and Planum6. For ROI material.B, the⇣ The values⌘ of the apparent depths shown in estimated permittivity is above2 10, which2.7 ⌧isf characteristictan of Rss 10Fig. 2v have= c/ beenp✏ computed neglecting the effect(3) of ",(2) and thus dense rocks (lava flows?| bedrock?)| · . 10 refer to the distance covered by an electromagnetic wave in free space during2 the same round-trip19 time. As such, appar- G 2 Ps =entPt depths overestimateRs the thickness of Lucus Planum(3) by a factor· comprised8⇡H ! between·| | p3 and 3, depending on the nature of the material through which the wave propagates (see e.g., Figure 3. MARSIS coverage over a shaded relief map of Lucus 15 Ulaby et al., 1986,2 Appendix E). Planum. Ground tracks are plotted as black lines. AnG estimate the dielectric permittivity2 2 for the different re- Pss = Pt 1 Rs · gions8⇡(H of+ Lucusz)! Planum· would| | provide· some insight on their ⇣ ⌘ nature2 and2.7⌧f atan more precise evaluation of their thickness. Fol- Rss 10 (4) | |lowing· the approach first presented in Picardi et al. (2005) 20 and used also in Watters et al. (2007), we estimate the left side of Eq. 1 by interpolating the topography beneath Lucus Planum from that of the surrounding area.

1 G 2 P = P R 2 (2) s t · 8⇡H ·| s| ✓ ◆

Figure 4. Apparent depth of subsurface echoes detected by MAR- 1 SIS, presumably originating at the base of Lucus Planum. The real depth is obtained dividing the apparent depth by the square root of the relative dielectric permittivity of the medium.

The depth of reﬂectors can be estimated from the round- trip time delay between surface and subsurface echo through the following relation: R. Orosei et al.: Radar sounding over Lucus Planum, Mars 3

c ⌧ z = The depth of reflectors can be estimated from(1) the round- 2ptrip" time delay between surface and subsurface echo through the following relation: 5 where z is depth, c the speed of light in vacuo, ⌧ the round- trip time delay between surface and subsurface echo, and " c ⌧ the realz part= of the relative dielectric permittivity of the Lucus (1) Planum material.2p" The values of the apparent depths shown in Fig. 2 have been computed neglecting the effect of ", and thus 5 z c ⌧ 10 refer towhere the distanceis depth, coveredthe by speed an electromagnetic of light in vacuo, wave inthe round- free spacetrip during time delay the same between round-trip surface time. and As subsurface such, appar- echo, and " ent depthsthe real overestimate part of the the relative thickness dielectric of Lucus permittivity Planum by of a the Lucus factor comprisedPlanum material. between Thep3 valuesand 3, depending of the apparent on the depths nature shown in Fig. 2 have been computed neglecting the effect of ", and thus of the material through which the wave propagates (see e.g., Figure 3. MARSIS coverage over a shaded relief map of Lucus 15 Ulaby10 etrefer al., to1986, the Appendix distance covered E). by an electromagnetic wavePlanum. in Ground tracks are plotted as black lines. An estimatefree space the during dielectric the permittivity same round-trip for the time. different As such, re- appar- gions ofent Lucus depths Planum overestimate would provide the thickness some insight of Lucus on their Planum by a nature andfactor a more comprised precise between evaluationp3 ofand their 3, depending thickness. onFol- the nature lowingof the the approach material first through presented which in the Picardi wave et propagates al. (2005) (see e.g., Figure 3. MARSIS coverage over a shaded relief map of Lucus 20 and15 usedUlaby also et in al., Watters 1986, et Appendix al. (2007), E). we estimate the left Planum. Ground tracks are plotted as black lines. side of Eq.An 1 byestimate interpolating the dielectric the topography permittivity beneath for the Lucus different re- Planumgions from of that Lucus of the Planum surrounding would area. provide some insight on their nature and a more precise evaluation of their thickness. Fol- Loss tangent B lowing the approach first presented in Picardi et al. (2005) 20 and used also in Watters et al. (2007), we estimate the left A side of Eq. 1 by interpolating the topography beneath Lucus C Planum from that of the surrounding area.

G 2 P = P R 2 (2) s t · 8⇡H ·| s| ✓ ◆

2 Figure 4. Apparent depth of subsurface echoes detected by MAR- G 2 SIS, presumably originating at the base of Lucus Planum. The real Ps = Pt Rs (2) · 8⇡H ·| | depth is obtained dividing the apparent depth by the square root of ✓ ◆ the relative dielectric permittivity of the medium.

2 2 G 2 Pss = Pt 1 Rs · 8⇡(H + z) · | | · both, areFigure found 4. withinApparent the deposits depth of subsurface located NW echoes of Apolli- detected by MAR- ✓ ◆ SIS, presumably originating at the base of Lucus Planum. The real 2 ⇣ ⌘ naris Patera, while no subsurface echoes could be detected in 25 Rss exp( 4⇡f tan⌧) (3) depth is obtained dividing the apparent depth by the square root of | | · the central section of Lucus Planum, in spite of several high- the relative dielectric permittivity of the medium. SNR observations. Subsurface reflections are common in the 2 2 Eastern and Northwestern sectors, in some cases to depths of 35 G 2 Pss = Pt 1 Rs more than 2000 m assuming a dielectric permittivity of about · 8⇡(H + z) · | | · both, are found within the deposits located NW of Apolli- ✓ ◆ 3 (Watters et al., 2007; Carter et al., 2009). 2 20 ⇣ ⌘ naris Patera, while no subsurface echoes could be detected in 25 Rss exp( 4⇡f tan⌧) (3)Because subsurface echoes were clustered in specific ar- | | · the central section of Lucus Planum, in spite of several high- The positions and strengths of subsurface echoes were ex- eas, Lucus Planum has been subdivided in three Regions of SNR observations. Subsurface reflections are common in the tracted manually from radargrams and mapped across Lu- Interest (ROI), as shown in the map below (Fig. 5) 40 Eastern and Northwestern sectors, in some cases to depths of 35 cus Planum, converting echo time delay to apparent depth To estimate the dielectric permittivity, the topography be- more than 2000 m assuming a dielectric permittivity of about (Fig. 4). The strongest subsurface echoes, resulting from neath Lucus Planum has been extrapolated from that of the 3 (Watters et al., 2007; Carter et al., 2009). 30 weak internal attenuation, strong subsurface reflectivity, or surrounding terrains. If this extrapolation is sufficiently ac- Because subsurface echoes were clustered in specific ar- The positions and strengths of subsurface echoes were ex- eas, Lucus Planum has been subdivided in three Regions of tracted manually from radargrams and mapped across Lu- Interest (ROI), as shown in the map below (Fig. 5) 40 cus Planum, converting echo time delay to apparent depth To estimate the dielectric permittivity, the topography be- (Fig. 4). The strongest subsurface echoes, resulting from neath Lucus Planum has been extrapolated from that of the 30 weak internal attenuation, strong subsurface reflectivity, or surrounding terrains. If this extrapolation is sufficiently ac- Loss tangent

• The rate of attenuation of subsurface echoes as a function of depth can be used to estimate the loss tangent of the Lucus Planum material

• The slope of the line of best ﬁt is related to the loss tangent, the ratio between imaginary and real part of the complex dielectric constant.

• The loss tangent estimates for ROIs A and B are between 0.004 and 0.006, while for ROI C is at around 0.002 scientific instruments • The difference between these values is signiﬁcant for ROIs A and C, while there is a large error bar on the estimate for ROI B.

• These values are compatible with porous rock, dry regolith or very

dust-laden ice Table 1. Dielectric properties of the subsurface material.

Crust Material Pore-Filling Material Andesite Basalt Water Ice Liquid Water

εr 3.5 7.1 3.15 88 tan δ 0.005 0.014 0.00022 0.0001

Table 2. Value ranges of the surfaceSource: geometric parameters. Picardi et al. (2004) 21

Large-Scale Model Small-Scale Model rms slope correlation length rms slope rms height

0.01-0.1 rad 200-3000 m 0.1-0.6 rad 0.1-1 m (0.57-5.7°) (5.7-34.3°)

It appears almost certain from morphologic and chemical evidence, as well as from SNC meteorites, that the martian surface is primarily basaltic. However, it could have a thin veneer of younger volcanics overlying a primitive crust. Whether this primitive crust is basaltic, anorthositic like the Moon, granodioritic like the Earth’s continents or some other kind of composition, is unknown. The NASA Pathfinder APXS analyses of rocks and soils confirm the basaltic nature of Mars’ surface. Chemical classifications of lavas show that the Barnacle Bill and Yogi rocks are distinct from basaltic martian meteorites. These rocks plot in or near the field of andesites, a type of lava common at continental margins on Earth. Although a multitude of different chemical compositions is present at the surface of Mars, it is necessary to select a few representative materials as most meaningful for electromagnetic studies. Given the above considerations about the nature of the martian crust, andesite and basalt were chosen because their dielectric constants are end-members of the range within which the martian surface materials may vary. The dielectric properties of the crust end-member materials, together with those of the water and ice filling the pores, are listed in Table 1. To summarise, the reference models representing the two most likely detection scenarios for a Mars orbital sounder at km depths are (Fig. 1):

Ice/water interface detection. According to the model, the porosity of the martian megaregolith is maximum at the surface and its decay with increasing depth is given by the exponential law in Eq. (1). The pores are filled with ice from the surface down to a depth below which liquid water is stable and becomes the pore- filling material. The change causes a discontinuity of the overall dielectric constant, which can be detected by the radar sounder. The ice/water interface is believed to be at a depth of between 0 m and 5000 m. Dry/ice interface detection. This model is based on the same assumptions as the ice/water model with respect to the megaregolith properties. However, the pore- filling material here is considered to be gas or some other vacuum-equivalent material up to a certain depth below which ice fills the pores. Hence the interface to be detected is between dry regolith and ice-filled regolith, expected to be at a depth of between 0 m and 1000 m.

These models will be used to estimate the penetration performance under typical MARSIS operating conditions.

3 45th Lunar and Planetary Science Conference (2014) 2672.pdf

THE DISTRIBUTION AND DIVERSITY OF LAYERING WITHIN THE MEDUSAE FOSSAE FORMATION. L. Kerber, Laboratoire de Métérologie Dynamique, 4 Place Jussieu, Paris, France ([email protected]).

Introduction: for identifying differences in composition, structure, The Medusae Fossae Formation (MFF) is a wide- and erosive state across the deposit, and (3) for identi- spread and voluminous formation which covers 2.1 x fying places where layering may preserve ancient fea- 106 km2 between 130-230ºE and 12ºS-12ºN [1-3]. As a tures. Layering was not seen by the SHARAD (Mars 45th Lunar and Planetary Science Conference (2014) 2672.pdf fine-grained, friable deposit, its surface is dominated SHAllow RADar sounder) instrument, meaning that by aeolian features such as yardangs [3-6] and a large any tens-of-meters layers that do exist either do not number of both fresh and indurated transverse aeolian have high permittivity contrasts or are discontinuous THE DISTRIBUTION AND DIVERSITY OF LAYERING WITHIN THE MEDUSAE FOSSAE ridges (TARs) [6]. Many hypotheses have been pro- [12]. In order to better document the occurrence of FORMATION. L. Kerber, Laboratoire de Métérologie Dynamique, 4 Place Jussieu, Paris, France posed for the formation of the deposit, including ig- ([email protected]). layering within the MFF, we examined 427 High Reso- nimbrites, volcanic ash fall, and wind-deposited ae- lution Imaging Science Experiment (HiRISE) images olian loess [7-9]. The deposition of the MFF began at Introduction: for identifying differences in composition, structure, spread across the formation, during which the occur- The Medusae Fossae Formation (MFF) is a wide- and erosive state across the deposit,the latestandSubsurface (3) in for the i denti- Hesperian [10], and, by virtue of natureits rence of layering was mapped [13]. HiRISE images spread and voluminous formation which covers 2.1 x fying places where layering mayfine-grained preserve ancient nature fe anda- gentle emplacement, the MFF were supplemented by Mars Reconnaissance Orbiter 6 2 • 10 km between 130-230ºE and 12ºS-12ºN [1-3]. As a tures. Layering was not seen bymay the preserve SHARADThe lackan (Mars important of subsurface record of Martian reﬂections history, in Contextthe central Imager (CTX),part of Mars Lucus Express Planum High Resolution can be fine-grained, friable deposit, its surface is dominated SHAllow RADar sounder) instrument,most directly the meaning result as a thatresult of several of the burial factors and :exhumat ion Stereo Camera (HRSC) and Mars Global Surveyor by aeolian features such as yardangs [3-6] and a large any tens-of-meters layers that doof existchannels either found do n inot its western regions [11]. While Mars Orbiter Camera (MOC) images where needed. number of both fresh and indurated transverse aeolian have high permittivity contrastslayering or are • discontinuo is notA high ubiquitousus topographic in the MFF, roughness examples of at scalesHere we comparabledescribe the results to theof this radar survey as it relates layering of different kinds are geographically wide- to the distribution and diversity of layering within the ridges (TARs) [6]. Many hypotheses have been pro- [12]. In order to better document the occurrencewavelength of causes scattering of the impinging pulse, resulting in weaker posed for the formation of the deposit, including ig- layering within the MFF, we examinedspread 427(Fig. High 1). TheReso- identification and characterization MFF. nimbrites, volcanic ash fall, and wind-deposited ae- lution Imaging Science Experimentof this (HiRISE) layeringsurface images is important and subsurface for three reasons: echoes (1) for olian loess [7-9]. The deposition of the MFF began at spread across the formation, duringdetermining which the the occur mode- of formation of the deposit (2) the latest in the Hesperian [10], and, by virtue of its rence of layering was mapped [13]. HiRISE• Roughness images at the base of the deposit is higher in its central part, (no fine-grained nature and gentle emplacement, the MFF were supplemented by Mars Reconnaissanceindication Orbiter of such trend in older surrounding terrains, though) may preserve an important record of Martian history, Context Imager (CTX), Mars Express High Resolution most directly as a result of the burial and exhumation Stereo Camera (HRSC) and Mars Global• Central Surveyor part of Lucus Planum consists of denser, more radar-attenuating of channels found in its western regions [11]. While Mars Orbiter Camera (MOC) images where needed. layering is not ubiquitous in the MFF, examples of Here we describe the results of this survey asmaterial. it relates layering of different kinds are geographically wide- to the distribution and diversity of layering within the spread (Fig. 1). The identification and characterization MFF. of this layering is important for three reasons: (1) for determining the mode of formation of the deposit (2)

Source: Kerber (2014)

Figure 1. The distribution of visible layers throughout the Medusae Fossae Formation. In the majority of cases there are only one or two massive layers, though in certain areas there are dozens of thin, meter-scale layers. Layers in the MFF are often discontinuous. In the eastern portions of the Medusae Fossae Formation, thick layers of dust lie on top of the formation. to be layers form swirling, occasionally intersecting Observations: Layering was observed across the banding, which does not appear to follow topography formation, though many HiRISE images showed no evi- (Fig. 2b). These areas are heavily jointed, with some dence for layering, and in others only one or two layer layer boundaries coincident with joints. In some re- boundaries were apparent in small parts of the scene. Layers gions, such as northern Lucus Planum, layered material are most common in the Western MFF near inverted fluvial which appears to be continuous with the MFF actually channels [11], where they form regular, stair-step benches consists of dust-covered lava layers (Fig. 2c). Figure 1. The distribution of visible layers throughout the Medusae Fossae Formation. In the majority of cases there are only (Fig. 2a). Their shapes in planform depend highly upon to- One exceptional exposure of layering occurs at the one or two massive layers, though in certain areas there are dozens of thin, meter-scale layers. Layers in the MFF are often discontinuous. In the eastern portions of the Medusae Fossae Formation, thick layers of dust lie on toppography. of the formation. Just south of Apollinaris Mons, what appear edge of Memnonia Planum, where <5 m layers can be to be layers form swirling, occasionally intersecting Observations: Layering was observed across the banding, which does not appear to follow topography formation, though many HiRISE images showed no evi- (Fig. 2b). These areas are heavily jointed, with some dence for layering, and in others only one or two layer layer boundaries coincident with joints. In some re- boundaries were apparent in small parts of the scene. Layers gions, such as northern Lucus Planum, layered material are most common in the Western MFF near inverted fluvial which appears to be continuous with the MFF actually channels [11], where they form regular, stair-step benches consists of dust-covered lava layers (Fig. 2c). (Fig. 2a). Their shapes in planform depend highly upon to- One exceptional exposure of layering occurs at the pography. Just south of Apollinaris Mons, what appear edge of Memnonia Planum, where <5 m layers can be Concluding remarks

• MARSIS can probe the subsurface of Lucus Planum

• Bottom reﬂectors of the contact between (AHth, Htu) and eHt, HNt (and whatever lies under lAv) are not continuous throughout the studied area

• Subsurface materials within Lucus Planum geologically heterogeneous to a certain extent