sign in sign up
<<
Home , Aonia Terra, Aram Chaos, Aromatum Chaos, Asimov (crater), Athabasca Valles, Atlantis Chaos

Research Collection

Doctoral Thesis

The Southern polar giant impact hypothesis for the origin of the dichotomy and the evolution of volcanism on

Author(s): Leone, Giovanni

Publication Date: 2016

Permanent : https://doi.org/10.3929/ethz-a-010652097

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library DISS. ETH NO. 22390

THE SOUTHERN POLAR GIANT IMPACT HYPOTHESIS FOR THE ORIGIN OF THE AND THE EVOLUTION OF

A thesis submitted to attain the degree of DOCTOR OF of ETH ZURICH

(Dr. Sc. ETH Zurich)

presented by

GIOVANNI LEONE PhD, Lancaster University

on 10.02.1967 citizen of Italy

accepted on the recommendation of

Prof. Paul James Tackley Prof. Olivier Bachmann Prof. Nicholas Lang Dr. Dave May

2016

… to my father and his loving memory.

2 Acknowledgements

Completing my second PhD and writing this thesis was an amazing journey that would not have been possible without the support and encouragement of many people at ETH, outside the ETH, and abroad. I am thankful to my family, my mother Rosalia and my sister Cetty, my wife Patrizia and my daughter Laura, who had the patience to support my dedication to this research.

My gratitude goes to my advisor Paul Tackley, who gave me the opportunity to work at ETH thus allowing me to get out from a difficult moment of my life, and my collaborators of research (Dave May, Taras Gerya and Guizhi Zhu), who provided and assisted me with their codes.

My gratitude goes to Nick Lang, the only one who accepted to be the external co-referee without prejudice on my Mars , and who reviewed my paper on with open mind.

Many thanks to Tim Eglinton for chairing my defense and for his much appreciated interest in my new ideas of research.

I am grateful to my former supervisor at Lancaster University, Lionel Wilson. As editor of the Journal of Volcanology and Geothermal Research, he allowed the publication of my independent papers on Mars against the prejudice of some reviewers.

I am also grateful to Andrew Dombard who solicited my paper about the Southern Polar Giant Impact for publication on the Geophysical Research Letters.

Very many thanks to Joseph Doetsch who completed the software necessary to produce the crucial Figure 3 of my paper on the volcanic alignments of Mars.

Very many thanks also to Elisabeth Läderach, André , Monika Bolliger, and Siegrid Trindler for being always kind and supportive with me. Special mention to the always helpful and Italian speaking Alfredo Picariello, he was my shield from mad bureaucracy.

My gratitude to Uli Kastrup, Gillian Grün, and Bettina Gutbrodt for the organization of my talk (with flowers) about held last July 2, 2014, in the C60 room at Focus Terra.

A special thought goes to Elisabeth Huwyler for her much appreciated advice and support in my job search and to Martina for her advice in the social support documents.

My greatest appreciation and gratitude goes to my friend Monika Kohler who kindly supported me with translations in every matter outside the ETH domain.

I also appreciated very much the Friday friendly conversations with Jean-Pierre Burg, Stewart Greenhalgh, and many others during my stay at ETH.

3 Summary Summary ...... 4

Abstract ...... 6 Sommario...... 6 Chapter 1 ...... 7 The Martian dichotomy and the volcanic history of Mars...... 7 1. The Martian dichotomy ...... 8 1.1 Connection with the current state of astronomical research...... 14 1.2 The volcanic history of Mars...... 16 1.3 Modification of the Martian dichotomy...... 18

Chapter 2 ...... 23 Technical details of the numerical method ...... 23 2.1 Model and Numerical Method...... 24 2.2 The I3ELVIS code...... 24 2.3 The StagYY code...... 38 2.4 Transferring composition and temperature from I3ELVIS to StagYY...... 41 2.5 Model parameters ...... 42

Chapter 3 ...... 45 Three-dimensional simulations of the southern polar giant impact hypothesis for the origin of the Martian dichotomy ...... 45 3.1 Introduction...... 47 3.2 Method...... 48 3.3 Results ...... 51 3.4 Discussion ...... 58 3.5 Conclusions ...... 59

Chapter 4 ...... 62 Alignments of Volcanic Features in the Southern Hemisphere of Mars produced by Migrating Mantle Plumes ...... 62 4.1 Introduction...... 64 4.2 Criteria for the identification and mapping of the volcanic features ...... 66 4.3 Description and interpretation of the alignments...... 89 4.3.1 Alignment 1...... 94 4.3.2 Alignment 2...... 97 4.3.3 Alignment 3...... 103 4.3.4 Alignment 4...... 108 4.3.5 Alignment 5...... 112 4.3.6 Alignment 6...... 120 4.3.7 Alignment 7...... 123 4.3.8 Alignment 8...... 128 4.3.9 Alignment 9...... 138 4.3.10 Alignment 10...... 143 4.3.11 Alignment 11...... 145 4.3.12 Alignment 12 ...... 148 4.4 Unaligned volcanic features ...... 152 4.5 Discussion ...... 153 4.5.1 The alignments ...... 153 4.5.2 The ages...... 157 4.6 Conclusions...... 162

4 Chapter 5 ...... 167 A network of tubes as the origin of Labyrinthus Noctis and Valles Marineris on Mars...... 167 5.1 Introduction...... 169 5.2 An overview of processes for the formation of the pit chains ...... 181 5.2.1 Magma-volatile interaction...... 183 5.2.2 Dike intrusions ...... 185 5.2.3 Karst dissolution...... 188 5.2.4 Dilational faulting ...... 190 5.2.5 Lava tubes...... 190 5.3 Lava vs. water for the formation of the ...... 191 5.4 A geological comparison between and Elysium volcanic provinces ...... 195 5.4.1 ...... 195 5.4.2 Elysium ...... 196 5.4.3 Circum- outflow channel mouths ...... 197 5.4.4 Valles Marineris ...... 198 5.5 Discussion ...... 201 5.6 Conclusion...... 205

Chapter 6 ...... 207 General overview and conclusions ...... 207 6.1 General overview and conclusions...... 208 6.1.1 Formation of the Martian dichotomy...... 208 6.1.2 Volcanism in the history of Mars ...... 209 6.1.3 The absence of Martian ...... 211 6.1.4 The extreme scarcity of ...... 212

References ...... 216 Giovanni Leone ...... 249 Education ...... 249 Additional training: ...... 250 - USGS, Astrogeology Center Flagstaff, Training Course with HiRISE imagery on ISIS – SOCET SET and ArcGIS software for Digital Terrain Models production, Sep 2013...... 250 - ETH Zurich, Training Course of Computational Magma Geodynamics, Feb 2014; Learning to Teach, May 2016...... 250 Theses ...... 250 Research Experience ...... 250 University Committees ...... 251 Teaching Experience ...... 251 Awards & Grants ...... 252 Skills & Activities ...... 252 Science Outreach ...... 253 Publication Highlights ...... 255 Book Chapters ...... 255 Peer Reviewed Journal Publications ...... 255 Conference Proceedings ...... 256

5 Abstract Until the first successful 4 mission, many speculated about the possible presence of liquid water on Mars. Many interpretations were just the result of optical illusions due to the poor resolution of the available images. However, this speculation survives even after that the Mars Reconnaissance Orbiter and several landers have returned the images of probably the most volcanic body of the Solar System after the and Io, which are still both active. The Martian dichotomy, the topographic difference between the southern highlands and the northern lowlands, and the distribution of the volcanoes according to twelve alignments on the southern highlands, suggest that some peculiar process must have shaped the during its history. Among the various hypotheses proposed, the impact with a lunar sized object on the Pole best explains the formation of the Martian dichotomy and the alignments of the volcanic structures observed on the surface of Mars. A series of 3-D simulations has shown that the Southern Polar Giant Impact (SPGI) melted the southern hemisphere of Mars into a large magma ocean that subsequently formed the southern highlands upon cooling. The thermal anomaly produced by the impact triggered mantle , with the formation of several mantle plumes migrating from the Equator to the South Pole of the planet, and the generation of a transient magnetic field that waned after the heat flux from the core was unable to maintain it. The heat flux continued its decline but still was able to support the volcanism until it reached the current level 3.5 Ga ago. The observation of lava flows inside the outflow channels like Valles Marineris and unaltered olivine in old craters, where water should have flowed and/or ponded for long time, casts several doubts on the survival of liquid water on the surface of the planet. Perhaps Mars was never the warmer and wetter world that was thought in the last two decades.

Sommario Molti credevano nella possibile presenza di acqua allo stato liquido su Marte fino alla prima riuscita missione . Molte interpretazioni erano solo il risultato d’illusioni ottiche dovute alla scarsa risoluzione delle immagini disponibili. Tuttavia, questa credenza sopravvive ancora oggi dopo che la Mars Reconnaissance Orbiter e alcuni landers hanno restituito le immagini del corpo piú vulcanico del Sistema Solare dopo la Terra e Io, che sono ancora attualmente attivi. La dicotomia marziana, cioè la differenza topografica fra gli altipiani meridionali e le pianure settentrionali, e la distribuzione dei vulcani allineati secondo dodici allineamenti lungo gli altipiani meridionali, indica che qualche processo peculiare deve aver modellato il pianeta durante la sua storia. Fra le varie ipotesi proposte, l’impatto con un oggetto di dimensioni lunari al Polo Sud è quella che spiega meglio la formazione della dicotomia marziana e gli allineamenti delle strutture vulcaniche osservate sulla superficie di Marte. Una serie di simulazioni in 3-D ha mostrato come il Gigantesco Impatto Polare Meridionale (GIPM) fuse l’emisfero meridionale di Marte in un grande oceano di magma che formó gli altipiani dopo il raffreddamento. L’anomalia termica prodotta dall’impatto ha iniziato la convezione del mantello con la formazione di diversi pennacchi di magma, migranti dall’Equatore al Polo Sud del pianeta, e di un campo magnetico transitorio che diminuí d’intensitá quando il flusso di calore proveniente dal nucleo non poté piú sostenerlo. Il flusso di calore ha continuato il suo declino ma è riuscito a mantenere il vulcanismo attivo fino ai valori odierni raggiunti 3.5 miliardi di anni fa. L’osservazione di flussi lavici e olivina inalterata sia nei canali di deflusso come Valles Marineris che in vecchi crateri l’acqua dovrebbe aver risieduto per lungo tempo solleva diversi dubbi sulla sua esistenza in forma liquida. Marte non è mai stato un mondo simile alla Terra cosí come è stato immaginato negli ultimi due decenni.

6 Chapter 1

The Martian dichotomy and the volcanic history of Mars.

7 1. The Martian dichotomy

The Martian dichotomy is the marked topographic difference between the southern highlands and the northern lowlands of Mars. The boundary between the highlands and the lowlands is located in proximity of the equator of the planet but it is neither parallel to it nor direct, the lowlands are often embayed in the highlands (see Fig. 1.1) and are separated by a stripe of colour named “transition topography” (Leone, 2015). This boundary was suggested to be heavily modified by different processes, aeolian, volcanic, glacial, and fluvial, that will be introduced in section 1.3. However, recent interpretations suggested that volcanic processes had a dominant role in the early history of Mars, have a dominant role today, while glacial and fluvial processes were negligible or non-existent at the equator of the planet likely for the whole history of Mars (Leone, 2016). The best examples are the

Valles Marineris, the longest outflow channel of Mars, and Kasei Valles, the widest outflow channel of Mars, which will be both described in detail in Chapter 5 (Leone, 2014). I will also show how the onset of the volcanism of Mars is a direct consequence of the formation of the

Martian dichotomy (Leone, 2016).

Figure 1.1: Planisphere of Mars showing the Martian dichotomy. The southern highlands cover nearly two thirds of the planet’s surface.

8 The hypotheses of formation for the Martian dichotomy should include at least three requirements (McGill and Squyres, 1991):

1) It must account for the shape of the lowlands and the highlands.

2) It must be physically consistent.

3) It must be compatible with the available geological and geophysical data.

The earliest works put forward several hypotheses: first order (or degree-1) mantle convection that existed prior to core formation (Wise et al., 1979); crustal thinning after the end of primordial bombardment (Mcgill and Dimitriou, 1990; Mcgill and Squyres, 1991); a giant impact forming the Borealis basin (Wilhelms and Squyres, 1984); multiple large (but not giant) overlapping impacts shaping the lowlands (Frey and Schultz, 1988).

Figure 1.2: Image taken from Andrews-Hanna et al. 2008 a) they speculate an interpolation of the dichotomy boundary below Tharsis and d) proposed an elliptical best fit for the Borealis basin, then e) interpreted the topographic gradation of the transitional terrains in as multiring impact features which are not present elsewhere around the Borealis basin.

9

Figure 1.3: Degree-1 mantle convection. Image taken from Keller and Tackley (2009). This process requires a timescale too long for the early formation of the Martian dichotomy and the resulting magmatic activity would change the initial dichotomy structure.

Later works have tried to elaborate on the previous hypotheses and suggested that the

Martian dichotomy may have been generated by a large oblique impact occurred on the northern hemisphere of the planet which formed the Borealis basin (Andrews-Hanna et al.,

2008; Marinova et al., 2008; Nimmo et al., 2008) (Fig. 1.2) or by endogenic processes like degree-1 mantle convection (Keller and Tackley, 2009) (Fig.1.3) which formed the southern highlands instead. However, also these models suffer of several problems and do not match all the above mentioned requirements: a) impacts so large create magma oceans and crustal flows (Reese and Solomatov, 2006) that are inconsistent with the topography, lack of significant volcanism, and the cratering rate of the lowlands (Frey, 2008); b) a set of results for several combinations of impact energies and angles (Marinova et al., 2008) is subject to the same objections; d) another simulation (Nimmo et al., 2008), trying to fit the Borealis basin with a vertical impact in 2-D through the same 1029 J energy requirements of Wilhelms and Squyres (1984), even placed the Martian dichotomy at the same age of the forming event around 100 Ma after the Calcium-Aluminum-Inclusions (henceforth CAI). As a brief note, the CAI have been found in the carbonaceous chondrite meteorites and thus are considered the oldest minerals known in the Solar System, formed between 4.567-4.571 Ga

(Amelin et al., 2002). The degree-1 mantle convection may indeed produce crustal thickening

10 consistent with the Martian dichotomy but only in particular conditions of sharp jumps in viscosity (Keller and Tackley, 2009) it is possible to obtain a dichotomy in slightly less than

100 Ma (Roberts and Zhong, 2006). However, this is still a process too long in a timescale that places the dichotomy formation early in Martian history (around 4.565-4.568 Ga or 2-3

Ma after CAI according to the results of my simulations described in Section 5), before the core completed its formation, as also evidenced by studies on 182Hf and 182W isotopic anomalies in the inner Solar System (Jacobsen, 2005; Kleine et al., 2002, 2004; Nimmo and

Kleine, 2007; Righter and Shearer, 2003; Solomon et al., 2005; Yin et al., 2002). The magnetic anomalies discovered by the (MGS) mission (Acuna et al.,

1999; Acuna et al., 2001), mostly concentrated in the southern highlands but also present in the lowlands and absent within the Hellas basin, thought to be formed 4.1 Ga ago (Lillis et al.,

2008) (Fig. 1.4), indicate that the magnetic field of the planet was already over at an earlier time than the estimates of half billion years given by Whaler and Purucker (2005) and

Williams and Nimmo (2004).

Figure 1.4: Magnetic map of Mars in polar (left) and cylindrical (right) projections, taken from Lillis et al. (2008). Magnetic anomalies are stronger in the southern highlands (, 195°W) but are also present in the northern lowlands (North projection). A big impact like the one that would have generated the Borealis basin should have erased any trace of magnetic anomalies in the crust as well as occurred in smaller impact basins (white circles) present in the lowlands (, , ) and in very old highlands basins (, ).

11 The quasi-circular depressions (QCD’s) present in the lowlands (Frey, 2006) also suggest a similar crustal age for both hemispheres (Fig. 1.5), although future further discovery of buried craters in the lowlands may suggest an older age, which are estimated to start formation during or immediately after the planet’s accretion so the crustal dichotomy is thought to be the most ancient feature on the planet (Carr and Head, 2010; Solomon et al.,

2005).

Figure 1.5: QCDs on Mars > 200 km in diameter over coloured MOLA topography (taken from Frey 2006). In blue are the lowlands, in red are the highlands. Quoting Frey, 2006 “Solid circles show “visible” (on images) features known to be impact basins. Dashed circles represent features not visible on images and are believed to be buried impact basins. (a) Equatorial views at 60°W, 300°W, and 180°W. (b) Polar views. Buried features outnumber visible basins by a factor 6 in the highlands and a factor 20 in the lowlands. Note the greater density of QCDs in the Southern Hemisphere, corresponding to the cratered highlands.”

Thus, another explanation must be found for the formation of the Martian dichotomy.

A hybrid model combining both endogenic and exogenic processes is potentially able to give a satisfactory explanation to the various problems of the single approach models. Previous authors (Reese et al., 2011; Reese and Solomatov, 2006) have suggested that a giant impact with a non-compositionally specified body between 0.1 – 1.0 lunar masses (≈ 800 -1700 km radius) may have formed and shaped the Martian dichotomy. The southern polar giant impact

(SPGI), which is the main focus of this thesis, melted the crust to form a hemispherical magma ocean and induced a thermal anomaly inside Mars (Reese et al., 2011; Reese and

12 Solomatov, 2006) rather than one in the northern hemisphere that created the Borealis basin.

Then, the cooling of such a large magma ocean formed the southern highlands uniformly thickening the crust of Mars. Such a single mechanism hypothesis of formation estimates a crust ~ 58 km thick below the highlands and ~ 32 km below the northern lowlands, figures consistent to the combined analysis of MOLA topography and MGS data ( et al., 2004).

A 2-D study of impact simulations obtained a hemispherical global magma ocean on Mars has been obtained with a mesosiderite composition -core (54% stone/silicates and 46% iron along radius) impactor of ~ 1000 km of radius (Golabek et al., 2011). Depending on the composition of the impactor, particularly important would be the percentage of iron in terms of radius, the radius of the impactor might be expected to give a significant contribution to the radius of the Martian core, estimated between 1600-1800 km by several studies of the Martian moment of inertia (Konopliv et al., 2006; Zharkov et al., 2009), thus raising important implications on the shortest timescale (3 Ma) required by the geochemistry of core formation and mantle differentiation given by Jacobsen (2005). Although the previous 2-D study

(Golabek et al., 2011) has obtained a mantle plume migration from the southern polar region to a sub-equatorial region roughly corresponding to Thaumasia (Hynek et al., 2011), it will be shown ahead in Chapters 3 and 4 that a plume migration from Thaumasia to Tharsis did not happen. This model was unable to fully explore the migration of the plumes along their paths due to the lack of the third dimension.

Iron diapirs, already present in the planet from previous meteoric contributions or as part of the initial composition, started sinking as liquid iron-alloys along cracks opened under the action of the gravity to contribute to the core formation (Stevenson, 2003), exactly in the opposite way as magma diapirs rise under the forces of buoyancy in a two phase flow (molten iron + solid silicate) in the host rock. A mechanism enhanced by the development of vertical

13 channels propagating from the diapirs of molten iron. Such a mechanism was however affected in the impacted area by the devastating effect of the iron core of the impactor in a way that will be shown ahead by the results of the first simulations. This scenario is also consistent with the core formation and onset of magnetic field processes on Mars (Stevenson,

2001) and implications on core formation inferred by experiments on the dihedral angle between liquid iron-alloy and crystalline silicates (Terasaki et al., 2005).

1.1 Connection with the current state of astronomical research

N-body simulations that studied the accretion time of the terrestrial (Kokubo and

Ida, 2000; Morishima et al., 2010; Thommes et al., 2008) generally agree with the formation of Lunar to Martian sized bodies in the inner Solar System within a timescale below 1-12 Ma after CAI, although there is a disagreement on the completion of the accretion process, as it will be shown below. Considering the uncertainties on the estimate of the CAI formation time

(Gilmour, 2002) and the uncertainties in the N-body simulations, the impactor of largest mass, say 1022 kg (the lunar mass is of the order of 1022 kg), that may have generated a large magma ocean on a Martian scale was available early in the history of the Solar System with starting timescales from 0.5 Ma (Kokubo and Ida, 2000), well within the limits placed in past studies, from lower limit 105-106 years (Tonks and Melosh, 1993) to upper limit 108 years

(Chambers and Wetherill, 1998), while other N-body simulation models place this timescale around 12 – 20 Ma (Morishima et al., 2010; Thommes et al., 2008).

Due to the presence of M-type like 16 Psyche as well as several others in the belt (Ockert- et al., 2010), the likely remnants of larger parent bodies in the 1-2

AU range which then migrated in the current position after giant impacts with protoplanets

(Goldstein et al., 2009), S-type and M-type end members impactors of various sizes (thus not only 50%-50% iron-silicate composition) have been explored. Indeed, some M-type asteroids

14 have been reported to contain various amounts of silicate (Fornasier et al., 2010; Hardersen et al., 2005; Shepard et al., 2010). Spectral and radar analysis of M-type asteroids has shown that seven of them have a good match with the 81% NiFe and 16% silicate composition of the

Landes iron meteorite (Bunch et al., 1972; Ockert-Bell et al., 2010).

Now, considering also that for impactor/target average mass ratio of 1:5 there is an accretion probability around 60% (Leinhardt and Richardson, 2002) and even assuming that a

Moon-sized impactor hit Mars, the accretion probability is higher because the mass ratio in the Moon-Mars case is 1:8.73. When the impactor/target mass ratio is less than unity the angle of impact has less effect (Leinhardt and Richardson, 2002; Teiser et al., 2011) and in the case of a Moon-sized impactor colliding with Mars the mass ratio is 0.11. A situation comparable to the 30-65% planet surface melting in case of impactor/target mass ratio of 0.14 and impact speed of 15 km s-1 or total melting if the ratio is > 0.4 at the same speed (Tonks and Melosh, 1993). N-body simulations have also provided the interesting result that the impactors’ velocity is statistically around the escape velocity of the target body (Agnor et al.,

1999), 5 km s-1 in the case of Mars. Thus, according to these figures, even a Moon-sized impactor could not be considered an upper limit for eventual simulations, although there are constraints on the core radius estimated through the moment of inertia of Mars and on the energy provided by the impact with such large bodies (Zharkov et al., 2009). All the above considerations decrease somehow the uncertainties on the impact speed and angle at least.

Scaling laws may provide a useful tool for calculating differential melt in oblique impacts

(Abramov et al., 2012).

15 1.2 The volcanic history of Mars

The largest volcanoes of Mars are also the largest of the Solar System and are mainly located in the Tharsis and Elysium regions, astride the equator of the planet. The research done in this thesis will show how the volcanism of Mars is linked to the SPGI that formed the

Martian dichotomy. According to the simulations, the formation of the first mantle plumes that emplaced and fed the first volcanoes could be placed in time as early as ~ 50 Ma after the SPGI (an absolute age of ~ 4.517 Ga ago). The first plumes formed in four points spaced 90 degrees apart around the equator of Mars and stationed for ~ 10-20 Ma before starting to move from their original points of formation. The distribution of all the volcanic centres observed on the surface of Mars shows that these plumes then migrated from the equator to the south pole of Mars forming volcanoes aligned according to loxodrome (aka rhumb lines) trajectories that will be described in detail in Chapter 4. The average time for a complete migration from the equator to the south pole for a plume was ~ 200 Ma with a total variation between ~ 100 and ~ 300 Ma bringing the average time of the first emplaced volcanoes back between 4.5 and 4.2 Ga ago. The SPGI model will show three main episodes of plume formation that characterised the first 500 Ma of the history of Mars (4.5-4.0 Ga) followed by a decline that did not form significant volcanism for other 500 Ma (4.0-3.5 Ga).

This decline likely caused the shut down of the magnetic field of Mars, occurred around 4.15-

4.07 Ga ago (Lillis et al., 2008).

Although the geological history of Mars was reconstructed mainly through the crater counts methods and the geochemical analysis of the Martian meteorites, recent in situ K-Ar measurements coupled with geomorphologic analysis provided new tools that constrained the ages estimated so far. The lander determined a K-Ar exposure age of 4.21 Ga at its landing site within the crater floor that is thus obviously older than the 3.7-3.5 estimated from the crater counts (Farley et al., 2014). The geomorphologic observation of the surface of

16 Mars has also shown that the Gale crater was filled by lava flows coming from Tyrrhenum

Mons (Leone, 2016). The crater counts that can be found in the literature for Tyrrhenum

Mons provided ages ranging from 3.7 to 3.4 Ga (Werner, 2009; Williams et al., 2009;

Robbins et al., 2011), somehow consistent with the formation of Gale crater. However, the K-

Ar measurement showed at the same time an older age for the filling of Gale crater, showed a clear inconsistency with the age estimated through the crater counts, and provided a constrain for the age of the activity of Tyrrhenum Mons as well. Also the geographic position of

Tyrrhenum Mons has important implications for the ages of the volcanism on Mars. It will be seen in Chapter 4 that Tyrrhenum Mons is located in a mid-advanced position along the tracks of migration of two mantle plumes so that its age, now 4.21 Ga according to the K-Ar measurements of its flows that reached Gale, is not only consistent with the ages inferred from the SPGI but can also be considered a later formed with respect to those located at higher around the equator. This means that larger volcanoes like ,

Elysium Mons, and Alba Patera could be even older despite the younger age estimated by the crater counts. Another profound implication of the link between Tyrrhenum Mons and Gale crater is related to the habitability of the . The craters reached by the landers

(i.e. , Gale) have shown lava filling that would make their floors very far from being habitable. Even a later filling by atmospheric or groundwater is very unlikely considering the simultaneous presence of unaltered olivine within the (i.e. McSween et al., 2006) of the old age of formation inferred from the K-Ar measurements. Eventual presence of water for periods between 100 and 10k years would the olivine into serpentine (Oze and

Sharma, 2007).

After 3.5 Ga ago the volcanic activity of the planet should not have had any significant episode, despite the wealth of crater counts that estimated otherwise, the studies in this thesis

17 will show well grounded reasons to think that the crater counts have margins of errors too large to be still considered a reliable method for age estimation.

1.3 Modification of the Martian dichotomy

The Martian dichotomy shows today the result of the various processes that modified it from the original formation. The processes suggested in the literature were impact, aeolian, volcanic, glacial, fluvial, and even marine. However, only the first three processes are incontrovertibly supported by the most updated observations and it is interesting to understand which one of them had a dominant role in the modifications and when occurred.

The aeolian processes were suggested as the current dominant agent of modifications in the absence of liquid water and lack of active volcanism on Mars, although it is difficult to reconstruct their role to the past because the techniques of extrapolation for wind abrasion rates are not well developed even on Earth ( et al., 2001). However, from the geomorphological and stratigraphic observation of the surface of Mars, it could be said with a good margin of certainty that the aeolian activity followed the volcanism in space and in time.

Impact processes were coeval and dominant together with volcanic processes in the early history of Mars (4.5-4.0 Ga) with extensive formation of large basins filled by lava in the main volcanic provinces of Mars (Leone, 2016). This view is confirmed (Carr and Head,

2010), although with substantial differences related to the crater count ages and to the now controversial role of water throughout the whole geologic history of Mars. The main outflow channels, the fluvial valleys, and even the fretted terrains of the dichotomy boundary observed around the equatorial regions of Mars were ascribed to the action of water (Parker et al., 1989). Even according to the mineralogical history of Mars reconstructed through

OMEGA data, the temporal sequence of the activities is volcanic → fluvial → glacial → aeolian (Bibring et al., 2006). However, the low pressure (~ 6 millibar) of the Martian

18 atmosphere, likely not changed during the Martian history (Gillmann et al., 2011), does not support the activities of water. Among the many other observations that do not support the glacial (and even more aqueous) activity of Mars is the sublimation of (dry) ice in just 4 sols that was observed at the Phoenix landing site ( et al., 2009). Other observations will be discussed in Chapter 4.

Some impact basins formed large bays along the dichotomy boundary (Frey, 2008).

These are Isidis, Chryse, Utopia and Amazonis. The superposition relationships at global scale seen on the map of Mars provide compelling evidence that these impacts occurred before the emplacement of the late stage volcanism along the boundary, a volcanism also estimated as early so far (e.g. Ivanov et al., 2012) although it will be shown how the estimates of absolute ages are affected by large uncertainties in the crater counts methods

(Robbins et al., 2013). The Isidis basin has even obliterated the transition topography and part of the highlands forming a large embayment of lowlands into the highlands domain. Also significant are the large basins of Hellas and Argyre that formed islands of lowlands inside the highlands.

Yardangs, wind eroded ridges typical of regions where winds blow constantly in the same direction all over the year (Ward and Greeley, 1984), they generally form fields along the transition topography in the Aeolis and Ares regions of Mars but they also occur in the

Iapygia region (Ward, 1979) and on the (Bridges et al., 2010) in the highlands

(Mandt and Leone, 2015). Yardangs form on friable and thus easily erodible volcanic terrains by the wind like ignimbrite deposits (Kerber et al., 2012; Mandt et al., 2008; Scott and

Tanaka, 1982). Fields of yardangs are also reported in the volcanic material of the Medusae

Fossae Formation (MFF) also located along the transition topography (Zimbelman and

Griffin, 2010). There is no report of yardangs from the northern lowlands of Mars, where volcanism is scarce.

19 The largest outflow channels are the Marineris-Shalbatana-Simud Valles system

(Kereszturi, 2010; Weitz et al., 2010), the (Pacifici et al., 2009), and Kasei Valles

( et al., 2010), all debouching in Chryse Planitia. These appeared to be the main routes for water from the highlands to the lowlands along with other fluvial valleys present along the transition topography. Among these: (Jaeger et al., 2007), Mawrth

Vallis (Loizeau et al., 2012), Ma’adim Vallis (Cabrol et al., 1998), and Al-Qahira Valles

(Penido et al., 2013), all discharging into , and discharging into . Outflow channels are also located within the highlands, although discharging into the Hellas basin. These are the Dao-Niger Valles system, Reull and

Harmakhim Valles (Glamoclija et al., 2011; Kostama et al., 2010). Another outflow channel in the highlands, ending into Argyre, is the Uzboi-Ladon-Morava system (Grant et al., 2011).

A few outflow channels are in the lowlands, Hrad and . Although water was suggested for the formation and the evolution of the outflow channels, many problems with both presence and stability of water on the surface of Mars were pointed out (Leverington,

2011; Leone, 2014).

Fretted terrains were interpreted as remnants of highlands eroded by fluvial activity and then modified by mass wasting following glacial erosion (Carr, 2001; Irwin et al., 2004;

Squyres, 1978). Fretted terrains are found along the mensae and in chaos formations in the transition topography (Nilosyrtis, Protonilus, Deuteronilus, , Hydraotes,

Aureum, etc.) but also in the highlands (East Labyrinthus Noctis, Gorgonum

Chaos) and in the lowlands (Acidalia Colles, Scandia Cavi, Galaxias Chaos, and an area northeast of in ) in regions where geothermal activity was the most plausible explanation (Chapman and Tanaka, 2002) suggesting that shallow intrusions might be involved in their formation (Head and Wilson, 2007), although not evolving in

20 effusive activity. However, this view was dismantled by recent observations based on high- resolution imagery coming from the Mars Reconnaissance Orbiter (Leone, 2014; 2016).

Lineated valley fills (LVF) and lobate debris aprons (LDA) were interpreted as the result of the glacial processes that modified the Martian dichotomy boundary in later epochs, mostly Amazonian (Morgan et al., 2009). Although the same processes that formed the fretted

(valleys) terrains have been invoked for the LVF (Carr, 1996), this hypothesis gained ground from the observation of terrestrial analogs (Marchant and Head, 2007). Although without direct observations other than the deposition of the brines at the Viking landing sites, it was thought that snow and ice were deposited everywhere in the equatorial and tropical regions during periods of high obliquity to become again stable at higher latitudes during low obliquities (Laskar et al., 2004). The erosional processes related to the transport and deposition of the ice were thus thought to be involved in the LVF and LDA formation observed at (Morgan et al., 2009). However, following observations at the Phoenix landing sites (Smith et al., 2009), it will be shown how the deposition of the brines has negligible erosional effects (Leone, 2016). Radar experiments mentioned buried in Arabia Terra and Tyrrhenum volcanic areas (Karlsson et al., 2015). It is just a pity that the team used a dielectric constant for water very similar to that of dacitic lava in their experiment so all that they might have discovered is just layers of lava flows piled one on top of the other, certainly not difficult to find in volcanic areas.

Even paleoshorelines have been interpreted in Chryse Planitia/Arabia Terra (Ghatan and Zimbelman, 2006; Webb, 2004), Deuteronilus Mensae (Parker et al., 1989), Nepenthes

Mensae (de Pablo and Pacifici, 2008), and Isidis Planitia/ (Erkeling et al., 2012) as the traces of the changing level of a putative ancient ocean supposed to fill the northern lowlands of Mars. This ocean then was thought to have disappeared raising several issues on its fate (Baker et al., 1991; Carr and Head, 2003; Clifford and Parker, 2001; Di Achille and

21 Hynek, 2010; Head et al., 1999). The interpreted paleoshorelines, from the maximum (and oldest) to the minimum (and youngest) level, are: Meridiani, Arabia, Ismenius and Aeolis,

Elysium, Mamers Valles, Deuteronilus, and Acidalia (Clifford and Parker, 2001). It was suggested from the position of the Meridiani paleoshoreline that the Martian ocean invaded even the highlands at its maximum level (Clifford and Parker, 2001). Villanueva et al (2015) estimated with great accuracy that this putative ocean filled the lowlands with 137 m of global equivalent layer (GEL) of water reaching the shores of the highlands during the

Noachian. Previous estimates after Mars Odyssey were only of 36 m (Christensen, 2006). But now the main and most critical question arises. Why the olivine deposited in the lowlands by the volcanic activity never became serpentine at the contact with the water of the putative ocean? This transformation should take only one hundred to 10’000 years, a very short time on geological scales. An alternative interpretation is that the paleoshorelines may simply be lava flow fronts (Erkeling et al., 2015).

22 Chapter 2

Technical details of the numerical method

23 2.1 Model and Numerical Method

The equations, model setup and numerical solution method are basically the same as in

Golabek et al. (2011) except in three dimensions (3D) rather than two dimensions (2D). We here give information that is specific to the present 3D experiments, and list all the parameters used. As a general overview, the I3ELVIS code is used to simulate the period between the impact and the end of core formation, then the thermal and compositional fields are transferred to the StagYY code, which is used to simulate Mars' long-term evolution to the present day.

2.2 The I3ELVIS code

The initial impact and post-impact 3D model setup is based on the simplified approach described in Golabek et al. (2011). For a given density distribution, the I3ELVIS code solves the following equations describing the conservation of mass, momentum, energy, and gravitional potential equation (Gerya, 2010).

The momentum equation, describing the balance of all the internal and external forces acting on a continuous medium, is given by

′ + = 0 (2.1) 𝜕𝜕𝜕𝜕𝑖𝑖𝑖𝑖 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕𝑗𝑗 − 𝜕𝜕𝜕𝜕𝑖𝑖 𝜌𝜌𝑔𝑔𝑖𝑖

where is the deviatoric stress, i and j are coordinates indices, and are spatial ′ 𝑖𝑖𝑖𝑖 𝑖𝑖 𝑗𝑗 coordinates, P𝜎𝜎 is the pressure, is the density, and is the i-th component𝑥𝑥 of𝑥𝑥 the gravity

𝑖𝑖 vector = ( , , ). 𝜌𝜌 𝑔𝑔

𝑔𝑔⃗ 𝑔𝑔𝑥𝑥 𝑔𝑔𝑦𝑦 𝑔𝑔𝑧𝑧

24 The extended Boussinesq approximation assumes the incompressibility of the fluid, in such a case a simplified constitutive law can be used:

= 2 (2.2) ′ 𝑖𝑖𝑖𝑖 𝚤𝚤𝚤𝚤 𝜎𝜎 𝜂𝜂𝜀𝜀 ̇

where is the shear viscosity and is the strain rate tensor.

𝑖𝑖𝑖𝑖 The x momentum𝜂𝜂 for the can be further𝜀𝜀̇ simplified to

x-Stokes (2 ) + 2 + (2 ) = (2.3) 𝜕𝜕 𝜕𝜕 𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜂𝜂𝜀𝜀𝑥𝑥𝑥𝑥̇ 𝜕𝜕𝜕𝜕 � 𝜂𝜂𝜀𝜀𝑥𝑥𝑥𝑥̇ � 𝜕𝜕𝜕𝜕 𝜂𝜂𝜀𝜀𝑥𝑥𝑥𝑥̇ − 𝜕𝜕𝜕𝜕 − 𝜌𝜌𝑔𝑔𝑥𝑥

and analogously for y, and z momentum equations. In the above, epsilon dot is the strain-rate tensor defined also as

= + (2.4) 1 𝜕𝜕𝑣𝑣𝑖𝑖 𝜕𝜕𝑣𝑣𝑗𝑗 𝜀𝜀𝚤𝚤𝚤𝚤̇ 2 �𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖�

where i and j are coordinate indices and is the spatial coordinate.

𝑖𝑖 The conservation of mass is given by𝑥𝑥

+ ( ) = 0 (2.5) 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 ∇ ∙ 𝜌𝜌𝑣𝑣⃑

The conservation of energy is derived from assuming that energy is conserved in a continuum and relates temperature changes to internal heat generation, which relates to the heat flux q to the gradient of temperature according to

25

= or = (2.6) 𝜕𝜕𝜕𝜕 𝑞𝑞⃗ −𝑘𝑘∇𝑇𝑇 𝑞𝑞𝑖𝑖 −𝑘𝑘 𝜕𝜕𝜕𝜕𝑖𝑖

where is the thermal conductivity and T is the temperature in °K. In its 3D Lagrangian form to balance𝑘𝑘 heat in a continuum and temperature changes due to generation of heat, is given by

x y z p = + (2.7) 𝐷𝐷𝐷𝐷 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝐷𝐷𝐷𝐷 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜌𝜌𝜌𝜌 − − − 𝐻𝐻

3 2 where r is the density (kg/m ), is the heat flux in its 3 components (W/m ), Cp is the heat capacity at constant pressure (J/kg/K)𝑞𝑞⃗ and H is the volumetric heat production (W/m3) which is decomposed into four terms, H = Hs+Ha+Hl+Hr. The shear heating term (Hs) given by

= (2.8) ′ 𝑠𝑠 𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝐻𝐻 𝜎𝜎 𝜀𝜀̇ and the term (Ha) given by

adiabatic heating = = + + (2.9) 𝐷𝐷𝐷𝐷 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝐻𝐻𝑎𝑎 𝛼𝛼𝛼𝛼 𝐷𝐷𝐷𝐷 𝛼𝛼𝛼𝛼 �𝜕𝜕𝜕𝜕 𝑣𝑣𝑥𝑥 𝜕𝜕𝜕𝜕 𝑣𝑣𝑦𝑦 𝜕𝜕𝜕𝜕 𝑣𝑣𝑧𝑧�

where is the deviatoric stress and is the strain rate (defined in Eq. 2.4), a is the thermal ′ 𝜎𝜎𝑖𝑖𝑖𝑖 𝜀𝜀𝑖𝑖𝑖𝑖̇ expansion coefficient, T is the temperature, P is the pressure, vx, vy and vz are the velocity components in the x,y,z direction respectively. The latent heat (Hl) depends on minerals transformations (melting or solidifications), in our model a value of 400 kJ kg-1 is assumed, latent heat of 200 kJ kg-1 for pure iron. In the model presented here, I considered time- dependent radiogenic heating

26

= exp( ) (2.10)

𝑟𝑟 0 𝐻𝐻 𝐻𝐻 −𝑡𝑡𝑡𝑡 where is the initial abundance, t is time, and l is the initial heating rate taken from a

0 review 𝐻𝐻based on earlier works that summarize radiogenic heating rates in the early Solar

System CI chondrite. The values used are those of 26Al, 60Fe, 235U, and 238U taken from the

Table 3 of Barr and Canup (2008).

The gravity vector is determined by the gravitional potential equation

= 4 ( , , ) (2.11) 2 𝑥𝑥 𝑦𝑦 𝑧𝑧 ∇ 𝜙𝜙 𝜋𝜋𝜋𝜋𝜌𝜌 where G is the and r is the density varying along the three dimensions.

The visco-plastic rheology used in the modelling is obtained through the combination of viscous and plastic rheological relationships under brittle/ductile transition physical conditions where solid state creep influences the rheology of the rocks as it is the main mechanism of deformation of the rocks under applied stress. Solid-state creep is the ability of crystalline substances to deform irreversibly under applied stress. There are different kinds of solid-state creep: grain size dependent diffusion creep and grain size independent dislocation creep. Stress is not much important in the former but it is important in the latter. The visco- plastic rheological model implies strong variations of viscosity with constant low viscosity

(1018 Pa s) of iron and molten silicate and high viscosity of solid silicate with temperature and strain rate dependent effective viscosity (heff), which will be defined and shown shortly ahead, computed according to experimentally determined flow laws. I use the dry olivine flow law for the mantle and plagioclase (An75) flow law for the crust (Ranalli, 1995).

27 The total deviatoric strain rate , which describes the deformation of the rocks, is

𝑖𝑖𝑖𝑖 defined as 𝜀𝜀̇′

= ( ) + ( ) (2.12)

𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝜀𝜀̇′ 𝜀𝜀̇′ 𝜀𝜀̇′ where ( ) = (2.13) 1 ′ 𝜀𝜀̇′𝑖𝑖𝑖𝑖 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 2𝜂𝜂 𝜎𝜎𝑖𝑖𝑖𝑖 and ( ) = 0 for < (2.14)

𝜀𝜀̇′𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝜎𝜎𝐼𝐼𝐼𝐼 𝜎𝜎𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 or ( ) = = ′ for = (2.15) 𝜕𝜕𝐺𝐺𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝜎𝜎𝑖𝑖𝑖𝑖 ′ 𝜀𝜀̇′𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡 𝜒𝜒 𝜕𝜕𝜎𝜎𝑖𝑖𝑖𝑖 𝜒𝜒 2𝜎𝜎𝐼𝐼𝐼𝐼 𝜎𝜎𝐼𝐼𝐼𝐼 𝜎𝜎𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦

= ′2 (2.16) 𝜎𝜎𝑖𝑖𝑖𝑖 𝜎𝜎𝐼𝐼𝐼𝐼 � 2

where is the deviatoric stress component, is the plastic flow potential, is the ′ 𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝐼𝐼𝐼𝐼 second𝜎𝜎 invariant of the stress and is the plastic𝐺𝐺 multiplier. The plastic multiplier𝜎𝜎 is a variable scaling coefficient that connects𝜒𝜒 components of the plastic strain rate to the local stress distribution in places where the condition = is reached, this coefficient is

𝐼𝐼𝐼𝐼 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 determined solving Eqs. 12.45-12.51 available in 𝜎𝜎Gerya𝜎𝜎 (2010) for known values of stress, strain rate, viscosity, and shear modulus. Effective creep viscosity, which depends on the effect of diffusion and dislocation creeps (Ranalli, 1995), is given by (Liao and Gerya, 2015)

= (2.17) 1 𝜂𝜂 1 1 �𝜂𝜂𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑+𝜂𝜂𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑� where is

𝜂𝜂𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = exp (2.18) 1 1−𝑛𝑛 𝐸𝐸𝑎𝑎+𝑃𝑃𝑃𝑃𝑎𝑎 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 2 𝐷𝐷 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑅𝑅𝑅𝑅 𝜂𝜂 𝐴𝐴 𝜎𝜎 � �

28 and is

𝜂𝜂𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 = 1 1−𝑛𝑛exp (2.19) 1 𝑛𝑛 𝑛𝑛 𝐸𝐸𝑎𝑎+𝑃𝑃𝑃𝑃𝑎𝑎 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 2 𝐷𝐷 𝐼𝐼𝐼𝐼 𝑛𝑛𝑛𝑛𝑛𝑛 𝜂𝜂 𝐴𝐴 𝜀𝜀̇ � � where P is the pressure in Pa, is a pre-exponential material constant (Pa-ns-1m-m), n is the

𝐷𝐷 stress/strain rate exponent, is𝐴𝐴 the activation energy (kJ mol-1), is the activation volume

3 -1 4 𝑎𝑎 𝑎𝑎 (cm mol ), scrit = 10 Pa is𝐸𝐸 the transition stress from diffusion to 𝑉𝑉dislocation creep (Turcotte and Schubert, 2002), T is the temperature in °K, R is the gas constant (8.314 J K-1 mol-1), and

is the second invariant of strain rate defined analogously to Eq. 2.16. The plastic strength

𝐼𝐼𝐼𝐼 of𝜀𝜀̇ dry rocks strongly increases with pressure to a limit of several GPa. At elevated stresses

(typically superior to 0.1 GPa) dry Peierls creep (Katayama and Karato, 2008) for the mantle and crust is used (Gerya, 2010). The flow law then becomes

= exp 1 (2.20) 𝑑𝑑 𝑓𝑓 2 𝐸𝐸𝑎𝑎+𝑃𝑃𝑉𝑉𝑎𝑎 𝜎𝜎𝐼𝐼𝐼𝐼 𝜀𝜀𝐼𝐼𝐼𝐼̇ 𝐴𝐴𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝜎𝜎𝐼𝐼𝐼𝐼 �− 𝑅𝑅𝑅𝑅 � − �𝜎𝜎𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃� � �

-2 -1 where APeierls is the Peierls pre-exponential material constant (Pa s ), and are the

𝐼𝐼𝐼𝐼 𝐼𝐼𝐼𝐼 second invariant of the deviatoric stress and strain rate, respectively, the exponents𝜎𝜎 𝜀𝜀 ̇ 0 < d <1 and 1 < f < 2 depend on the geometry of the obstacles that limit the dislocation motion

(Gerya, 2010). The viscosity is then treated as effective viscosity and calculated through the relationship between and

𝐼𝐼𝐼𝐼 𝐼𝐼𝐼𝐼 𝜎𝜎 𝜀𝜀̇

= = exp 1 𝑓𝑓 (2.21) 𝜎𝜎𝐼𝐼𝐼𝐼 𝜎𝜎𝐼𝐼𝐼𝐼 𝐸𝐸𝑎𝑎+𝑃𝑃𝑉𝑉𝑎𝑎 𝜎𝜎𝐼𝐼𝐼𝐼 𝑑𝑑 2 𝜂𝜂𝑒𝑒𝑒𝑒𝑒𝑒 2𝜀𝜀̇ 𝐼𝐼𝐼𝐼 𝐴𝐴𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝜎𝜎𝐼𝐼𝐼𝐼 �� 𝑅𝑅𝑅𝑅 � ∙ � − �𝜎𝜎𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃� � �

29 The plastic yield strength of the rock depends on the stress applied on it.

𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 𝜎𝜎

= + sin( ) (2.22)

𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 𝜎𝜎 𝐶𝐶 𝜗𝜗 𝑃𝑃

where C is the cohesion or rock strength assumed to be 100 MPa at P=0, is the effective internal friction angle ( ( ) of 0.3 is used for the mantle and 0.2 is used𝜗𝜗 for the crust). In addition, visco-plastic 𝑠𝑠𝑠𝑠𝑠𝑠rheology𝜗𝜗 of the silicate assumes the viscosity limitation in form

h ≤ (C+FP)/ (2eII) (2.23)

where eII is the second strain rate invariant (Pa), P is the pressure (Pa), F is the internal friction coefficient (F=0.3 is used for the mantle and F=0.2 is used for the crust), C=100 MPa is the rock strength at P =0.

The effective density of the partially molten silicates depends on melt fraction , composition c, temperature T, pressure P and changes linearly with the amount of melt𝜑𝜑 fraction as

( , , , ) = ( , , ) ( , , ) ( , , ) (2.24)

𝑒𝑒𝑒𝑒𝑒𝑒 𝑠𝑠𝑠𝑠𝑠𝑠 𝑠𝑠𝑠𝑠𝑠𝑠 𝑙𝑙𝑙𝑙𝑙𝑙 𝜌𝜌 𝑐𝑐 𝑃𝑃 𝑇𝑇 𝜑𝜑 𝜌𝜌 𝑐𝑐 𝑃𝑃 𝑇𝑇 − 𝜑𝜑 �𝜌𝜌 𝑐𝑐 𝑃𝑃 𝑇𝑇 − 𝜌𝜌 𝑐𝑐 𝑃𝑃 𝑇𝑇 � where and are the densities of the solid and liquid silicates, respectively. Densities of

𝑠𝑠𝑠𝑠𝑠𝑠 𝑙𝑙𝑙𝑙𝑙𝑙 solid mantle𝜌𝜌 and𝜌𝜌 crust are self-consistently calculated with the Perple_X package (Connolly,

2005), using the minimization of Gibbs free energy for the corresponding P–T conditions to determine stable minerals for a Mars-like mantle composition (Khan and Connolly, 2008) and basaltic crust. Latent heat associated with phase changes is also considered in the numerical

30 model. In highly viscous flows, where the inertial forces are negligible if compared to 𝐷𝐷𝑣𝑣𝑖𝑖 the viscous resistance, the conservation of momentum 𝜌𝜌can𝐷𝐷𝐷𝐷 be described by the Stokes equations (with introduction of the pressure P) using the extended Boussinesq approximation.

The equations 2.1, 2.5, 2.7-2.9, 2.11-2.24 are then discretized in their Lagrangian form and solved through finite differences (FD) on a 3D staggered grid with marker-in-cell techniques combined with multi-grid iterative solvers. The employed numerical code

I3ELVIS (Gerya and Yuen, 2007; Gerya, 2010) combines conservative finite differences on a fully staggered grid and marker-in-cell techniques with multigrid solver.

(a)

(b)

Figure 2.1: (a) Elementary volume (cell) of 3D staggered grid used for discretisation of continuity, Poisson and Stokes equations in the case of incompressible viscous flow with variable viscosity.”; (b) Stencil of 3D staggered grid used for the discretization of the temperature equation with variable thermal conductivity (From Gerya, 2010).

31 In Fig. 2.1 there is a representation of the elementary 3D grid cell used for the discretization of the equations. Pressure, gravity, deviatoric stress and strain rate are located in the cube at the centre of the cell; velocity and gravity components are located in the middle of the faces of the cell; shear stresses and strain rates are located in the middle of the edges of the cell; viscosity is located in the middle of the edges corresponding to shear stresses; heat fluxes qx, qy and qz are located in the same positions of the viscosity; all the other material properties and the temperature are located at the cell corners, which are the basic grid nodes.

The indexing of the various locations of the material properties above described follows these criteria:

Cell corners (Grid nodes): Nx, Ny, Nz

Cell centres: Nx – 1, Ny – 1, Nz – 1

Cell edges: Nx, Ny, Nz -1; Nx, Ny – 1, Nz; Nx -1, Ny, Nz

Cell external nodes (Boundary conditions): Nx, Ny, Nz +1; Nx, Ny + 1, Nz; Nx + 1, Ny, Nz

The Eulerian computational domain is equivalent to 8000x8000x8000 km and is resolved with a regular rectangular grid of 293x293x293 nodes and contains 100 million randomly distributed Lagrangian markers to represent different materials. The "Spherical-Cartesian" approach, in which a spherical body is contained within a Cartesian mesh, is employed for modeling a 3D self-gravitating body with a free planetary surface (Gerya, 2010; Lin et al.,

2009). The momentum, mass and heat conservation equations and equation for gravitational potential are discretized and solved on a non-deforming Eulerian grid whereas the advection of transport properties including viscosity, temperature etc. is performed with moving

Lagrangian markers. The free surface boundary condition at the planetary surface is implemented by using a “sticky” atmosphere (Schmeling et al., 2008) with low density (1

32 kg/m3) and viscosity (1018 Pa s). The upper and lower viscosity cutoff limits for all materials are 1024 Pa s and 1018 Pa s, respectively.

The discretized heat conservation equation is

, , , , , , , , , , , , , , 0 + + + = , , (2.25) 𝑇𝑇𝑖𝑖 𝑗𝑗 𝑙𝑙− 𝑇𝑇𝑖𝑖 𝑗𝑗 𝑙𝑙 𝜕𝜕𝜕𝜕𝑥𝑥𝑖𝑖 𝑗𝑗 𝑙𝑙 𝜕𝜕𝜕𝜕𝑦𝑦𝑖𝑖 𝑗𝑗 𝑙𝑙 𝜕𝜕𝜕𝜕𝑧𝑧𝑖𝑖 𝑗𝑗 𝑙𝑙 𝑖𝑖 𝑗𝑗 𝑙𝑙 𝜌𝜌𝑖𝑖 𝑗𝑗 𝑙𝑙𝐶𝐶𝑃𝑃 Δ𝑡𝑡 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝐻𝐻𝑖𝑖 𝑗𝑗 𝑙𝑙

where , , is the temperature of reference for the temperatures located in the other nodal 0 𝑖𝑖 𝑗𝑗 𝑙𝑙 points and𝑇𝑇 calculated through linear interpolation

for = , , , , , , , , , , , , , , , , (2.26) 𝜕𝜕𝜕𝜕𝑥𝑥 �𝑘𝑘𝑖𝑖 𝑗𝑗−1 𝑙𝑙 + 𝑘𝑘𝑖𝑖 𝑗𝑗 𝑙𝑙��𝑇𝑇 𝑖𝑖 𝑗𝑗 𝑙𝑙 − 𝑇𝑇𝑖𝑖 𝑗𝑗−1 𝑙𝑙� �𝑘𝑘𝑖𝑖 𝑗𝑗 𝑙𝑙 + 𝑘𝑘𝑖𝑖 𝑗𝑗+1 𝑙𝑙��𝑇𝑇 𝑖𝑖 𝑗𝑗 +1 𝑙𝑙 − 𝑇𝑇𝑖𝑖 𝑗𝑗 𝑙𝑙�

𝜕𝜕𝜕𝜕 1 1 1 − 1 1 1 ∆𝑥𝑥𝑗𝑗− �∆𝑥𝑥𝑗𝑗− + ∆𝑥𝑥𝑗𝑗+ � ∆𝑥𝑥𝑗𝑗+ �∆𝑥𝑥𝑗𝑗− + ∆𝑥𝑥𝑗𝑗+ � 2 2 2 2 2 2

and analogously for and . The variable thermal conductivity k is indicated with indexes 𝜕𝜕𝜕𝜕𝑦𝑦 𝜕𝜕𝜕𝜕𝑧𝑧 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 i, j, and l, along the axes x, y, and z, respectively, as well as the other parameters in the equation. Discretization is done in a Lagrangian form, which does not include the advective term because it is solved with markers-in-cell techniques. The advected material properties are then interpolated for the Eulerian nodes through a weighted-distance averaging according to the relationships:

= ( ) (2.27) 𝑚𝑚 𝑚𝑚 ( ) 𝑡𝑡+∆𝑡𝑡 ∑ 𝑇𝑇 𝑤𝑤𝑚𝑚 𝑖𝑖 𝑇𝑇𝑖𝑖 ∑𝑚𝑚 𝑤𝑤𝑚𝑚 𝑖𝑖

( , , ) = 1 × 1 × 1 (2.28) Δ𝑥𝑥𝑚𝑚 ∆𝑦𝑦𝑚𝑚 ∆𝑧𝑧𝑚𝑚 𝑚𝑚 𝑖𝑖 𝑗𝑗 𝑙𝑙 1 1 1 Δ𝑥𝑥�𝑗𝑗− � ∆𝑦𝑦�𝑖𝑖− � ∆𝑧𝑧�𝑙𝑙− � 𝑤𝑤 � − 2 � � − 2 � � − 2 �

33

where t+Dt is the solution at the time Dt associated to each of the variables calculated in the

th th stencil, ( , , ) is the statistical weight of the m marker for the i,j,l node calculated through

𝑚𝑚 𝑖𝑖 𝑗𝑗 𝑙𝑙 𝑤𝑤 which are the distances of the marker from the nodes i,j,l of the cell. A Runge-

𝑚𝑚 𝑚𝑚 𝑚𝑚 ΔKutta𝑥𝑥 Δ 𝑦𝑦advectionΔ𝑧𝑧 scheme is then used to move the Lagrangian markers of the material at different points using the second order in time and the fourth order in space.

As an initial condition for the differentiation and core formation model after the giant impact, the planet is fully accreted but not differentiated yet. Thus, several hundreds of iron diapirs were initially randomly distributed throughout the planetary body while there is initially no core. This approach, in which the extended accretion history is compressed into an initial condition, has the advantage of conserving the potential energy released by core formation (Golabek et al., 2011). The resulting mantle temperature profile has the benefit of being spatially variable, yielding an early onset of mantle convection. The impact is approximated by a thermal anomaly in the planetary interior containing a spherical iron diapir formed from the core of the impactor. Then the thermal anomaly is parameterized depending on the size of the impactor and its iron fraction.

The parameterization of the impact is based on the formation of the isobaric core, the nearest spherical region surrounding the impacted site of a planet where the shock pressure is almost uniform. The radius of the isobaric core is (Monteux et al., 2007)

𝑖𝑖𝑖𝑖 𝑟𝑟

= 3 (2.29) 1 3 𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖𝑖𝑖 𝑟𝑟 𝑟𝑟 where is the radius of the impactor. As the shock wave moves away from the isobaric

𝑖𝑖𝑖𝑖𝑖𝑖 core, the𝑟𝑟 pressure decays according to

𝑑𝑑 𝑃𝑃

34 = (2.30) 𝑟𝑟𝑖𝑖𝑖𝑖 𝑛𝑛 𝑑𝑑 0 𝑑𝑑 𝑃𝑃 𝑃𝑃 � � where is the initial shock pressure at the moment of the impact, d is the distance from the

0 isobaric𝑃𝑃 core, and n is the decay exponent constant that has to be determined experimentally through the use of hydrocodes as the velocities required to obtain significant melting are not achievable in laboratories

The temperature generated by the dissipation of the impact energy is

𝑑𝑑 𝑇𝑇

= + (2.31) 𝑟𝑟𝑖𝑖𝑖𝑖 𝑛𝑛 𝑑𝑑 0 𝑑𝑑 𝑇𝑇 𝑇𝑇 ΔΤ � � where is the temperature in the isobaric core at the time of the impact and is the thermal

0 anomaly𝑇𝑇 associated with the distance d. Assuming that the densities of theΔΤ target and the impactor are similar and that the impact velocity is comparable to the escape velocity of the target, the thermal anomaly can be calculated by the following relationship

= (2.32) ( ) 2 4𝜋𝜋 Υ 𝜌𝜌𝜌𝜌𝑟𝑟 ΔΤ 9 𝑓𝑓 𝑛𝑛 𝐶𝐶𝑃𝑃

where r is the radius of the target, G is the gravitational constant, and are the density and

𝑃𝑃 the specific heat of the target, is the efficiency of kinetic to thermal𝜌𝜌 energy𝐶𝐶 conversion that should be 0.5 given the impactΥ velocity of 5 km s-1 used in the model, ( ) is a function that represents the effectively heated volume normalized by the isobaric core𝑓𝑓 volume𝑛𝑛 ( ( ) = 1 if only the isobaric core is heated). If , the integration of equation 2.26 gives𝑓𝑓 a 𝑛𝑛value of

𝑖𝑖𝑖𝑖𝑖𝑖 ( ) ~ 2.7 corresponding to 37%𝑟𝑟 of the≪ heating𝑟𝑟 released in the isobaric core. Due to the large𝑓𝑓 𝑛𝑛 amount of heat released by the , the viscosity of the molten silicates is

35 expected to be low and, as a consequence, the high Rayleigh (and Nusselt) number will make heat transfer and subsequent cooling of the magma ocean highly effective.

The effective thermal conductivity of molten silicates is estimated through the soft turbulence model and the heat flux q is calculated as

( ) = 0.089 (2.33) 1 0 𝜅𝜅 𝑇𝑇−𝑇𝑇 3 𝐿𝐿 𝑞𝑞 𝑅𝑅𝑅𝑅 where is the thermal conductivity, L is the depth of the magma ocean, Ra is the Rayleigh number,𝜅𝜅 and T is the temperature at depth L, and is the temperature at the surface of the

0 planet. The Rayleigh number is given by 𝑇𝑇

( ) = 2 3 (2.34) 𝛼𝛼𝛼𝛼 𝑇𝑇−𝑇𝑇0 𝜌𝜌𝑙𝑙𝑙𝑙𝑙𝑙 𝐶𝐶𝑃𝑃 𝐿𝐿 𝑅𝑅𝑅𝑅 𝜅𝜅𝜂𝜂𝑙𝑙𝑙𝑙𝑙𝑙

where is the thermal expansivity and g is the gravity. At last, the effective thermal conductivity𝛼𝛼 is calculated by

= 3 1 (2.35) . ( ) − 𝑞𝑞 2 1 𝛼𝛼𝛼𝛼𝐶𝐶𝑃𝑃 2 2 𝜅𝜅𝑒𝑒𝑒𝑒𝑒𝑒 �0 089� 𝑇𝑇−𝑇𝑇0 𝜌𝜌𝑙𝑙𝑙𝑙𝑙𝑙 �𝜂𝜂𝑛𝑛𝑛𝑛𝑛𝑛�

where is the cut-off viscosity taken as 1017 Pa s.

𝑛𝑛𝑛𝑛𝑛𝑛 Partial𝜂𝜂 melting of the mantle, melt extraction and percolation toward the bottom of the forming basaltic crust is implemented in a simplified manner. According to our model, magma added to the crust is balanced by melt production and extraction in the mantle.

However, melt percolation is not modeled directly and considered to be nearly instantaneous.

The standard (i.e. without melt extraction) volumetric degree of mantle melting M0 changes

36 with pressure and temperature according to the linear batch-melting model (Golabek et al.,

2011). Lagrangian markers track the amount of melt extracted during the evolution of each experiment. The total amount of melt, M, for every marker takes into account the amount of previously extracted melt and is calculated as

M = M0 − Sn Mext (2.36)

where Sn Mext is the total melt fraction extracted during the previous n extraction episodes.

The rock is considered non-molten (refractory) when the extracted melt fraction is larger than the standard one (i.e. when Sn Mext M0 If M >0 for a given marker, the melt fraction Mext =

M is extracted and Sn M ext is updated. > ). The extracted melt fraction Mext is assumed to propagate much faster than the rocks deform. Melts produced at depths are moved to the surface and added to the bottom of the forming crust. In order to ensure melt volume conservation and account for mantle compaction and in response to the melt extraction, melt addition to the bottom of the crust is performed at every time step by converting the shallowest markers of mantle into crustal markers. The local volume of these new crustal markers matches the local volume of extracted melt computed for the time step.

Following Golabek et al. (2011), basaltic melts are assumed to be only extracted from relatively shallow (<300 km depth) mantle regions with low degree of melting (M0<0.2).

I have estimated the timescale at which the magma ocean reaches the Martian highlands extension performing the first runs for a varying combination of initial parameters, studying the effect of a) the radiogenic heating through the onset of the impact time event, b) the initial isothermal peridotite solidus temperature in the range 1100-1300 K and c) the compressibility of the molten silicates on the evolution of the model. It is also possible to observe the timing of the core formation when all the iron and the heavier materials are completely settled down.

37 A discussion of the preliminary results is available in the next section. The results of the

I3ELVIS calculations are stored in an output format with the file extension ”prn”. The prn files, suitably converted into a vtr format (via a Python script) are readable by ParaView where the fields of interest (i.e. composition, density, temperature and viscosity) are extracted and then visualized. In the I3ELVIS output, the different materials (rock/metal types) are indicated with an integer that is then recalled in the threshold menu of ParaView to be visualized on the screen. In this way, the user can follow the evolution of the chosen materials

(e.g. molten silicates, solid silicates, iron) in the output of the experiments.

2.3 The StagYY code

The physical model and parameters used in the present 3-D experiments are identical to those in the 2-D experiments of Golabek et al. (2011), which is straightforward because

StagYY can model both a full 3-D spherical shell used here and the 2-D spherical annulus

(Hernlund et al. 2008) used in Golabek et al. (2011). The physical parameters including the rheology, radioactive heating, solidus etc. in StagYY are adjusted to match as well as possible those used in the I3ELVIS part of the experiment. As in I3ELVIS, melt that is shallower than a certain depth is instantaneously moved to the crust. Golabek et al. (2011) detail the StagYY treatments of multiple composition-dependent phase transitions, heat-producing element fractionation, and increase in solidus with degree of depletion.

The StagYY code solves the equations for the conservation of mass, momentum, and energy (Tackley, 2008)

( ) = 0 (2.37)

∇ ∙ 𝜌𝜌𝑣𝑣⃑

38 ′ = (2.38) 𝜕𝜕𝜕𝜕𝑖𝑖𝑖𝑖 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕𝑗𝑗 − 𝜕𝜕𝜕𝜕𝑖𝑖 𝜌𝜌𝑔𝑔𝑖𝑖

= + • ( ) + + (2.39) 𝐷𝐷𝐷𝐷 ′ 𝑝𝑝 𝐷𝐷𝐷𝐷 𝑟𝑟 𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝜌𝜌𝐶𝐶 −𝛼𝛼𝛼𝛼𝛼𝛼𝑣𝑣 ∇ 𝑘𝑘∇𝑇𝑇 𝜌𝜌𝜌𝜌 𝜎𝜎 ∶ 𝜀𝜀̇

and for composition = 0 (2.40) 𝐷𝐷𝐷𝐷 𝐷𝐷𝐷𝐷 where the variables are the temperature T, composition C, velocity v, and pressure p. is the ′ 𝑖𝑖𝑖𝑖 deviatoric stress tensor and is the strain rate tensor. H is the internal heating rate. 𝜎𝜎Another

𝑖𝑖𝑖𝑖 difference is that StagYY 𝜀𝜀doeṡ not use the Boussinesq approximation, but instead the compressible anelastic approximation, discretizing the equations on the yin-yang grid (see

Figure 2.2).

Figure 2.2. Basic Yin-Yang grid. (a) Two identical component grids, Yin and Yang. (b) Yin and Yang combined to cover a spherical surface with partial overlap (Kageyama and Sato, 2004).

In the yin-yang grid each subgrid acts as boundary condition for the other and, for computational efficiency, the interpolation weights and the points are pre-calculated once and stored. The advantage of this grid is that it maintains a nearly constant spacing and the grid

39 lines are orthogonal, allowing the implementation of finite differences for the solution of the equations. The solution of the equations is then obtained by iterations through multi-grid cycles (F or V-cycles) in which the residue (the error) is relaxed through a hierarchy of nested fine and coarse grids with different spacing to accelerate the convergence rate of the solver.

The multi-grid techniques are also used in the I3ELVIS code but the difference is that the latent heat effects due to phase transitions are individually specified in StagYY, as effective heat capacity and thermal expansivity, whereas in I3ELVIS they are included via pre- calculated tables of density using the Perple_X package. The advection term of the heat equation is computed using a finite volume (FV) scheme while markers/tracers are instead advected using the second-fourth order Runge-Kutta scheme as well as in I3ELVIS but taking into account the correction term for the spherical geometry.

Other physical assumptions have been retained in the 3D model. For example, the viscosity law is

( , ) = exp (2.41) 𝐸𝐸𝑎𝑎+𝑃𝑃𝑉𝑉𝑎𝑎 𝑘𝑘𝑘𝑘 𝜂𝜂 𝑃𝑃 𝑇𝑇 𝐴𝐴 � � where the values of Ea and Va are given in Table 1 and A is a factor that gives

h (P = 0, T = 1500K) = h0 (2.42)

Phase transitions in both olivine and garnet/pyroxene mineralogical systems are considered and calculated through the phase function

𝑖𝑖𝑖𝑖 Γ

= 1 + tanh (2.43) 1 �𝑑𝑑−𝑑𝑑𝑖𝑖𝑖𝑖�− 𝛾𝛾𝑖𝑖𝑖𝑖�𝑇𝑇−𝑇𝑇𝑖𝑖𝑖𝑖� Γ𝑖𝑖𝑖𝑖 2 � � 𝑊𝑊𝑖𝑖𝑖𝑖 ��

40

where and are the depth and temperature of the ith phase transition of the mineralogical

𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 system𝑑𝑑 j, is𝑇𝑇 the Clapeyron slope, and is the width of the phase transition. The

𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 dependency𝛾𝛾 of the viscosity on the composition𝑊𝑊 is give by

( , , ) = ( , ) (2.44) Γ𝑖𝑖𝑖𝑖𝑣𝑣𝑗𝑗 𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝜂𝜂 𝑐𝑐 𝑃𝑃 𝑇𝑇 𝜂𝜂 𝑃𝑃 𝑇𝑇 ∑ �Δ𝜂𝜂 � where j = 1 indicates the olivine system and j = 2 the garnet/pyroxene system, is the

𝑗𝑗 fraction of composition j in each grid, and the factors give factors of 20 and 30 𝑣𝑣increase

𝑖𝑖𝑖𝑖 in viscosity over the phase transitions correspondingΔ𝜂𝜂 to Earth’s 410 and 600 km discontinuities, respectively.

The olivine mineralogical system has a volume fraction of 0.55. The dry olivine rheology and the initial fraction of 0.15 in the mantle are also applied here as well with the assumption that the erupted melt immediately solidifies in place. The thermal anomaly generated by the impactor triggers a transient superplume producing additional melt that should develop into volcanism on a timescale of a few hundreds Ma, but such an evolution must be followed through the various tests.

2.4 Transferring composition and temperature from I3ELVIS to StagYY

StagYY (Tackley, 2008) also uses a joint Eulerian-Lagrangian approach, with a fixed

Eulerian fully staggered grid and moving Lagrangian markers. Thus, the markers from the

I3ELVIS simulation are read into StagYY, transferring compositional information with no loss of accuracy. The initial CMB temperature is based on the mean temperature of the core from the I3ELVIS simulation. Two significant differences in the discretizations are

41 straightforwardly addressed: (i) StagYY treats temperature advection and diffusion purely on the Eulerian grid (using a finite-volume scheme) rather than on markers; therefore the transfer step involves local averaging of marker temperatures to the appropriate locations on the grid.

(ii) StagYY uses the yin-yang spherical mesh to model a spherical shell, the outer boundary of which corresponds to the surface of Mars and the inner boundary the core-mantle boundary, instead of embedding a sphere in a Cartesian mesh. This is handled using a straightforward coordinate transformation (from Cartesian to spherical polar coordinates) and discarding of markers that are outside the mantle (such tracers normally track 'air' or 'metal').

The same I3ELVIS .prn files, suitably converted through other C scripts, provide the input files readable by the StagYY code. The choice of the output file to be converted coincides exactly with the best composition visualization match of the Martian dichotomy obtained with the I3ELVIS code and with the best density profile (when the core is already formed). The exact time when this match is achieved through the I3ELVIS code becomes the starting time of the StagYY code run which will show the onset of mantle convection and the development of the plumes for a time span extending for 4.5 Ga or until any sign of volcanic activity on the surface has ceased.

2.5 Model parameters

The parameters of the model are defined in the following table:

Parameter Symbol Value Units

Radius of Mars r 3.389 × 106 m

Density of uncompressed solid mantle r 3500 kg m -3

42 Density of uncompressed molten mantle r liq 2900 kg m-3

Density of uncompressed iron r Fe 8000 kg m-3

Mantle basalt fraction after core formation f 0.15

Cohesion c 1.00 × 108 Pa

Friction coefficient f 0.2-0.3

9 Peierls stress sP 9.1 × 10 Pa

-1 Activation energy Ea 532 kJ mol

-6 3 -1 Activation volume Va 8.00 × 10 m mol

4 Dislocation creep onset stress s0 3.00 × 10 Pa

Power law exponent n 3.5

Latent heat of melting L 400 kJ kg -1

-1 -1 Heat capacity of silicate cP 1000 J kg K

-1 -1 Heat capacity of iron cP-Fe 1000 J kg K

Thermal expansivity of solid silicate a Si-sol 3.0 × 10-5 K-1

Thermal expansivity of molten silicate a Si-liq 6.0 × 10-5 K-1

Thermal expansivity of solid iron a Fe-sol 1.0 × 10-5 K-1

Thermal expansivity of molten iron a Fe-liq 1.0 × 10-5 K-1

-1 -1 Thermal conductivity of solid silicate k Si-sol 3 W m K

5 -1 -1 Thermal conductivity of molten silicate k Si-liq 3.0 × 10 W m K

2 -1 -1 Thermal conductivity of solid iron k Fe-sol 1.0 × 10 W m K

5 -1 -1 Thermal conductivity of molten silicate k Fe-liq 3.0 × 10 W m K

5 -1 -1 Effective thermal conductivity of molten materials Keff 3.0 × 10 W m K

Crustal radiogenic heating enrichment factor L 1

43 The temperature and the viscosity of the sticky atmosphere have been considered constant all over the surface at values of 273 K and 1018 Pa s, respectively, and we used initial isothermal temperature profiles close to the surface solidus temperature of peridotite (T = 1200-1400 K).

44 Chapter 3

Three-dimensional simulations of the southern polar giant impact hypothesis for the origin of the Martian dichotomy

45 Three-dimensional simulations of the southern polar giant impact hypothesis for the origin of the Martian dichotomy.

Giovanni Leone1, Paul J. Tackley1, Taras V. Gerya1, Dave A. May1, and Guizhi Zhu1

1ETH Zurich (Institute of Geophysics, ETH Zurich, 8092 Zurich, Switzerland; e-mail:

[email protected]).

Published in the Geophysical Research Letters, vol. 41(24), 8736-8743.1

Abstract

We demonstrate via numerical simulations that the impact of a ~ lunar-sized body with

Mars is capable of creating a hemispherical magma ocean that upon cooling and solidification resulted in the formation of the southern highlands and thus the Martian dichotomy. The giant impact may have contributed a significant amount of iron to the Martian core and generated a deep thermal anomaly that led to the onset and development of the volcanism in the southern highlands. Our model predicts several mantle plumes converging to the South Pole from the equatorial regions as well as new plumes forming in the equatorial region and also an absence of significant large-scale volcanism in the northern lowlands. The core heat flux evolution obtained from our numerical models is consistent with the decline of the magnetic field. We argue that such a scenario is more consistent with a range of observations than a northern giant impact (excavating the Borealis basin) for the formation of the Martian dichotomy.

1 Movies and supplementary material are available at the following website: http://onlinelibrary.wiley.com/doi/10.1002/2014GL062261/abstract

46

3.1 Introduction

Previous studies have suggested that the Martian dichotomy may have been generated either by endogenic processes like degree-1 mantle convection (Keller and Tackley, 2009;

Zhong and Zuber, 2001) or by exogenic (impact) processes: multiple impacts in the northern hemisphere (Frey and Schultz, 1988) or a giant impact occurring on the northern hemisphere of the planet forming the Borealis basin (Andrews-Hanna et al., 2008; Marinova et al., 2008;

Nimmo et al., 2008; Wilhelms and Squyres, 1984). However, these hypotheses have several problems, for example: (i) despite the apparent difference in cratering rates between highlands and lowlands, the quasi-circular depressions (QCD’s) in the lowlands hint at a similar crustal age for both hemispheres (Frey, 2006), which are estimated to start formation during or immediately after the planet’s accretion; therefore the crustal dichotomy is the most ancient geologic feature on the planet (Carr and Head, 2010; Solomon et al., 2005); (ii) multiple impacts occurring only on the northern hemisphere to shape the lowlands are statistically unlikely; (iii) impacts as large as those proposed by Marinova et al (2008), Nimmo et al

(2008) and Andrews-Hanna et al (2008) tend to create a deep thermal anomaly, upwelling and magmatism that continues for a much longer time period than the initial impact-generated magma, inconsistent with both topography and the evident lack of significant volcanic edifices in the northern lowlands (Tanaka et al., 2008).

An alternative hypothesis to the Northern Polar Giant Impact (NPGI) is that a Southern

Polar Giant Impact (SPGI) with a body between 0.1 – 1 lunar masses (≈800 - 1700 km radius) generated a hemispherical magma ocean (the Australis magma ocean (AMO)) that solidified to form the thicker crust of the southern hemisphere, as investigated analytically by (Reese et al., 2011; Reese and Solomatov, 2006; Reese and Solomatov, 2010). The resulting mantle

47 thermal anomaly may also have induced degree-1 convection leading to further magmatism in the southern hemisphere, something that the analysis of (Reese et al., 2011; Reese and

Solomatov, 2006; Reese and Solomatov, 2010) did not cover. Golabek et al. (2011) tested this using a suite of 2-D simulations going from the immediate post-impact time to the present day, finding that it is indeed a viable hypothesis. However, since the formation of the dichotomy is an inherently 3-D process (Keller and Tackley, 2009), we have conducted a series of 3-D simulations of post-giant impact evolution of a Mars-sized planet.

3.2 Method

The physical principles and assumptions adopted in these 3-D experiments are essentially the same as those in the previous 2-D experiments (Golabek et al., 2011). We used the

I3ELVIS code (Gerya and Yuen, 2007) to simulate in 3-D the period immediately following the giant impact. We then used the StagYY code (Tackley, 2008) to analyze the long-term evolution over 4.5 Ga to the present day, including mantle convection, crustal growth, and core-mantle boundary (CMB) and surface heat flux. This involved the transfer of temperature and composition fields from I3ELVIS to StagYY.

The effect of the impact is parameterized as in the previous 2-D study (Golabek et al.,

2011) and, based on earlier works (Monteux et al., 2007; Reese and Solomatov, 2010; Senshu et al., 2002; Tonks and Melosh, 1993). The impactor is assumed to have hit the South Pole and to have created an isobaric core - a region of approximately the same shock pressure and temperature (Fig. 3.1b) (Golabek et al., 2011) and, as in the 2-D research, the impact is assumed to have happened when the planet was fully accreted but not yet differentiated (Fig.

3.1a). The impactor’s core is placed at the base of the heated region (Fig. 3.1b).

Regarding impact velocity, previous estimates suggested 50% planet surface melting in case of an impactor/target mass ratio of 0.14 at a speed of 15 km s-1 (Tonks and Melosh, 1993) but

48 N-body simulations have shown that the impactors’ speed is more frequently around the escape velocity of the target (Agnor et al., 1999), and rarely more than twice the escape velocity. Mars’ escape velocity of 5 km s-1 is also slow enough to avoid bouncing or disruption processes during the giant impact (Stewart and Leinhardt, 2012). Oblique impacts were not considered because they are less effective than vertical ones (Pierazzo and Melosh,

Figure 3.1. Composition (a, d, g), temperature (b, e, h) and density (c, f, i) fields for the immediate post impact time and for the post-impact period, the numbers from 0 to 8 on each axis of the temperature and density fields refer to the model box (8000 km). Panels a, b, and c are at the giant impact time of 4 Ma after CAI; d, e, and f are at 4.0003 Ma after CAI; g, h, and i are at 5.3468 Ma after CAI. The coloured rectangles for the composition field shows a scale where the colors indicate the specific material type used within the I3ELVIS program: 5 (blue) is solid silicate, 10 (pale blue) is molten silicate metal, 25 (pale red) is molten iron, 30 (red) is solid iron. In panel g the molten phases are mixed together because the differentiation process is still in its initial phase, so that it is quite difficult to discern them, particularly in the inner part of the planet. The small red circles in the panels a, c, d, and f are the iron cores of smaller impactors that hit the planet before the giant impact.

49

2000; Reese and Solomatov, 2010; Stewart and Leinhardt, 2012). The effect of higher impact velocities of up to 10 km s-1 (Le Feuvre and Wieczorek, 2008) with smaller impactors was instead explored; N-body simulations indicate that this is at the upper end of the likely range of velocities (Agnor et al., 1999; Morishima et al., 2010).

We have performed our 3-D runs for impactor sizes ranging from 1008 to 2000 km radius

(instead of the ~400-1000 km in the 2-D experiments), focusing on mesosiderite-type composition (50% iron by radius) but also including a case with siderite composition (80% iron in radius) and a case with a pure stony impactor. We neglect nickel due to its average 7-

15 wt% (McCoy et al., 2011) in impactors' cores. The higher iron content case is interesting because of the possible presence of M-type asteroids like 16 Psyche as well as several others in the asteroid belt (Ockert-Bell et al., 2010) - likely remnants of larger parent bodies in the 1-

2 AU range which migrated in the current position after giant impacts with protoplanets

(Goldstein et al., 2009).

We tested impact times from 0 to 5 Ma after CAI, when impactors up to Moon-size were already available (Agnor et al., 1999), then settled on an impact time of 4 Ma (the 2-D experiments used 5 Ma), when radiogenic heating was weaker than during accretion due to the decay of the short-lived radionuclides (26Al and 60Fe) and the Martian core was not yet completely formed (Sramek et al., 2010). This value is also above the earliest time of 3 Ma indicated by the geochemistry (Jacobsen, 2005) but still within an upper bound that places the core formation around 15 Ma (Righter and Shearer, 2003). Later additions to the core are not precluded if the new impactors' cores merge quickly without equilibrating with the silicate part of the planet, although an impactor of the size considered here might cause significant re- equilibration. The short-lived radiogenic heating is strong until 3 Ma so that any newly-

50 formed crust would be re-melted; the new crust is stable from 4 Ma onwards when this effect decays.

3.3 Results

Here we present all cases, including the one that was the most successful in matching the observations of Mars. The effects of varying impactor size, velocity and composition are also discussed. Our best cases are a meso-siderite (50% iron by radius) of 2000 km radius (Fig. 3.2 and Fig. 3.3) and a siderite (80% iron by radius) of 1600 km radius, both with an impact speed of 5 km s-1 (Fig. 3.2 and Fig. 3.4).

Fig. 3.1 shows the processes of core formation and massive melting in the impact hemisphere for the 2000 km meso-sideritic impactor, also shown in Supplementary Movie 1, while Supplementary Movie 2 shows the development of the mantle plumes for the 1600 km sideritic impactor. The impactor’s core reaches the planet center and merges with the planet’s forming core (Fig. 3.1d, 3.1e, and 3.1f), resulting in the differentiated structure of the planet.

After 1.3 Ma post impact, pockets of molten silicate are still visible beneath the cooled crust

(Fig. 3.1g), while the halo around the core is formed by the differentiation of the iron from the silicates. Core formation occurs very rapidly - in our simulations the planet’s core is formed by ~5 Ma after CAI (Fig. 3.1g, 3.1h, and 3.1i), i.e. about 1 Ma after the impact. However it could be that even shorter timescales are realistic, which are here not resolved due to the applied minimum viscosity cut-off (see section 2.2).

A larger meso-sideritic impactor (2000 km radius, 50% iron by radius, impact speed 5 km s-1) produces a hemispherical magma ocean consistent with the extent of the Martian dichotomy (Fig. 3.1g-3.2a). For comparison, in the 2-D experiments (Golabek et al., 2011) a much smaller impactor (1000 km radius, 50% iron by radius, 5 km s-1) produced a hemispherical dichotomy - this difference can be understood if the important criterion is the

51 ratio of impactor kinetic energy (which is proportional to impactor mass) to planet mass: in 2-

2 3 D this scales as (rimpactor/rplanet) whereas in 3-D it scales as (rimpactor/rplanet) ; thus in 3-D a larger impactor is required. A smaller, mesosideritic, 1600 km diameter impactor is less efficient in producing a dichotomy (Fig. 3.2c and 3.2d), although additional iron in its core would increase the shear heating and thus the extent of the magma ocean (Fig. 3.3).

Figure 3.2. a) Surface topography 4.5 Ga after the southern polar giant impact (SPGI) at 5 km s-1 for the impactor of 2000 km radius. The y axis indicates the , the x axis the , and the colobar altimetry, which is shifted about 3 km above Mars values due to a different reference level. b) Crustal thickness distribution formed by the 2000 km impactor, which is around 50 km below the highlands and around 30 km below the lowlands, values consistent with the average crustal thicknesses estimated through gravity measurements (Neumann et al., 2004). c) Surface topography 4.5 Ga after SPGI at 5 km s-1 for the impactor of 1600 km of radius; the values are comparable to those obtained for the 2000 km impactor. d) Crustal thickness distribution formed by the 1600 km impactor. e) Surface topography 4.5 Ga after SPGI at 7.5 km s-1 for the

52 impactor of 1221 km of radius. f) Crustal thickness distribution formed by the 1221 km impactor. g) Surface topography 4.5 Ga after SPGI at 10 km s-1 for the impactor of 1008 km of radius. h) Crustal thickness distribution formed by the 1008 km impactor. i) Surface topography 4.5 Ga after SPGI at 5 km s-1 for the stony impactor of 1750 km of radius. j) Crustal thickness distribution formed by the 1750 km impactor. k.-m.) South polar views of the 2000 K temperature isosurface: k) 157 Ma after the impact, l) 290 Ma after the impact, m) 388 Ma after the impact with a siderite of 1600 km. These show mantle plumes converging towards the South Pole (center of the image) as well as new plumes forming increasingly further from the South Pole.

Figure 3.3. Evolution of the Martian dichotomy, from 0.25 Ga to 3.5 Ga after CAI, formed by an impactor of 2000 km of radius and 50% of iron in radius. The first column shows the surface topography while the second column shows the crustal thickness. a) Surface topography 0.25 Ga after CAI. a) Surface topography 0.25 Ga after CAI. The dichotomy has already formed, the highlands are indicated by red, the transition topography by yellow and green, and the lowlands by blue. There is a striking similarity in latitude with the result obtained by the impactor of 1600 km and 80% of iron. b) Crustal thickness 0.25 Ga after CAI, new crust has formed in the southern hemisphere and it is almost comparable with the previous experiment. c) Surface topography 0.5 Ga after CAI. Although the latitude is also comparable to the corresponding one in the 1600 km impactor, the border of the dichotomy appears more frayed. d) Also the crustal thickness is comparable 0.5 Ga after CAI. e) Surface topography 1.0 Ga after CAI, the first “fjords” start to appear along the border. f) Crustal thickness 1.0 Ga after CAI, the frayed pattern that started to appear on the topography was already present in the crustal

53 thickness since the beginning. g) Surface topography 1.5 Ga after CAI, the transition topography starts to protrude into the lowlands but is less developed than the one produced by the previous experiment. h) Crustal thickness 1.5 Ga after CAI, the crustal production under the transition topography starts to appear. i) Surface topography 2.0 Ga after CAI, the transition topography is more extended in the northern polar area with respect to the previous experiment. j) Crustal thickness 2.0 Ga after CAI, patches of crust become more developed under the transition topography in the northern polar area and in the equatorial stripe. k-n) No significant variations are observed from 2.5 to 3.5 Ga after CAI.

Figure 3.4. Evolution of the Martian dichotomy from 0.25 Ga to 3.5 Ga after CAI. The first column shows the surface topography while the second column shows the crustal thickness. a) Surface topography 0.25 Ga after CAI. The dichotomy has already formed. The highlands are indicated by red, the transition topography by yellow and green, and the lowlands by blue. b) Crustal thickness 0.25 Ga after CAI. New crust has formed in the southern hemisphere. c) Surface topography 0.5 Ga after CAI. The topography in the highlands has become

54 heterogeneous while the transition topography (pale blue) decreases its extent. d) Crustal thickness 0.5 Ga after CAI. The growth of the new crust under the highlands occurs mainly at mid-high latitudes with pockets concentrated in the western hemisphere and an isolated ridge in the polar area of the eastern hemisphere. e) Surface topography 1.0 Ga after CAI. The highlands have stabilized at around 20 degrees of latitude south while the transition topography starts to expand towards the lowlands modified by volcanic processes occurring along the equatorial regions. f) Crustal thickness 1.0 Ga after CAI. The crust under the highlands grows following the pattern established in the previous 0.5 Ga. A few pockets of crust appear under the northern polar region. g) Surface topography 1.5 Ga after CAI. There is a vigorous expansion of the transition topography after the volcanic activity and the appearance of a topographic “island” of likely volcanic origin in the northern polar region. h) Crustal thickness 1.5 Ga after CAI. The growth of the crust under the highlands has become more homogeneous, decreasing the number of pockets, while the topographic island in the northern hemisphere is not followed by crustal growth underneath. The number of pockets of new crust increases in the northern hemisphere instead. i) Surface topography 2.0 Ga after CAI. The expansion of a Tharsis-like and an Elysium- like feature astride the Equator at the dichotomy boundary has begun. The two features are at an angular distance of roughly 90 degrees to each other, similar to Tharsis and Elysium. j) Crustal thickness 2.0 Ga after CAI. The pockets of new crust increase both their extent and number under the southern and the northern hemispheres. k) Surface topography 2.5 Ga after CAI. The Tharsis-like and Elysium-like features start to take shape while sub- polar Scandia-like features start to appear. l) Crustal thickness 2.5 GA after CAI. The crustal growth is now stabilized both under the southern and the northern hemispheres. m) Surface topography 3.5 Ga after CAI. The topographic island at the North Pole has disappeared while the Scandia-like features remain as well as the Tharsis-like and Elysium-like features. The dichotomy has now acquired the final configuration that can be seen today. n) Crustal thickness 3.5 Ga after CIA. The crustal thickness has also acquired the final configuration.

Increasing the impact velocity while decreasing the radius such as to maintain a constant kinetic energy produced almost the same extent of thickened crust, both for a 1221 km

(radius) meso-siderite at 7.5 km s-1 and for a 1008 km meso-siderite at 10 km s-1 (Fig. 3.2e and 3.2f and Fig. 3.2g and 3.2h). A stony impactor of 1750 km radius at 5 km s-1 was less effective (Fig. 3.2i and 3.2j). A radius of 2000 km at 5 km s-1 is required in order to reasonably match the latitude of the highlands with a meso-sideritic impactor (Fig. 3.2a and

3.2b). Our results for both the 2000 km and 1600 km impactor show a similar average crust thickness of about 60 km below the highlands and 30 below the lowlands (Fig. 3.2b and

3.2d), which is similar to the 58 km and 32 km, respectively, estimated by the combined analysis of MOLA topography and MGS gravity data (Neumann et al., 2004). Another consequence of such a giant impact is a residual asymmetrical thermal anomaly inside Mars that develops into active volcanism in the southern hemisphere (Fig. 3.1h), with plumes that might even have contributed to form minor volcanic features in the northern lowlands (Garvin et al., 2000), but not enough to produce crust comparable to the highlands (see Supplementary

55 Movie 3). Further details on the development of crustal thickness and surface topography over time can be found in Fig. 3.3 and in Fig. 3.4.

Visualisation of mantle temperature iso-surfaces during the long-term evolution phase

(Supplementary Movie 2) reveals several mantle plumes in the southern region 157 Ma after the impact (Fig. 3.2k), which slowly migrate towards the South Pole of Mars due to the large- scale flow induced by the impact thermal anomaly: upwelling in the southern polar region results in lateral flow towards the South Pole in the deep mantle and away from the South

Pole in the shallow mantle. With time, new plumes form in the equatorial region at an angular distance of about 20 degrees from the first ones and also migrate towards the pole (Fig. 3.2l,

290 Ma after the impact) a process that is also continuing 388 Ma after the impact (Fig.

3.2m).

The global CMB heat flux following the giant impact drops from 100 to 60 mW/m2 in the first 0.1 Ga, to 50 mW m-2 in the first 0.4 Ga and to 20 mW m-2 by 0.6 Ga, then declining to

10 mW m-2 around 1 Ga (Fig. 3.5). This large drop is sufficient to shut off the geodynamo without the assistance of additional impacts (as proposed by (Roberts et al., 2009)) or any other special mechanism being necessary. Quantitatively, a lower bound on core heat flow for a planetary dynamo to operate is given by the heat conducted down the core adiabat. For

Earth this has recently been estimated to be 10-16 TW (de Koker et al., 2012; Gomi et al.,

2013; Pozzo et al., 2012), corresponding to a flux of 66-105 mW m-2. Scaling to the lower this leads to a critical heat flux somewhere in the range 20-40 mW m-2, which is passed at around 4 Ga in our simulations.

56

Figure 3.5. Surface (red line) and core-mantle boundary (blue line) heat flux evolution along the whole history of Mars for the different impactors of meso-sideritic composition: a) 2000 km; b) 1750 km; c) 1600 km; d) 1221 km; e) 1008 km. No particular variation in the heat flux trend is observed for the different impactors, the thermal evolution of the planet is neither affected by the size of the impactor nor by the strength of the initial thermal anomaly. The different thermal anomalies are evidently dissipated by different degrees of volcanism during the initial phase while the long-term thermal evolution of the planet follows the normal decay of the radiogenic heating.

57

This timeline is thus consistent with the 4.15-4.07 Ga estimated from the map of the crustal magnetic field (Lillis et al., 2008), although there is some uncertainty in the absolute age. It is also consistent with the time of changes in the Mars surface environment inferred by mineral alteration (Bibring et al., 2006), which may be linked to loss of the atmosphere following magnetic field shutdown. The surface global average heat flux, after an initial peak of 30-35 mW m-2, decreases to 20 mW m-2. This is consistent with the global average heat flux values estimated from lithospheric flexures for present day, which indicate 14-22 mW m-

2 (Ruiz et al., 2010) with the lower end estimated for Isidis (Dehant et al., 2012).

3.4 Discussion

Mars is a planet that shows no evident traces of plate tectonics and the geological record visible on the surface dates back to its most ancient times (Carr and Head, 2010). Although

Sleep [1994] proposed a brief episode of plate tectonics in the northern hemisphere, no evidence of any relic subduction zone has been found which, together with concerns about the timescale involved, make this scenario unlikely (Watters et al., 2007). Volcanic activity was clearly important early in its history; for example a recent analysis of HiRISE images of

Valles Marineris has led to the hypothesis that it formed through erosion by lava flows rather than tectonic force, and that the putative shorelines in Chryse Planitia are the lava flow fronts coming from Valles Marineris (Leone, 2014). The lack of significant ongoing volcanic activity, such as that observed on Earth today, and the asymmetric position of the volcanic features, mostly located in the southern hemisphere, indicate that internal heat sources

(primordial and radioactive) are insufficient to sustain activity over geological time. This conclusion is expected given the small size of Mars relative to Earth (Breuer and Moore,

2007).

58 According to the geo-chrono-stratigraphic map of Mars (Tanaka et al., 2013), the oldest volcanic features are located between the southern polar region and the Equator, in particular upper terrains until the Equator except for Tharsis and Elysium. The geologic ages estimated for the volcanic features would thus suggest a progressive sequence of emplacement in space and time from the South Pole to the Equator, consistent with the appearance of new plumes progressively closer to the Equator in our model, even though there is a tendency for already-formed plumes to migrate from the Equator towards the South

Pole. The majority of the volcanic provinces are of age, except the South Pole and the Tharsis and Elysium Rises that are indicated as Amazonian. The young age inferred might likely be related to the geological processes that the volcanic centers underwent during their history, volcanic, aqueous, glacial, and aeolian processes that continued (volcanic) or started

(aqueous, glacial, aeolian) after the original time of emplacement. For example, although its latest lava flows appear young, the bulk material of Olympus Mons might date back to

Hesperian or Noachian times (Isherwood et al., 2013).

We find that a larger impactor is needed in these 3-D simulations than was required in the previous 2-D cylindrical calculations (Golabek et al., 2011), which is because the ratio of impactor mass to target mass scales differently with dimensionality, and also in the analytical study of Reese et al. (2010), which is mainly because they assumed a highe impact velocity.

3.5 Conclusions

Our model results indicate that the south polar giant impact (SPGI) is a viable hypothesis to explain the origin of the crustal dichotomy, and arguably explains better: a) the origin and the morphology of the Martian dichotomy; b) the distribution, the timing and the onset of the volcanic features and c) the decline of the magnetic field, than other hypotheses. Our

59 assumptions and consequent results are consistent with a range of geochemical, geological, volcanological and astronomical results that are briefly summarized here:

(i) Short time for core formation of 5-7 Ma after CAI, dependent on the size of the

impactor (the larger the size the shorter the time) which is within the 3-15 Ma

estimated by the geochemistry.

(ii) Extent of the highlands in latitude consistent with that observed on the map of

Mars.

(iii) Crustal thickness consistent with that inferred from the gravity measurements.

(iv) Presence of significant volcanism mainly at the Equator and in the southern

hemisphere.

(v) Migration of mantle plumes in the southern hemisphere.

(vi) Decline time of the transient magnetic field consistent with the observation of the

magnetic anomalies on both highlands and lowlands.

(vii) Decline time of the volcanism consistent with geological observations.

(viii) Impactor velocity similar to the escape velocity of the target.

(ix) Impact velocity avoiding disruption of the target.

(x) Dichotomy match with either impactor’s meso-sideritic or sideritic composition

We consider the impact time of 4 Ma after CAI as a lower bound for the dichotomy formation - the giant impact could have happened any time between 4-15 Ma after CAI, but not later as it would raise a problem with the geochemistry of the mantle-core differentiation.

If the giant impact occurred later, it might have geochemically re-equilibrated the core and mantle, thus altering the above mentioned times, although this is not necessarily the case - for

60 large impacts the impactor core can drop quite rapidly into the target core, scenario C in

(Nimmo and Agnor, 2006).

We find that a larger impactor is needed in these 3-D simulations than was required in the previous 2-D cylindrical calculations (Golabek et al., 2011), which is because the ratio of impactor mass to target mass scales differently with dimensionality, and also in the analytical study of (Reese et al., 2010), which is mainly because they assumed a higher impact velocity.

Further refinement of the available models could still be possible thanks to the seismic and thermal data that will come from the InSight Mission to Mars, providing new information on the size of the planet’s core and its surface thermal flux, and to the observation of the distribution of the volcanic centres on the surface.

Acknowledgements: GL acknowledges the support of the ETH Research Commission, Grant ETH-03 10-1. The authors thank G. J. Golabek for the impact subroutine of the I3ELVIS code and for a Matlab subroutine aimed at the visualization of the I3ELVIS output data. The authors wish also to thank two anonymous reviewers for the helpful comments that improved this paper. The data will be available upon request.

61 Chapter 4

Alignments of Volcanic Features in the Southern Hemisphere of Mars produced by Migrating Mantle Plumes

62 Alignments of Volcanic Features in the Southern Hemisphere of Mars produced by

Migrating Mantle Plumes

Giovanni Leone

ETH Zurich (Department of Earth Sciences, NO H28, Sonneggstrasse 5, 8092 Zurich, Switzerland; e-mail: [email protected]).

Published in the Journal of Volcanology and Geothermal Research, vol. 309, 78-95.

Abstract

Mars shows alignments of volcanic landforms in its southern hemisphere, starting from the equatorial regions and converging towards the south pole, and visible at global scale. These composite alignments of volcanoes, , shields, vents, heads of valley networks and massifs between the equatorial regions and the southern polar region define twelve different lines, fitted by rhumb lines (loxodromes), that I propose to be the traces of mantle plumes.

The morphology of the volcanic centres changes along some of the alignments suggesting different processes of magma emplacement and eruptive style. The diameters of the volcanic centres and of the volcanic provinces are largest at Tharsis and Elysium, directly proportional to the number of alignments starting from them. A minor presence of unaligned volcanic features is observed on the northern lowlands and on the highlands outside the 12 major alignments. The heads of channels commonly interpreted as fluvial valleys are aligned with the other volcanic centres; unaltered olivine is present along their bed-floors, raising severe doubts as to their aqueous origin. Several hypotheses have tried to explain the formation of

Tharsis with the migration of a single mantle plume under the Martian lithosphere, but the discovery of twelve alignments, six starting from Tharsis, favours the hypothesis of several

63 mantle plumes as predicted by the model of the Southern Polar Giant Impact (SPGI) and provides a new view on the formation of the volcanic provinces of Mars.

4.1 Introduction

Leone et al. (2014) found that the Southern Polar Giant Impact (SPGI) model for the early history of Mars specifically predicts, as a consequence of the impact, the formation of mantle plumes originating at four points located near the equator and 90 degrees of longitude apart and migrating from the equator towards the South Pole. Since mantle plumes are commonly the sources of volcanism, I examined all of the existing volcanic features on Mars documented in the literature and recorded their locations. I also examined large numbers of

CTX and HiRISE images in order to: (a) identify lava fields that might have radiated from central volcanic sources which have been masked by later impact craters or other modification processes and (b) identify other features that might be related to plumes. These other features specifically include the sources of what are generally assumed to be fluvial features formed by crustal deformation due to the presence of a mantle plume that caused fracturing that released water from deep (Bargery and Wilson, 2011); here I argue that these features are channels largely formed by erupted lava from deep or shallow magma reservoirs (Leone, 2014), but the correctness or otherwise of this interpretation does not detract from using these features as indicators of crustal stresses induced by mantle plumes, and so I included these locations. Given that I identified four volcanic provinces near the equator (Tharsis, Elysium, Syrtis Major and Western Arabia Terra) located 90 degrees apart as the expected four plume track starting points in the SPGI model, I have then assigned by eye the documented volcanic and related centres to plume tracks radiating from these locations towards the South Pole and checked using a Matlab programme if they formed alignments. Although this method of selecting candidates for each plume track can be

64 criticised as not proving statistically that it is the only way of assigning track members with minimal outliers, it is justified by the underlying physical requirement of the SPGI model that plumes do not meander at random in the mantle.

Three large volcanoes of Tharsis (Arsia, Pavonis and Ascraeus Montes) are aligned so perfectly that it would be amazing if this alignment formed just accidentally. Another group of massifs, the , was studied by Ghatan and Head (2002) who concluded

(without any statistical analysis) that these edifices were aligned, were not the remnants of rims or the result of tectonism, but instead were produced by sub-glacial volcanic eruptions. Aligned volcanic plains along the Dorsa Argentea formation were interpreted by Zhong (2009) and Hynek et al. (2011) as formed by a single mantle plume migration to Tharsis due to one-plate lithosphere rotation. These authors concluded that a similar trend does not exist elsewhere on Mars. However, the several alignments of volcanic features along different directions converging to the South Pole of Mars described in this paper neither supports this hypothesis (Leone et al., 2013) nor that of hypothetical superplumes below Tharsis and Elysium as suggested by Baker et al. (2007).

The SPGI model explains the formation of the Martian dichotomy as an alternative to other hypotheses, such as degree-1 mantle convection (Zhong and Zuber, 2001), multiple impacts

(Frey and Schultz, 1988) and a single giant impact in the northern hemisphere (Andrews-

Hanna et al., 2008; Marinova et al., 2008; Nimmo et al., 2008). The SPGI model suggests instead a vigorous phase of volcanism in the first 500 Ma after the giant impact, characterized by the migration of mantle plumes, and a phase of decreasing volcanism in the subsequent

500 Ma until the heat flux of the planet reached the current values when no active volcanism is observed. This sequence of events is in agreement with other petrological (Baratoux et al.,

2013) and thermal evolution (Baratoux et al., 2011; Morschhauser et al., 2011) studies but there is disagreement on the age when the volcanism ceased. Although there is again a general

65 agreement on the fact that the crust formed early in Martian history and it might have reached thicknesses of ~ 100-150 km upon mantle cooling and a heat flow decrease from 38 to 26 mW m-2, the age ranges from Amazonian (Baratoux et al., 2011) to Noachian-Hesperian

(Hauck and Phillips, 2002; Grott et al., 2005; Leone et al., 2014). In any case, a thick crust would make the rise of magma to the surface difficult, so that active volcanism would not happen even if the planet were still internally hot today (Baratoux et al., 2011).

The specific goals of this paper are: 1) the development of a global inventory of all the volcanic eruptive centres, including their diagnostic characteristics; 2) the assessment of potential alignments; 3) the comparison between crater counts and SPGI plume model ages to deduce plume migration rates; and 4) the first verification on the surface of Mars of the SPGI model for the formation of the Martian dichotomy and the onset of the volcanism.

4.2 Criteria for the identification and mapping of the volcanic features

A rich literature served as a basis for the identification of the volcanic features indicated in

Table 1 and the collection of their crater counts ages listed in Table 2 ( and Hiller,

1981; Hodges and Moore, 1994; Plescia, 1994; Maxwell and Craddock, 1995; Dohm and

Tanaka, 1999; Dohm et al., 2001a; Stewart and Head, 2001; Tanaka and Kolb, 2001; Ghatan and Head, 2002; Anguita et al., 2006; Head et al., 2006; Fassett and Head, 2008; Tanaka et al., 2008; Hoke and Hynek, 2009; Werner, 2009; Williams et al., 2009; Fenton and Hayward,

2010; Murray et al., 2010; Robbins et al., 2011; Xiao et al., 2012; de Pablo et al., 2013;

Isherwood et al., 2013; Michalski and Bleacher, 2013; Platz et al., 2013; Richardson et al.,

2013; Tanaka et al., 2014). Some other papers described the morphometric properties of the volcanic features (Plescia, 2004; Acocella, 2007; Grosse et al., 2012), and others showed the morphology of the typical lava flows on Mars (Keszthelyi et al., 2000; Jaeger et al., 2007;

Keszthelyi et al., 2008; Jaeger et al., 2010).

66 Feature Lateral Name Diameter Distance Latitude Longitude Cluster No cluster Feature Latitude Longitude Along Cross Lateral (km) (km) path path distance (km) (km) (km) Alignment 1 Olympus Mons 610.13 18°20'0"N 133°12'0"W Biblis 168.60 1112.85 2°21'17.45"N 123°49'00.16"W Unnamed vent 0.37 821.08 7°51'00.06"S 114°23'13.40"W Phoenicis Lacus 36 337.65 11°59'59.77"S 110°31'11.71"W Claritas 29 953.78 22°51'30.62"S 103°21'13.89"W Warrego Rise cluster 721.46 40°5'0"S 93°5'0"W 531.87 491.97 Thaumasia Fossae 22.00 383.43 46°32'19.17"S 92°15'44.20"W Aonia Mons 27.07 430.14 53°19'0"S 87°55'0"W 502.94 60°14'04.13"S 97°03'10.55"W Aonia Tholus 53.69 576.54 59°7'0"S 80°5'0"W

Cavi Angusti 640.04 1138.36 78°09'37.05"S 74°47'15.78"W Average distance 412.85 Hemispherical length 6475.29

Alignment 2 181.57 31°42'0"N 150°6'15.75"E 401 453.53 24°38'0"N 146°45'15.75"E Unnamed 60.54 2353.48 8°11'23.7"S 123°45'38.35"E

crater 269.77 1311.95 21°49'0"S 106°20'5.75"E Hadriaca Anseris Mons 52.51 355.64 29°47'0"S 86°46'0"E 241.21 Patera 30°12'01.02"S 92°47'22.02E Unnamed Amphitrites mons 23.61 2095.53 45°28'0"S 55°7'0"E 539.49 Patera 58°41'38.81"S 60°52'03.64"E Unnamed Malea mons 10.67 1285.75 58°19'0"S 37°36'0"E 535.85 Patera 63°32'26.38"S 51°35'00.98"E Unnamed Pityusa mons 26.99 539.69 60°56'0"S 28°53'0"E 501.69 Patera 66°52'57.11"S 36°51'36.12"E Unnamed mons 45.97 197.85 63°29'0"S 17°39'0"E Sisyphi Montes 200 473.19 68°37'0"S 11°35'0"E 617.4 667.42 Unnamed Sisyphi Cavi 423.63 976.35 71°54'41.74"S 4°31'40.06"W 207.2 Mons 66°25'44.93"S 2°56'06.75"W Sisyphi Tholus 27.52 859.19 75°40'49.11"S 18°32'27.13"W Average distance 838.63 Hemispherical distance 10902.15

Alignment 3 Ascuris Issedon Planum 617.66 40°35'08.85"N 80°46'32.36"W 307.65 Paterae 38°07'09.27"N 90°15'31.00"W Uranius Issedon Patera 114 1032.07 27°20'0.0"N 92°82'16.13"W 414.99 Tholus 36°03'22.72"N 94°49'56.44"W Ceraunius 128.58 275.20 23°59'13.28"N 97°15'10.96"W 105.84 Tholus 26°15'02.36"N 97°34'14.73"W 456.4 1132.71 11°11'40.08"N 104°24'1.53"W 366.53 786.81 0°55'14.87"N 112°35'37.81"W

68 470 731.75 9°10'12"S 120°16'12"W Sirenum Tholus 53.9 2131.32 35°0'0"S 215°14'9"E Sirenum Mons 122.86 264.53 38°7'0"S 212°7'9"E 105.58 561.50 46°52'49.92"S 153°03'43.10"W Phaethontis 87.97 133.62 48°25'14.22"S 155°30'2.36"W Unnamed colles 123.75 652.91 55°11'15.22"S 164°10'30.75"W Chronius Mons 56.14 866.98 61°28'40.31"S 178°4'38.79"E Ulyxis Rupes 383.09 618.13 68°46'23.89"S 159°59'27.01"E Thyles Rupes 548.75 542.80 70°28'0"S 133°57'0"E Promethei Planum 831.28 839.22 79°20'0"S 88°3'0"E Average distance 754.97 Hemispherical length 10569.55

Alignment 4 Unnamed Mons 14.03 15°3’15.75”S 179°21’35.82”E Unnamed Mons 13.53 97.94 15°54’36.26”S 177°57’5.14”E Apollinaris Tholus 32.39 161.51 17°38’21.44”S 175°45’5.66”E Zephyria Tholus 35.95 202.13 19°48’0.97”S 172°56’9.86”E Unnamed mons 53.61 293.14 24°36’10.83”S 171°57’0.36”E

69 Unnamed Tholus 68.52 585.35 23°54’45.12”S 165°03’16.22”E Unnamed Tholus 106.43 299.41 35°4’8.36”S 159°46’40.58”E Unnamed Tholus 68.34 412.48 41°34’26.32”S 156°39’29.36”E Electris Mons 104.47 295.28 45°38’13.04”S 152°40’2.34”E Eridania Mons 143.29 873.77 57°4’0.07”S 137°34’21.5”E 123.27 537.91 65°27’10.37”S 130°31’47.75”E Thyles Montes 380 222.64 68°58’10.37”S 126°59’47.75”E 112.69 216.65 72°17’10.37”S 122°37’47.75”E Average distance 322.94 Hemispherical length 4198.21

Alignment 5 Arabia Terra calderas cluster 38°41’0”N 5°31’42.75”E 360.54 256.12 Mamers Mamers Valles cluster 871.79 30°40’16.63”N 19°48’49.76”E 210.54 102.00 414.3 Valles 35°22’25.13”N 25°31’15.33”E Unnamed filled crater 18.5 421.37 28°16’34.44”N 26°18’4.74”E Locras Valles 351.32 1829.82 6°19’29.01”N 47°56’26.87”E Schroeter 291.59 687.22 1°54’19.15”S 55°59’38.7”E Naro Vallis 442.72 441.69 6°25’32.15”S 61°57’47.70”E - Hellas ridge 700 1044.92 14°46’42.67”S 77°40’54.31”E 375 Tyrrhenum Mons 269.77 1671.41 21°24’0”S 106°38’5.75”E Unnamed 69.03 2131.23 37°46’29.01”S 138°38’26.87”E

70 tholus 88.96 1055.32 48°59’32.45”S 157°47’10.06”E

Rossby 80.42 410.12 47°31’41.44”S 167°54’57.13”E Average distance 960.44 Hemispherical length 10564.89

Alignment 6 Elysium Mons 401 24°38’0”N 146°46’15.75”E Albor Tholus 158.38 382.80 19°4’0”N 150°15’15.75”E Tholi cluster 1263.51 2°55’55.09”N 164°25’25.45”E 367.89 308.48 275.4 873.58 8°34’44.13”S 174°10’47.56”E Unnamed mons 13.54 534.92 16°5’49.26”S 179°16’30.36”E Unnamed mons 17.90 52.10 16°51’47.89”S 179°44’25.36”E Unnamed tholus 84.22 776.88 29°0’0”S 175°20’15.75”W Unnamed Simois mons 48.81 467.85 35°20’0”S 169°48’15.75”W 200.68 Colles 37°19’29.69”S 173°12’01.72”W Unnamed tholus 24.02 224.84 38°41’41.29”S 167°35’55.1”W Unnamed tholus 52.67 161.07 40°37’50.49”S 165°08’8.73”W Unnamed tholus 70.28 143.93 42°35’52.39”S 163°13’23.41”W Average distance 443.77 Hemispherical 4881.48

71 length

Alignment 7 548.02 40°12'0"N 109°48'32.75"W Jovis Jovis Fossae 58.07 1256.31 19°00'59.94"N 112°07'30.78"W 313.79 Tholus 19°06'10.37"N 112°05'47.75"W Pavonis Mons 366.53 1076.73 0°55'14.87"N 112°35'37.81"W Claritas cluster 1395.83 22°39'57.35"S 113°27'34.23"W 1037.92 420.54 Unnamed crater 75.58 1328.26 44°58'28.38"S 115°55'35.76"W Unnamed channels 71.24 641.39 55°38'48.98"S 118°49'51.86"W Unnamed channel 90.58 438.18 62°40'06.83"S 119°51'13.44"W 104.31 67°44'59.60"S 115°22'53.33"W Australe Scopuli 504.58 1117.46 81°49'10.88"S 121°13'13.76"W Average distance 906.77 Hemispherical length 7254.16

Alignment 8 Unnamed channel 60 23°45'40.08"N 17°26'00.04"W Unnamed Coogon crater 55 154.82 21°04'40.60"N 17°26'55.65"W 251.48 North 19°06'07.46"N 21°29'10.61"W Coogoon 634.63 247.74 18°41'05.60"N 14°25'08.18"W 444.79 373.54 South 16°04'09.35"N 19°25'09.71"W crater near 71 503.48 10°38'48.18"N 11°17'47.83"W

72 Thymiamata 120 336.24 4°58'12.93"N 10°04'04.46"W 320.14 6°28'15.69"N 4°54'21.31"W 145.21 453.16 2°20'45.12"S 8°23'30.76"W Bashkaus Marikh Samara Valles 661.84 1734.60 31°37'27.41"S 8°54'18.70"W 870.31 Vallis 24°43'53.23"S 5°48'07.16"E 90.84 1166.19 49°36'33.97"S 1°45'06.02"E 200.15 46°58'18.16"S 4°55'29.34"E Unnamed mons 22.30 382.21 55°50'0"S 4°25'0"E Unnamed mons 39.63 261.63 59°56'0"S 1°40'0"E Unnamed tholus 28.48 105.83 61°47'0"S 2°23'0"E Unnamed mons 26.08 104.52 63°27'0"S 3°24'0"E Unnamed mons 11.60 101.19 65°09'0"S 2°56'0"E Unnamed mons 38.34 69.98 66°15'0"S 3°52'0"E Unnamed mons 38.96 142.41 68°38'0"S 4°58'0"E Unnamed colles 123.56 334.70 74°14'0"S 5°05'0"E Australe Montes 41.92 223.02 77°29'0"S 12°40'0"E Australe Montes 49 96.68 79°06'0"S 12°18'0"E Australe Montes 83.13 129.85 81°11'0"S 16°0'0"E Australe Montes 49.61 127.17 82°55'0"S 25°22'0"E Average distance 326.03 Hemispherical 6675.42

73 distance

Alignment 9 Olympus Mons 610.13 18°20'0"N 133°12'0"W Unnamed Vent 1.433 1074.76 6°30'32.56"N 120°12'33.31"W Pavonis Mons 366.53 543.70 0°47'0"N 112°20'0"W 1190.31 926.56 6°55'0"S 99°0'0"W Unnamed tholus 118.01 1226.97 15°46'7.08"S 79°52'40.49"W Unnamed Unnamed mons 111.19 1150.94 20°24'16"S 60°16'48.57"W 141.76 mons Unnamed mons 182.62 872.21 23°54'0"S 45°0'0"W Erythraea Fossa 155.19 865.41 27°35'13.31"S 29°32'00.55"W Vulcani Pelagus 8.70 831.03 34°40'58.36"S 15°00'03.71"W Average distance 712.98 Hemispherical length 5585.79

Alignment 10 Meroe Nili Patera 67.51 8°58'0"N 67°10'5.75"E 147.87 Patera 6°58'44.91"N 68°45'55.40"E Nakhtong Vallis 61.08 1631.87 2°52'29.01"S 42°15'26.87"E Naktong Verde Vallis 133 721.20 6°57'53.48"S 29°10'50.35"E 213.39 Vallis 2°52'29.01"S 42°15'26.87"E Brazos Valles 387.51 352.82 7°20'58.71"S 22°49'37.47"E

74 Evros Vallis 358.01 501.39 11°24'31.63"S 15°27'47.31"E Unnamed crater 55.37 631.09 15°25'42.43"S 5°22'38.45"E Peta 75.75 882.80 21°15'31.30"S 9°05'15.42"W Unnamed crater 32.46 989.86 29°42'22.09"S 25°04'08.35"W Average distance 713.88 Hemispherical length 5711.03

Alignment 11 Labeatis Mons 42.78 37°27'45.12"N 75°87'18.14"W Head Labeatis Fossae 20 504.05 28°44'27.32"N 81°58'12.89"W Tharsis Tholus 149.3 1138.15 13°14'19.10"N 90°41'40.20"W Unnamed vent 10 957.13 2°07'12.99"N 101°20'09.04"W Noctis Labyrinthus 1190.31 383.63 6°19'20.52"S 101°12'29.27"W cluster 358.04 10°19'40.85"S 106°40'57.42"W 524.06 274.59 Claritas cluster 1005.87 32°24'19.73"S 116°44'48.78"W 1014.07 701.61 Unnamed mons 179.64 1150.11 49°17'0"S 134°3'0"W Chico Valles cluster 1206.87 67°12'0"S 158°35'0"W 193.27 152.36 Average distance 744.87 Hemispherical length 6703.85

Alignment 12 Olympus 610.13 18°20'0"N 133°12'0"W

75 Mons Head of valley 1.48 671.82 9°10'18.57"N 126°53'10.81"W Ulysses Patera 57.86 483.82 2°56'52.30"N 121°25'17.81"W Unnamed vent 0.63 631.45 3°53'33.88"S 113°16'21.15"W Syria Planum cluster 811.93 10°19'40.85"S 106°40'57.42"W 524.06 274.59 Unnamed Solis massif 63.48 1717.55 31°09'58.24"S 78°11'04.61"W 191.49 Lacus 29°23'20.05"S 89°45'04.55"W Unnamed massif 100.05 150.85 32°48'00.13"S 76°01'05.04"W Unnamed Ogygis crater 55 637.25 40°13'00.11"S 67°28'07.34"W 332.11 Regio 46°46'50.13"S 70°01'53.35"W Unnamed crater 61 79.38 40°43'45.30"S 65°51'55.16"W Unnamed Argyre Mons 60.58 932.14 50°18'58.64"S 48°01'40.07"W 581.61 crater 53°57'25.25"S 66°18'14.82"W 19.09 127.33 52°00'41.62"S 45°03'53.03"W Unnamed Lodwar 33.39 203.72 55°05'32.38"S 43°19'02.34"W 510.12 mons 65°03'56.53"S 39°57'16.15"W Average distance 537.27 Hemispherical length 4660.15

Unaligned - Highlands Unnamed crater 63.46 22°18'12.13"S 137°42'28.28"E Herschel 297.92 14°28'46.91"S 129°53'17.31"E Unnamed crater 35.58 41°12'23.68"S 171°13'50.86"E Unnamed 47.55 45°06'19.08"S 175°15'20.00"W

76 massif Unnamed massif 124.48 46°08'40.00"S 178°12'03.29"E

Unaligned - Lowlands Scandia Tholi 398.27 73°54'49.55"N 158°43'20.12"W Abalos Colles 235.83 76°49'42.79"N 71°39'33.98"W Hyperboreus Labyrinthus 111.97 80°17'07.78"N 59°35'13.50"W

Table 1. List of the volcanic features (geographic data points) of the alignments. The diameters indicated in column 2 were taken from the Gazetteer of the for the named features or geodetically measured with ArcGIS. The geodetic distance between consecutive features is indicated in column 3. The coordinates of the features are provided in columns 4 and 5. Columns 6 and 7 indicate whether the volcanic feature is also the centroid of a cluster of surrounding similar volcanic features and the eventual maximum along-track and cross- track extent of the cluster. The column 8 indicates the lateral distance of the laterally farthest feature belonging to the alignment. Finally, the column entitled “lateral features” indicates the names (if available) of the laterally shifted volcanic features, along with their coordinates, assigned to the alignment.

77 Feature Name Main ages of formation and activity from crater counts (Ga)

Alignment 1

Olympus Mons 4.0-3.67-2.54 (Isherwood et al., 2013; Neukum et al, 2004) Biblis Tholus 3.68-2.80 (Werner, 2009; Plescia, 1994) Unnamed vent 1.5-1.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Phoenicis Lacus 3.2 (Scott and Tanaka, 1986; Dohm et al., 2001a; Tanaka et al., 2014) Claritas 3.7-3.5 (Dohm et al., 2001a) Warrego Rise 4.04-3.7 (Dohm et al., 2001b; Xiao et al., 2012) Thaumasia Fossae 3.7-3.5 (Tanaka and Scott, 1986; Dohm et al., 2001a) Aonia Mons 4.1-3.7 (Anguita et al., 2006) Aonia Tholus 4.1-3.7 (Anguita et al., 2006) Cavi Angusti 3.7 (Tanaka and Kolb, 2001)

Alignment 2

Hecates Tholus 3.8-3.5 (de Pablo et al., 2013; Robbins et al., 2011; Werner, 2009) Elysium Mons 3.7-3.0 (Werner, 2009; Robbins et al., 2011; Pasckert et al., 2012) Unnamed crater 3.9-3.7 (Tanaka et al., 2014) Tyrrhenum Mons 3.7-3.4 (Werner, 2009; Robbins et al., 2011; Williams et al., 2009) Hadriaca Patera 3.66-3.55 (Werner, 2009; Robbins et al., 2011; Williams et al., 2009)

Anseris Mons 3.6 (Scott and Carr, 1978; Scott and Tanaka, 1986; Tanaka et al., 2014) Unnamed mons 3.6 (Scott and Carr, 1978; Scott and Tanaka, 1986; Tanaka et al., 2014) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Sisyphi Montes 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Sisyphi Cavi 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Sisyphi Tholus 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010)

Alignment 3

Ascuris Planum 3.9 (Tanaka et al., 2014) Uranius Patera 3.7-3.19 (Werner, 2009; Robbins et al., 2011) 3.59 (Robbins et al., 2011) Ascraeus Mons 3.3-1.5 (Neukum and Hiller, 1981; Murray et al., 2010; Robbins et al., 2011) Pavonis Mons 3.0-2.5 (Scott et al., 1998; Neukum and Hiller, 1981) Arsia Mons 3.54-3.2 (Neukum and Hiller, 1981; Werner, 2009) Sirenum Tholus 3.99 (Xiao et al., 2012) Sirenum Mons 4.05 (Xiao et al., 2012) Li Fan 3.8 (Tanaka et al., 2014) Phaethontis 3.8 (Tanaka et al., 2014) Unnamed colles 3.8 (Tanaka et al., 2014) Chronius Mons 3.8 (Tanaka et al., 2014) Ulyxis Rupes 3.7 (Tanaka et al., 2014) Thyles Rupes 3.7 (Tanaka and Kolb, 2001; Tanaka et al., 2014) Promethei Planum 3.5 Tanaka and Scott, 1987

Alignment 4

Unnamed Mons 3.9 (Tanaka et al., 2014) Unnamed Mons 3.9 (Tanaka et al., 2014) Apollinaris Tholus 3.9 (Tanaka et al., 2014) Zephyria Tholus 3.9 (Scott and Tanaka, 1986; Stewart and Head, 2001) Unnamed mons 3.8 (Tanaka et al., 2014) Unnamed Tholus 3.8 (Tanaka et al., 2014) Unnamed Tholus 3.8 (Tanaka et al., 2014) Unnamed Tholus 3.8 (Tanaka et al., 2014) Electris Mons 3.8 (Greeley and Guest, 1987) Eridania Mons 3.8 (Greeley and Guest, 1987) Byrd 3.8 (Greeley and Guest, 1987) Thyles Montes 3.7 (Tanaka and Kolb, 2001; Tanaka et al., 2014) Burroughs 3.0 (Scott and Carr, 1978; Tanaka et al., 2014)

Alignment 5

Arabia Terra calderas 3.9-3.7 (Michalski and Bleacher, 2013) Mamers Valles 3.8-3.7 (Mangold, 2012) Unnamed filled crater 3.8 (Scott and Tanaka, 1986; Tanaka et al., 2014) Locras Valles 3.9-3.7 (Maxwell and Craddock, 1995) Schroeter 3.8-3.6 (Head et al., 2006) Naro Vallis 3.8-3.6 (Head et al., 2006) Huygens-Hellas ridge 3.8-3.6 (Head et al., 2006) Tyrrhenum Mons 3.7-3.4 (Werner, 2009; Robbins et al., 2011; Williams et al., 2009) Unnamed tholus 3.7-3.5 (Greeley and Guest, 1987) Huggins 3.7-3.5 (Greeley and Guest, 1987) 3.7-3.5 (Greeley and Guest, 1987)

Alignment 6

Elysium Mons 3.7-3.0 (Werner, 2009; Robbins et al., 2011; Pasckert et al., 2012) Albor Tholus 3.53-1.6 (Robbins et al., 2011; Werner, 2009) Cerberus Tholi 3.8-3.7 (Scott and Tanaka, 1986; Scott and Carr, 1978) Apollinaris Mons 3.72-3.52 (Werner, 2009; Robbins et al., 2011) Unnamed mons 3.65 (Greeley et al, 2005) Unnamed mons 3.65 (Greeley et al, 2005) Unnamed tholus 3.7 (Tanaka et al., 2014) Unnamed mons 3.7 (Tanaka et al., 2014) Unnamed tholus 3.7 (Tanaka et al., 2014)

79 Unnamed tholus 3.7 (Tanaka et al., 2014) Unnamed tholus 3.7 (Tanaka et al., 2014)

Alignment 7

3.8-3.6 (Neukum and Hiller, 1981; Ivanov and Head, 2006; Werner, Alba Mons 2009) Jovis Fossae 3.7 (Neukum and Hiller, 1981) Pavonis Mons 3.0-2.5 (Scott et al., 1998; Neukum and Hiller, 1981) Claritas 4.04-3.7 (Dohm et al., 2001a; Xiao et al., 2012) Unnamed crater 3.8-3.7 (Dohm et al., 2001a) Unnamed channels 3.7-3.6 (Scott and Tanaka, 1986) Unnamed channel 3.5 (Scott and Carr, 1978; Tanaka et al., 2014) Australe Scopuli 3.0 (Scott and Carr, 1978; Tanaka et al., 2014)

Alignment 8

Unnamed channel 4.1-4.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Unnamed crater 4.1-4.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Mawrth Vallis 4.0-3.9 (Loizeau et al., 2012) crater near Thymiamata 3.9 (Scott and Carr, 1978; Scott and Tanaka, 1986) Crommelin 3.8 (Wilhelms, 1974) Miyamoto () 3.74 (Hynek and Phillips, 2008) Bashkaus and Samara Valles 3.8-3.6 (Grant and Parker, 2002; Bouley et al., 2010) Maunder 3.7 (Tanaka et al., 2014) Unnamed crater 3.7 (Tanaka et al., 2014) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed tholus 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed mons 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Unnamed colles 3.7-3.5 (Ghatan and Head, 2002; Fenton and Hayward, 2010) Australe Montes 3.7-3.5 (Scott and Carr, 1978; Tanaka et al., 2014) Australe Montes 3.7-3.5 (Scott and Carr, 1978; Tanaka et al., 2014) Australe Montes 3.7-3.5 (Scott and Carr, 1978; Tanaka et al., 2014) Australe Montes 3.7-3.5 (Scott and Carr, 1978; Tanaka et al., 2014)

Alignment 9

Olympus Mons 4.0-3.67-2.54 (Isherwood et al., 2013; Neukum and Hiller, 1981) Unnamed Vent 1.5-1.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Pavonis Mons 3.0-2.5 (Scott et al., 1998; Neukum and Hiller, 1981) Noctis Labyrinthus 3.2-3.0 (Tanaka et al., 2014) Unnamed tholus 3.7 (Scott and Carr, 1978; Scott and Tanaka, 1986; Tanaka et al., 2014)

80 Unnamed mons 3.7 (Dohm and Tanaka, 1999) Unnamed mons 3.7 (Grant et al., 2011) Erythraea Fossa 3.7-3.5 (Buhler et al., 2011; Scott and Tanaka, 1986) Vulcani Pelagus 3.8 (Scott and Tanaka, 1986)

Alignment 10

Nili Patera 3.9-3.7 (Michalski and Bleacher, 2013) Unnamed mons 3.7 (Bouley et al., 2009) Verde Vallis 3.72 (Hoke and Hynek, 2009) Brazos Valles 3.71 (Fassett and Head, 2008) Evros Vallis 3.63 (Hoke and Hynek, 2009) Unnamed crater 3.6 (Scott and Tanaka (1986) Peta 3.68 (Hoke and Hynek, 2009) Unnamed crater 3.6 (Scott and Tanaka (1986)

Alignment 11

Labeatis Mons 4.1-3.9 (Scott and Tanaka, 1986; Tanaka et al., 2014) Head Labeatis Fossae 3.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Tharsis Tholus 3.54-3.24 (Werner, 2009; Robbins et al., 2011) Unnamed vent 1.5-1.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Noctis Labyrinthus 3.2-3.0 (Tanaka et al., 2014) Syria Planum 3.7-2.8 (Richardson et al., 2013) Claritas 4.04-3.7 (Dohm et al., 2001a; Xiao et al., 2012) Unnamed mons 3.59 (Platz et al., 2013) Chico Valles 3.0 (Tanaka et al., 2014)

Alignment 12

Olympus Mons 4.0-3.67-2.54 (Isherwood et al., 2013; Neukum and Hiller, 1981) Head of Valley in Gigas Sulci 3.5-3.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Ulysses Patera 3.94-3.73 (Werner, 2009) Unnamed vent 1.5-1.0 (Scott and Carr, 1978; Scott and Tanaka, 1986) Syria Planum 3.7-2.8 (Richardson et al., 2013) Coracis Fossae North 3.9-4.09 (Scott and Tanaka, 1986; Xiao et al., 2012) Coracis Fossae South 3.9-3.92 (Scott and Tanaka, 1986; Xiao et al., 2012) Phrixi Regio 3.9 (Scott and Tanaka, 1986) Argyre Mons 3.8 (scott and Tanaka, 1986) Mari 3.7 (Scott and Tanaka, 1986) Lodwar 3.7 (Scott and Tanaka, 1986)

Table 2. List of the volcanic features (geographic data points) of the alignments. The ranges of age, or the single ages, are those taken from the literature that is indicated in brackets..

The recognized volcanic landforms were thus considered geographic data points, that is, source (or eruptive) data points for the assessment of the alignments. Although very many

81 volcanic features were already available from the literature and thus incorporated in this paper, many other new volcanic features were discovered and only some of which have been officially named (e.g. Aonia Mons, Aonia Tholus, Eridania Mons, Sirenum Mons, Sirenum

Tholus). Some volcanic features were partially or totally destroyed by impact craters and only the presence of surrounding lava flows has made it possible to infer the presence of the original structure (Xiao et al., 2012). The geomorphic criteria for the identification of the geographic data points are those indicated for the morphometric properties of the volcanic features of high relief (Plescia, 2004; Grosse et al., 2012) and low relief (Acocella, 2007;

Bleacher et al., 2009; Richardson et al., 2013). The typical features are shown in Figure 4.1.

The identification of the largest volcanic features (i.e. Alba and Olympus Mons) was already possible at global scale 1:25M on the Topographic Map of Mars (USGS, 2003) where the main directions of some alignments were already indicated by the largest volcanoes.

However, after this first observation, the refinement of the mapping was performed at regional scale looking for every occurrence all over the surface of the planet, on maps of 1:5000000, of smaller shields, calderas and massifs between 200 and 50 km in diameter. The final refinement was done at local scale using CTX and HiRISE imagery again on the whole surface of Mars looking for massifs, low shields and every volcanic feature that might appear as a source point for lava flows or heads of valleys between 50 and 1 km diameter that are not immediately evident in THEMIS imagery. All the surface of Mars covered by any available imagery was observed, and all the volcanic features outside any alignment were listed in the unaligned section. The lowest limit of 1 km was determined from the diameter of the heads of small valleys and of the vents on top of the smallest low-relief volcanic landforms encountered on the surface of Mars but even vents of hundreds of meters were also observed in the CTX images of Tharsis. All the diameters of the volcanic landforms in the alignments were then listed in Table 1. In the case of clusters of volcanic features, the along-path and

82 cross-path distances are given. The diameters of the named features were taken from the

Gazetteer of the Planetary Nomenclature, and the diameters of the unnamed features were geodetically measured through ArcGIS. The observations were supported by the Mars Orbiter

Laser Altimeter (MOLA) dataset for the volumetric calculations and the analysis of the altimetry mentioned in the next sections. These datasets are already easily accessible on the web so it was not necessary to show them all due to the limited space in figures. Priority was given to visible images, and THEMIS mosaics were used where gaps in CTX data coverage was found.

The best way to unambiguously discern uncertain features with weakly supportive characteristics and alternative interpretations like craters and heads of valley networks (Fig.

4.1a, Fig.4.1b and Fig. 4.1f) was to identify lava flows in the inter-crater plains and to track them back to their sources. Some valley networks probably were initially incised directly from the flows according to a plausible mechanism of the preferential pathways explained by

Bleacher et al. (2015) and then subsequent channelization of lava deepened them further, so that the valley networks appeared draped with lava. The heads of the valleys formed directly from lava flows were excluded from the alignments because they were formed away from the eruptive points. Other features like fields of vents (Fig. 4.1c), calderas (Fig. 4.1d), minor shields (Fig. 4.1e), Noctis Labyrinthus (Fig. 4.1g), major shields (Fig. 4.1h) and massifs (Fig.

4.1i) with either flat-tops and/or dissected flanks were considered unambiguous.

The figures shown in the next section are meant to illustrate all the new features identified that are not already mentioned in the literature and to show how lava flows can be helpful in the identification of volcanic features that have ambiguous, partial, or no interpretation yet.

Where possible, features commonly considered of tectonic origin like fossae and labyrinthi

(Fig. 4.1g) were analysed with MOLA data to assess a possible volcanic origin in addition to

83

Figure 4.1. Typical features considered as geographic data points for the alignments. a) Peta crater, probably a volcanic shaped by nested impact craters merged together by the infill of volcanic material that eroded the dividing rims, the Paraná Valles originated from its western slope (very near to the rim) with likely overflow from the central caldera; b) the Rossby crater is a probable volcanic caldera of circular shape instead, located on a topographic high, which also shows overflows on the western rim from which the Drava Valles originated; c) unnamed vent on Syria Planum; d) some of the circum-Hellas calderas, their rims are evidently smoothed by abundant overflows and do not show volcanic structure; e) volcanic calderas on tholus structure; f) head of spreading from volcanic mons structure; g) Noctis Labyrinthus is a network of collapsed lava tubes from which originated the Valles Marineris, although considered as an eruptive point not all the labyrinthi structures might be considered eruptive sources as likely being collapsed lava tubes where lava flowed; h) Olympus Mons, one of the largest shield volcanoes of Mars and the highest of the Solar System; i) Chronius Mons, volcanic massif with flat top, similar to other massifs located in the Sisyphi Montes.

the methodology used by Hodges and Moore (1994) and by Leone (2014). Particular care has been placed on distinguishing calderas from melt generating impact craters, the latter being excluded for their different significance, although both might be surrounded by lava flows.

Generally calderas have a low and smooth rim due to periodic lava overflows, concentric inner and outer depressions, and an irregular sub-circular shape (Fig. 4.1d). An impact crater has higher rims and a more circular shape, sometimes characterized by a central peak, with distinguishable ejecta (Xiao et al., 2012). Other key observations were the MOLA profiles

84 showing flat floors in putative impact craters, compared to the profiles published in a work specifically dedicated to the characterization of the topography of the impact craters (Mahanti et al., 2014). The MOLA profiles are not shown for reasons of limited space in the figures but they are all available on the MOLA website and they can all easily be checked for every mentioned feature. Many craters with tributaries and distributaries located near large volcanic centres and showing infill of lava flows were not considered as geographic data points, although they might also have an underground feeder, because the origin of the fluid is located elsewhere and there is no way to unequivocally demonstrate a volcanic internal eruptive point. High-elevation impact craters partially filled by lava flows, as typified by lunar crater Plato (Thompson et al., 2006), characterized by ridged plains and/or surrounded by radial valleys and/or sinuous channels like Rimae Plato were considered geographic data points because lava cannot cross high crater rims from surrounding surface flows. Floor- fractured craters (Schultz and Glicken, 1979) analogous to those observed on the Moon

(Schultz, 1976) showing evident characteristics of eruptive points were also considered as geographic data points but eventually excluded whenever there was evidence of external feeding (i.e. surface flows coming from other sources).

Last but not least, it is worth to mention the chaos terrains and why they are excluded from being geographic data points for the alignments. There is a heterogeneous literature regarding the specific formation of chaos terrains on Mars, including intrusive activity

(Chapman and Tanaka, 2002; Meresse et al., 2008; Rodriguez et al., 2005), erosive activity by groundwater (Harrison and Chapman, 2008; Zegers et al., 2010), volatile sublimation

(Pedersen and Head, 2011), tectonic activity (Spagnuolo et al., 2011), and excavation of magma (Sharp, 1973). The latest geologic analysis of the Valles Marineris (Leone, 2014) has shown that the chaos terrains are mostly located within the chasmata, wide channels carved and eroded by the lava flows erupted from the Tharsis Montes via Labyrinthus Noctis, where

85 also the heads of other nearby outflow channels (e.g. Echus and Juventae ) are connected with other chasmata through underground feeders and lava tubes. It will also be shown that other chaos terrains not located within chasmata but still on the highlands like

Gorgonum Chaos and are connected at the surface to the lava flows of volcanic landforms. Thus, the chaos terrains cannot be considered as formed by direct eruptive activity related to the migration of a mantle plume, and so are not included here.

Although lava fields and lava channels are obviously features of volcanic origin, they were not considered as geographic data points for the alignments. The reason for this important distinction is related to the evidence that lava fields and lava channels take different directions for thousands of kilometres while, again, the aim of this research is understanding whether there is correlation between the locations of the eruptive points and the migration of the mantle plumes that fed such activity.

The centroid (or geometric centre) for all the identified volcanic features was pinpointed with its longitude and latitude according to the following criteria where the lava erupted to the surface:

· In the case of volcanic structure with central crater or caldera (patera), the geometric

centre of the central main crater.

· In the case of field of vents or impact destroyed structure, the geometric centre of the

cluster or the geometric centre of the radial directions from where the lava flows

spread.

· In the case of massifs, the highest peak or the head of the main lava channel.

· In the case of elongated features (fossae, fluvial valleys, labyrinthi), the head of the

main lava channel.

· In the case of clusters of massifs or calderas, the geometric centre of the whole cluster

was selected.

86

Figure 4.2. A) Global map of the volcanic features described in the literature, the volcanic features are indicated by a white dot. B) Global map of the newly-discovered volcanic features. C) The CTX image P01_001353_1510_XN_29S136W, centred at 28.27 S – 136.42 W and taken on , shows a crater 50 km in diameter half buried by lava flows (the white arrows indicate the latest flow fronts); D) Global union map of the previously known and the newly-discovered volcanic features; the white rectangles labelled b

87 and d show the locations of the areas imaged in the panels c and e. E) The CTX image G02_018853_1837_XN_03N195W, centred at 3.77 N – 195.87 W and taken on Elysium Planitia, shows only the remaining rim of a half buried crater larger than 50 km. F) Global union map of the interpreted alignments; the colour of the dots refers to the legend indicated in the centre-left part of the figure. G) Global union map of the interpreted alignments, the geodesic lines were traced from the start to the end geographic data points with the End Arc Segment tool of ArcGIS.

The centroids are indicated by a single colour in Fig. 4.2a, to exclude any bias in interpretation, which show the locations of features generally accepted in the literature as being the sources of volcanic materials. Fig. 4.2b shows the locations of the features that I have newly identified as volcanic sources, as described in the next section. Fig. 4.2d shows the combination of both of these two data sets, and Fig. 4.2f shows how all of the features were assigned to the identified alignments by a coloured dot: purple for Alignment 1, yellow for Alignment 2, green for Alignment 3, red for Alignment 4, orange for Alignment 5, blue for Alignment 6, for Alignment 7, grey for Alignment 8, dust violet for Alignment 9, turquoise for Alignment 10, heliotrope for Alignment 11, and Sahara sand for Alignment 12.

At last, Fig. 4.2g shows the map of the interpreted alignments in which the start and end points for each alignment are connected by a curve traced with the End Arc Segment tool of

ArcGIS.

88 4.3 Description and interpretation of the alignments

All the volcanic features found on Mars were drawn on ArcGIS maps (Fig. 2). These

ArcGIS maps are in cylindrical projection to obtain better accuracy towards the equator of the planet where the main volcanic features are located. The polar distortion typical of this projection was compensated with a stereographic projection in Fig. 3c to increase the accuracy towards the South Pole. All the fields of vents and all the low shields of Tharsis were not added to these maps to avoid a confused figure at global scale but they belong to the alignments, this is shown in Fig. A.12 of the Appendix. The most striking result that comes out from Figs. 4.2f and 4.2g is the radial pattern of several alignments spreading from Tharsis.

Olympus Mons is the start of three alignments; its magma reservoir was likely fed by all these plumes, and this explains its size and height. It is interesting to note how no volcanoes spread and align northward as predicted by the SPGI model. Tharsis is the origin of the largest number of plumes (six) that cross each other at different times forming alignments of volcanoes, low-shields, and vents. Elysium follows with the origin of two alignments and it is also interesting to note that the size of the volcanic province is proportional to the number of the plumes formed; the reason for such a difference in number of originated plumes between

Tharsis and the other provinces is still unknown.

Given the relatively short distance among nearby alignments within Tharsis, it might seem difficult to assign the volcanic features to the various alignments calling for specific statistical analyses. However, Fig. 4.2d shows that the radial directions formed by the larger and the smaller volcanoes spreading from Tharsis indicate the alignments so clearly even in the single-coloured map that they were already easy to interpret in Fig. 4.2f without any statistical analysis. A geodesic tracking of the alignments seemed appropriate to understand whether they were real and was thus done. Two main guides must be used when tracking an alignment: the start and the end of the alignment to establish its direction and the lateral distance of any other volcanoes from the track of the alignment to infer the size of the plume or any eventual interaction with another sub-parallel alignment. The start and the end points were already indicated by the SPGI model, namely the large features starting from the previously mentioned four points located 90 degrees apart and obviously the South Pole for the primary alignments, thus providing the directions in which to observe the alignments even through an apparently confusing cluster of volcanoes that forms the bulge of Tharsis. All the other regions of Mars, much less cluttered than Tharsis, showed the alignments in a clearer way.

The most compelling type of alignment might consist of the largest number of volcanoes of similar type and size (e.g., Tharsis Montes, Sisyphi Montes), with (not necessarily) similar along-track spacing following a geodesic line with little or limited cross-track deviation. The minimum number of volcanic features found per alignment was 8 along Alignment 10 and the maximum number was 21 along Alignment 8 (Tables 4.1 and 4.2). Now the main issue is the assessment of the maximum or minimum lateral spacing to define a reasonable low cross- track deviation. Obviously, the looser the constraints, the greater the chance of including in one alignment the features clearly belonging to another alignment near or at the crossing points. Choosing a constraint as tight as possible to increase the degree of confidence would leave out geologically related features slightly exceeding this tight limit. The SPGI model has shown that the plumes may have different sizes and may include related volcanic features at some distance, between 150 and 600 km, from the plume track, this range being inferred from the average lateral distance of the volcanic landforms away from the alignments (see Table

4.1 for a comprehensive list of along track and lateral distances) and from the lateral width of the fields of vents spreading from Olympus Mons, Alba Mons and Syria Planum. Such a grading system that would provide a qualitative sense of strong vs. weak alignment candidates can be resolved through the choice of a tight constraint of a geodesic line to verify the alignment, which is the first priority, and then the estimation of the size of the plume can be obtained through the observation of the field of vents in figure 4.14 and through the

90 observation of the maximum cross track (lateral) distance of the volcanic features located outside the alignments.

The assessment of the alignments is based on the verification of a set of aligned geographic data points already visible by eye and located along the directions indicated by the above- mentioned compulsory start and end points for each alignment using the Matlab polyfit tool.

A strong geological constrain is already provided by the SPGI model, which shows that the plumes do not meander at random all over the mantle. All the coordinates of the geographic data points along the alignments were included in a database and a number indicating the alignment (1, 2, 3…12) was assigned to each pair of coordinates. The assumption at the base of the Matlab script is that I indicate all the coordinates of the volcanic centres that should belong to an alignment and then the polyfit tool checks if these coordinates are fit with a line.

If the chosen coordinates (and thus volcanic centres) show a good fit then the initial assumption is verified. The polyfit tool of Matlab performs the best fit (in a least squares sense) through all the geographic data points of each alignment according to a chosen geodesic line (i.e. rhumb line, great circle, small circle). I have chosen a rhumb line (also known as loxodrome) because I observed that the alignments follow lines at constant angle with the meridians. A rhumb line appears as a straight line on a Mercator projection map.

Great circles do not have constant angles with the meridians and were thus excluded, and small circles were excluded because they are partial cases of great circles that do not pass through the centre of a sphere. The test with Matlab showed that indeed the rhumb lines fit the alignments and the results are shown in Fig. 4.23a.

The lateral distance of the offset volcanic features was then measured geodetically with

ArcGIS from the centroids of the aligned volcanic features. Although the alignments were fitted with rhumb lines, a mantle plume is evidently not as thin as a line and the fit would be much better with a stripe as large as the plume but the graphical result would then be much worse because the alignment would be obscured. The sum of all the consecutive along-track

91 distances gave the total hemispherical length of the alignments. The average along-track distance was calculated according to a simple arithmetic mean in order to have some statistics along the alignments. This average distance was useful to estimate whether there might be fields of vents or low-relief volcanic centres hidden in areas covered by the thick lava fields of Tharsis and Elysium, the main places where a distance larger than average was measured, and to understand whether tracing the alignments along different directions than those of the interpreted alignments might give similar results.

Two or more coloured dots are superposed in Fig. 4.2f on the volcanic centres where two or more alignments cross each other. Unfortunately, only the latest dot added is displayed on the ArcGIS maps but the volcanic centres belonging to more than one alignment are listed in

Table 1 anyway. An important clarification regarding the crossing points is that the mantle plumes do not arrive at the crossing point at the same time; the crossing points are just formed by the tracks left by migrating plumes at a different time. Mantle plumes in numerical convection experiments never pass through each other but can only merge or create bifurcations (Leone et al., 2014). The few scattered volcanic features that do not belong to any alignment were indicated with a dot of dark violet colour on the highlands or olivine yellow colour if included in the clusters of volcanic features identified in the lowlands (see legend of

Fig. 4.2).

A distinction between primary alignments and secondary alignments is here introduced.

The difference between main alignments and secondary alignments is that the latter merge with the former or do not reach the South Pole, even though all of them start from the volcanic provinces located along the transition topography (Leone, 2015) or along the

Martian dichotomy boundary. Often the highlands do not directly border the lowlands but are connected with a stripe of declining topography from 0 to -3 km height located in the middle.

This stripe was defined as the transition topography by Leone (2015). A primary alignment does not necessarily have a larger average along-track distance, nor must it necessarily be

92 longer than a secondary: for example the secondary Alignment 5 is longer than the primary

Alignments 1, 7, 8, and 11 (see Table 4.1).

Observations at regional scale allowed the analysis of the alignments in more detail showing both geomorphological similarities and differences related to different eruptive styles. The primary Alignments 1, 3, and 7 reach the South Pole through the volcanic plains of Argentea Planum, Promethei Planum, and , respectively, and share similarities, but also have differences that are described in detail in the Appendix. They cross each other along with another three secondary alignments under the Tharsis bulge contributing to the formation of the largest volcanic province of Mars. Alignment 2 is another main alignment reaching the South Pole from Elysium through Hellas, characterized by parallel massifs and calderas (Fig. 4.2a). Alignment 8 reaches the South Pole via Australe

Montes from the head of Mawrth Vallis following a north-south direction, different from the common NE-SW or NW-SE directions of all the other alignments but similar to the direction of Alignment 7. The last main alignment in order of recognition is Alignment 11. In contrast, alignments 4, 5, 6, 9, 10 and 12 are all secondary alignments because:

· Alignment 4 merges with the main Alignment 3 at Thyles Rupes even though it

seems to reach the South Pole.

· Alignment 5 crosses Alignment 4 and seems to merge with Alignment 3 but it does

not reach the South Pole.

· Alignment 6: same as Alignment 5.

· Alignments 9 and 12 do not reach the South Pole.

· Alignment 10 crosses Alignments 5 and 8 and does not reach the South Pole.

93 A detailed description of the volcanic features included in each alignment is available in the Appendix. The images provided in each section of the Appendix are meant to support the volcanic origin of the newly discovered features2.

4.3.1 Alignment 1

Alignment 1 was tracked from Olympus Mons through the volcanic centres of Biblis

Tholus, Phoenicis Lacus, , , Thaumasia Fossae, Aonia Mons,

Aonia Tholus, and Cavi Angusti as far as the South Pole. The size of the volcanic centres along the alignment is the largest at Olympus Mons and strongly decreases from Biblis

Tholus onwards.

Olympus Mons is regarded as a volcano with a long eruptive history because the latest flows appear young (Neukum et al., 2004). However, the age of first emplacement and its bulk material might date back to Hesperian or even to Noachian times (Isherwood et al.,

2013) as it will be shown in the discussion section. Consistently with such older ages Biblis

Tholus has an age spanning from 3.68 Ga (Werner, 2009) to 2.8 Ga (Plescia, 1994), 3.35 Ga

(Robbins et al., 2011), and 3.65 Ga (Neukum and Hiller, 1981), likely postdating the formation of Olympus Mons.

After Biblis Tholus Alignment 1 headed towards the Warrego Rise through Phoenicis

Lacus coasting the Claritas Fossae to the east (Fig. 4.3a). An unnamed vent on a low shield

(Fig. 4.3.b and Fig. 4.3d) and a couple of heads of valleys were found along the track that crosses other alignments in the area surrounded by Arsia Mons, Pavonis Mons and Noctis

Labyrinthus (Fig. 4.3c, Fig. 4.3e and Fig. 4.3f). Of course, other low shields and vents were found near or inside the intermixed zone formed by the various migrating plumes but more aligned with Alignments 3 and 7. The low shield in Fig. 4.3d and the head of the valley near

2 The subchapters 4.3.1 – 4.3.12 included in this thesis were published as Appendix in the Supplementary Material of the original version of the paper.

94

Figure 4.3. A) MOLA context image for the Alignment 1 in the Tharsis region of Mars, the pink dots indicate the eruptive points, the white rectangles labelled b and c refer to their respective panels of this figure. B) The THEMIS mosaic centred at 7°51'20.99"S and 114°20'25.79"W shows an unnamed low shield with a vent on top, the white rectangle labelled d refers to the respective panel where further detail of the feature will be provided. C) The THEMIS mosaic centred at 13°03'37.78"S and 108°53'15.32"W shows the area of Phoenicis Lacus where one of the eruptive points is located; the pink dot at the head of the valley indicates the eruptive point; the valley is carved on an original pit chain that develops to the southeast as far as an unnamed channel that will be detailed in panel f; the white rectangles labelled e and f refer to the respective panels of this figure. D) The CTX image P11_005427_1745_XI_05S114W shows the vent on top of the unnamed low shield; a sort of pit chain develops to the east of the vent and a thin channel departs from the vent to the south. E) The mosaic of the CTX images P10_005084_1671_XN_12S110W and B09_013048_1670_XI_13S110W, centred at 12°19'23.52"S and 110°27'13.83"W, provides detail of the valley that likely formed from the evolution of a pit chain; several other pits are visible nearby, sign of circulation of magma underground. F) The mosaic of the CTX images G16_024454_1653_XI_14S107W, P11_005229_1655_XI_14S108W, D22_035781_1653_XI_14S108W, D15_032946_1652_XI_14S108W, F04_037548_1653_XI_14S108W, P11_005440_1652_XI_14S108W, and B08_012626_1653_XI_14S109W, centred at 14°18'49.17"S and 108°04'54.27"W, shows un unnamed channel that forms directly from the lava flows coming from Syria Planum; the white curvy lines indicate the lava flow fronts while the white arrows indicate the direction of lava movement on the surface, the pit chain observed in the image is the terminal end of the valley observed in the previous panel.

95 Phoenicis Lacus shown in Fig. 4.3e were considered likely eruptive points belonging to

Alignment 1 but the valley in Fig. 4.3f was excluded because it formed directly from the lava that flowed downhill from Syria Planum.

Other volcanic centres along the alignment, most of them within the Warrego Rise cluster, were already identified in the work of Xiao et al. (2012). The Warrego Rise was already interpreted as a centre of magmatic-driven uplift, with a formation that dates back from the

Noachian to the Early Hesperian (Dohm et al., 2001b). Out of the Tharsis lava flow fields there are some newly interpreted volcanic centres: heads of valleys perpendicular to (and cut by) the Thaumasia Fossae, starting from a topographically high unnamed crater located just west of the crater ; the newly named Aonia Mons and Aonia Tholus in a region of

Noachian-Hesperian age. The latter were already included (but not specifically described) in the orogenic activity postulated at a time (4 Ga) when the magnetic field of Mars was already on the wane (Anguita et al., 2006).

The alignment ends with the interesting structures of Cavi Angusti (Ghatan et al., 2003) and the confining Angustus Labyrinthus, also known as “Inca City” (Sharp, 1973). These features were interpreted as mobergs (Allen, 1979), the result of dike emplacement (Mutch et al., 1976), or the result of ice melting where volcanic activity was likely involved (Ghatan et al., 2003). The volcanic hypothesis is the most plausible for the formation of the confining larger volcanic plains of Argentea Planum, characterized by the volcano-ice interactions in the Hesperian (3.7 Ga) Dorsa Argentea formation (Carr and Head, 2010; Tanaka and Kolb,

2001), and Parva Planum. The dark albedo patches found at Inca City also show seasonal variations as a result of defrosting that uncovers underlying dark material ( et al.,

2002), likely volcanic material.

96 4.3.2 Alignment 2

Alignment 2 was tracked to the Australe – Sisyphi Montes starting from the large volcanoes of Hecates Tholus and Elysium Mons. The volcanic domes located northeast of the

Phlegra Montes, originally interpreted as erosional remnants of volcanic fields (Greeley and

Guest, 1987; Tanaka et al., 2005), were recently interpreted as the surface manifestation of intrusive bodies of magma (Michaut et al., 2013). The Phlegra Montes, the Montes, the Hibes Montes, the Cerberus Tholi, and the Tartarus Montes are all characterized by reliefs morphologically similar to the domes studied by Michaut et al. (2013), generally small and no higher than 2 km as revealed by MOLA data, and embayed and eroded by the lava fields of

Elysium and Hecates. These small reliefs are generally confined within the lava fields, their heights are similar to the heights of the flow fields seen on Tharsis, so they are likely the remains of older flows eroded and partially covered by the younger lava flows erupted from

Elysium and Hecates (Leone, 2014). Thus, the original interpretation of erosional remnants of volcanic fields is still preferred and maintained in this paper. For this reason these features were not considered geographic data points for the Alignment 2.

The dimensions of Hecates Tholus and Albor Tholus, compared to the larger Elysium

Mons, may suggest the alternative hypothesis that they could also be secondary cones.

However, if the respective positions of Hecates Tholus within Alignment 2 and Albor Tholus within Alignment 6 were due to random magma rise from secondary (or stand-alone) conduits of Elysium, their respective and aligned positions in Alignment 2 and Alignment 6 would be an amazing coincidence. Another argument in support to the hypothesis of the alignments is the older formation of Hecates Tholus and the younger of Albor Tholus with respect to

Elysium Mons, 3.8 Ga estimated for Hecates (de Pablo et al., 2013), 3.7 for Elysium and 3.53 for Albor Tholus (Werner, 2009), as inferred from the crater counts.

97 Figure 4.4. A) MOLA context image for the region where is located the Herschel crater and an unnamed crater at the source of Licus Vallis, the unnamed crater is located on a topographic high and thus cannot receive supply of lava flows from the Tyrrhenum Mons lava fields; the white curvy lines indicate an unnamed lava channel that flooded Gale crater starting from the lava overflows of the Herschel crater. B) Crop of the CTX image B17_016336_1728_XN_07S236W covering the floor of the unnamed crater, infill of lava flows is visible in the image, the white arrows indicate the direction of movement of the lava flows, the white solid lines indicated the fronts of the lava flows. C) The THEMIS mosaic shows the area south of the crater Herschel, the white arrows indicate the direction of movement of the lava flows on the surface, the white rectangle labelled d refers to the next panel of the figure; several craters appear flooded and some of them, including Herschel, have the southern rim eroded confirming the direction of provenance of the lava flows. D) The mosaic of the CTX images P19_008477_1590_XN_21S231W, P12_005761_1591_XN_20S231W, P13_006117_1584_XN_21S231W, and B01_010046_1589_XI_21S230W, centred at 20°18'53.92"S and 128°40'45.36"E, shows the lava flow fronts indicated by the curvy white solid lines; the white arrows indicate the direction of advancement of the lava flow fronts.

The subsequent volcanic centre after Elysium is an unnamed crater on a Middle Noachian

(Tanaka et al., 2014) topographic high of ~ 60 km of diameter located northwest of the

Herschel crater and thus unreachable by the lava flows coming from Tyrrhenum Mons (Fig.

4.4a). Both the unnamed crater and the Herschel crater appear filled with volcanic lava, as shown by the CTX images (Fig. 4.4b, Fig. 4.4c, and Fig. 4.4d), but the unnamed crater is not reached by the lava fields of Tyrrhenum Mons while the Herschel crater is breached and

98 flooded by the lava fields (Fig. 4.4a, Fig. 4.4c and Fig. 4.4d). The lava overflows from the north-eastern side of the Herschel crater form an unnamed valley that breaches the south- western rim of Gale crater and floods it (Fig. 4.4a). Curiosity landed exactly on this lava flooding and took samples of basaltic material at the Cumberland hole (Grotzinger et al.,

2014), all the implications for this observation will be more evident in the discussion section.

The distance between Elysium Mons and the unnamed crater is above 2000 km, a distance above the average for Alignment 2 (see Table 1), suggesting that there might statistically be a very low-height volcanic centre (either field of vents or head of fossa) buried under the lava fields of Elysium Planitia.

The next large volcanic feature along Alignment 2 is Tyrrhenum Mons, a large volcano of

Late Noachian – Early Hesperian age (3.7-3.4 Ga) as inferred by the crater counts (Werner,

2009; Williams et al., 2009; Robbins et al., 2011). Although Tyrrhenum Mons and Elysium

Mons are far away from each other, the extreme limits of their lava fields confine and interact between the Amenthes Fossae and the along the Martian dichotomy boundary. The eastern Tyrrhenum lava fields flooded the crater Herschel (Fig. 4.4c), which then flooded the Gale crater through Farah Vallis, so Tyrrhenum Mons directly flooded Gale with its lava. This fact should not be surprising considering the wider extent of the Elysium and Tharsis lava fields where lava travelled for thousands and thousands of km within channels.

From Tyrrhenum Mons the volcanic features seem to divide into two parallel branches: 1)

Anseris Mons, followed by a unnamed massif located in South-Western Hellas after the

Alpheus Colles (Fig. 4.4b), as far as the Sisyphi Montes; 2) Hadriaca Patera, the central caldera of , followed by Malea Patera, , Amphitrites Patera, and

Pityusa Patera on the other side of Hellas (Fig. 4.2a and Fig. 4.2c). Different mechanisms of formation have been suggested for the massifs and the calderas, mostly related to the different density of magma bodies rising in the lithosphere (Williams et al., 2009). This observation

99 suggests that there might be a bifurcation in the mantle plume that formed Alignment 2 with one of the branches about 500 km far away from the other or that the plume is just 500 km wide. Unfortunately, there is no observable field of vents departing from Elysium to make a comparison. One of the most intriguing observations along Alignment 2 is both the symmetry and continuity of the features on both sides of and inside Hellas (Fig. 4.2a and Fig. 4.2c), an alignment that has not been obliterated by the impact that formed the basin but that followed its formation in time. Anseris Mons might not be genetically related to the impact that formed the Hellas and the Hellespontus Montes, although it is located in the rim of the Hellas basin as well as Hadriacus Mons.

The combined THEMIS and CTX mosaics showed dissected flanks on the slopes of

Anseris Mons (Fig. 4.5b and Fig. 4.5c) and the continuity of these incisions with the channels observed on the surrounding lava fields (Fig. 4.5d and Fig. 4.5e). The same channels were not observed on the Hellas Montes, except for a mountain devastated by an impact crater where the channels seem more originated by debris flows caused by lava erosion at the base of the mountains rather than by lava sapping (Fig. 4.5f and Fig. 4.5g). A geomorphological study of the banded terrain around the Alpheus Colles excluded a volcanic process for their formation only because of the lack of evident volcanic features but the viscous characteristics of the material required for their formation are more similar to lava than ice as suggested in the Diot et al. (2014) paper. The key position of an unnamed massif on the south-western side of

Hellas’ floor (Fig. 4.5a and Fig. 4.5h) excludes an origin through the impact that formed the basin and it is also a valuable geomorphological connection between Anseris Mons and the series of aligned massifs heading to the Sysiphi Montes. The age of the unnamed massif could be included between the Early Noachian formation of Hellas and the Middle Noachian

Pityusa Patera, somewhere between 4.1 and 3.8 Ga. The ages inferred from the crater counts for the Circum-Hellas calderas (Williams et al., 2009) are quite heterogeneous, Hadriaca

Patera (3.66 Ga), Amphitrites Patera (3.63 Ga), Malea Patera (3.81 Ga), Pityusa Patera (3.80

100

Figure 4.5. A) MOLA context image for the Hellas basin, the white rectangles labelled b, f, and h refer to their respective panels in this figure. B) The THEMIS mosaic shows the area surrounding Anseris Mons; the white curvy line indicates a broad flow front spreading from its north-eastern slope, this was inferred from the south-west north-east direction of the flow front, the nearest volcanic centre (Hadriacus Mons) is located to the east; Anseris Mons shows deep channels on its southern and north-eastern slopes that continue on the surrounding lava fields, the same channel development is not visible on nearby reliefs and all over the Hellespontus and Hellas Montes, the white rectangles c, d, and e refer to the panels where further detail of the channels is shown. C) The mosaic of the CTX images F03_036908_1500_XI_30S273W and

101 D14_032517_1495_XN_30S273W, centred at 29°55'41.45"S and 86°41'44.12"E, provides an overview of the southern and north-eastern slopes of Anseris Mons; the white arrow indicating the channel on the left shows a debris flow but the eastern channel on the main central amphitheater and the channel on the northeastern slope continue on the surrounding lava fields. D) The CTX image P22_009441_1498_XN_30S272W shows the continuation of the northeastern channel on the surrounding lava fields, the white arrows indicate the fainter segment; the sinuous morphology of the channel suggests that a fluid, likely lava, and not a debris flow must have carved it. E) The CTX image D14_032517_1495_XN_30S273W shows the connection between the channel on the southern slope of Anseris Mons and its continuation on the wider channel located on the southern lava fields, the white arrows indicate again the fainter segment; also in this case the morphology of the channel is not compatible with debris flow but with a fluid, likely lava. F) The THEMIS mosaic shows a sample area south- east of the Hellas Montes that displays a good concentration of high reliefs located near wide valleys (i.e. ), not all the reliefs show channels on their slopes, the only relief showing some channels is a mountain devastated by the large impact crater Sebec located on the top left of the image (white rectangle labeled g). G) The mosaic of the CTX images P21_009256_1396_XI_40S259W, B18_016759_1395_XI_40S260W, F01_036090_1395_XN_40S260W, P14_006461_1403_XN_39S261W, and P16_007173_1395_XN_40S261W, centred at 39°58'10.29"S and 99°55'09.91"E, shows details on the remaining part of the mountain; a landslide is visible on the western side near to Reull Vallis; some channels are visible on the north-eastern slope but they end up into a debris apron, unlike Anseris Mons no channel develops on the surrounding lava fields. H) Mosaic of the CTX images P19_008638_1342_XI_45S303W, P14_006489_1344_XN_45S304W, P19_008348_1341_XI_45S304W, and P19_008493_1335_XI_46S305W, centred at 45°27'22.69"S and 55°07'17.11"E, shows a dome of ~15 km of diameter and of likely volcanic origin rising from the floor of Hellas.

Ga), quite beyond the migration times estimated by the SPGI model (Leone et al., 2014), but these crater count ages were substantially confirmed by the work of Werner (2009). However, as it will be shown in the discussion section, the whole stratigraphic column might be brought back in time.

The Sysiphi Montes are alignments of massifs of likely volcanic origin (Ghatan and

Head, 2002), which the crater counts place at Hesperian (3.7-3.5 Ga) age (Fenton and

Hayward, 2010), crossing each other in direction east-west and in direction north-south. The north-south direction coincides with the crossing point of Alignment 8 that will be described in the next section 3.8 ahead. The migrating plume that formed Alignment 2 ends its journey towards the South Pole at Sisyphi Tholus to the west of the Sysiphi Montes – Australe

Montes chains, which are both found at the end of the Alignment 8 (Fig. 4.2a).

102 4.3.3 Alignment 3

Alignment 3 was tracked from the centre of Ascuris Planum (Table 1), an eroded and lava- flooded plain remaining of a previous upper (Scott and Tanaka, 1986) to middle Noachian

(Tanaka et al., 2014) large massif located on Western and crossed by the Tempe and . The remnants of this old and once large massif are still visible in the chain of reliefs located west and north of Ascuris Planum. Following the pattern of the Tempe

Fossae back to Ascraeus Mons it was still possible to see them passing nearby Uranius Mons.

Different ages were inferred from the crater counts for Ascraeus Mons, 3.3 Ga (Neukum and

Hiller, 1981) and 1.5-1.0 Ga ((Murray et al., 2010); Robbins et al., 2011). Although Uranius

Mons, together with the nearby Ceraunius and Uranius Tholus, may appear as one of the secondary cones of Ascraeus Mons the age inferred from the crater counts for Uranius Mons is between 3.7 Ga (Werner, 2009) and 3.43-3.19 Ga (Robbins et al., 2011) suggesting that it formed before it. Ceraunius Tholus, located between Uranius Mons and Ascraeus Mons, has an inferred oldest age of 3.59 Ga (Robbins et al., 2011) and thus still older than Ascraeus

Mons, 3.3-1.5 (Neukum and Hiller, 1981; Murray et al., 2010; Robbins et al., 2011), but also younger than Uranius Mons; Uranius Tholus is located between Uranius Mons and Ceraunius

Tholus, although a bit shifted laterally (105.84 km as can be seen in Table 1), and has an inferred (for the shield at least) age of 3.56-3.50 Ga (Werner, 2009; Robbins et al., 2011). The crater count ages here suggest that these features are not secondary with respect to Ascraeus

Mons but rather primary and that follow the sequence of emplacement according to the direction of Alignment 3. The size of the plume that formed Alignment 3 can be inferred from the lateral geodetic distances of the Issedon Paterae (307.65 km) and the Issedon Tholus

(414.99 km), which are shifted on the same side of Uranius Tholus.

Pavonis Mons follows Ascraeus Mons along the alignment and has an estimated age of

3.0-2.5 Ga (Neukum and Hiller, 1981; Scott et al., 1998) but also older ages of 3.56 Ga

(Werner, 2009) as well as much younger of 0.86 Ga (Robbins et al., 2011); almost the same

103 applies for Arsia Mons with its 3.2 Ga (Neukum and Hiller, 1981), 3.54 Ga (Werner, 2009), and 0.1 Ga (Robbins et al., 2011). However, given the too high variation in the inferred ages from the crater counts, the oldest ages for Pavonis Mons and Arsia Mons were chosen for the graph that will be shown in Fig. 4.25. Also along this alignment there might be a very low- height volcanic centre buried by the lava fields along the slope of Arsia Mons, perhaps fields of vents somewhere below Daedalia Planum, as there are 2131 km of distance between Arsia

Mons and the next volcanic centre of Sirenum Tholus.

The aligned ancient highlands volcanoes of Sirenum Tholus and Sirenum Mons were considered older than the Tharsis volcanoes, the ages inferred for Sirenum Tholus and

Sirenum Mons from the available crater counts were of 3.99 Ga and of 4.05 Ga respectively

(Xiao et al., 2012) although they would have been formed after Arsia Mons according to the spatial-temporal direction of the alignment, Sirenum Mons appears half-smashed by an impact crater while the edge of Sirenum Tholus appears shaped by several impact craters. To the east of Sirenum Tholus, not only along the alignment but also laterally to it, there are many large impact craters on Daedalia Planum that are half-buried by the massive lava flow fields of Arsia Mons suggesting that there might be many others completely buried. A close up of one of these craters is shown in Fig. 4.2b while the possible implications of this observation will be further discussed in section 4.5.2 ahead.

The next features along the alignment are Phaethontis, a crater of Late Noachian age (~ 3.8

Ga, Tanaka et al., 2014) originating the Tader Valles together with the nearby crater Li Fan, and unnamed calderas of Late Noachian age located south of the Copernicus crater (Fig.

4.6a). Phaethontis is a crater located on a topographic high, in a region where the Mars

Odyssey Neutron Spectrometer indicates low water content (Christensen, 2006), and shows overflows along its rim (Fig. 4.6b and Fig. 4.6d). The nearby and preceding Li Fan crater was also considered eruptive point of Alignment 3 because showed direct connection with the heads of Tader Valles (Fig. 4.6c).

104

Figure 4.6. A) Context MOLA image for the Terra Sirenum region of Mars, focussed on two craters located on a topographic high from which originate the Tader Valles, the white rectangles labeled with the letters b, e and h refer to their respective THEMIS panels of this figure, the coloured dots refer to the volcanic features belonging to their respective alignment (green for Alignment 3 and blue for Alignment 6). B) THEMIS mosaic showing the source area of the Tader Valles located between the craters Phaethontis and Li Fan, the white rectangles labelled c and d refer to the next panels of this figure. C) Crop of the CTX image P08_004321_1317_XI_48S152W taken on the southern rim of the crater Li Fan, the black arrows indicate the possible ways for the overflows to carve a branch of the Tader Valles. D) Crop of the CTX image D12_031893_1294_XI_50S154W, taken on the eastern rim of the crater Phaethontis where the Tader Valles originate, the white lines indicate the latest lava flow fronts and the white arrows the direction of the flows towards the Tader Valles. E) THEMIS mosaic showing a branch of the Tader Valles spreading mainly to the north following the local topography (white arrow) but also to the east where some flooding is visible in the filled crater (black arrow); the white rectangle labelled f refers to the next panel of this figure. F) Crop of the CTX image B11_013788_1321_XN_47S149W taken along the main northern route of the lava flows likely channelized through the Tader Valles, a filled crater is visible on the right side of the image, the white rectangle g refers to the next panel. G) Blow up of the same CTX image B11_013788_1321_XN_47S149W focusing on the edge of the latest lava flow indicated by the white arrows; the image shows clearly another previous episode of flooding covering the crater’s ejecta, the white curvy line indicates the front of such a flow. H) THEMIS mosaic showing another branch of the Tader Valles heading to the south, a white circle indicates a filled crater, the question mark indicates uncertainty on the interpretation of another filled crater that might instead be a lava flow front; the white rectangle labelled i refers to the next panel of this figure. I) Crop of the CTX image G15_024047_1303_XI_49S152W taken on the southern Tader Valles; the likely lava infill of the plains is still visible inside and also partially overflowing from the bed-floor of the valleys.

The presence of craters filled by lava flows (Fig. 4.6e, Fig. 4.6f, Fig. 4.6g, Fig. 4.6h) along the bed-floor of the Tader Valles (Fig. 4.6i) confirms that this area was characterized by lava flooding coming from the eruptive points of Phaethontis and Li Fan. Although located at the confluence between Alignment 3 and Alignment 6, Ptolemaeus was not considered eruptive

105 point because of its breached eastern and western rims. It is already visible from the MOLA altimetry of Fig. 4.6a that it got its infill from the discharge of Hipparcus.

Figure 4.7. A) Context MOLA image for the Terra Sirenum region of Mars, the white rectangles labelled b and g indicate the areas shown in panel b and g, the green dots indicate the centroids for the volcanic features of Alignment 3. B) THEMIS mosaic showing a wide area with two volcanic calderas and a vent (white arrows) that filled the surrounding craters with lava flows, the white rectangles labelled c and f are the areas covered by the THEMIS mosaics shown in panels c and f, the white rectangles labelled d and e are the areas covered by the CTX images shown in panels d and e, the green dot shows the centroid for the cluster of multiple eruptive points located in the area; however, lava flooding is widespread (the main directions of the flows are indicated by the black arrows), as also shown by the nearby nested craters where the lava infill eroded their dividing rims, and might cover additional eruptive points located northwest of the centroid. C) the THEMIS mosaic shows a long and sinuous lava channel starting from the vent indicated by the green dot. D) The CTX image G13_023243_1234_XI_56S160W covers the area between the vent and the calderas shown in panel b, the white arrow indicates a section of the sinuous lava channel, the black arrows indicate the directions of the flows erupted by the calderas, the flow fronts are indicated by the white solid lines, one of the calderas is recognizable for the overflows indicated by the black arrow in the panel, more detail is provided in panel f. E) The CTX image G03_019498_1240_XN_56S157W covers the area located to the north-east of the calderas, lava channels and flooded craters are also visible in the image. F) THEMIS mosaic showing in detail the area of the calderas, the morphological characteristics of the calderas are similar to those of Siloe and Euphrates Patera located in Arabia Terra (Michalski and Bleacher, 2013), in the image are clearly visible both the inner and the outer depressions caused by the magma withdrawal following the eruption that formed the lava flows, the black arrows indicate the direction of the lava flows, the white lines indicate some of the flow fronts recognizable in the image, there are flooded craters along the directions of the flows. G) THEMIS mosaic showing the inter- crater plains between Copernicus and the topographic rise where the calderas are located; these plains and Copernicus itself are flooded by the lava erupted from the calderas and the vent, it was possible to trace back the path of the lava from the main flows directions indicated by the black arrows, the white rectangle labelled h indicates the area where the inlet of the lava channel to Copernicus is located, part of the rim of Copernicus was eroded away by the lava flooding. H) THEMIS mosaic showing the lava channel heading to Copernicus, the white arrows indicate the path of the channel.

106

The next eruptive point along the Alignment 3 is located south of Copernicus, where an unnamed vent was taken as centroid for a wider volcanic area located among the craters

Copernicus, , and Liu Hsin (Fig. 4.7a and Fig. 4.7b). The area is characterized by a topographic rise (Fig. 4.7b) on which is located a cluster of several volcanic calderas of

Noachian age (Scott and Tanaka, 1986). The lava erupted by the vent and the calderas flooded the rise and contributed to the formation of an inter-crater volcanic plain just south of

Copernicus. This volcanic plain was already interpreted as a mixture of lava flows and impact (Scott and Tanaka, 1986). The position of the vent coincides with the head of a valley that was interpreted as a likely lava channel (Fig. 4.7c). Two calderas showing overflows heading towards the north were immediately recognized at ~ 50 km south from the vent (Fig.

4.7d). Other lava channels, flooded craters, and (probably) a flooded caldera are located ~ 150 km to the east of the vent (Fig. 4.7e). The lava flooding then contributed to the infill of

Copernicus and of the other surrounding impact craters (Fig. 4.7g). The calderas located to the south of the vent, an overview is available in Fig. 4.7f, were recognized for: a) their typical inner and outer concentric depressions similar to those observed within the Siloe and

Euphrates calderas in Arabia Terra by Michalski and Bleacher (2013); b) lack of observable ejecta typical of impact craters; and c) presence of lava flows and lava channels. The wide extent of the flooding excludes impact melting for the formation of both the channels and the lava flows. The flooding on the topographic rise followed two main directions northwards

(black arrows in Fig. 4.7b): one directed towards the Copernicus crater and one directed to the east of it (Fig. 4.7g). A lava channel is recognizable on the flows directed towards Copernicus

(Fig. 4.7h). The floor of Copernicus was already interpreted as a geologic unit formed by lava flows of low viscosity (Scott and Tanaka, 1986). The age of these volcanic units is estimated as Middle Noachian (Scott and Tanaka, 1986) and Late Noachian-Early Hesperian in the geo- chrono-stratigraphic map of Mars (Tanaka et al., 2014).

107 An interesting feature surrounded by several volcanic plains is the isolated Chronius Mons

(Fig. 1i), located between Terra Cimmeria and Terra Sirenum. Chronius Mons was listed with the number 45 among the volcanoes of Xiao et al. (2012) but the information on its age is missing from their work, the only information available (~ Late Noachian) came from the geo-chrono-stratigraphic map of Tanaka et al. (2014). Its importance simply lays on the fact that it is a volcanic centre that provided a connection to the continuity of Alignment 3 towards the South Pole and a possible contribution to several inter-crater plains that will also be observed in sub-section 3.11. No other volcanic centre is seen around as far as one thousand of km, other than those belonging to other alignments, thus excluding a widespread and random formation of volcanoes outside the alignments in this region.

The alignment ends at the South Pole, after the volcanic features located at the Ulyxis and

Thyles Rupes (Xiao et al., 2012; Ackiss and Wray, 2014), where flow fronts and plains of likely volcanic origin are recognizable south of the Promethei Rupes (Tanaka and Scott,

1987).

4.3.4 Alignment 4

Alignment 4 was tracked from the unnamed massifs located to the east of Ma’adim Vallis, near the border between the highlands and the transition topography, then Apollinaris Tholus,

Zephyria Tholus, one unnamed flat-topped topographic high, three unnamed tholi, Electris

Mons, Eridania Mons, the crater Byrd, the Thyles Rupes, and the crater Burroughs to the end into Alignment 3 immediately after the crater Rayleigh. Alignment 4 is the only alignment that does not start from any of the four main points where all the other alignments start. The angle formed by the directions of Alignment 4 and Alignment 3 is roughly 30 degrees (Fig.

4.8) and is comparable to that observed between two merging plumes in the SPGI model

(Leone et al., 2014).

108 Figure 4.8. a) ArcGIS image of the Alignment 4 merging Alignment 3 on Promethei Planum immediately after the craters Burroughs and Rayleigh, the angle formed by the directions of the two alignments is roughly 30 degrees, the white rectangle labelled b refers to the image shown in panel b; b) THEMIS mosaic showing the area of the crater Byrd, the white arrow show the western side of the crater characterized by overflows and by valleys.

109 Figure 4.9. A) MOLA context image for the territory east of Ma’adim Vallis, a small group of reliefs are the remains of highlands volcanoes demolished by impacts, the coloured dots refer to the respective eruptive points (red for Alignment 4 and blue for Alignment 6), the black rectangle labelled b is the area contained in the THEMIS mosaic shown in panel b; the white arrows indicate the main regional direction of movement of the lava floodings. B) THEMIS mosaic showing the crossing point between Alignment 4 (red dot) and Alignment 6 (blue dot), the white arrows indicate the direction of the channels that spread from the relief indicated by the red dot; the white rectangle labeled c refers to the next panel c of this figure. C) A zoom into the CTX image D12_031986_1628_XI_17S180W shows channels departing from an unnamed relief devastated by an impact crater, the white arrows indicate the direction of the channels carved on the surrounding plains. D) MOLA context image for a flat-topped topographic high located in northern Terra Cimmeria among several ridged plains interpreted as extensive lava flows erupted with low effective viscosity from many sources at high rates (Greeley and Guest, 1987); the white arrows indicate the main direction of the flows in direction of Al-Qahira Vallis, the white rectangle labelled e indicates the next panel. E) The mosaic of the CTX images B18_016651_1562_XN_23S194W, F04_037327_1564_XN_23S194W, B11_014106_1565_XN_23S194W, B11_013750_1569_XN_23S195W, B18_016506_1566_XN_23S195W, P18_007922_1557_XN_24S195W, B19_016928_1564_XN_23S196W, B12_014172_1575_XN_22S196W, G18_025288_1552_XN_24S196W, D07_030088_1552_XN_24S197W, G22_026910_1570_XN_23S198W, G02_019051_1552_XN_24S198W, G04_019697_1528_XN_27S194W, and B20_017495_1524_XN_27S196W, centred at 24°35'25.70"S and 164°02'19.83"E, shows pit chains on the lava fields located south west from the relief; the white arrows show the main directions of the lava inferred from the flow fronts indicated by the curvy white lines, the red dot indicates the eruptive point inferred by the directions of the lava flow fronts, the white rectangle labeled f indicates the area covered in the next panel f. F) The mosaic of the CTX images F04_037327_1564_XN_23S194W, B11_014106_1565_XN_23S194W, B11_013750_1569_XN_23S195W, and B18_016506_1566_XN_23S195W, centred at 23°24'57.52"S and 164°59'30.04"E, provides a close-up on the eruptive source area of the unnamed flat-topped relief; the white arrows indicate the main directions of the lava flows, the red dot the inferred eruptive point, and the white curvy lines indicate the lava flow fronts. G) MOLA context image for a couple of unnamed tholi located in central Terra Cimmeria, the white rectangles labelled h and I refer to the next panels of this figure, the red dot is the centroid for the feature. H) The mosaic of the CTX images B18_016770_1429_XI_37S201W, B19_016981_1432_XN_36S201W, B19_017192_1443_XI_35S202W, and B18_016625_1434_XN_36S202W, centred at 35°08'42.16"S and 158°11'45.51"E, shows the western side of the

110 unnamed tholi; the watershed, or probably better say the lava-shed (dashed white line), is inferred from the position of the heads of the many channels visible in the image; the white curvy lines indicate likely lava flow fronts. I) The mosaic of the CTX images D14_032778_1434_XN_36S198W, G14_023864_1442_XI_35S198W, P11_005272_1441_XN_35S199W, P16_007263_1421_XN_37S199W, B18_016704_1441_XI_35S200W, B20_017482_1444_XI_35S200W, and B19_017126_1444_XN_35S200W, centred at 35°42'11.14"S and 160°09'38.12"E, shows the eastern side of the unnamed tholi; the red dot is the centroid of the feature, the white curvy lines indicate likely lava flow fronts .

The territory located east of Ma’adim Vallis is a well-imaged area at THEMIS resolution, although not completely at CTX resolution, and it is characterized by unnamed massifs aligned with Apollinaris Mons along the direction northwest-southeast and with Apollinaris

Tholus along the direction northeast-southwest at the likely crossing point between Alignment

4 and Alignment 6 (Fig. 4.9a and Fig. 4.9b). The massif indicated by the blue dot in Fig. 4.9a is likely belonging to the Alignment 6 not only for its alignment with Apollinaris Mons and with the tholi in the Atlantis region but also because the Alignment 4 shows no laterally offset volcanic features from the start to the end (Fig. 4.2c). The massifs show channels starting directly from their slopes and continuing on the surrounding volcanic lava plains (Fig. 4.9b and Fig. 4.9c), already interpreted as volcanic plains by Greeley and Guest (1987). The heads of the unnamed valleys located east of the unnamed massif at the start of the alignment were not considered geographic data points because were originated by the flooding coming from the south (Fig. 4.9a), the same flooding that came from the chaos terrains located in the middle of the Ariadnes Colles and that are also at the origin of Ma’adim Vallis. Ma’adim

Vallis was thought to be among the oldest features of Mars, dating back to 4.4-4.3 Ga

(Neukum and Hiller, 1981), but later estimates suggested an age around the Noachian –

Hesperian (3.7 Ga) boundary (Irwin et al., 2004). The presence of unaltered olivine-phyric at Gusev crater (McSween et al., 2006), where Ma’adim Vallis discharged its lava flows (Greeley et al., 2005), suggests a pure volcanic origin. This fact is quite remarkable considering that Gusev was regarded as a place where water should have ponded for long time according to Cabrol et al. (2003).

111 The next centre along the alignment is Apollinaris Tholus, a volcanic centre with similar characteristics to Zephyria Tholus, a middle Noachian (3.9 Ga; Scott and Tanaka, 1986) stratovolcano that originated the Durius Valles (Stewart and Head, 2001). Following these volcanoes along the alignment, there is a couple of unnamed volcanic domes (tholi) in Terra

Cimmeria that were obliterated but not completely destroyed by impact craters (Fig. 4.9d and

Fig. 4.9g). Their identification was possible thanks to the presence of pit chains, suggesting presence of likely lava tubes with collapsed roofs, and the lava flows found on their flanks

(Fig. 4.9e and Fig. 4.9f). These domes are quite large, between ~ 70 and 100 km (Table 1), but only the one in central Terra Cimmeria is rich of channels along its slopes (Fig. 4.9h and

Fig. 4.9i). Both contribute to the infill of the surrounding ridged volcanic plains providing a valuable connection between Zephyria Tholus and the recently named Electris and Eridania

Montes.

The crater Byrd shows overflows and valleys on its western side (Fig. 4.8b), geomorphologically similar to others seen in confirmed volcanic centres, but no good HiRISE coverage in support to the interpreted volcanic origin are yet available in this area. The same lack of high-resolution imagery occurs in the area of the Thyles Montes near the crater

Burroughs. Although a volcanic origin for the Thyles Rupes region has already been mentioned (Ackiss and Wray, 2014), a better coverage of CTX imagery (at least) in this area would be needed to provide further insight in this region of Mars.

4.3.5 Alignment 5

Alignment 5 was tracked from the calderas of Arabia Terra, the centroid is on the largest unnamed caldera located in the middle of the triangle formed by Ismenia Patera, Euphrates

Patera, and Siloe Patera, through the heads of the Mamers Valles, the heads of the Locras

Valles, the crater Schroeter, the head of the Naro Vallis, the long Huygens-Hellas system of

112 ridges, Tyrrhenum Mons (where Alignment 5 crosses Alignment 2), as far as the craters

Huggins and Rossby located beyond Alignment 3 (Fig. 4.2a). The crater Huggins is taken as the nearest named reference point to an unnamed volcanic cone located nearby to the east, the

Rossby crater is instead the source of the Drava Valles. Volcanic features in Acidalia Planitia were not considered geographic data points.

Volcanism in Acidalia Planitia was already recognized in the Viking images and interpreted as the result of volcano-ice interactions (Allen, 1979; Frey and Jarosewich, 1982) although ground ice was not found at the latitudes of the Viking landing sites, more details in the discussion section. Subsequent interpretations were mostly based on mud volcanism

(Tanaka, 1997; Farrand et al., 2005; Oehler and Allen, 2010) or Early Hesperian volcanic flows (Head and Kreslavsky, 2002). The CRISM detection of unaltered olivine and clinopyroxene showed that the bedrock of southern Acidalia Planitia, exposed by impact craters, is made of basaltic lava (Salvatore et al., 2010) and confirmed the scarcity or absence of water in the lowlands at middle northern latitudes, a finding already confirmed by the Mars

Odyssey Neutron Spectrometer data (Christensen, 2006). Before the MRO imagery, the origin of these basaltic plains was less clear as there are not visible vents (Head et al., 2002) and/or high volcanic features in Acidalia Planitia but only mounds and widespread cones without evident eruptive activity (Farrand et al., 2005). The observation of tear eyes shaped structures

(TESS) from Chryse Planitia to Acidalia Planitia led to the conclusion that the basaltic plains might have been originated by flows coming from the circum-Chryse outflow channels, as suggested by Tanaka (1997), with the important difference that these are likely not mudflows as they would imply the presence of water that is not supported by the presence of unaltered olivine. In fact, the presence of lava flows in the circum-Chryse outflow channels (Leone,

2014) and the detection of unaltered olivine both on Chryse and Acidalia Planitia (Ehlmann et al., 2010; Salvatore et al., 2010) weaken the mudflows hypothesis. The Acidalia Colles and the Ortygia Colles are also made of mounds geomorphologically too similar to those observed

113 on the lava fields of Elysium, the chaotic terrains of Valles Marineris (Leone, 2014), and the mensae terrains to be considered eruptive centres.

The Arabia Terra calderas, with (65 km of diameter) similar in dimension to

Nili Patera (Michalski et al., 2013), are among the smallest of Mars. However, Eden Patera has recently and suddenly become a new class of volcanic feature, a supervolcano (Michalski and Bleacher, 2013). Although Eden Patera is similar to but still smaller than Nili Patera (75 km) and even dwarfed by Pityusa Patera with its 225 km of diameter, these latter calderas have not been included in the Michalski and Bleacher (2013) list of supervolcanoes. It is also difficult to understand the need to establish a new class of super-volcanic features given that

Eden Patera is already a caldera (patera)! Aside from this issue, the interpretation of these features as calderas is quite remarkable and suggests the likely presence of volcanic activity.

Furthermore, the observation of these calderas grouped in a cluster (Table 1) elongated towards the direction of the Alignment 5 (Fig. 4.2a) also suggests that they could have been formed by the passage of a migrating plume.

The next significant feature along the alignment is the source area of the Mamers Valles, located around the crater (Fig. 4.10a), suggested as formed from melting of subsurface ice by a warm blanket of ejecta (Mangold, 2012). The head of the Eastern Mamers Valles, leading to Ismenius Lacus, originated from a point located 250 km northeast from Cerulli well outside its ejecta (Fig. 4.10d and Fig. 4.10g). The channel connecting the western slope of

Cerulli with the Western Mamers Valles was formed on an area where Mars Odyssey data showed scarce abundance of water (Christensen, 2006) and does not seem to overflow directly from Cerulli (Fig. 4.10b and Fig. 4.10e).

114 Figure 4.10. A) Context image for the Cerulli crater and the Mamers Valles source region obtained through a mosaic of THEMIS Daytime Infrared images, the white rectangles labelled b, c, and d refer to the respective panels in this figure. B) THEMIS mosaic showing the source area of the Western Mamers Valles, the orange dot is the chosen centroid for the various sources of the area: heads of valleys, floor-fractured craters (all indicated by white arrows); the white rectangle labelled e indicates the panel e of this figure where further detail of one of the heads of valleys is given. C) THEMIS mosaic showing the source of two unnamed valleys along the alignment 5, the orange dot indicates the position of the source that will be imaged in detail in panel f (area covered by the white rectangle). D) THEMIS mosaic showing the source of the Eastern Mamers Valles, the white rectangle labelled g refers to the respective panel of this figure where further detail of one of the heads of the valleys is given. E) The mosaic of the CTX images B07_012476_2116_XN_31N339W, B11_013966_2126_XN_32N339W, B17_016155_2132_XI_33N339W, B21_018001_2134_XN_33N338W, and F04_037398_2113_XN_31N338W, centred at 31°36'22.44"N and 20°58'19.83"E, shows the heads of several incisions near the rim of the crater Cerulli (white arrows). F) The mosaic of the CTX images B18_016524_2082_XN_28N333W and P06_003391_2071_XN_27N333W, centred at 28°13'37.54"N and 26°22'23.97"E, shows a filled crater (indicated by the circular dashed line) as eruptive point for two diverging heads of valleys; the southern head is visible in the image (white arrow), the northern head is just out of the image but is visible in the context of the panel c. G) The mosaic of the CTX images P17_007874_2163_XN_36N333W, B01_009944_2159_XN_35N333W, B17_016102_2159_XN_35N333W, F03_036989_2155_XN_35N334W, and B02_010366_2161_XI_36N334W, centred at 35°32'34.73"N and 26°17'02.46"E, shows the heads of the Eastern Mamers Valles (white arrows); the white curvy line indicates a lava flow front. H) The mosaic of the CTX images G23_027192_2145_XI_34N343W, P22_009562_2143_XN_34N343W, F05_037807_2147_XN_34N344W, G15_024080_2152_XN_35N345W, B21_017632_2152_XN_35N344W, D17_033772_2153_XN_35N346W, and G23_027324_2156_XN_35N346W, centred at 34°44'48.27"N and 15°05'57.63"E, shows the centroid (orange dot) for a series of eruptive points (white arrows) located west of Ismenius Cavus; the centroid is located on a dome in the middle of a crater from which several flows spread concentrically (white curvy lines); the morphology of the dome is not the typical peak generated by an impact; the crater is edged by several incisions, the white rectangle labeled i refers to the next panel of the figure. I) A zoom into the CTX image G15_024080_2152_XN_35N345W, provides a close-up on one of the incisions edging the crater; the channel indicated by the white arrow shows filling all over the length of its bed-floor.

115 The ~ 300,000 km3 of volume, a lower estimate not including the volcanic infill of the valleys, removed from the Mamers Valles and from the connected Deuteronilus would require up to two or three orders of magnitude of water (Andrews-Hanna and Phillips, 2007;

Leone, 2014; Leverington, 2011) that cannot simply be justified by the scarce amount in the limited area of the Cerulli impact. Even considering the whole ~ 40,000 km2 of the area covered by Cerulli and its ejecta as pure water ice, which is already unrealistic, the column of water needed to carve the valleys would still be deep 7.5 km raising serious problems about the of the crust that should host it. Melting of water ice does not seem a viable hypothesis and thus an alternative volcanic source must be invoked. The volcanic hypothesis is instead supported by the presence of a series of pit chains spreading radially from Cerulli and by a widespread lava flooding (Fig. 4.10b, Fig. 4.10c, and Fig. 4.10f). The lava flooding is evident along the alignment with several flow fronts that spread concentrically from a filled crater that shows incisions along its rim (Fig. 4.10h). Not all the craters in the area show incisions along their rims thus making the hypothesis of localized eruptive points much stronger. A close up in one of the incisions along the crater from which the lava flooding spread shows a filling that does not seem of aeolian nature (Fig. 4.10i), a possible mechanism of formation could be mass wasting caused by the channelization of the lava flooding back to the crater following magma withdrawal. In this case, the crater could be a volcanic caldera.

This hypothesis should not be surprising because of the presence of the nearby Arabia Terra calderas.

The next feature along the alignment is an unnamed topographic high, devastated by impact craters and characterized by the heads of the Locras Valles (Fig. 4.11a to Fig. 4.11g), from which the Indus Vallis and the Cusus Valles originated (Fig. 4.11a, Fig. 4.11h, and Fig.

4.11i).

116 Figure 4.11. A) Elevation context image for the Locras-Cusus-Indus Valles system based on MOLA data, the white rectangles labelled b and c refer to the THEMIS mosaics shown in the respective panels. B) THEMIS mosaic for the sources of Locras and Cusus Valles, the white rectangles labelled d, e, and h refer to their respective panels of this figure. C) THEMIS mosaic for the additional sources of Locras Valles located to the west of the centroid, the westernmost of these sources (white rectangle labelled g) is the point where the lateral spacing from the centroid is measured. D) THEMIS mosaic of the Locras Valles centroid source area (orange dot), the white arrows indicate additional heads of the valley network, one of these is imaged in panel f (white rectangle). E) The mosaic of the CTX images D13_032426_1913_XI_11N313W, D17_033771_1915_XI_11N313W, G13_023235_1914_XI_11N312W, and D04_028668_1909_XI_10N312W, centred at 10°37'43.15"N and 47°07'03.87"E, shows an inter-crater plain where the Locras Valles discharged their fluid; the white arrows indicate the points where the carved channels of the Locras Valles continue into sinuous filled channels that flooded craters and the surrounding plains. F) The CTX image P04_002586_1879_XI_07N314W shows the detailed view of one the heads of the Locras Valles located at around half way between the centroid and the westernmost head. G) The CTX image D19_034826_1904_XN_10N318W shows the westernmost of the heads of the Locras Valles. H) The THEMIS mosaic shows the easternmost of the heads of the Locras Valles and the head of the Cusus Valles (white labelled arrows), the head of the Cusus Valles is the other extreme point where the lateral spacing from the centroid is measured, the white rectangle labelled i refers to the next and last panel of the figure. I) The CTX image B19_016866_1897_XN_09N307W shows one of the two points where the Cusus Valles get out from a filled crater along their way from the source.

The area was already studied on Viking images for the dendritic development of several valley networks, with some ambiguity of interpretation on the nature of the upland plains

(volcanic vs. sedimentary), concluding that the valleys formed in two distinct time intervals between 2-3 Ga ago after the release of groundwater or magmatic volatile that contributed to

117 the destruction of the impact craters (Grant and Schultz, 1990). However, another study based on more recent THEMIS and MOLA data concluded that not even the valleys approaching the morphology of terrestrial networks (i.e. Indus Vallis), which are Noachian in age, satisfy the scaling relationships of perennial fluvial channels (Som et al., 2009). The analysis of the CTX images also revealed that platy-ridged valley infill, likely formed by lava flows and not eroded (incised) by any later fluvial action, appears from the dust also in the median course of the Locras Valles (Fig. 4.11e).

The crater Schroeter, from which the Tisia Valles originated with the contribution of the unaligned crater Huygens, is at the beginning of two major systems of broadly arcuate ridges that can be traced for 600-700 km, perhaps extending farther in the subsurface, consistent with the emplacement of shallow dikes (Head et al., 2006). The direction of this system of ridges follows the same direction of Alignment 5. Since the subsurface extent of the ridges is unknown and might be much longer than expected, probably more than 700 km, the centroid for this long Huygens-Hellas system is taken conservatively at the end of the northern longest ridge (Table 1). A maximum distance of 375 km between the two ridges is measured near the crater Schroeter, a minimum distance of 160 km is measured in proximity of the centroid. The head of the Naro Vallis starts from three nested (and filled) craters located near the sector C of the ridge system. The sector F is superposed on smooth plains of Hesperian age interpreted as composed of lava flows (Head et al., 2006), these plains are then connected with the inter- crater plains of Oenotria Plana to the north and to plains gradating towards Hellas to the south. The width and the geometry of these dikes, as well as the inferred effusion rate, are consistent with the volcanic flooding of the pre-existent Noachian topography (Head et al.,

2006). The Noachian age of the flooding is also consistent with the 3.9 Ga of the Locras

Valles (Maxwell and Craddock, 1995) and the 3.81 Ga (Fassett and Head, 2008) of Naro

Vallis.

118 The Tyrrhenum Mons, already mentioned in the Alignment 2 subsection, is at the centre of an important volcanic province located between Terra Tyrrhena and Terra Promethei. A labyrinth and a valley network, Tyrrhenus Labyrinthus and Vichada Valles respectively, are on its western slope while the Circum-Hellas system of outflow channels and valleys (Dao,

Niger, Reull, Waikato, and ) are on its southern slope.

Figure 4.12. A) ArcGIS image of the end of Alignment 5 crossing Alignment 4, this suggests that Alignment 5 does not merge Alignment 4, the white rectangles labelled b, c, and d refer to the respective panels in this figure. B) The CTX image P20_008714_1423_XN_37S221W shows the summit of an unnamed tholus along the Alignment 5 after Tyrrhenum Mons, the channels carved on its slope start directly from the summit, the black and white lines follow the fronts of a nearby impact crater ejecta and of lava overflows respectively. C) THEMIS mosaic showing the area of an unnamed volcanic cone located east of the crater Huggins, the white arrows

119 indicate the direction of the lava flows coming out from the cone. D) THEMIS mosaic of the Rossby crater and the Drava Valles system, the white rectangle labelled e refers to the next panel. E) The CTX image G11_022598_1317_XN_48S192W covers the part of the crater rim where the main overflows originate, the white arrows indicate the overflows imaged in the next panel f. F) Blow up of the same CTX image G11_022598_1317_XN_48S192W focusing on the main overflows, the white arrows indicate the points where the lava flows form channels, the white line indicates the most advanced of the latest lava flows.

The next volcanic centre, located beyond the inter-crater volcanic plains fed by the eastern lava fields of Tyrrhenum Mons, is an unnamed domical mountain (tholus) surrounded by impact craters (Fig. 4.12a and Fig. 4.12b). Its slopes are dissected with channels that terminate into a southern volcanic plain interpreted as ridged plains material unit Hr and characterized by extensive lava flows of low viscosity (Greeley and Guest, 1987). This unnamed tholus provides the connection towards the last volcanic centres of the alignment: a volcanic cone located east of the Huggins crater (Fig. 4.12c); the Rossby crater, from which the Drava Valles originated (Fig. 4.12d, Fig. 4.12e and Fig. 4.12f) and contributed with lava flows to the inter-crater plains fed also by other volcanic centres of Alignment 4 (Fig. 4.12a).

4.3.6 Alignment 6

Alignment 6 was tracked from Elysium Mons, an important volcano of Mars already mentioned in Alignment 2, through the volcanic centres of Albor Tholus, the low shields near

Cerberus Tholi, Apollinaris Mons, and a series of unnamed montes and tholi aligned through the Atlantis region ending into Alignment 3 (Fig. 4.13a). The Tartarus Montes and the Hibes

Montes were not included in the alignment because they could be the remnants of the lava fields erupted by Elysium as already explained for the Phlegra Montes in the Alignment 2 section. However, a few low shields located near the Hibes Montes and near the crater

Tombaugh (Vaucher et al., 2009) were included. Small cones were already identified in

Elysium Planitia but not suggested as effusive centres (Keszthelyi et al., 2008). Apollinaris

Mons, already studied for the origin of and the

120 (Kerber et al., 2011) and for its possible role in the dispersal of ashes on longer distance

(Kerber et al., 2012), is an important geographic data point for the direction of Alignment 6.

Between Alignment 4 and Alignment 6 there are several chaos terrains crossed by the

Sirenum Fossae. These are , Atlantis Chaos and other unnamed two chaos terrains ending to the Ariadnes Colles. To the east of the chaos terrains there are Late

Noachian (~ 3.7 Ga) volcanic plains (Tanaka et al., 2014b) surrounding a series of unnamed and aligned tholi. The tholi are separated by a short spacing (Table 1) and end the Alignment

6 into Alignment 3 (Fig. 4.13a). The volcanic plains seem not fed by the , there is no trace of overflows although the fossae are filled with lava flows (Fig. 4.13d). The nearby tholi fill several impact craters with their flows (Fig. 4.13b, 13e, and Fig. 4.13j), the rims of the impact craters appear breached along the main directions of the flows (Fig. 4.13b).

The chaos terrains show several episodes of flooding (Fig. 4.13c), they appear flooded predominantly from the south (where the crater Copernicus is located) to the north (Fig. 4.13f and Fig. 4.13g). Probably the lava flows flooding Atlantis Chaos continued toward the volcanic plains to the northeast but an impact crater obscures the location of the passage (Fig.

4.13l). The channels of the tholi seem to give a contribution volumetrically less significant than the flooding of the chaos terrains coming prevalently from the south (Fig. 4.13b, Fig.

4.13j). Even the contribution from the Simois Colles, interpreted as a volcanic feature (Scott and Tanaka, 1986; Capitan and Van de Wiel, 2011), seems not much significant (Fig. 4.13g and Fig. 4.13m). A detailed volumetric and areal study that may shed further on this issue would be appropriate in an eventual follow-up paper. MOLA profiles do not show vertical displacement between the two sides of the Sirenum Fossae, an example is shown in

Fig. 4.13k, suggesting that some of the fossae might be erosional and not tectonic. CTX mosaics show breakouts along the Sirenum Fossae, but not fissure eruptions (Fig. 4.13d), and partially collapsed lava tubes in continuity with lava channels that do not originate from the

121 fossae (Fig. 4.13d and Fig. 4.13h), a scenario already seen along the (Leone,

2014).

Figure 4.13. A) ArcGIS image of the Alignment 6 merging the Alignment 3, the alignment of unnamed tholi at the end of Alignment 6 leads towards the crater Phaethontis; the grey dot marks an unaligned volcanic feature, the white arrow indicates the location of the half buried crater shown in Fig. 1b. B) THEMIS mosaic of the Gorgonum and Atlantis chaos regions of Mars, the white labelled rectangles indicate the areas imaged in the respective panels of the figure, the blue dots indicate the volcanic centres, the white arrows indicate the main directions of the lava flows on the surface, the question mark on one of the white arrows indicates uncertainty on the way out from Atlantis Chaos. C) The mosaic of the CTX images B19_016914_1399_XI_40S172W and B17_016281_1407_XN_39S171W, centred at 39°01'51.62"S and 171°45'19.82"W, shows the superposition of lava flows on the southwestern side of Gorgonum Chaos indicating several (at least three) episodes of flooding, the white solid lines indicate the flow fronts. D) The mosaic of the CTX images P14_006563_1438_XN_36S166W and P13_005996_1437_XI_36S166W, centred at 36°35'41.12"S and 166°20'01.10"W, shows a segment of the Sirenum Fossae passing in front of the southern and breached side of the crater Mariner, no visible lava flows come out from the fossae, the fossae appear filled with remains of flowing lava inside developing lateral breakouts; just south of the Sirenum Fossae lava channels and lava tubes are visible on a lava field, the lava tubes are observed in continuity with one of the channels and bifurcate along their way to the east, the white rectangle labelled h refers to the panel h of this figure where further detail of the lava tubes is provided. E) The THEMIS mosaic centred at 38°33'00.11"S and 166°54'08.63"W shows the flooded plains surrounding two unnamed tholi belonging to the alignment; the white solid lines indicate the lava flow fronts, the white arrows indicate the main directions of the lava flows, the western fronts head towards Gorgonum Chaos along the Sirenum Fossae; no discernible vents are observed in the area, channels are visible on the eastern tholus (see blow up box on the top right corner), it is plausible that the volcanic source of the

122 flooding might be the same that formed the tholus in this area. F) The THEMIS mosaic centred at 40°44'00.67"S and 173°10'11.12"W shows the lava flooding coming from Copernicus; two lava channels are visible on the image, one heading towards the north west in direction of Atlantis Chaos and one to the north-east directly towards Gorgonum Chaos; from the size of the lava flows, this southern flooding appears to be the most important contribution to the chaos terrains. G) The THEMIS mosaic centred at 37°29'51.12"S and 174°31'12.58"W shows that the Simois Colles contribute to the flooding of the surrounding chaos terrains, although this contribution does not appear comparable to the southern contribution (white arrow heading towards Atlantis Chaos); the source of the channels that convey some lava flows to the chaos terrains is located in the main northern crater and is imaged in the white rectangle that refers to panel m. H) The HiRISE image ESP_028887_1430, centred at 36°38'20.92"S and 166°11'18.45"W, shows in detail the bifurcation of the lava tubes seen in panel d; the segment of the tube located west of the bifurcation shows a collapsed roof while the roofs are still intact in the segments to the east of the bifurcation. I) THEMIS mosaic showing the area to the north-east of Atlantis Chaos, there are several flooded plains surrounding one of the tholi (located within the white rectangle labelled j) belonging to the alignment 6. J) The THEMIS mosaic centred at 28°56'40.47"S and 175°02'20.52"W shows the unnamed tholus belonging to the alignment 6, a couple of channels start from the south-western caldera, it is still possible to distinguish both inner and outer concentric depressions at the available resolution (curvy line), the channels discharge directly on the lava flooded plains located at the foothills of the tholus, the flows of the southern channel cover a fossa (dashed line) and flood a crater located at the bottom of the image; fronts of lava flows are also spreading from the northeastern caldera (solid lines), a circular white line indicates the inner depression. K) MOLA profile crossing the Sirenum Fossae and the craters Galap and Toile, located east of Gorgonum Chaos, shows no vertical displacement in correspondence of the Sirenum Fossae. L) The THEMIS mosaic centred at 32°14'61.08"S and 175°37'13.32"W shows an impact crater located on a probable passage out of Atlantis Chaos to the northern flooded plains, the impact ejecta are superposed to lava flows in an adjacent crater to the east and cover the northeastern rim of Atlantis Chaos, a lava flooded area is also visible between the rims of the chaos and the larger impact crater. M) The mosaic of the CTX images B19_017125_1419_XI_38S173W and B17_016136_1432_XN_36S173W, centred at 37°18'45.24"S and 173°02'22.27"W, shows a lava channel starting from a probable vent located at the centre of the northern crater of the Simois Colles and heading towards Gorgonum Chaos.

4.3.7 Alignment 7

Alignment 7 was tracked from Alba Mons through the volcanic centres of Jovis Fossae,

Pavonis Mons (where it crosses Alignment 1 and Alignment3), a cluster of massifs west of

Claritas Fossae, an unnamed crater, an isolated and large massif along the Icaria Fossae, a couple of heads of channels, the (offset) crater Steno, and a cluster of calderas at the heads of the Chico Valles as far as the South Pole.

Alba Mons is a polygenic volcano probably formed 3.8 Ga ago (Neukum and Hiller, 1981;

Ohman and McGovern, 2014) and it is crossed and surrounded at the same time by several systems of fossae. The fossae surrounding Alba Mons are the Tractus Fossae and the Cyane

Fossae. They seem to originate from Tharsis (Keszthelyi et al., 2008; Spagnuolo et al., 2008) and have been reinterpreted as an asymmetric rift zone resulting from plume volcanism

(Spagnuolo et al., 2008). Volcano-tectonic processes related to magma rise and dike emplacements were also invoked for Alba Mons (Wilson and Head, 2002; Cailleau et al.,

123 2005). The fossae crossing Alba Mons are found on its eastern slope, Phlegeton and Acheron

Catenae for example, and might be related to the surface manifestation of shallow dikes (Scott et al., 2002; Ohman and McGovern, 2014; Leone, 2014). There are also other systems of fossae originating directly from Alba Mons, these are the Alba Fossae and the Tantalus

Fossae, explained as formed by the late stage volcanic loading of the summit and the solidification of the magma reservoir (Ivanov and Head, 2006).

Figure 4.14. Map of vents in the Tharsis region of Mars, the vents are indicated by the yellow dots, the white arrows indicate the direction of advancement of the alignments observed in the fields of vents. A detailed description is provided in the text.

A map of the Tharsis region (Ken Tanaka, personal communication) shows several alignments of vents (Fig. 4.14). Aligned fields of vents were already published in works regarding Pavonis Mons (Bleacher et al., 2009) and Syria Planum (Richardson et al., 2013) but these works did not show them in the whole context of Tharsis. Figure 14 shows a complete overview of Tharsis providing further insight in all the alignments of the vents and in the likely size of the plumes inferred from the width of the fields of vents. The aligned fields of vents located south of Pavonis Mons and Ascraeus Mons, and even inside Arsia

124 Mons (Bleacher et al., 2009), follow the directions of the hemispherical Alignments 3 and 7.

The field of vents south of Pavonis Mons follows the direction of Alignment 7 towards

Claritas Fossae. The field of vents located south of Ascraeus Mons, as well as that inside

Arsia Mons, is aligned with the direction of Alignment 3. A chain of vents is visible north of

Labyrinthus Noctis and within Syria Planum ending at Phoenicis Lacus along the direction of

Alignment 11. Vents are not observed along the direction of Alignment 3 towards Daedalia

Planum and Sirenum Tholus probably because this is the main direction of emplacement of lava flows according to the slope of Arsia Mons. Nevertheless, as seen for other alignments, a volcanic centre or many other vents might likely be buried under the flow field of Arsia

Mons, otherwise the distance between Arsia Mons and the next volcanic centres would be statistically too high (see Table 1).

The next feature after Pavonis Mons along this alignment is a cluster of volcanic massifs located west of Claritas Fossae (Xiao et al., 2012). The centre of the cluster, which is the selected geographic data point on the basis of geometrical criteria, is an unnamed massif located at the coordinates indicated in Table 1. The cluster is characterized by several lines of massifs oriented in direction north-south and embayed by the massive lava flows erupted by

Arsia Mons. The cluster ends up at the extreme limit of the Arsia Mons flow fields after which the inter-crater volcanic plains of are crossed by the Icaria Fossae. For these volcanic plains a formation age of 3.59 Ga was determined (Platz et al., 2013).

Considering the high uncertainty in the crater counts methods, this is an estimate that essentially confirms the one made by Dohm et al. (2001a) who placed the volcanic plains

(their Hnf unit) of Icaria Planum in the Early Hesperian (3.7 Ga).

West of Icaria Planum there is an unnamed crater that shows several channels on its southern slope (Fig. 4.15a and Fig. 4.15b), these channels end up into the volcanic inter-crater plains where wide lava flow fronts spread concentrically from the channels (Fig. 4.15a). This unnamed crater shows overflows almost all over its rim and a flooded inner crater with a

125 central pit where is placed the centroid, a filled crater is also visible on its south-eastern slope

(Fig. 4.15b). A close view of the southern rim shows that the channels were formed directly from the overflows (Fig. 4.15c). Both the unnamed crater and the related volcanic plains are located within the Npl2 unit of Dohm et al. (2001a), which is aged Late Noachian (3.8-3.7

Ga).

Figure 4.15. A) MOLA context image for an unnamed crater located west of Icaria Planum along Alignment 7, the white rectangle labelled b refers to the next panel b in this figure. B) The THEMIS mosaic centred at 44°44'28.98"S and 116°17'18.71"W shows a series of channels flowing out from the southern rim of an unnamed crater; the unnamed crater is nearly centred by another smaller crater with a central pit where the centroid of the feature is inferred, the white curvy line indicates the southernmost extent of its ejecta; overflows are visible almost all over the rim of the unnamed crater and a large channel is carved on its southwestern rim; another channel is running in south-west direction along its western slope but does not seem starting from the crater, the head of this channel is indicated by a white arrow, the overflows coming from the western rim of the unnamed crater partially fill the bed-floor of the channel. C) The CTX image P16_007260_1366_XN_43S116W provides a close-up on the overflows of the southern rim of the unnamed crater; the channels are forming directly from the overflows. D) MOLA context image for a series of channels along the Alignment 7, the white rectangles labelled e and f refer to the areas of interest covered in the subsequent panels e and f; a few ghost craters (dashed circles) are also visible in the area. E) The THEMIS mosaic centred at 55°26'31.46"S and 118°25'22.26"W shows a series of channels located south of the crater ; the centroid (black dot) is taken at the head of the central and deepest channel; the southernmost of the channels seems coming from the easternmost of two

126 unnamed and nested craters but the lack of high res imagery has not allowed a better survey. F) The THEMIS mosaic centred at 62°27'33.37"S and 118°19'37.17"W, shows a channel located around 150 km south east of the crater Dokuchaev; the channel is quite isolated and does not seem related to the activity of the nearby fossa. G) MOLA context image for the southern polar plains of Parva Planum; the white rectangle labelled h covers a wide area including the crater Steno from which some flows flooding the northern Parva Planum seem to originate; the white sinuous line follows (and indicates) a long crossing the whole plains and inferred to start (dashed part of the same line) from a series of domes located just north of the Australe Scopuli (white rectangle labelled i). H) The THEMIS mosaic centred at 67°53'11.78"S and 108°25'36.80"W shows the area south east of the crater Steno; the area shows two main flow fronts east of Steno and one secondary located west of Steno, the white arrows indicate the main directions of the lava flowing downhill; a series of filled (solid circles) and/or ghost craters (dashed circles) are left behind the main flow fronts; the eruptive source has been inferred by the directions of the lava flows and was placed within the crater Steno with some degree of uncertainty indicated by the question mark, however the topographic high on which Steno is located leaves little alternative choices. I) The CTX image B11_013959_0961_XN_83S114W shows a little chain of small massifs and domes of likely volcanic origin; the largest massif, where the centroid is located, shows a channel on its eastern slope.

The next eruptive points along the alignment are located south of the crater Brashear and south-east of the crater Dokuchaev, where ghost craters also appear from the volcanic plains and where the valley networks end (Fig. 4.15d). The CTX coverage of the area is not complete yet so THEMIS mosaics showed the location of the channels (Fig. 4.15e and Fig.

4.15f). The channels south of Brashear formed on a 3 to 6 km topographic high (Fig. 4.15d) that is unreachable from the lava flooding eventually provided by the fossae and from the water provided by the putative basal melting coming from the southern polar areas. The channels contributed to the filling of the volcanic plains located at the foothills of the topographic high. These plains were interpreted as smooth Hpl3 unit formed by thick inter- bedded lava flows of Early Hesperian age (3.7-3.6 Ga) covered by a veneer of aeolian deposits (Scott and Tanaka, 1986), so it could not be excluded that the channels that contributed to their formation might have their same age. The eruptive point located southeast of the crater Dokuchaev does not seem related to the nearby fossa (Fig. 4.15f).

The last eruptive points of Alignment 7 were found around the southern polar region (Fig.

4.15g): one, although the position is still uncertain, was inferred on a topographic high obliterated by the 104.31 km offset (Table 1) impact crater Steno (Fig. 4.15g and Fig. 4.15h); the other was found on a chain of massifs and domes of probable volcanic origin located just northwest from the Australe Scopuli (Fig. 4.15g and Fig. 4.15i).

127

4.3.8 Alignment 8

As already seen for Alignment 5, the Acidalia region is not a suitable candidate for the origin of the Alignment 8 as well. Thus, the alignment was tracked through the head of

Mawrth Vallis and the heads of the Coogoon Valles, including several other eruptive points in

Western Arabia Terra like the crater , a flooded crater near Thymiamata, a possibly buried vent located near the crater Vernal, the crater Crommelin a floor-fractured crater attached to the Miyamoto crater (although some other vent cannot be excluded in Meridiani planum), the heads of Marikh and Bashkaus Valles, a floor-fractured crater near the crater

Maunder, ending with a series of aligned massifs in continuity with the Sisyphi – Australe

Montes.

Mawrth Vallis formed from an unnamed crater located north of the crater Trouvelot (Fig.

4.16a, Fig. 4.16b, and Fig. 4.16e) and it was always considered a fluvial valley of astrobiological and climatological significance for the presence of phyllosilicates and their their putative formation in aqueous environment (Farrand et al., 2009; McKeown et al., 2009;

Michalski et al., 2010; Loizeau et al., 2012; Bishop et al., 2013). The age of formation inferred from the crater counts for the main layered unit is located between 4.0 and 3.9 Ga, followed by putative fluvial activity 3.8 Ga ago, ending between 3.7 and 3.6 Ga ago (Loizeau et al., 2012). Mawrth Vallis today shows evident lava flows fitting the width of the bed-floor along the direction of the stream, nearby flows (coming probably from Becquerel) cover only part of its initial bed-floor instead (Fig. 4.16e). The presence of lava flows is also confirmed by the unaltered olivine unequivocally found along the bed-floor of Mawrth Vallis, although with sporadic serpentine (Ehlmann et al., 2010). Other eruptive points feeding tributaries were found along Mawrth Vallis (Fig. 4.16c and Fig. 4.16f) and some others at the heads of the

Coogon Valles (Fig. 4.16d). The eruptive point of Becquerel was inferred from the directions of the lava flow fronts in the area but also from the presence of pit chains (Fig. 4.16g).

128

Figure 4.16. A) MOLA context image for Western Arabia Terra where the source area of Mawrth Vallis is located, the white dots are the eruptive sources in the region, the white rectangles labelled b and g refer to the respective panels of this figure. B) The THEMIS mosaic centred at 20°11'43.11"N and 18°00'44.81"W shows the source areas of Mawrth Vallis and Coogoon Valles, the white curvy lines indicate lava flow fronts, the white rectangles labeled c, e, and d, refer to the respective panels of this figure. C) The THEMIS mosaic, centred at 22°17'17.85"N and 18°11'01.41"W, is focused on the source areas of small tributaries of Mawrth Vallis and on the flooding of the Oyama crater likely caused by the Mawrth Vallis overflows; the white dots are the sources of the tributaries, the arrows indicate the direction of the flows from the source to the channels, the white rectangle labeled f refers to the panel F of this figure. D) The mosaic of the CTX images B02_010315_1984_XI_18N019W, P22_009748_1974_XN_17N020W, and B06_011871_1972_XI_17N020W, centred at 16°10'21.30"N and 18°42'00.66"W, shows the source area of the southern Coogoon Valles; the white dot indicates the head of the valley, the white curvy lines indicate the lava flow fronts, the white arrows indicate the direction of the flows; it is important to notice how a flow front and a pit chain are spreading radially from the head of the valley. E) The mosaic of the CTX images G09_021800_1991_XI_19N013W, G03_019255_1992_XI_19N013W, P14_006689_1989_XN_18N013W, P07_003709_1968_XN_16N014W, P17_007546_2005_XN_20N014W, and P12_005621_2003_XI_20N014W, centred at 18°38'31.68"N and 13°50'27.23"W, shows the source area of Mawrth Vallis; the black arrow indicates the point where Mawrth

129 Vallis starts its way from the unnamed crater; the white curvy line indicates a lava flow front likely coming from Becquerel, the white arrow indicates the direction of the movement of the flow front. F) The mosaic of the CTX images B02_010394_2026_XI_22N016W, G02_019031_2021_XI_22N017W , F05_037795_2015_XN_21N017W, P18_007968_2023_XN_22N018W, F01_036305_2025_XN_22N018W, centred at 22°02'11.33"N and 17°39'14.16"W, shows the source area of two small tributaries of Mawrth Vallis; the course of the channels is evidenced by the white lines for better visibility. G) The THEMIS mosaic centred at 22°06'22.19"N and 8°51'46.88"W shows the Becquerel crater area; several flow fronts spread radially from the crater Becquerel and several pit chains are visible in the area; the intense flooding has filled nearby craters (solid circles), some of them became ghost craters (dashed circles); the white rectangle labelled h focuses on an area where a filled crater is crossed by a channel originated from the Becquerel lava flows and it is imaged in the next panel h. H) The THEMIS mosaic centred at 24°11'41.25"N and 8°12'21.55"W, shows in detail a filled crater crossed by a channel that originated from the Becquerel lava flows; the white arrows indicate the directions of entrance to and exit from the crater; the filled craters are indicated by a solid circle, the ghost craters by a dashed circle. I) MOLA context image for the Meridiani Terra region, the white dots indicate the eruptive points in the region, the white rectangles labelled k, j, and l indicate the respective panels of this figure. J) The THEMIS mosaic centred at 10°29'06.52"N and 11°07'27.77"W shows a filled unnamed crater located west of Thymiamata; the crater does not show evident tributaries but lava overflows on its northern and its western rim, nearby old craters neither show similar flooding nor evident overflows, for this reason the unnamed crater is very likely an eruptive point. K) The THEMIS mosaic centred at 6°00'02.29"N and 5°06'47.97"W shows the crater Vernal and surrounding area; the white dot indicating the eruptive point has a question mark showing uncertainty on its position, the current position was inferred by the concentric spreading of the lava flow fronts; Vernal is likely not the eruptive point because no evident overflows are visible along its rims, it shows only an eroded western rim where the lava floods likely came from; the lava flooding is also shown by the presence of filled (solid circles) and ghost (dashed circles) craters. L) The mosaic of the CTX images D14_032560_1850_XI_05N010W and D04_028736_1852_XI_05N010W, centred at 4°59'24.35"N and 10°08'20.36"W, shows details of the stratified mound located within the crater Crommelin; the crater located on the mound near the centre of the image does not show evident ejecta and it is not completely filled by the lava; the lava flows on top of the mound, and of the local stratigraphic column of course, seem to come from this crater instead; this crater could be the possible eruptive point.

A channel formed on the lava flows spreading from Becquerel crosses a lava filled crater

(Fig. 4.16h). No incision is seen along the channel direction within the crater thus suggesting no fluvial erosion but just transport of lava that filled the crater before coming out as distributary.

The next eruptive points were tracked southward from the Mawrth Vallis source area in a filled crater located west of Thymiamata, in the crater Crommelin, and in a likely buried vent located near the crater Vernal (Fig. 4.16i). The names of the nearby craters and/or locations provided the names for the geographic data points shown in Table 1. A widespread mantle of lava flows, dated about 3.8 Ga ago (Wilhelms, 1974), filled pre-existing craters and was already observed since the post-Mariner age but the location of some eruptive sources could barely be inferred only now with the higher resolution of the CTX images and through the directions of the lava flows. This method brought to the finding of the filled crater located near Thymiamata, where overflows from the filled crater were also observed (Fig. 4.16j).

130 Also the geographic data point taking the name of the Vernal crater was not directly observed but was inferred from the geometric centre of radially spreading lava flow fronts in the area

(Fig. 4.16k), some of them heading to the west towards Oxia Palus. Volcanic processes have already been studied in Oxia Palus, a region located between Ares Vallis to the west and

Meridiani Planum to the east, where floor-fractured and filled craters were ascribed to volcanic processes (Greeley and Spudis, 1978).

The next volcanic centre along the alignment was not easy to find because of the widespread presence of stratifications of platy-ridged lava flows on Meridiani Planum that also filled the crater Miyamoto. Fluvial processes to explain its morphology were also proposed (Andrews-Hanna et al., 2010; Newsom et al., 2010) but, as usual, without a detailed explanation of the potential sources of water in the area. The presence of an unnamed and externally isolated crater of about 50 km on its western rim, well visible in the Fig. 1 of the

Newsom et al. (2010) paper, indicates the possible presence of underground igneous activity.

This crater appears geomorphologically similar to the floor-fractured craters seen on the

Moon by Schultz and Glicken (1979), the fractures look channels already at THEMIS resolution, a morphology that will also be seen within Maunder and nearby craters. Another interpretation of the Meridiani Planum layered deposits, estimated to be formed around 3.74

Ga ago without significant volumetric contribution of the nearby valley networks, suggests eolian deposits cemented by groundwater or sulphur-rich volcanic processes (Hynek and

Phillips, 2008), an hypothesis that will be dismantled by the next observations.

A thorough check between Meridiani Planum and has shown that there is a large drainage basin (Fig. 4.17a) previously interpreted as a fluvial network formed by precipitation-recharged groundwater sapping (Grant and Parker, 2002), where water was believed to have ponded in a warmer and wetter paleo-climate from the Late Noachian (3.8-

3.7 Ga ago) to the Early Hesperian (3.7-3.6 Ga ago), but there are now several doubts about the fluvial origin. Nearly all the channels discharge directly on Meridiani Planum and do not

131 carve the terrains out there but they appear to contribute to its flooding instead (Fig. 4.17b).

Last but not least, many floodings have the unequivocal morphology of the lava flows at the source, along the course, and at the mouth of the valleys (Fig. 4.17c, Fig. 4.17e, Fig. 4.17f,

Fig. 4.17h, Fig. 4.17i, Fig. 4.17j, Fig. 4.17k, and Fig. 4.17l).

Figure 4.17. A) MOLA context image for the drainage basin located between Noachis Terra and Meridiani Planum at the crossing point between Alignment 8 and Alignment 10; the white rectangles refer to the respective panels of this figure, the white arrow indicates the direction of the lava drainage. B) The THEMIS mosaic centred at 5°05'24.78"S and 4°18'07.77"W shows the mouths of the channels discharging into Meridiani Planum; the channels are evidenced by thin lines and their mouths are imaged in panels d, g, and j; the dashed circles indicate ghost craters, the solid circles indicate filled craters; the white arrows indicate the direction of movement of the lava. C) The THEMIS mosaic centred at 5°05'24.78"S and 6°10'10.51"E shows the lower

132 course of Evros Vallis and the drainage network coming from a basin located just west of the crater Schiaparelli; all the channels are indicated by the thinner lines and start from or continue through the lava flow fronts indicated by the solid and thicker curvy white lines; the dashed lines indicate reconstructed segments of the channels; there is the usual presence of filled and ghost craters along the direction of lava movement on the surface. D) The mosaic of the CTX images P12_005858_1767_XI_03S001W and P13_006214_1765_XN_03S001W, centred at 3°57'16.21"S and 1°36'23.25"W, shows details of the Evros Vallis mouth; the white arrows indicate the directions of the fluid in the channel that are coincident with the mouth of Evros Vallis and with an unnamed filled crater. E) The THEMIS mosaic centred at 13°20'01.41"S and 4°18'07.77"W shows the mid-lower course of Bashkaus Valles and Marikh Vallis; also in this panel the channels are indicated by thin solid lines to be differentiated by the lava flow fronts (thicker white lines) while the white arrows indicate the lava movement on the surface; the filled and the ghost craters are all located along the directions of the channels or along the flooding on the surface; although located along the direction of the lava flooding some craters are not filled, this means that these craters formed after the last flooding event; the filled crater on the right side of the image contains a smaller ghost crater inside it, this means that this smaller crater formed before the last flooding event that covered it completely. F) The THEMIS mosaic centred at 13°20'01.41"S and 6°10'10.51"E shows the middle course of Marikh Vallis and the middle-high course of Evros Vallis; the crater Mädler is not only filled by Marikh Vallis but also by a couple of communicating craters that receive surface flooding from the south. G) The CTX image F02_036542_1763_XI_03S004W shows the mouth of Marikh Vallis; an impact crater deviated the course of the channel indicating that it was still active at the time of the impact. H) The THEMIS mosaic centred at 21°25'21.86"S and 4°18'07.77"W shows the middle-high course of the Bashkaus Valles; one branch crosses the Noachis Basin, the other branch passes between the Noachis Basin and the crater ; the latter branch is characterized more by surface flooding rather than channel forming erosion, this explains the higher number of filled and ghost craters along its course; part of the flooding passes through the Newcomb crater; on the left side of the image is visible the crater Peta belonging to the Alignment 10, from which the Paraná Valles originate. I) The THEMIS mosaic centred at 25°15'21.44"S and 6°10'10.51"E shows the source of Marikh Vallis (white dot); the source is located on a topographic high from which several channels (thin white lines) spread in different directions; the white arrows indicate the direction of movement of lava both in channels (thin white lines) and in surface flow fronts (thick white lines); also lava flow fronts coming from the source of the Bashkaus Valles and from the crater Newcomb contribute to Marikh Vallis. J) The mosaic of the CTX images B01_010051_1761_XI_03S009W, B02_010262_1772_XI_02S009W, and P21_009273_1770_XI_03S008W, centred at 3°19'12.62"S and 9°17'17.26"W, shows the mouth of the Bashkaus Valles located west of the crater Miyamoto; the bed-floor of the Bashkaus Valles is indicated by a thin white line; the small rectangle along the bed-floor indicates the location of the image shown in the blow-up box, the texture of the lava flows is clearly visible in the bed-floor of the Bashkaus Valles. K) The THEMIS mosaic centred at 29°36'07.67"S and 4°18'07.77"W shows the source area of the Bashkaus Valles; the source point is located on a topographic high that originates also the Samara Valles (white dot); the dashed line indicates a reconstructed segment of the channel. L) The THEMIS mosaic centred at 13°15'04.02"S and 13°51'06.36"E shows the source area of Evros Vallis; the source point (turquoise dot) is located at the head of Evros Vallis but it is part of a topographic high that also feeds a drainage basin located west of the Schiaparelli crater (see panel 17a for a context view); Evros Vallis has two tributaries on its high course, one of them sided by several filled craters and one ghost crater.

The hypothesis of atmospheric precipitation was already questioned by the hypsometric study of the valley networks, which concluded that groundwater sapping is more consistent with the morphology of the valleys rather than surface runoff (Luo, 2002), and now more arguments (mostly mineralogical) against both surface runoff and even groundwater sapping are available and will be accounted for in the discussion section. The ages of the terrains at the border between Noachian and Hesperian were essentially confirmed by a test on different age-dating techniques on the nearby Parana Valles (Bouley et al., 2010). The Alignment 8 in this region shows only two source points: one at the head of Marikh Vallis and one at the head

133 of the Bashkaus Valles. The Samara Valles share the same eruptive point with the Bashkaus

Valles (Fig. 4.17k) and then discharge into the Morava Valles (where also the Parana and the

Loire Valles discharge) that become Ares Vallis after passage through Margaritifer and Iani

Chaos. Unaltered olivine was found along Ares Vallis and at its sources (Ehlmann et al.,

2010) suggesting that lava had no contact with water for more than 10k years (Oze and

Sharma, 2007) or never. No other source points than those mentioned here and in Table 1 for

Alignment 8 were found along the whole drainage basin from Meridiani Planum to the heads of Bashkaus Valles and Marikh Vallis. It is worth of notice that the valleys formed on or directly from the lava flows present in the region. Evros Vallis and Marikh Vallis have segments of their courses characterized by flooding regime on the surface before becoming channelized again. The source of Evros Vallis was found aligned to the Alignment 10 that will be shown further ahead.

The next feature encountered along Alignment 8 is an unnamed crater located in Noachis

Terra that shows incisions on its central uplift (Fig. 4.18a, Fig. 4.18b, and Fig. 4.18c). Further to the south a cluster of several apparently floor-fractured craters is suggestive of volcanic activity as observed on the Moon by Schultz and Glicken (1979), the higher resolution images showed that the fractures are in reality channels. The cluster of these craters is almost centred between the degraded craters Asimov and Maunder at the centre of several inter-crater plains

(Fig. 4.18d). The degraded geomorphology of the crater Asimov was explained as the result of basal ice melting during a period of favourable obliquity estimated in > 8 Ma ago (Morgan et al., 2011) but there are problems with this hypothesis. First, it is difficult to understand why such a climate-induced basal melting of ice was limited to this crater and not extended to the whole Noachis Terra (to nearby craters like or for example) and, second, why the global buried cryosphere should be only limited to the area of Asimov, Maunder, and to the two other nearby unnamed craters (Fig. 4.18a). An alternative explanation is given by the

134 passage of the plume that formed Alignment 8 exactly below the cluster of Asimov and

Maunder.

Figure 4.18. A) MOLA context image for several source/eruptive points located in Noachis Terra, the white rectangles labelled b, d, and j refer to their respective panels in this figure. B) The THEMIS mosaic centred at 41°39'28.40"S and 7°21'43.57"Wshows an unnamed crater with a central uplift that fills it nearly completely, the white rectangle labelled c refers to the next panel of this figure. C) The CTX images B08_012596_1383_XI_41S007W and P16_007256_1383_XN_41S007W, centred at 41°07'51.01"S and 7°19'03.41"W, shows a couple of incisions on the slopes of the central uplift of the unnamed crater. D) The THEMIS mosaic centred at 49°23'10.85"S and 5°15'29.50"E shows the cluster of eruptive points located in the craters Maunder, Asimov, and an unnamed smaller one; the white arrows indicate the direction of lava movement on the surface; the white curvy lines indicate lava flow fronts; the white rectangles labelled e, f, and g refer to their respective panels of this figure. E) The CTX image P15_007005_1304_XN_49S352W shows a network of channels emplaced on lava fields within a filled crater; the bed-floors of the channels are highlighted

135 with thin white lines; the channels developed along the directions of expansion of the lava field; the thick white curvy lines indicate the lava flow fronts; the white arrows indicate the directions of movement of the fluid (likely lava) into the channels inferred from the local topography; the white dot indicates the inferred source point; the black arrows and the black perimeter indicate aeolian deposits. F) The THEMIS mosaic centred at 47°02'55.53"S and 4°57'55.22"E shows details of the crater Asimov; the white dot, placed on the central depression of the crater, indicates the possible source point; several channels spread radially from the central depression and are indicated by the thin white lines; the thick curvy lines indicate lava flow fronts spreading from Asimov; the crater has a large circular moat of likely erosional origin due to lava, the shape of the southern moat reminds that of the troughs in Valles Marineris; the moat nearly isolates the central part of the crater forming a sub-circular mesa. G) The THEMIS mosaic centred at 49°42'28.01"S and 2°42'24.08"E shows the crater Maunder and two nearby craters; the crater Maunder has a moat limited to its eastern side as well as its central mesa; the shape of the moat is once again similar to that of the Valles Marineris troughs, this also applies to the moat within its nearest crater; the two smaller nearby craters are connected by a channel similar to those observed within Asimov and to the one inside Maunder; the presence of the moat and of the channels within Maunder suggest post-impact igneous activity, an activity confirmed by some lava flows on its eastern side; the long range of the flow fronts excludes the possibility that they could be Maunder’s ejecta. H) The mosaic of the CTX images P03_002100_1302_XN_49S357W and P13_006135_1302_XN_49S358W, centred at 49°37'52.14"S and 1°48'04.99"E, provides a close up on the channels within Maunder; the centroid of the feature (white dot) is located on a central uplift from which two channels spread in opposite directions (white arrows), streaks of black material of likely igneous origin are also visible in the image. I) The mosaic of the CTX images G15_024226_1298_XN_50S356W and B21_018002_1306_XI_49S356W, centred at 49°22'53.36"S and 4°01'30.80"E, provides a close up on the lava flows and on channel connecting the two smaller craters seen in panel g; the terrain seen in the image is clearly not of aeolian and/or sedimentary activity, surely not located in a topographic low where sediments can accumulate, thus (for exclusion) it could be of igneous origin also confirmed by the typical morphology of the a’a lava flows similar to those already seen in other locations of Mars; the channel observed in the image is directly connected to the moat of the southern crater and to the depression of the northern crater already seen in panel g, it could be formed by the leakage of the lava from the flow into the craters as also suggested by the other smaller nearby incisions seen on the crater’s rim; the morphology of the channel and the nearby smaller incisions excludes any hypothesis of tectonic activity; last but not least, the large volume of the moats excludes a suitable amount of water hidden underground between the two craters, without considering that water is unstable on the surface of Mars and would evaporate immediately when coming to the surface. J) The THEMIS mosaic centred at 49°14'44.45"S and 2°58'42.13"W shows the crater ; both the crater Roddenberry and nearby territories show lava fields properly indicated in the image; the white arrow indicates an incision on the crater’s rim through where lava could have leaked in from the nearby territories; the white rectangles indicate the next panels where blow up images of the lava fields will be shown. K) The CTX image P13_006188_1302_XN_49S005W provides detail on the front of the lava field located inside the crater Roddenberry; no lines were put on the image to indicate the lava flow front because it is already well discernible. L) The CTX image F01_036252_1324_XN_47S001W provides detail on the lava field superposed on the Maunder’s crater overflows/ejecta; in this case the curvy white line differentiates the superposed lava field from the overflows/ejecta coming from Maunder.

The directions of the lava flow fronts in Fig. 4.18d show how the eruptive points following

Alignment 8 should be centred right at Maunder, Asimov, and at the unnamed crater located to the east of Maunder and to the south-east of Asimov, where several channels and overflows can be seen (Fig. 4.18d and Fig. 4.18e). A close up on one of these channels can also be seen in Fig. 4.18e. The channels run along the directions of expansion of the lava fields and it is quite difficult to find water in the same places where lava outpoured before, it will be shown in the discussion section how many places are simply unreachable by the polar basal melting and how the cycle of deposition-sublimation of the brines is unable to create significant underground glaciers. Furthermore, the observation of the Asimov’s moat (Fig. 4.18f) shows

136 a striking geomorphological similarity with incipient chasmata (i.e. Coprates and Candor

Chasma; see Leone, 2014), a morphology also partially observed within Maunder and nearby craters (Fig. 4.18g). The moat separates the rim from a central plateau run by several channels

(Fig. 4.18f, Fig. 4.18g and Fig. 4.18h). The geomorphological analysis of the available CTX images has shown that Maunder is not the source of the inter-crater volcanic plain extending to the north of it, the direction of the flows coming out from Asimov would make it a better candidate (Fig. 4.18f). There is evidence of overflows from Maunder’s rim that covered its western ejecta but they are embayed by lava flows coming from the north-west (Fig. 4.18j and

Fig. 4.18l), probably the same flows that flooded Roddenberry from the north-east (Fig. 4.18j and Fig. 4.18k). Better coverage of CTX images would be needed to identify eventual other possible source points for the formation of the inter-crater lava fields but the mentioned points provide already a pretty satisfactory explanation at the moment.

South of Maunder, at 381 km of distance (Table 1), Alignment 8 continues along a chain of massifs heading to the Sysiphi Montes. The Sysiphi Montes have already been studied in the past and several mechanisms for their formation have been evaluated with the conclusion that a volcanic origin best explains their morphology and origin (Ghathan and Head, 2002). It is interesting to note how these volcanic massifs are aligned with considerable continuity with the Australe Montes to the south and also how they cross the alignment of massifs of

Alignment 2 coming from the east. The Australe Montes have been classified as knobby material, unit k in the geologic map of the Mare Australe area of Mars, interpreted as knobs either formed by ejecta material or rims of older crater material (Condit and Soderblom,

1978). However, there are several elements that favour an alternative interpretation of volcanic origin: a) the continuity of alignment with the volcanic Sysiphi Montes, including the average distance of about 100 km (see Table 1); b) the presence of Sysiphi Cavi between the Sysiphi Montes and the Australe Montes, a volcanic origin was proposed for the morphologically similar Cavi Angusti (Ghatan et al., 2003); c) if the Condit and Soderblom

137 (1978) hypothesis of the remains of an old impact basin rim was true, the > 2000 km length of the Sysiphi – Australe alignment would make it one of the largest impact basins of Mars, even larger than Hellas, but there is no topographic evidence for such a structure; d) the average height of 1500 m of the Australe Montes, provided by MOLA data, is consistent with the average height of 1000-1500 m (Ghatan and Head, 2002) of the Sysiphi Montes suggesting a similar origin by the same migrating plume.

4.3.9 Alignment 9

Starting from Olympus Mons, as well as Alignment 1, this alignment was tracked through

Pavonis Mons, Noctis Labyrinthus, a unnamed volcanic massif in Sinai Planum, a couple of unnamed volcanic massifs in Thaumasia Planum, an unnamed massif (heavily destroyed by impact craters) near the head of , and the head of the Erythraea Fossa. The tracking was facilitated by the previous identification of volcanic centres in Thaumasia

Planum by Xiao et al. (2012). Of course, no need of particular interpretation to identify the volcanic centres located at the beginning of the alignment, particularly for Olympus and

Pavonis Montes that were already included in Alignment 1 and Alignment 3 respectively.

The volcanic massifs in Sinai Planum and Thaumasia Planum as far as the unnamed massif located north of the head of Nirgal Vallis were already mentioned in the work of Xiao et al.

(2012). The latter massif is located on a local topographic high and its valleys are tributaries of Her Desher Vallis (Fig. 4.19a, Fig. 4.19b, and Fig. 4.19c). A closer look through available

CTX and THEMIS images revealed once more the presence of lava flows at the heads of Her

Desher Vallis (Fig. 4.19c) and Nirgal Vallis (Fig. 4.19d) and that these valleys formed directly from lava flows. The Her Desher Vallis tributaries coming from the unnamed massif also formed from lava flows (Fig. 4.19e). The presence of lava was observed even inside one of the craters located on top of the unnamed massif (Fig. 4.19f). Both Her Desher and Nirgal

Vallis, although the latter more evidently, are tributaries of , which in turn ends

138 up into crater to the north and into Argyre to the south. The presence of unaltered olivine was already confirmed along Nirgal Vallis (Ehlmann et al., 2010) and inside Holden crater (Rogers and Bandfield, 2009).

Figure 4.19. A) MOLA context image for the Southern ; the violet dots refer to the source points of Alignment 9, the turquoise dots refer to those of Alignment 10; the white rectangles labelled b, g, and i refer to their respective panels of this figure. B) The THEMIS mosaic centred at 25°27'36.51"S and 45°45'24.97"W shows the source areas of Her Desher Vallis and Nirgal Vallis; many channels spread radially from a topographic high devastated by impact craters, the channels located on its southern slope contribute to Her Desher Vallis, the centroid (violet dot) for the topographic high is located on the top between two craters; the white rectangles labelled c, d, and e refer to their respective panels of this figure. C) The mosaic of the CTX images P17_007587_1556_XN_24S048W, P20_009024_1556_XN_24S048W, and P13_006229_1552_XN_24S048W, centred at 24°39'44.59"S and 48°25'08.35"W, shows the head of Her Desher Vallis; as already seen in the Noachis Terra drainage basin, the valley forms directly on or from the lava flows;

139 the white curvy lines are the lava flow fronts, the white arrows indicate the direction of lava movement on the surface. D) The mosaic of the CTX images D21_035304_1527_XN_27S045W, P11_005385_1525_XN_27S045W, B01_010092_1525_XN_27S045W, P20_008958_1525_XN_27S045W, P12_005596_1525_XN_27S046W, and G18_025322_1523_XI_27S046W, centred at 27°14'31.51"S and 45°40'25.64"W, shows the head of Nirgal Vallis; as well as already seen for Her Desher Vallis, the head of Nirgal Vallis and those of its tributaries form directly on or from the lava flows; the lava flow fronts are indicated by the white curvy lines, the white arrows indicate the direction of lava movement on the surface. E) The mosaic of the CTX images F04_037427_1545_XN_25S044W and B12_014417_1551_XN_24S045W, centred at 25°02'24.77"S and 45°04'13.45"W, shows the heads of some channels located on the southern slope of a topographic high north of Nirgal Vallis and Her Desher Vallis; again, as already seen in other valleys, the channels form directly from the lava flows; the channels are indicated by a thin white line, the lava flow fronts are indicated by a thick white line, the direction of lava movement is indicated by the white arrows. F) The CTX image D20_035159_1561_XN_23S045W provides detail of the western crater floor on top of the topographic high; the image shows clearly a lava field on the floor of the crater with several areas covered by aeolian deposits easily recognizable by the presence of ripples and . G) The THEMIS mosaic centred at 28°12'25.47"S and 30°17'08.35"W shows the Erythraea Fossa and its source area; several channels, as well as several lava flow fronts, spread from the source of the fossa indicated with a violet dot; the channels are evidenced by white thin lines while the lava flow fronts by thick white lines; the directions of the lava flows are indicated by the white arrows; the white rectangle labelled h refers to the next panel of this figure. H) The mosaic of the CTX images B17_016223_1527_XI_27S029W and D03_028249_1522_XN_27S029W, centred at 27°50'41.39"S and 29°06'49.03"W, shows the source area of the Erythraea Fossa; the channels indicated by thin lines spread from the terrains around a crater visible at the centre of the image, the violet dot indicating the source point is located within the crater mostly for geometric reasons (i.e. located at the centre of the four heads of the channels) because the crater is likely not the source point; only the northern and the western channels head towards the Erythraea Fossa. I) The THEMIS mosaic centred at 29°38'18.84"S and 23°42'38.01"W shows the last source point of the Alignment 10; the source points are located within two craters (the southern larger and the northern smaller) from which several channels spread towards the east and the north; the channels are evidenced by the usual thin white lines where the course is visible on the surface; the dashed lines indicate the course of the channels where interrupted by impact craters, the direction is inferred from the direction of the end of the last segment and the beginning of the next segment, the question mark indicates uncertainty on the reconstruction of the channel; another lateral source point with respect to the main alignment contributes to one of the channels with a tributary; the key points of this image will be observed in more detail in the next panels j, k, and l (white rectangles). J) The mosaic of the CTX images F05_037756_1492_XN_30S024W and B17_016302_1505_XI_29S025W, centred at 30°10'07.84"S and 24°46'12.50"W, shows the eastern rim and part of the floor of the southern crater; the floor of the crater is draped with lava and several incisions are visible on the walls, probably formed by the withdrawal of lava within the crater; several overflows (white thick lines indicate the fronts) are visible along the eastern rim and a channel (indicated by a thin white line) forms directly from these overflows. K) The mosaic of the CTX images B18_016658_1522_XN_27S024W, B19_016869_1518_XN_28S025W, and B17_016302_1505_XI_29S025W, centred at 29°36'07.22"S and 24°47'02.14"W, shows the northern and the eastern rims of the northern crater; the centroid of the feature is located at the geometric centre of the crater where a depression is also present; several incisions present on the craters’ walls converge radially towards the central depression; the northern channel visible on the image coincides with one of the incisions on the rim; overflows are visible along the northern rim; again, it is plausible that these incisions were formed by the withdrawal of the lava that formed the overflows and the channels; considering that the crater is located on a topographic high, it is unlikely that these incisions could be formed by surface flows heading towards the crater. L) The mosaic of the CTX images B08_012636_1489_XN_31S021W, P03_002048_1494_XN_30S021W, and G16_024319_1495_XN_30S021W, centred at 30°49'02.11"S and 21°11'56.18"W, shows an unnamed crater and several channels that form on its western slope; the crater has a central ring-shaped mound and some platy-ridged flows just east of it; there are overflows on the western slope of the crater so the centroid for the feature was placed at the centre of the ring.

The analysis of the topography showed that the received a contribution from the Erythraea Fossa (Fig. 4.19g) but also that this contribution is too small to justify the volume of the Ladon Valles. Erythraea Fossa is roughly 600 km3 (considering its maximum values of width in the lower course) while Ladon Valles is ~ 4000 km3 (considering both its minimum depth and width), the Holden crater does not show any evident distributary, so it

140 would be reasonable to think that the Ladon Valles received contributions directly from the

Mare Erythraeum (and thus from the Tharsis volcanoes) through the lava flow fields located to the north-north-west. On this regard, the lava flows coming from Tharsis must have crossed the crater and indeed there are channels conveying lava into it. This area was heavily misinterpreted due to the dogmatic belief that water “must” have been present on

Mars. The delta of Eberswalde was interpreted as formed by alluvial deposits but the origin of the water remained speculative without any credible explanation (Malin and Edgett, 2003).

The of the alluvial channels was explained with aeolian deflation (Pain et al.,

2007), although an alternative explanation could have included the typical shape of lava tubes visible on the lava fields, but the volcanic option was not considered. Lava flows were also observed at the head of the Erythraea Fossa (Fig. 4.19h), interpreted as an Hesperian fluvial system (Buhler et al., 2011), which ends up into the Ladon Valles (Fig. 4.19a and Fig. 4.19g).

The Ladon Valles, an important channel located downhill of the Holden crater and of Nirgal

Vallis, were dated from 4.0-3.7 Ga (Neukum and Hiller, 1981) to 3.7 Ga (Grant, 2000). The last feature of the alignment is the region of Vulcani Pelagus, several channels formed on the slopes of a topographic rise located northwest of the crater (Fig. 4.20a). From the observation of the heads of the channels, spreading north and south away from a couple of attached craters at the centre of Fig. 4.20b, the highest topographic point does not seem to be the source of the flows. Although a clearly discernible eruptive point was not found, the many lava flows on which the channels are emplaced seem to spread from a point indicated by a coloured dot in Fig. 4.20b. If the heads of the channels did not form from the lava flows directly and thus can be taken as source points according to the criteria specified so far, the dot should be moved about 100 km to the west along the same direction of the alignment.

However, there is the well-founded possibility that these channels formed from the lava flows coming from the spreading point indicated by the dot in Fig. 4.20b.

141 Figure 4.20. A) MOLA context image for the region of Vulcani Pelagus located at the end of Alignment 9, the white rectangle labelled b refers to the next panel of this figure. B) The THEMIS mosaic centred at 35°20'11.43"S and 16°54'13.50"W shows the area west of Vulcani Pelagus where several channels form directly from the lava flows; the white curvy lines indicate the lava flow fronts and the white arrows indicate the direction of lava movement on the surface; the solid circles indicate filled craters, the white rectangles labelled c and d indicate the areas imaged in the next panels. C) The mosaic of the CTX images B06_012003_1451_XN_34S018W, B07_012425_1451_XI_34S018W, and P11_005450_1445_XN_35S018W, centred at 34°48'27.91"S and 18°33'00.03"W, shows the confluence of several channels formed from/on the encounter of lava flows coming from opposite directions; the white curvy lines indicate the lava flow fronts, the white arrows indicate the directions of lava movement on the surface; the dashed circles indicate likely ghost craters. D) The CTX image P15_006953_1470_XN_33S015W shows a close up on a flooded crater located near the inferred centroid for Vulcani Pelagus; the white curvy line indicates a lava flow front and the white arrow its direction of movement; a feature of probable erosional origin, perhaps an eroded pit chain is visible near the northern rim of the crater. E) MOLA context image for an unnamed elongated feature near the head of one of the

142 Labeatis Fossae; the white rectangle labelled b refers to the next panel of the image. F) The THEMIS mosaic centred at 28°38'17.71"N and 82°25'55.47"W shows the area surrounding the elongated feature; the white thinner line indicates the limit of the ejecta of the impact crater located to the north-west of the feature, it is clearly visible that the elongated feature is immediately out of range of the ejecta, this is important to establish that the terrains indicated by the white thicker line might be overflows and not crater ejecta; the white rectangle labelled c refers to the next panel of this figure. G) The CTX image G21_026233_2090_XN_29N082W shows details of the south-western perimeter of the elongated feature; the white thick line indicates the limit of the overflows coming out from the elongated feature; it is possible to see that one of the Labeatis Fossae cuts through the overflows and thus postpones the emplacement of the overflows; the white rectangle d refers to the next panel of this figure. H) The same CTX image G21_026233_2090_XN_29N082W provides a close up of the rim of the elongated feature where some overflows are visible; the floor of the elongated feature might possibly be draped with lava flows.

In this case the head of the channel would not be the source point but, regardless of this fact, it would not be out of the alignment anyway if considered as source point. The only difference would be in the 100 km of along track distance, a nearly insignificant distance if compared to the average several hundreds of km measured along the alignments. At last, there are several flows coming from opposite directions along the course of the channels (Fig.

4.20b and Fig. 4.20c), it is clearly visible how these flows flooded pre-existing craters (Fig.

4.20b, Fig. 4.20c, and Fig. 4.20d), and this observation may suggest additional invisible vents further on the west along the alignment.

4.3.10 Alignment 10

The tracking of this alignment was facilitated by the previous identification of volcanic centres in by Xiao et al. (2012) but also by the decreasing age of the valley networks from Terra Sabaea to Margaritifer Sinus (Hoke and Hynek, 2009). Starting from

Meroe and Nili Paterae before the crossing point with Alignment 5 in Terra Sabaea, the alignment is tracked through the heads of Naktong Vallis, an unnamed massif at the head of

Verde Vallis, an unnamed massif in Sinus Sabaeus at the head of the Brazos Valles, an unnamed (and heavily bombarded) massif at the head of Evros Vallis, an unnamed crater contributing to Marikh Vallis, the crater Peta at the head of the Erythraeum Chaos and the

Paranà Valles, and an unnamed crater located on a topographic high.

143 The activity of the Syrtis Major calderas was dated from Late Noachian to Early Hesperian

(3.9-3.7 Ga) age (Michalski and Bleacher, 2013) and is characterized by a 2 km deep central depression perhaps caused by the collapse of a shallow magma reservoir located into the mega- of Mars (Hiesinger and Head, 2004). Nili Patera in particular has shown traces of differentiated (felsic) lava (Christensen et al., 2005; Wray et al., 2013) that might support such a view.

The observation of the volcanic centres along this alignment has led to the surprising discovery that the heads of several fluvial valleys, most of them located near highlands volcanoes, are aligned with the Syrtis Major calderas through Terra Sabaea (Fig. 2a and Fig.

2c). It is interesting to note how the formation and the ages of these valleys, estimated from crater counts, are decreasing in space and in time (Hoke and Hynek, 2009) according to the direction of the alignment. The ages were collected in Table 1 and show a decreasing order from the Syrtis Major as far as the heads of the valley networks located in Margaritifer Sinus, a time span of 210-260 Ma is observed between the formation of the valleys in Terra Sabaea and the valleys in Margaritifer Sinus (Hoke and Hynek, 2009). The presence of the volcanic massifs listed by Xiao et al. (2012) provided a good point of reference for the probable volcanic origin of Verde Vallis, Nakhtong Vallis and other unnamed valleys in Terra Sabaea so these features were not imaged in this paper. Other features not included in the Xiao et al

(2012) paper, which are located from Evros Vallis westwards, were imaged in this paper.

The head of Evros Vallis is located on the southern slope of an unnamed massif located south of the crater Schiaparelli (Fig. 4.17a and Fig. 4.17l). The course of Evros Vallis was followed as far as its mouth in Meridiani Planum (Fig. 4.17b, Fig, 17c, and Fig. 4.17f) and revealed a surface lava flooding limited to the segment located to the east of the crater Mädler

(Fig. 4.17c and Fig. 4.17f). Furthermore, the same massif from which Evros Vallis originated was also a contributing source for another drainage basin located just west of the lava-flooded crater Schiaparelli (the direction is indicated by the white arrow in Fig. 4.17a). The lava flows

144 coming from this basin directly formed two unnamed valleys that discharged into Meridiani

Planum (Fig. 4.17c).

The next point along the alignment is a crater located on a topographic high (Fig. 4.17a) that contributed to Marikh Vallis with some channels and to the flooding of the crater Mädler through a chain of filled craters (turquoise dot in Fig. 4.17f), a morphological similarity already seen along Alignment 8 where another topographic high formed Marikh Vallis (Fig.

4.17i).

The Paraná Valles originated from the crater Peta (Fig. 4.17a and Fig. 4.17h) to drain into

Erythraeum Chaos. Although not directly imaged in this paper, it is possible to see on any map of Mars that Erythraeum Chaos discharged to Morava Valles through the Loire Valles.

As already mentioned in the Alignment 8 section, the Morava Valles become Ares Vallis where Ehlmann et al (2010) found unaltered olivine.

South-east of the Erythraea Fossa there is the last eruptive point of the alignment, at the centre of several flow fields (Fig. 4.19i), chosen as centroid for many other heads of valleys which contribute to several large inter-crater plains extending as far as Argyre to the south and to the valley networks of Oltis and Himera Valles to the north (Fig. 4.19a). Both these valleys discharge into the Morava Valles – Ares Vallis system as well.

4.3.11 Alignment 11

Alignment 11 starts with Labeatis Mons (Table 1), a Noachian massif (Tanaka et al., 2014) located on Western Tempe Terra, surrounded by the curvilinear .

Near the head of one of the Labeatis Fossae there is an elongated feature of about 20 km that shows overflows all along its southern perimeter (Fig. 4.20e, Fig. 4.20f, Fig. 4.20g, and

Fig. 4.20h).

145 The next main feature, Tharsis Tholus, has an inferred age for the shield between 3.54 Ga

(Werner, 2009) and 3.24 Ga (Robbins et al., 2011), and it is located at a distance of 531.29 km from Alignment 3. Another minor feature along the alignment is an unnamed vent at

2°07'12.99"N and 101°20'09.04"W, located north of Noctis Fossae and west of Fortuna

Fossae, with the same morphological characteristics of several nearby other vents studied by

Hauber et al. (2009). These low shields do not occur everywhere on Tharsis and, as it can be seen from Figure 22 of Hauber et al. (2009) and as already seen in Fig. 4.14, they are distributed according trends following the main alignments crossing Tharsis. Alignment 11 crosses regions with volcanic centres already seen in other alignments, Noctis Labyrinthus,

Syria Planum, and the western Claritas massifs before reaching the volcanic plains extending between the Icaria Fossae. The origin of these volcanic plains might be ascribed both to the

Icaria Fossae and to an unnamed massif of about 150 km of diameter heavily bombarded by impact craters (Fig. 4.21a). Several channels spread from this massif in almost all the directions (Fig. 4.21b). The morphology of these channels is similar to many others observed along other alignments (Fig. 4.21c and Fig. 4.21f). A closer view on the Icaria Fossae has shown possible skylights (Fig. 4.21d) and breakouts (Fig. 4.21e), features that have nothing to do with tectonics, suggesting that these fossae might be remains of lava tubes or erosional lava channels. The last feature of Alignment 11 is the system of channels of the Chico Valles that feeds two main volcanic plains (Fig. 4.21g). The sources of the channels were located from the radial spreading of the channels (Fig. 4.21h, Fig. 4.21i, Fig. 4.21k, and Fig. 4.21l) and from the directions of the main lava flow fronts observed in the region (Fig. 4.21j). The

Icaria Fossae seem not to be involved in the formation of the northern Chico Valles (Fig.

4.21l). The flows come both from the east and from the west encountering at the foothills of the Chico Valles (Fig. 4.21j).

146 Figure 4.21. A) MOLA context image for a massif of ~150 km of diameter located north of the craters and along Alignment 11, the white rectangle labelled b refers to the next panel of this figure. B) The THEMIS mosaic centred at 50°35'42.83"S and 132°44'55.70"W shows several channels spreading from an unnamed massif devastated by impact craters and located between Aonia Terra and Terra Sirenum; the white thin lines indicate the channels and the white arrows indicate the directions of the flows; the violet dot indicates the centroid for the feature, the white rectangles refer to their respective panels of this figure. C) The CTX image F04_037549_1307_XN_49S131W provides detail of the channels spreading from the massif imaged in panel b. D) The CTX image G12_022899_1279_XN_52S130W, taken on an unnamed crater located north of the crater Hussey, shows a black hole that might also be a skylight in the middle of another one of the Icaria Fossae; unfortunately, the resolution of the image is not enough to confirm it but, if the black hole is an eventual skylight, it would reveal that these fossae might be collapsed lava tubes. E) The CTX image D16_033369_1289_XN_51S132W, taken on the Icaria Fossae crossing the same unnamed crater of the previous panels c and d, shows lava appearing from the dusty cover and a breakout of the Icaria Fossae that also fills a nearby smaller crater; the overall morphology of lava flowing in the fossa instead of spreading laterally from it suggests an erosional rather than a tectonic origin. F) The same CTX image F04_037549_1307_XN_49S131W shows other channels located to the south respect with those seen in panel c. G) MOLA context image for the source areas of the Chico Valles located near the end of Alignment 11, the white rectangles labelled f, g and h

147 refer to the respective panels of this figure. H) The THEMIS mosaic centred at 66°54'23.71"S and 143°58'33.33"W shows the area of the central Chico Valles; the white thin lines indicate the course of the channels that are part of the Chico Valles; the centroid for the feature was inferred from the radial spreading of the channels and it is indicated by the violet dot; the dashed line indicates the segment of a channel interrupted by an impact crater; the solid circle indicates a filled crater, the dashed circle indicates a ghost crater; the white arrows indicate the directions of the flows, the Chico Valles spread in almost all directions that will be followed in the next panels. I) The THEMIS mosaic centred at 67°33'03.59"S and 133°59'09.00"W shows the area of the central-eastern Chico Valles; the channels visible in the figure come from the area imaged in the previous panel and follow the eastern slope along the direction indicated by the white arrows; a ghost crater is present along the course of one of the channels; this channel ends up in a feature that appears to be a lava tube (black arrow) but the resolution is not enough to discern it so a question mark was placed. J) The THEMIS mosaic centred at 69°09'51.11"S and 162°15'20.51"W shows the southern end of the inter-crater volcanic plain located east of the crater and west of the Chico Valles; the area shows many filled and ghost craters; the flow fronts come both from the east and from the west, the filled crater Charlier is located along the direction of a wide flow front but the western source area is likely located further to the west and out of the sight of the image. K) The THEMIS mosaic centred at 72°26'42.30"S and 148°28'23.37"W shows several valleys coming from the same topographic rise that formed the Chico Valles; the centroid for the feature (violet dot) was inferred from the different directions of flow of the valleys indicated by the white arrows; although the centroid coincides with a crater, the resolution is not enough to understand whether this might be the likely source that formed the valleys; the northernmost of the channels ends up in a lava field that shows possible lava tubes (black arrow), the question mark is placed only because the resolution is not enough to confirm it. L) The THEMIS mosaic centred at 63°21'43.34"S and 138°41'49.96"W shows the northern Chico Valles and the southern end of the Icaria Fossae; the source of the northern channels is devastated by impact craters, one of the channels comes from a filled crater; one of the Icaria Fossae crosses a filled crater and terminates well before a small crater where the source point (violet dot) of one of the channels is located; from the resolution of this image it seems that the Icaria Fossa does not feed the crater where the source point is located.

However, the source of the western flows was not clear, the flows might come from west of (and flood) Charlier or even come from the far away Chronius Mons and/or from the Ulixis

Rupes. The alignment ends at the South Pole towards the Ultimi Scopuli but the presence of the polar cap makes difficult to observe eventual volcanic features there.

4.3.12 Alignment 12

As well as Alignment 1 and Alignment 9, Alignment 12 was tracked from Olympus Mons.

The next eruptive point is the Noachian (3.92-3.73 Ga) Ulysses Patera (Werner, 2009), located after some aligned vents on both sides of the (Fig. 4.14). Ulysses

Patera ranges from 3.92-3.73 Ga (Werner, 2009), 3.75 Ga (Neukum and Hiller, 1981), to 3.4

Ga for the flanks (Plescia, 1994) while the lowest age reported is 1.92 Ga (Robbins et al.,

2011). It is also visible a chain of vents crossing Ulysses Patera along Alignment 12, passing between Pavonis and Arsia Mons (Fig. 4.14). The next geographic data point of this alignment is found along the region of Syria Planum, characterized by the southern expansion

148 of the lava tubes coming from Labyrinthus Noctis (Leone, 2014), in a field of vents aligned mainly along the direction of Solis Planum towards the Coracis Fossae indicating three volcanic episodes for the total duration of ~ 900 Ma from 3.7 Ga to 2.8 Ga ago (Richardson et al., 2013). Some of the vents are located on top of low shields (Richardson et al., 2013).

Along Solis Planum Alignment 12 does not show vents or any other low shield likely due to a thicker cover of lava flows, the next features are found into two volcanic massifs near the

Coracis Fossae indicated by Xiao et al (2012). These massifs, together with the Coracis

Fossae, have contributed to the lava flooding of large parts of Ogygis Regio, Phrixi Regio and

Bosporus Planum. However, this is not the only contribution that these regions received. A few craters located within Ogygis and Phrixi Regio contributed to the lava flooding forming some valley networks (Fig. 4.22a and Fig. 4.22b). The lava flows and the channels formed directly from these craters can be seen in Fig. 4.22g and Fig. 4.22i. Other source points were individuated within Argyre Planitia suggesting that this large impact basin was already formed before the migration of the Alignment 12 plume reached it. The most unambiguous volcanic centre is Argyre Mons, which shows a tholus shape with central caldera and a main channel starting from its south-western rim (Fig. 4.22c). Although the Nereidum and the

Charitum Montes are associated to the formation of the Argyre impact basin in Noachian age

(Hiesinger and Head, 2002), the mountains located inside Argyre Planitia (Argyre Mons,

Octantis Mons, and Oceanidum Mons) do not seem to be related to the same event but to a likely volcanic origin. However, given the relatively short distance (~ 65 km) that separates

Oceanidum and Octantis Montes from the (Fig. 4.22d), they were not taken as source points because of the uncertainty related to such a short distance. The same principle is applied to the although they look aligned with Argyre Mons and with other low relief source points (Fig. 4.22a). Source points instead include low relief features like the filled crater Mari, from which the sinuous pit chain Octantis Cavi is directly

149 formed (Fig. 4.22f), and the crater Lodwar, from which a thin channel directly starts (Fig.

4.22j).

Figure 4.22. A) MOLA context image for the regions of Phrixi Regio, Ogygis Regio, Aonia Terra, Argyre Planitia basin, and Argentea Planum; the sand dots indicate the source points; the white rectangles labelled b, c, d, e, and f refer to their respective panels of this figure; the thin curvy lines indicate the course of the channels that likely carried lava from the volcanic plains of Argentea Planum and Sisyphi Planum to flood the floor of Argyre Planitia. B) The THEMIS mosaic centred at 44°37'37.64"S and 67°09'52.07"W shows valley networks located southwest and north of Ogygis Regio; the source points (Sahara sand dot) are individuated within craters from which several valleys spread almost radially mainly along the north and west directions in Ogygis Regio and to the north and to the east in Phrixi Regio; a distributary comes out from the flooded craters east of that received lava from the eruptive point that will be shown in panel e. C) The THEMIS mosaic centred at 50°22'17.79"S and 48°07'22.89"W shows Argyre Mons and its valley network; a channel comes directly from its central crater and it is indicated by a white arrow. D) The THEMIS mosaic centred at 55°36'44.63"S and 41°32'39.39"W shows the network of lava tubes in Southern Argyre Planitia; the network

150 spreads from the debouchment of Surius Vallis and develops toward the north and the east; the white rectangles refer to the respective panels of this figure; the panels k and l will show why the sinuous structures are not eskers; the Sahara sand dot indicates a probable eruptive point located in the crater Lodwar. E) The THEMIS mosaic centred at 53°43'24.89"S and 65°12'37.11"W shows the source point located around 165 km south east of the crater Douglass; the source point (Sahara sand dot) is located inside a crater from which many channels spread radially in almost all directions; the white rectangle labelled h refers to the respective panel of the figure. F) The THEMIS mosaic centred at 52°24'45.57"S and 46°15'44.77"W shows the filled crater Mari and the Octantis Cavi; Octantis Cavi is a pit chain that originates from Mari and then follows a sinuous pattern; the crater Mari has a morphology already seen in other source points like some of the Arabia Terra calderas. G) The mosaic of the CTX images B11_013772_1329_XN_47S072W, B07_012282_1329_XN_47S073W, B10_013627_1329_XN_47S073W, and B09_013271_1328_XI_47S074W, centred at 47°17'59.76"S and 73°23'01.83"W, provides a close up of the valley network on the slope of an unnamed crater located west of Ogygis Regio, lava flows and the channels formed among them are clearly visible in the image. H) The mosaic of the CTX images B16_015974_1251_XN_54S066W and F05_037850_1271_XN_52S067W, centred at 54°03'58.75"S and 66°56'38.60"W, provides a close up of the channels spreading from an unnamed crater located around 165 km southeast of Douglass; the Sahara sand dot indicates the centroid of the source point. I) The mosaic of the CTX images B12_014286_1392_XN_40S066W and P15_006889_1417_XN_38S067W, centred at 40°12'35.29"S and 66°19'31.09"W, shows the origin of a valley network from two craters in Phrixi Regio; one of the branches of the valley network forms directly from the overflows of the crater located on the western limb of the image (white curvy line). J) The CTX image D14_032812_1242_XN_55S043W provides detail of the crater Lodwar; the Sahara sand coloured dot indicates the centroid for the feature and the probable source point; the thin white line indicates a channel that spreads from the crater. K) The CTX image D15_032891_1239_XI_56S039W, taken on the confluence between Auxo Dorsum and Charis Dorsum, shows why the features are lava tubes and not eskers; a breakout and collapsed segments are visible in the image, features that are not observed in eskers; the final part of the tube is completely collapsed and evolved into a channel. L) The CTX image G12_022922_1247_XI_55S037W, taken on the final end of both Auxo Dorsum and Charis Dorsum, shows the typical morphology of the channel and not that of the esker.

The dorsa located near Octantis and Oceanidum Montes (Fig. 4.22d) were interpreted as eskers formed by putative glacial activity 3.6 Ga ago (Bernhardt et al., 2013), although they are morphologically more similar to the lava tubes present on Chryse Planitia (Leone, 2014).

Other authors provided different interpretations, lava flows or wrinkle ridges (Tanaka and

Scott, 1987), exhumed igneous dikes (Carr and , 1980), lacustrine barriers (Parker,

1994), glacial moraines (Hiesinger and Head, 2002), but the prevailing view remained that of the eskers (Howard, 1981; Kargel and Strom, 1992; Hiesinger and Head, 2002; Kargel, 2004;

Banks et al., 2009). The interpretation and the formation of the eskers relies on the presence of water brought into the basin by the valleys located along the Argyre’s rim, an hypothesis that is not certainly supported by the widespread presence of unaltered olivine (Poulet et al.,

2007; see also Ehlmann et al. (2010) for unaltered olivine in Uzboi Vallis). The process of valley formation bringing water into the basin was also interpreted as post-dating the volcanic infill (Hiesinger and Head, 2002). The length of the valleys and their origin from the volcanic plains of Argentea Planum and Sisyphi Planum (see white lines in Fig. 4.22a) and the radial

151 spreading of the main dorsa (in this case, lava tubes) from Surius Vallis (Fig. 4.22d) show that the valleys indeed contributed to the volcanic infill of Argyre Planitia and post-dated the infill produced by the volcanic centres inside Argyre. However, the infill brought by the valleys was the lava flooding of the volcanic plains of Argentea Planum and Sisyphi Planum and thus not formed by water at all. This hypothesis is also confirmed by the observation of lava tubes like Pasithea Dorsum embaying and surrounding Oceanidum Mons and of the volcanic infill embaying also the Chalce Montes (Fig. 4.22d). Furthermore, a closer view of the Auxo

Dorsum and Charis Dorsum already at CTX resolution showed that the morphology of these dorsa is more compatible with lava tubes rather than eskers, collapsed sections in proximity of a breakout and terminal sections as channels are clearly visible in the available images (Fig.

4.22k and Fig. 4.22l).

Also Alignment 12 has features that mark a lateral distance that can be useful to infer the size of the plume, these are: 1) an unnamed crater located west of Ogygis Regio from which several channels spread in direction south-west and north (Fig. 4.22b and Fig. 4.22g); 2) an unnamed crater from which many channels spread in almost all the directions and located southeast of Douglass (Fig. 4.22e and Fig. 4.22h); an unnamed massif of the same size of those observed along the Sisyphi Montes located northeast of the crater Phillips (Fig. 4.22a).

All the distances from the loxodrome were geodetically measured with ArcGIS and reported in Table 1.

4.4 Unaligned volcanic features

Other than the aligned volcanic features described so far there are several volcanic features that do not seem to belong to any alignment at hemispherical scale so they are listed here as unaligned. These volcanic features are indicated with an olivine yellow coloured dot on the lowlands and with a dark violet dot on the highlands in Fig. 4.2f; most of them are already reported in the literature. Among these are the crater Huygens, a olivine-bearing (and thus

152 lava) flooded crater located on a topographic high from which many channels spread radially

(Ackiss et al., 2014), and three massifs located a few hundreds of km west of Copernicus and devastated by impact craters; one of them is included in the Xiao et al (2012) list. On the lowlands there are: the Borealis Volcanic Field (Hodges and Moore, 1994), where volcanic units may have formed as the result of sub-glacial eruptions analogous to Icelandic volcanoes

(Fagan et al., 2010); the Hyperboreus Labyrinthus; the Hyperborei Cavi; the Scandia Cavi

(Tanaka et al., 2008) that may have formed after subsurface heating by analogy with the Cavi

Angusti and Angustus Labyrinthus seen at the end of Alignment 1 in the southern polar region; and the volcanic cones reported in a small region at the mouth of

(Garvin et al., 2000). Evidence for Noachian volcanic flooding was reported in Noachis Terra but related to the genesis of Hellas (Rogers and Nazarian, 2013).

4.5 Discussion

4.5.1 The alignments

Fig. 4.23a shows how the loxodromes fit all the alignments. Comparison of Fig. 4.2a and

Fig. 4.2b shows that many volcanic features are not included in the current literature, particularly in the southern hemisphere of Mars and have been identified here for the first time. The most important result coming from the observation of both figures combined together is the completion of a picture that shows all the alignments in their entirety so that the majority of the volcanic landforms in the southern hemisphere of Mars cannot be considered to be randomly distributed. There are evident and wide empty spaces outside the various alignments that rule out a random distribution of the volcanic features (Fig. 4.2d and

Fig. 4.2f). The volcanic centres often appear slightly shifted laterally on both sides of the fitting loxodrome running from the first to the last point of the alignment. The movies of the

153 Figure 4.23. A) Diagram showing the best fit of the alignments obtained using loxodromes (rhumb lines); the colour of the symbols is the same indicated in the ArcGIS map of Fig. 2c. B) 3D view of the loxodromes obtained through the best fit for the alignments. C) Southern polar view of the alignments converging to the South Pole of Mars.

154 migrating plumes in the SPGI model given by Leone et al. (2014) show these slight lateral fluctuations, although the causes are not well understood yet. About 15% of the total features appear shifted from the alignments with a lateral distance indicative of the size of the plume

(no cluster – lateral distance column of Table 1). These shifted positions are likely related to a lateral heterogeneity in crust thickness that forces magma to find lateral ways to reach the surface. The observation of the fields of vents in Fig. 4.14 has shown that migrating plumes leave a stripe behind them, which could be 150 – 600 km wide, from which the volcanoes emerge. The fact that only 15% of the total features appear shifted shows that: a) the rise of the magma in the crust occurs in a quite straightforward way after the passage of the plume and b) the lateral heterogeneities are not so diffuse. The loxodromes that best connect the alignments were also observed in 3D (Fig. 4.23b) and from a southern polar view (Fig. 4.23c).

A study of the crossing points was also done to resolve the issue of the assignment of some features to one alignment rather than another. The low-relief volcanic centres and the vents might be in ambiguous positions at the crossing points of the alignments but the largest volcanic centres are hardly in ambiguous positions and provide reliable geographic data points for the assessment of the alignments. Even situations of ambiguity within the crossing points can be resolved with a focused observation. For example, there is a small massif within the cluster of the Sisyphi Montes, indicated with a white arrow in Figs. 4.24a and 4.24b near the crossing point between Alignment 2 and Alignment 8, at a geodesic distance of 78.04 km from its nearest neighbour belonging to Alignment 2 and at 109.47 km from its nearest neighbour belonging to Alignment 8. The geodesic distance between two neighbour massifs both belonging to Alignment 2 in the same cluster is 235.59 km. Applying the tighter constraint of 78.04 km to increase the confidence in Alignment 2, one of these two neighbour massifs would be excluded. So, comparing the 235.59 km of the two massifs belonging to the same alignment with the 109.47 km of the massifs belonging to different alignments, it can be

155 easily seen that the massif indicated with the white arrow in Fig. 4.24 might belong both to

Alignment 2 and to Alignment 8 on the basis of the shortest distance.

Figure 4.24. A) Diagram showing the best fit of the alignments obtained using loxodromes (rhumb lines); the colour of the symbols is the same indicated in the ArcGIS map of Fig. 2c. B) 3D view of the loxodromes obtained through the best fit for the alignments. C) Southern polar view of the alignments converging to the South Pole of Mars.

156 As can also be seen from the regional view in Fig. 4.24a, all the other massifs have been easily assigned to their respective alignments even if near to the crossing point. Of course, the interpretation is much easier away from the crossing points where the lateral distance of the offset between volcanic centres is much greater. The Issedon Paterae are located between parallel ridges at 907.10 km from Alba Mons and 307.65 km from Ascuris Planum, the latter being an average distance within Alignment 3 (Table 1, no cluster – lateral distance column).

There is a general absence of large, high relief structures outside the maximum lateral distance measured for the alignments; this means unequivocally that the waning phase of the

Martian volcanism did not produce any large, high relief feature visible on a global scale.

4.5.2 The ages

The ages inferred from the crater counts showed a clear decreasing trend from the Equator to the South Pole along all of the alignments that do not cross the huge lava fields of Tharsis and Elysium (Table 2 and Fig. 4.25). In all the other cases, the young age, which I argue is too young, estimated for the Tharsis and Elysium lava fields clearly alters the trends.

Furthermore, the range of ages (4.0-3.67-2.54 Ga) estimated for Olympus Mons (Table 1), which is the starting point of three alignments, forced me to consider the age of 3.67 Ga as a reasonable compromise for the trend of all the alignments in which it is involved, but the initial emplacement could be much older than this (Isherwood et al., 2013). A range of ages

(3.7-3.0) was also found for another important centre, Elysium (Table 2). As will be shown later, these ages are not only too young for the first emplacement of volcanics at Olympus

Mons and Elysium but also come out from a compromise with their suggested but unrealistically long activity estimated to last until Amazonian time. As will be discussed later, there is also the possibility, ignored in the current literature, that abnormal eruption rates may

157 have built these volcanoes faster than normal at much earlier times, a possibility predicted by the SPGI.

Figure 4.25. Graph showing the ages and the relative trends within each alignment for all of the 12 alignments tracked on Mars; a clear descending trend of ages is visible only on those alignments that do not cross the lava fields of Tharsis and Elysium.

As a result, the overly young ages estimated for Olympus Mons and Elysium Mons changed the trends that otherwise would have been as predicted by the SPGI in all the

158 alignments. In fact, only Alignment 8, that has an old starting age (4.1-4.0 Ga) and is the only one among the primary alignments that does not cross the lava fields of Tharsis and Elysium, shows a clear trend of decreasing age with distance along track (Fig. 4.25). A clear decreasing trend of ages was particularly observed along the secondary Alignment 10, over all of the ages estimated for the formation of the valley networks from Terra Sabaea to Margaritifer

Sinus with a timespan of 210-260 Ma (Hoke and Hynek, 2009), but also along Alignments 4 and 5 (Fig. 4.25). It is important to notice how all these alignments do not cross the lava fields of Tharsis and Elysium.

Dividing the plume migration timespan by the number of volcanic centres along

Alignment 10, an emplacement rate of a volcanic centre every ~ 37 Ma is obtained. Applying this calculation to all the alignments gives an average emplacement rate of a volcano every

16-50 Ma with an average of ~ 33 Ma. Given that Alignment 10 has an hemispherical geodesic distance of ~ 5700 km, covered in 210-260 Ma, the plume advanced at an average velocity at the surface of 21-27 km Ma-1. Provided that at least the initial volcanic centres could be accurately dated with in situ K-Ar measurements, this calculation could be easily applied to the hemispherical geodesic distances of all the other alignments. One could then compare the calculated ages to the crater count ages available in Table 2 in order to obtain an improved map of the absolute ages. The fastest plume was the one migrating along Alignment

2 with a maximum velocity at the surface of 118 km Ma-1. The differences in ages along other alignments range from about 100 to 300 Ma between the Equator and the South Pole with an average of 200 Ma (Table 2). This timespan is consistent with the average ~ 200 Ma predicted for the migration of the plumes in the SPGI model for an impactor of 1600 km radius with

80% of the radius consisting of iron (Leone et al., 2014). Although the crater counts ages are consistent with the SPGI ages as order of magnitude, there is a mismatch in absolute ages.

According to the SPGI model the phase of the plume migration started about 50 Ma after the giant impact (about 4.45-4.35 Ga ago) and lasted for 500 Ma with an additional 500 Ma of

159 widespread (no plume migration) volcanism of minor intensity (Leone et al., 2014), a total timespan that is consistent with other studies (Baratoux et al., 2013; Morschhauser et al.,

2011; Spohn et al., 2001). The phase of plume migration is also consistent with the general lack of volcanic centres of large dimensions observed outside the alignments. According to the crater counts collected in the literature, the volcanism lasted on average about 500-600 Ma starting from the Noachian (4.1-4.0 Ga ago) and ending in the Hesperian (3.5 Ga ago) with the exception of some Amazonian lava flows and a few volcanoes mostly located on Tharsis and Elysium (3.0 – 0.2 Ga ago). Despite the uncertainties related to the crater counts (Robbins et al., 2013), which will be mentioned shortly, there is a substantial agreement around the

Noachian – Hesperian period in the literature (Table 2). Despite the jump back in time between Arsia Mons and Sirenum Tholus, just outside the lava fields of Daedalia Planum, the ages maintain their decreasing trend both within and outside Tharsis. An example of this trend can be observed even within the lava fields of Tharsis through the descending oldest ages from Uranius Patera to Ascraeus Mons. The SPGI model, which also considers all the possible thermal contributions since the accretion of the planet, suggests that the heat flux of the planet was high early in its history supporting volcanism only from 4.5 to 3.5 Ga (Leone et al., 2014), with the strongest phase in the first 500 Ma, and is substantially in agreement with other thermal models (e.g., Grott et al., 2005; Hauck and Phillips, 2002). The majority of the volcanic provinces are estimated from Noachian to Hesperian age, thus perfectly within this time interval, except for some regions around the South Pole and the lava fields of

Tharsis and Elysium Rises that are indicated as Amazonian (Tanaka et al., 2014). Other authors have even suggested that the lava flows on Daedalia Planum might be as young as

175 Ma as inferred from the SNC meteorites (Lang et al., 2009). How to explain such a discrepancy?

The young age inferred for the Amazonian geological units is mainly the effect of the final abundant lava flooding that buried large craters and thus masked the original time of

160 emplacement of Tharsis and Elysium. If this final flooding happened just after the end of the

Late Heavy Bombardment (LHB), which is more consistent with the thermal models, of course large craters on the lava fields would not be expected. Furthermore, there are other important considerations: (a) there are admitted discrepancies so large between the spectra of the Tharsis lava flows and those of the shergottites that any connection between them is tenuous if not impossible (Lang et al., 2009); (b) the reduced flux of impactors after the Late

Heavy Bombardment (4.1-3.9 Ga ago) could hardly produce large basins (Bottke et al., 2007) and can thus explain the apparently young lava flows of Tharsis; c) the (putative) mudflows of the Cumberland sample, a basaltic rock containing olivine studied by the Curiosity rover within the Gale crater (Grotzinger et al., 2014), were dated at 4.21 Ga through more accurate

K-Ar concentrations with respect to a younger (3.7-3.5 Ga) formation age of Gale crater estimated through the less accurate crater counts (Farley et al., 2014); this shows how all the other known ages might be pushed back in time as well; d) the ages of the geo-chrono- stratigraphic map of Mars (Tanaka et al., 2014), for example, rely on a crater database that has high uncertainties regarding the quasi-circular depressions (Robbins and Hynek, 2012). Thus, regarding all the craters that are buried under lava flows or other sedimentary material, several examples of these craters have been found both on Elysium Planitia (Fig. 4.2d) and

Daedalia Planum (Fig. 4.2b), suggesting that others can be completely buried and thus invisible in the images and in the counts; e) last but not least, the SPGI model revealed that some plumes remained stationary for about 40 Ma with a peak in eruption rate under their origin points before starting their migration, a process that may explain why Olympus Mons,

Alba Mons, and Elysium Mons are so large compared with all the other volcanoes following them along the alignments. These various lines of evidence raise the issue that the latest lava flows might not be as young as shown in the geo-chrono-stratigraphic map of Mars made by

Tanaka et al. (2014), pushing the clock back in time along the whole stratigraphic column.

Unfortunately, although the crater counts may seem a useful and supportive method to

161 estimate the absolute ages, they are still a very uncertain method and should be replaced by in situ K-Ar measurements as already done at Gale crater. There is objectively no way to count the craters buried under the huge pile of lava erupted by Tharsis and Elysium. However, in many places the crater counts were helpful to understand that a decreasing trend of ages exists from the equatorial to the southern polar regions of Mars. A volcano can hardly have a long period of activity from probably the Early Noachian to Late Amazonian if not supported by adequate heat. Such a long period of activity is neither supported by the heat flux of the SPGI model nor by the crustal thickness estimated for Mars (Baratoux et al., 2011; Leone et al.,

2014). Even the magma produced by a still internally-hot Mars, which might not be the case, can hardly overcome the thick crust already built in the first stages of evolution of the planet.

4.6 Conclusions

The observation of the whole map of Mars has shown twelve alignments of volcanic features developing in the southern hemisphere of Mars for several thousands of kilometres.

The radial directions formed by the smaller features outside Tharsis and the largest features inside Tharsis provided clear landmarks to track the alignments within the Tharsis cluster. All the other alignments starting from Elysium, West Arabia Terra and Syrtis Major were identified from the symmetrical distribution of the volcanic features following northwest – southeast or northeast – southwest directions. All the alignments were verified through loxodromes (rhumb lines) in different projections. A test on different directions than those identified revealed a poorer fit both for the minor number of the features per alignment and for the much higher along-track distance among the volcanic features. The general lack of large high relief volcanic features comparable to those observed at the Equator and in the southern hemisphere, both away from the alignments and in the northern hemisphere, rules out a random formation of the volcanic landforms. These alignments are likely the result of

162 several plumes originated within the latitudinal stripe 40°N-10°N and then migrated to the south pole of Mars after the SPGI that formed the Martian dichotomy. No similar migration was observed towards the northern hemisphere. The SPGI model showed that almost all the plumes originated from four broader points located 90 degrees of longitude apart, which is more or less the same angular distance separating Tharsis, Elysium, Syrtis Major, and West

Arabia Terra. It is worth to notice that eleven out of twelve migrating plumes originated exactly from these locations. The peak eruption rate generated by the SPGI in some plumes formed the largest volcanoes found at the beginning of the alignments. The same peak eruption did not happen in all the alignments as also observed in the SPGI model. The largest volcanic province of the solar system, Tharsis, was formed by the contribution of six different migrating plumes that fed the magma reservoirs of the volcanic features and left trails of vents and low shields that cross each other around the largest volcanoes. There is a proportional extent of the volcanic provinces to the number of plumes originated or, in the case of

Tyrrhenum Mons, to the strength and the size of the plumes that formed the volcanic features.

It could be objected that Pavonis Mons is too small with respect to the number of the plumes that crossed it but there is also a significant drainage of lava that formed Kasei Valles and

Valles Marineris, the largest outflow channels of Mars, from its eastern flank. Why a higher number of plumes originated on Tharsis compared to other provinces is not understood yet.

After the strong initial peak of eruption rate the plumes lost their strength during the migration and this explains the decreasing size of the volcanic features observed along the alignments towards the South Pole. All the migrations occurred in different times within the first 500 Ma of the Martian history. The migration of the plumes did not occur at the same velocity everywhere, some plumes completed their migration in ~ 100 Ma and others in ~ 300

Ma with an average of ~ 200 Ma, the fastest plume was the one that formed the Alignment 2.

The aligned volcanic features seen only in the southern hemisphere with the largest volcanoes generally seen at the start of the alignments confirm that an asymmetrical process, a giant

163 impact favoured over degree-1 mantle convection, must have taken place early in the history of the planet. This process was transient and its end coincided with the shut down of the magnetic field of Mars. The phase of migration of the plumes was actually followed by a decline of the volcanism that has not generated any other large feature on Mars, this is also confirmed by the lack of features far away from the alignments. A long-term radiogenic heating source supporting a prolonged volcanism is not consistent with the lack of activity observed on Mars today (Leone et al., 2014) and it is not consistent with the disproportionate size of the oldest and largest volcanic features compared to the size of the planet. The largest volcanic features located in a limited stripe of latitude of the planet, the twelve alignments in the southern hemisphere of Mars, the few unaligned features in the north, and the lack of active volcanism observed today, all suggest an exogenic event much stronger and more focused on the southern hemisphere with respect of a random distribution of radiogenic elements such as the one observed on Earth today for example. Such a disproportion between the largest volcanoes on Mars and the largest volcanoes on Earth, compared to the respective sizes of the planets, requires an anomalous heat source for Mars that can only be given by a giant impact like the SPGI. A prolonged volcanism until the Amazonian age is very unlikely and it is neither supported by the SPGI nor by any other thermal model of Mars, not only for the insufficient production of heat after that the effect of the SPGI vanished but also for the thickness of the (early formed) crust that would not allow magma rise even in a still hot Mars today. I note here that the level of heat flux inferred by several studies of lithospheric flexure today (Ruiz et al., 2010) is perfectly consistent with that reached by the planet at the end of the migration of the plumes (Leone et al., 2014).

The analysis of the ages obtained by the crater counts, compared to those of the SPGI model and to the available K-Ar measurements directly on Mars, led to some important conclusions: a) there is a decreasing trend of ages in volcanism evident only in the alignments not crossing the lava fields of Tharsis and Elysium, consistent with the migration of mantle

164 plumes in space and in time, although there is a mismatch in absolute ages; b) the older ages of some smaller volcanic features located along the alignments of Tharsis excluded that they could be secondary cones, this is the case of Tharsis, Uranius, Jovis and Ceraunius Tholus,

Biblis and Ulysses Patera, and raises doubts on the much younger age of their larger and perfectly aligned neighbours (Ascraeus, Pavonis, and Arsia Montes); c) the heads of many fluvial valley networks are aligned with other volcanic centres and their ages of formation follow a decreasing trend along the alignments so they are likely the result of the passage of the same plumes; d) the decrease in impactors’ flux after the LHB altered the geologic clock on which the crater counts are based; e) last but not least, the average velocity of migration of the plumes and the average emplacement rate of the volcanic centres favours a SPGI model with a sideritic impactor of 1600 km of radius rather than a meso-sideritic impactor of 2000 km of radius. The ascending trends of ages observed along the alignments crossing the lava fields of Tharsis and Elysium are obviously due to the intrinsic uncertainties related to the crater counts that made those lava fields much younger than really they are. The discrepancy between the age measured through K-Ar concentrations and the age estimated through crater counts at Gale crater supports this hypothesis. The ages inferred from the SPGI model could provide a useful reference timeline that can be compared with the existing ages inferred from crater counts available in the literature or with future radiometric and exposure age dating methods for calibration.

In conclusion, the discovery and the observation of the volcanic alignments, other than being a successful test for the SPGI model, provided a new and alternative view for the history of Mars.

Acknowledgements: The author was supported by the ETH Research Commission Grant ETH-03 10-1. Discussions with Luca dal Zilio and Stijn Vantieghem were helpful in the initial development of the Matlab code used for the study of the alignments. Many thanks to Joseph Doetsch who completed and greatly improved the Matlab code that made the figures of the important assessment of the alignments. The author also thanks the editor Lionel Wilson, Shijie Zhong and five other anonymous reviewers for the comments that stimulated

165 the improvement of this paper. The copyright for the CTX images is acknowledged to the Malin Space Science Systems, CRISM to Johns Hopkins University – Applied Physics Laboratory, HiRISE to University of , other copyrights are indicated directly on the images where applicable.

166 Chapter 5

A network of lava tubes as the origin of Labyrinthus Noctis and Valles Marineris on Mars.

167 A network of lava tubes as the origin of Labyrinthus Noctis and Valles Marineris on

Mars.

Giovanni Leone

ETH Zurich (Department of Earth Sciences, NO H28, Sonneggstrasse 5, 8092 Zurich, Switzerland; e-mail: [email protected]).

Published in the Journal of Volcanology and Geothermal Research, 2014, vol. 277, 1-8.3

Abstract

The role of lava tube networks and lava channels is reassessed as the primordial stage of the volcano-erosional processes that formed the Labyrinthus Noctis-Valles Marineris system instead of a tectonic origin. The combined use of CTX, CRISM, HiRISE imagery, and

MOLA profiles has provided valuable insight in the evolution of pit chains into fossae first and then chasmata later due to mass wasting processes caused by the erosional effect of the lava flows that draped Valles Marineris and other outflow channels. Although a quantitative evaluation of eruption rates is difficult even with digital terrain models (DTMs) because of the mixing between new flows and paleoflows, a comparison with Elysium and other Tharsis outflow channels suggests that the availability of lava supply is correlated to their widths. The images of ubiquitous lava flows rather than sporadic light-toned deposits strengthen the role of lava over that of water in the erosional processes that formed Labyrinthus Noctis and carved Valles Marineris like many other outflow channel on Mars. The erosional evolution of the outflow channels shows an increasing trend of age and a decreasing trend of depth from the sources on Tharsis to the mouths at Chryse Planitia. This finding, coupled with the observation of lava flows mantling Chryse Planitia, may have profound implications for the water inventories thought to have filled the lowlands with an ocean.

3 The supplementary images described in the text are available at the following website: http://www.sciencedirect.com/science/article/pii/S0377027314000316

168

5.1 Introduction

The largest outflow channels are characterized by deep chasmata, elongated, steep-sided depressions typically seen on the largest volcanic provinces of Mars, Tharsis and Elysium, and particularly in Labyrinthus Noctis and Valles Marineris. Similarities to terrestrial rift zones have been suggested for the origin of Labyrinthus Noctis (Bistacchi et al., 2004;

Masson, 1977) and Valles Marineris on Mars (Mege and Masson, 1996a; Schultz, 1991;

Wilkins and Schultz, 2003); also combined volcano-tectonic processes (Dohm et al., 2009;

Lucchitta, 1992) and even rotation of Thaumasia (Anguita et al., 2001), but hypotheses of formation through lava tubes and their possible evolution to pit chains, fossae, and then chasmata have received less or no consideration. Although pit chains and lava tubes have already been observed mainly on the large volcanic structures of Tharsis (Leveille and Datta,

2010), they were not proposed as a likely source feature for the formation of Labyrinthus

Noctis and Valles Marineris with the exception of a first observation of (Mccauley et al., 1972) and Viking (Carr et al., 1977) images, a suggestive idea elaborated later (Tanaka,

1997) but never developed further. Tectonic processes like rift zones generated by lithospheric movements are difficult to explain on a planet characterized by a stagnant lid since its early stages of formation (O'rourke and Korenaga, 2012); the map of Mars neither shows subduction zones nor tectonic plates but only a marked dichotomy between the northern lowlands and the southern highlands. A way to solve this problem has been tried with the interpretation of tectonic lineaments as the surface manifestation of dike swarms

(Ernst et al., 2001; Scott et al., 2002; Wilson and Head, 2002). Although tectonism (i.e. faulting) related to magma movements and presence of dike swarms could be plausible in volcanic activity, erosional processes related to lava tube roof collapse first and circulation of fluids afterwards (mainly lava likely followed by either atmospheric and/or groundwater) better explain the evolution of pit chains into chasmata, particularly in the Labyrinthus

169 Noctis-Valles Marineris system. One visually immediate reason is that neither lateral lava flows nor traces of lava fountaining are observed along both chasmata and fossae as result of dikes rupturing and erupting to the surface, lava flows coming out from shallow dikes should have covered previous lineaments and consequent superposition relationships would be seen otherwise. Another problem with faulting is the absence of significant vertical displacements in the MOLA profiles along the sides of both V-shaped fossae and (flat floored) trapezoidal shaped chasmata in Valles Marineris; the images in this paper will show abundant erosive and mass wasting processes instead. Furthermore, several problems with water have already been evidenced in the literature (mostly Leverington, 2011 and references therein) and will also be briefly summarized in this paper. If chasmata were related only to water flowing on the surface, they should also be seen in other non-volcanic areas of Mars. Despite the presence of various volcanic features in the southern highlands and in the southern polar area (i.e. Ghatan and Head, 2002), pit chains have never been reported, although they might exist where intrusive volcanic activity is observed. Nevertheless, the abundance of pit chains on Tharsis and Elysium seems related to the particularly strong volcanic activity observed on these main

Rises. So more attention should have been placed on subterranean circulation of magma within Valles Marineris, as evidenced by previous studies of (Dohm et al.,

2009) and by the erosive effect of lava flowing in lava tubes (Greeley et al., 1998). A problem with flood excavation in Valles Marineris was raised by et al. (2009) because of the absence of connection between with an outflow channel, although they acknowledged that a pure tectonic origin is difficult to reconcile with its morphology, and suggested a karstic removal of mixtures of salts and water ice. Aside that these mixtures would need a way out as well as any other fluid, they suggested a collapse process of formation for chasmata after mixture removal but ignored the possibility of underground connection through lava tubes, although the circulation of subsurface magma was already a well-known process, see Greeley et al. (1998) for example. The presence of opaline and Fe-

170 sulfate mineralogy in several Valles Marineris chasmata confirms that alteration activity in an aqueous environment took place on already emplaced basaltic materials (Weitz et al., 2010); however, the available images will show that this is not a widespread process as it should be expected by the amount of water needed to carve the outflow channels. Indeed, light-toned deposit (LTD) outcrops occur occasionally both on isolated topographic highs and lows, in gently dipping layers postdating chasm formation (Grindrod et al., 2012) or superposed on landslides, making the understanding of the depositional process quite difficult.

The combined HiRISE, CTX, CRISM imagery, and MOLA profiles show that:

a) There are curvilinear pit chains in Labyrinthus Noctis not associated to faults (Fig.

5.1) and even perpendicular to them. Pit chains have been genetically associated to

linear faults (Wyrick et al., 2004) but they are mainly formed by the collapse of lava

tubes, as also evidenced by studies on Earth and Mars (Cushing et al., 2007;

Keszthelyi et al., 2008; Leveille and Datta, 2010).

Figure 5.1. The CTX image P21_009184_1631_XN_16S097W, centred at 16.94 S – 97.06 W and taken at the extreme southern end of Labyrinthus Noctis in southeastern Syria Planum, shows a curvilinear pit chain that is not associated to faults. The white inset, the HiRISE image ESP_026036_1630, provides insight into one of the pits.

b) There are tube-fed flow fields and skylights between Arsia and Pavonis Montes (Fig.

5.2 and Fig. 5.3), exactly uphill along the direction of the main Labyrinthus Noctis-

171 Valles Marineris axis, an observation relevant to the next point and to the diagrams

that will be shown in Fig. 5.4.

Figure 5.2. The CTX image P20_008934_1756_XI_04S112W, centred at 4.52 S – 112.48 W, shows a combination of open lava channel and enclosed lava tube along a flow field located southeast of Pavonis Mons suggesting both surface and subsurface circulation of lava. The white inset, the HiRISE image PSP_009501_1755, provides insight of the lava tube in direction southeast towards Labyrinthus Noctis.

Figure 5.3. The CTX image B12_014380_1778_XN_02S118W, centred at 2.22 S – 118.78 W and taken between Arsia and Pavonis Montes, shows a lava tube departing laterally in correspondence of a lava channel enlargement. Pit chains can be observed along the tube due to the collapse of the roof following lava withdrawal (white arrow). The white inset, the HiRISE image ESP_014380_1775, shows a couple of skylights over the lava tube.

172

c) Incipient chasmata of western Valles Marineris are directly connected to the eastern

end of Labyrinthus Noctis and, as well as Athabasca Valles (Jaeger et al., 2007) and

other outflow channels that will be mentioned shortly, are lava-draped. This

connection suggests that Valles Marineris is genetically associated and would have

advanced through Labyrinthus Noctis, maybe widening its channels into chasmata as

far as Arsia and Pavonis Montes, if erosional processes would have continued (Fig.

5.4).

Figure 5.4. Diagram showing the stages of formation of Valles Marineris and the outflow channels. The first stage is characterized by the erosion of the floor due to turbulent lava flowing in the tube and then collapse of the roof after the flow ceased. The second stage is characterized by new eruptions in which lava flows are channelized in the collapsed tubes, drape their floors and erode their walls causing landslides. The third stage is

173 characterized by new flows so massive to enlarge the channels creating ridges and mesas in the paleoflows, this is the formation of the chaos terrains. When the volcanism fades away, the last episodes of lava flows carve narrow channels in the previous lava fields.

d) The nearby Kasei Valles is also lava draped and shows bifurcated pit chains that

suggest lava circulation underground (Fig. 5.5). The association between pit chains

and sinuous lava tubes is also evident at Uranius Dorsum (Fig. 5.6), a chain of mounds

and cratered cones (Chapman et al., 2010) formed on lava floods so massive to even

fill the Kasei lateral channels (Fig. 5.7 and Fig. 5.8) and to make the floor of Kasei

Valles wider than that of Valles Marineris. Lava flows are often seen filling Kasei

Valles and Valles Marineris, they embay and erode knobs and mesas of mensae-type

in Labeatis Mensae (Fig. 5.7 and Fig. 5.8) and chaos-type terrains in Aromatum

Chaos, , , and Aram Chaos (Fig. 5.15 ahead) as far as

Chryse Planitia (Fig. 5.12).

Figure 5.5. The CTX image P19_008524_1858_XN_05N078W, taken on upper and centred at 5.85 N – 78.58 W, shows a bifurcation of pit chains on the remnants of mesas embayed by lava flooding. The white inset, obtained by a blow up of the same image, shows how the lava flows reach the top of the mesa and fill pre-existing channels (indicated by the white arrows).

174

Figure 5.6. The CTX image B02_010238_2031_XN_23N078W, taken along Uranius Dorsum and centred at 23.15 N – 78.25 W, shows a portion of its chain of mounds and cones in association with a short curvilinear pit chain (white arrow). The white inset, the HiRISE image ESP_025534_2020, shows a blow up of the Uranius Dorsum.

Figure 5.7. The CTX image B01_010040_2091_XN_29N073W, centred at 29.13 N – 73.52 W and taken between Labeatis Fossae and Labeatis Mensae, shows knobs surrounded by lava fields and channels filled by lava flows. The white arrows indicate one of the latest flow fronts while the black arrows indicate the direction of the lava entering and filling the channels. Unfortunately, no HiRISE and CRISM coverage is available in the field of the CTX image.

175

Figure 5.8. The CTX image B01_010185_2099_XN_29N071W, centred at 30 N – 71.75 W and taken in Labeatis Mensa where the Uranius Dorsum ends, shows a knob surrounded by lava fields and channels filled by lava flows that are connected uphill. The smaller white inset, the HiRISE image PSP_010185_2100, shows a knob surrounded by lava fields. The white arrows indicate the moat surrounding the knob. This knob, along with many others scattered all over Labeatis Mensa, indicates where the mesa was before being submerged and/or eroded away by massive lava flooding. The larger white inset, a blow up of the same CTX image, shows that the lava flooding was likely so massive to fill the channels uphill and nearly connect with other channels carved by lava coming from the Labeatis Fossae on the west. The white arrows indicate the flow front and the black arrow the direction of the lava entering the channel.

Figure 5.9. The HiRISE image PSP_007725_2075, centred at 27.11 N – 136.24 E, taken on a breakout from the main channel of Granicus Valles, shows superposition relationships among flat lava paleobedfloors indicated by the white arrows, the CTX image P03_002398_2071_XN_27N223W shows several other breakouts (again indicated by the white arrows) along the course of the Granicus Valles.

176

e) Maja and Shalbatana Valles, two smaller and sinuous outflow channels parallel to

Kasei Valles, are also lava draped (Kereszturi, 2010; Leverington, 2011) (See also Fig.

5.13). These two outflow channels decrease their size as they get farther from Tharsis

and do not have the typical dendritic morphology of fluvial valleys with tributaries but

rather show breakouts (Fig. 5.9) commonly seen along lava channels (i.e. along Kasei

and Granicus Valles).

f) The slopes of Elysium also show lava draped incipient chasmata and sinuous outflow

channels similar to those visible on the slopes of Tharsis. Breakouts and landslides are

present along the linear Cerberus Fossae (Fig. 5.10 and Fig. 5.11) and the Valles

Marineris (Fig. 5.14) walls. The position of the boulders on the floor of Granicus

Vallis does not suggest fluvial transport but rather mass wasting. Even the presence of

boulders at the Pathfinder landing site, thought to be indicative of water flooding

transport, now appears less obvious (Hobbs et al., 2011).

Figure 5.10. The CTX image P02_001896_1961_XN_16N197W, taken at Grjota Vallis and centred at 16.21 N – 197.19 W, shows a particular of the point where one of the fossae characterizing the system of the Cerberus Fossae passes through the Tartarus Montes. The right white inset, the HiRISE image ESP_018075_1955, provides insight of the channel passing through the mesa. The left white inset, the HiRISE image PSP_006287_1955, shows an incipient breakout on the west.

177

Figure 5.11. The CTX image P18_008212_1885_XN_08N198W, centred at 8.6 N – 198.6 W and taken where one of the Cerberus Fossae passes amid hills and mounds of likely volcanic origin, shows a breakout along the main channel characterized by landslides and mass wasting. The white inset, the HiRISE image ESP_032791_1895, shows collapse faults on the landslide crown and a block nearly falling downhill (black arrow). The left tip of the breakout ends with a pit chain.

Figure 5.12. The CTX image G05_020232_2052_XN_25N046W, taken on Chryse Planitia and centred at 25.29 N – 46.65 W, shows lava flows at about 300 km of distance from the mouth of Kasei Valles. The white inset, the HiRISE image ESP_029568_2060, shows a closer look of the flow front indicated by the white arrows.

178

Figure 5.13. The CTX image P15_006783_1911_XN_11N059W, taken on the median course of and centred at 11.19 N – 59.41 W, shows lava flow fronts indicated by the white arrows. The white inset, the HiRISE image ESP_024373_1895, shows a closer look into the lava flows embaying the mesas and enlarging a channel carved through them by previous flows.

Figure 5.14. The CTX image G10_022091_1772_XI_02S035W, centred at 2.86 S – 35.32 W and taken on the main channel of the Valles Marineris between Aurorae and Hydraotes Chaos, shows a series of landslides on the left side of the chasm. Although no CRISM observation is available here, the lava flow material is exposed by the black ejecta of a nearby impact crater on the chasm floor. The white inset, the HiRISE image PSP_003578_1770, provides a particular of the scarp, carved by the passage of lava, between two landslides.

179

Figure 5.15. The CTX image B05_011726_1825_XI_02N019W, taken on a supposed fluvial valley discharging from Aram Chaos to Ares Vallis and centred at 2.57 N – 19.13 W, shows a lava channel instead. The white inset, the HiRISE image ESP_011304_1825, shows the classic texture of lava flooding on the floor of the channel followed by a thin, not ubiquitous, veil of LTDs left by the subsequent passage of water on emplaced lava.

g) The MOLA profiles show that the depth of Valles Marineris and the other outflow

channels reaches its minimum at the mouth near Chryse Planitia because of the

maximum stratification of the lava flows channelized there. It is beyond the scope of

this paper to accurately quantify the volumetric percentages of lava erupted, which

would be difficult even with digital terrain models (DTMs) obtained from high-

resolution images. The available sample imagery will show that the lava flows caused

landslides (Fig. 5.16) and covered (Fig. 5.17) their material perhaps re-distributing it

all along the course of the outflow channels as far as the floor of Chryse Planitia. So

the total volume of the infill must take into account the volumetric sum of both mixed

contributions, which now would be difficult to separate.

h) The geochronostratigraphic map of Mars (Tanaka et al., 2013) confirms that the upper

course of Valles Marineris and the other outflow channels is younger than its lower

course. Together with the observation of point c it may also suggest the hypothesis

that the outflow channels may have been developing and advancing from the mouth to

180 the , both in terms of depositional and erosional history. This mechanism of

formation does not require lava travelling all the way through the channel, although it

is widely known from the observations of other planetary bodies that lava flows can

travel in channels for long distances even on cold surfaces.

i) Last but not least, only occasional presence of LTD is observed (Fig. 5.19, Fig. 5.20,

Fig. 5.21, and Fig. 5.22) for the amount of water expected for the formation of the

outflow channels.

All these points will be crucial to understand the various stages of outflow channel evolution that will be discussed later.

5.2 An overview of processes for the formation of the pit chains

Pit chains are quite common on both Tharsis and Elysium Rises and several mechanisms have been proposed for their formation: magma-volatile interaction (Mege and Masson,

1996b), dike intrusions (Scott and Wilson, 2002; Wilson and Head, 2002), karst dissolution

(Spencer and Fanale, 1990), dilational faulting (Wyrick et al., 2004), and collapse of lava tubes (Cushing et al., 2007; Okubo and Martel, 1998). Which one of these mechanisms is the most plausible for the formation of the pit chains is not reassessed in this work because the various proponents have more or less documented their arguments with several examples.

However, the various mechanisms proposed are analyzed here and compared to the observations in order to assess which one is the most plausible for the formation of

Labyrinthus Noctis and Valles Marineris.

181

Figure 5.16. The CTX image P08_004159_1652_XI_14S056W, centred at 14.98 S – 56.80 W, provides insight on the mass wasting processes that have widened the walls of . Landslides cover the chasm floor. The white arrow indicates what remains of the central ridge from where a landslide covering the chasm floor originated. The black rectangle is the area imaged in the CRISM observation 0000B30B.

182

Figure 5.17. The CTX image G18_025362_1661_XN_13S058W, centred at 13.98 S – 58.11 W, shows an area of the Coprates Chasma where landslides and debris aprons falling down from the chasm wall have been covered by lava flows. The white arrows in the white inset indicate likely traces on the floor of the material fallen from the chasm wall. The black arrows in the black insets indicate points where the lava flows erode the base of the mesas and the wall of the chasm thus causing the landslides.

5.2.1 Magma-volatile interaction

The interaction of magma with (i.e. water) produces explosive phreatomagmatic eruptions leaving on the ground specific features like rootless cones, formed by the passage of lava on volatile-rich ground, cinder cones or maars (Mege and Masson,

1996b) in a disordered pattern (Scott and Wilson, 2002). Rootless cones have been observed in the Valles Marineris but not in association with pit chains. Another product of magma volatile interaction is thought to be chaos terrains (Chapman and Tanaka, 2002; Head and

Wilson, 2007; Meresse et al., 2008); an example is the formation of (Leask

183 et al., 2006) and other chaos terrains, that start to appear only in the lower course of Valles

Marineris and Kasei Valles making other questions arise. Why are the chaos terrains almost

Figure 5.18. The CTX image P16_007258_1686_XN_11S066W, centred at 11.50 S – 66.10W and taken in west Coprates Chasma, shows “windows” (indicated by the white arrows) in the thickness of the lava flows that draped the chasm floor. The white line indicates an alignment of mounds, the relics of an ancient ridge that further divided Coprates Chasma in different catenae and that has been nearly eroded away by the lava flows.

184 confined within the lower course of the largest outflow channels? If the mechanism of formation of chaos terrains is related only to magma-cryosphere interaction, why are lava flows seen carving and enlarging the channels between the knobs and the mesas both embaying and eroding their bases? Why do CRISM, CTX and HiRISE images not show widespread traces of LTD, generally associated to water activity, but only widespread activity of lava flows? And why are chaos terrains not observed in other volcanic provinces located at higher southern latitudes supposed to be ice-rich? The Sisyphi Montes for example, a chain of aligned mountains supposed to be formed by magma-ice interaction (Ghatan and Head, 2002), show neither pit chains nor chaos terrains nearby. Is it possible that dikes and sills reached the correct depth to form the mountains but not to form pit chains and chaos terrains? So, magma- volatile interaction seems unlikely for the formation of pit chains in Labyrinthus Noctis and

Valles Marineris.

5.2.2 Dike intrusions

Pit chains have also been explained as the product of collapse into voids created by leakage of volatiles from the top of a dike (Scott and Wilson, 2002). According to this mechanism of formation, pit chains, grabens and troughs (fossae) are all part of the same dike emplacement. The absence of associated lava flows has been justified by the low atmospheric pressure that would simply favour pyroclastic activity with dispersal of the products at varying distance depending on the size of the clasts (Wilson and Head, 1994). However, there are several facts that must be taken into account: a) several pit chains head into fissures perpendicularly, bifurcate in the absence of fractures (Fig. 5.5), and even cross each other perpendicularly while the model of formation suggests a general directional parallelism with the grabens, the faults and the troughs (Scott and Wilson, 2002); b) the atmospheric pressure on Mars has been almost the same throughout the history of Mars (Gillmann et al., 2011); c) lava flows are seen everywhere but not coming out from pit chains and no trace of pyroclastic

185 deposits has ever been found nearby. Although this mechanism does not seem plausible for the curvilinear pit chains seen in Labyrinthus Noctis and the bifurcate pit chains in Kasei

Valles, it cannot be completely excluded for the formation of some straight pit chains and fossae along Alba Mons (i.e. Acheron Catena).

Figure 5.19. The CTX image P01_001588_1699_XN_10S069W, centred at 10.12 S – 69.76 W and taken in eastern Melas Chasma, shows a portion of the chasm floor where a landslide is superposed on stratifications of lava flows. An isolated mesa is all that remains of the dividing ridge, traces of which are also visible in Coprates Chasma. LTDs, likely the alteration product of basalt in an aqueous environment (Bibring et al., 2006; Weitz et al., 2010), are also visible in part of the chasm floor.

186

Figure 5.20. The CTX image P15_006744_1699_XN_10S072W, centred at 10.11 S – 72 W and taken in central Melas Chasma near the border with , shows a landslide superposing on lava flows draping the chasm floor. The white arrow indicates that the LTDs are superposed on the landslide suggesting that the alteration processes of the igneous material in an aqueous environment occurred after the mass wasting processes.

187

Figure 5.21. The CTX image B19_016884_1713_XN_08S077W, centred at 8.81 S – 77.27 W and taken on eastern , shows another landslide covering the chasm and thus superposed on the lava flows draping its floor. LTDs are superposed on the landslide as well as already seen in Melas Chasma suggesting that their deposition is subsequent to the mass wasting processes.

5.2.3 Karst dissolution

This mechanism/model is based on collapse due to removal of ice or salts present underground and requires that ice bodies must be aligned with the main trend of the outflow channel (Spencer and Fanale, 1990), a similar mechanism suggested by Adams et al. (2009).

These authors concur that there are several problems with a tectonic origin because of the

188 morphologically uniform erosional style of the canyon walls. An active hydrological cycle is required to recycle the water necessary to remove the salts but there are many problems related to this issue that will be discussed later. Another problem is finding a plausible explanation for the availability of stronger acids, sulfuric, hydrochloric, and carbonic necessary to dissolve the carbonates confined in the chasmata only, particularly difficult when the proposed origin for the acids is by solution of volcanic gases in rainwater (Spencer and

Fanale, 1990), but there is a more serious problem related to the widespread lack of carbonates (Wyrick et al., 2004) despite the orbital detection of carbonate bearing rocks on

Nili Fossae (Ehlmann et al., 2008). For these reasons, this mechanism cannot be taken in consideration.

Figure 5.22. The CTX image G16_024427_1740_XN_06S092W, centred at 6.05 S – 92.76 W and taken on the volcanic plain between eastern Labyrinthus Noctis and the Tithonium-Ius Chasmata, shows again LTDs superposed on landslide material. The white inset, the HiRISE image ESP_024427_1740 on the left, shows that these deposits are found mainly in localized ponds. The white arrows indicate nearby ponds where deposits are not present.

189 5.2.4 Dilational faulting

Another mechanism has been claimed as the most plausible if compared with the others described so far, it is the presence of pit chains associated with faults explained in terms of dilational faulting (Wyrick et al., 2004). According to this mechanism, the pit chains form as a consequence of low tensile stress in interlayered ashes and lava flows. The dilational segments then produce the pits with a typical pattern parallel to the faults as reported in many examples from Alba Mons to Valles Marineris (Wyrick et al., 2004). However, this mechanism of formation does not explain: a) the pit chains in absence of faults and b) the perpendicularity to faults and/or multidirectional pattern of the pit chains. So this mechanism cannot be a viable one for the formation of Labyrinthus Noctis and Valles Marineris.

5.2.5 Lava tubes

The most plausible mechanism of formation for the pit chains seems to be the collapse of lava tubes (Cushing et al., 2007; Keszthelyi et al., 2008; Leveille and Datta, 2010), often seen in association with skylights branching perpendicularly from lava channels in the Tharsis area

(Fig. 6.3), despite arguments against the propagation of magma perpendicularly to the topographic slope (Wyrick et al., 2004). Pit chains are observed intersecting existing faults perpendicularly while lava tubes with skylights branch perpendicularly from lava channels aided both by the combined gravity and topography where slope breaks or bottlenecks slow down or stop the lava flow. See also a diagram describing this process in Fig. 12 of the work by Greeley (1971). If the eruption rate is well sustained, the lava flooding erodes the lateral channels evolving them into the large-scale perpendicular enlargements and lateral chasmata of the Valles Marineris (like the axis Melas-Candor-Ophir). Several works have shown the importance of eruption rates over other factors in developing long lava flows and networks of lava tubes (Calvari and Pinkerton, 1999; Guest et al., 2012; Pinkerton and Wilson, 1994;

190 Wadge, 1978; Walker, 1973) and the increase of pressure due to magma injection underground contributes to the development of a network of lava tubes (Glaze et al., 2005;

Guest et al., 2012). The end of magma supply decreases the pressure in lava tubes favouring the subsequent collapse of the tube roofs. Collapsed tubes can then be easy pathways to channelize lava flows coming from subsequent eruptive episodes.

5.3 Lava vs. water for the formation of the outflow channels

Although coverage is not yet available for all the surface of Mars, the HiRISE imagery offers new insight into the geologic evolution of the pit chains and into the processes that formed Labyrinthus Noctis and Valles Marineris on Mars thanks to a resolution up to 25 cm/pixel (McEwen et al., 2007). However, the better coverage of CTX images, combined with available HiRISE images, CRISM mineralogy, and MOLA profiles will show how pit chains could evolve into a complex network of channels to the erosional stage of the mensae and chaos terrains observed along the lower course of Kasei Valles and Valles Marineris.

Even if the build-up of the Tharsis Rise has created a set of faults in N-S direction, as well as on Alba Mons, this is not the factor contributing to the development of Labyrinthus Noctis otherwise a similar labyrinth-shaped structure would have been formed on the fractures surrounding Alba Mons or , where pit chains and fossae are also present. A comparison between the Tharsis bulge, 5000 km in diameter and 10 km in average height

(excluding volcanic peaks), and Alba Mons, 2000 km in diameter and 6 km in height (Carr,

2006), gives a rough but immediate idea of the difference in volume of lava erupted. Despite the smaller diameter of the single Tharsis volcanic structures compared with Alba Mons, the

Tharsis bulge is the maximum expression of volcanic activity on Mars. The main reason for the missed labyrinth on Alba Mons and the difference in the degree of evolution of the outflow channels is thus the higher availability of magma supply, and thus erupted lava, along the eastern flank of Tharsis that carved Valles Marineris original lava tubes deeper than any

191 other outflow channels before being in part drained away to Chryse Planitia and in part distributed along its course. The absence of a labyrinth and the presence of 1.5-2.5 km deep fossae (Harmakhis, Niger and ) in other significant volcanic provinces like the bigger (than Alba Mons) 2200 km north-eastern Hellas – south-western

(with the 6 km high Tyrrhenus Mons), the 1 km deep in the smaller (than Alba

Mons) Syrtis Major, the 1.2 km deep Olympica Fossae at Alba Mons, supports such an hypothesis. Nevertheless, if deep chasmata and fossae were ascribed to water only, either hydrothermal and/or runoff processes, they would be seen either on the smaller volcanic provinces or on other non-volcanic provinces too.

Only rough volumetric estimates can be done about Valles Marineris and its system of outflow channels without accurate and complete DTMs but they are still useful for the arguments that will be discussed ahead. Dividing the channels in better measurable segments has given an estimate of about 9.85 × 106 km3 for Valles Marineris (from eastern Labyrinthus

Noctis to southern Chryse Planitia, including the laterally connected chaos terrains) and 3.05

× 106 km3 for Kasei Valles; Maja, Shalbatana and Ares Vallis were not included in the count for practical measuring reasons and for their smaller volumetric contribution. The erosion rate of lava flowing in lava tubes estimated from vertical decrease of flow level observed in skylights, depending on steep slope that favours turbulent condition, has been quantified from

4 to 10 cm/d (Peterson and Swanson, 1974; Greeley et al., 1998) while experiments at current

Martian conditions estimate a lower end erosion rate for water (if not frozen) at 17.28 cm/d

(Conway et al., 2011). Regardless of the difference in erosion rate between lava and water, which would both vertically carve Valles Marineris in a relatively short time (about 2700 years of continuous lava flow or less for water at the above rates), it is difficult to understand how the largest amount of water on Mars (Carr, 1986; Leverington, 2011) can be concentrated mainly on Tharsis given the random external supply during the early ages of the planet (Carr and Head, 2010). Even considering an unrealistic 40:60 /water ratio, more than 19.35

192 × 106 km3 of water supply should have been available uphill between the sources of Valles

Marineris and Kasei Valles. But water volumes up to two or three orders of magnitude are more realistic estimates from the analysis of the Martian outflow channels - chaos systems

(Andrews-Hanna and Phillips, 2007; Leverington, 2011). It is also difficult to understand how such an amount of water can be replenished underground at a 10 km high hydraulic head

(Leverington, 2011) while unlikely to reside in the atmosphere for long periods of time throughout the history of Mars (Gillmann et al., 2011). The amount of water estimated in the

Martian atmosphere is about 10-1 km3 and the hydrogen map of the Mars Odyssey shows a poor abundance between the Tharsis Montes and Labyrinthus Noctis (Christensen, 2006).

Even assuming the source region abundances of water applicable to the whole mantle, from 130 to 250 ppm (corresponding from 5.3 × 107 km3 to 1.3 × 108 km3), equivalent to a global water layer of 370-860 m, the amount degassed would only be 400 m (McCubbin et al., 2010) while the minimum volume of water necessary to carve the outflow channels is equivalent to a global layer of about 300-500 m (Leverington, 2011). Other models suggested larger water inventories (up to 2700-3000 m) with a more or less degassed proportion (Lunine et al., 2003; Scambos and Jakosky, 1990), but another problem with water is the absence of rivers/outflow channels starting from the snowpack of the Tharsis Montes, which have been suggested to recharge the heads of Labyrinthus Noctis – Valles Marineris source aquifers

(Harrison and Grimm, 2004).

If uncertainties on the amount of water exist (Carr, 1986; Leverington, 2011), less uncertain is the presence of lava. From the available imagery it is possible to see that lava flows are distributed all along the outflow channels, erode their walls and form stratifications that decrease the depth and increase the width of the channels. Considering that the outflow channels of Mars are associated to a volcanic province, there are only two possibilities to explain the scaling size of the outflow channels observed planet-wide: a) nearly all the water of the planet is concentrated only in volcanic provinces proportionally to their size or b) the

193 eruption rates of lava are proportional to the volcanic province size. The latter seems more likely because of the many problems that aqueous processes suffer with respect to the volcanic processes for the formation of the outflow channels (Leverington, 2011). These problems are once more summarized briefly here:

· Lack of compelling evidence for sufficient water inventories necessary to the

excavation of the outflow channels at their hydraulic heads.

· Lack of compelling evidence for plausible recharge mechanisms of the aquifers.

The observation of volcanic eruptions also in atmosphere-less terrestrial planets shows that lava is less subject to environmental conditions than water and can travel for long distances in shallow underground tubes. The longitudinal section of the Ainahou Ranch lava tube, taken from the work of Greeley et al. (1998) as an example, explains the mechanism of formation, although on a smaller scale than seen on Mars. If the lava tube is emplaced on a steep slope, which seems to be the case of Valles Marineris and other outflow channels, the flow is likely to be turbulent and can erode the substrate. This is also the first stage shown in the diagram of

Fig. 5.4 in which the erosion occurs mainly vertically and the maximum depth of the tube is reached in its distal part. When the eruption stops and lava ceases to circulate in the tube, the first pit chains start to appear due to the subsidence of the tube roof that will lead to its collapse. This is the beginning of the erosion that will be continued by new lava flooding episodes, which form landslides that enlarge the tubes into fossae and then chasmata (Fig.

5.4). The landslide material is then covered by new lava flows and contributes to the infill along the channel (i.e. Fig. 5.17). Assuming that lava flows mantled the whole Chryse

Planitia, a rough estimate of the volume infill, 1600 km of diameter by 1 km of reasonable but still rough average flow thickness, gives 2 x 106 km3. If this volume is compared to the volume of rock carved away from the Valles Marineris (again 9.85 × 106 km3), this means that the main part of the flows and the landslide material has been distributed along the outflow channels. And these numbers do not take into account yet the other Circum-Chryse

194 outflow channels. This explains the low depth and the large width of Kasei Valles, particularly in the Echus Chasma sector, compared to the Valles Marineris. This also means that Kasei Valles is at a more advanced erosional stage than Valles Marineris; a particular that will be noticed in the next section is that fewer knobs and mesas survived the huge lava flooding in chaos terrains located in lower Kasei Valles compared to those located in lower

Valles Marineris.

5.4 A geological comparison between Tharsis and Elysium volcanic provinces

The combined imagery taken along the beds of Elysium, Valles Marineris and other circum-Chryse outflow channels will show in this section an overview of the various stages of the erosional processes that formed them and will provide further insight into how lava played a fundamental role for their formation.

5.4.1 Kasei Valles

Presence of pit chains has been found at Echus Chasma (Fig. 6.5) or through the chain of mounds of Uranius Dorsum (Fig. 6.6), interpreted as a lava tube (Chapman et al., 2010).

Echus Chasma shows bifurcations (Fig. 6.5) and perpendicular intersections of pit chains in relics of mesa terrains, suggesting that lava tubes formed networks underground in the Tharsis slopes (paleoflows). Echus Chasma also shows massive lava flooding that embayed and covered mesa terrains and pre-existing paleoflows filling its flat floor to 1.5 km depth, comparable to the 1 km deep Kasei Valles at its mouth near Chryse Planitia. Lateral channels along Echus Chasma show variable depth and sections from 600 m to 1 km. An intersection of fossae from different directions occurs in Labeatis Fossae and chaos terrains occur in

Labeatis Mensae. The chaos terrains are similar to those visible in the low course of Valles

Marineris or Deuteronilus Mensae or those reported in Northern Tempe Terra, thought to be

195 formed by geological processes like ice melting (Chuang and Crown, 2005). Again, the same problems expressed before about the distribution of water on the planet can be applied to ice, regardless of the arguments that explain how ice could be kept frozen beneath the rock debris

(Chuang and Crown, 2005). Another MOLA profile taken along the in the median course of the Kasei Valles shows its maximum depth, from 3 to 3.2 km. Lava fills lateral channels for several km uphill in this sector (Fig. 5.7 and Fig. 5.8). Both the main

Kasei upper and lower channel floor are lava draped, as confirmed by the pyroxenes detected, as well as near their confluence in lower Kasei Valles.

5.4.2 Elysium

The Elysium volcanic province is the second largest after Tharsis with its 2500 km diameter and the nearly 14 km high Elysium Mons (Carr, 2006). Crossing relationships at

Stygis Catena between a pit chain and a fossa show that circulation of both underground and surface lava occurred on the Elysium slopes as well, although on a smaller scale because the size of its chasmata is not comparable to those seen on Tharsis, still supporting the hypothesis that the size of the erosional features (pit chains, fossae, chasmata) is proportional to the amount of lava available to carve them. Although the slopes of Elysium Mons between

Patapsco and Iberus Vallis are mantled by a thick cover of iron phyllosilicates (red colour in the CRISM images), the presence of pyroxenes (both high and low-Ca) appears from the dusty flat floor and the walls of those supposed to be fluvial valleys, giving another view on their mechanism of formation. The typical fluvial valley shows tributaries and a V-shaped floor often covered by boulders transported by the water stream. None of these geomorphologic characteristics are observed on the Elysium outflow channels. Granicus

Valles in particular shows breakouts typical of lava channels (Fig. 5.9) and iron phyllosilicates deposited both on top of the mesas and on the floor of the channels parallel to the wind stream, inferred by local wind streaks, suggesting that the presence of iron

196 phyllosilicates is due more to eolian transport rather than water deposition. Even the presence of the boulders on the channel walls and none on the bed-floor is suggestive of mass wasting processes rather than fluvial transport. As it will be seen along other outflow channels as well, travel in lava tubes allows magma to reach long distances and the thermal erosion of lava can even open a breach in a natural obstacle, the breach is then enlarged by successive lava flooding becoming a channel. An example can be seen at the Cerberus Fossae where several fossae cross the Tartarus Montes and at Grjota Vallis in particular where fossae make their way perpendicularly through mesas (Fig. 5.10). A MOLA profile shows a flat floor in correspondence to a channel crossing the Tartarus Montes. Landslides and breakouts forming sub-parallel curvilinear channels, on a smaller scale than those that will be observed on the

Valles Marineris chasmata, can be seen on the Cerberus Fossae as well (Fig. 5.11) making less plausible an origin through gas cavities over a dike. Furthermore, CRISM data where available, indicate high-Ca pyroxene inside the channels suggesting passage of lava.

5.4.3 Circum-Chryse Planitia outflow channel mouths

The observation of Valles Marineris from its mouth in Chryse Planitia to its source in

Labyrinthus Noctis, as well as smaller outflow channels, has provided a good direct evidence of the erosive processes described in the diagram of Fig. 6.4. Several MOLA profiles have been taken at the outlets of the outflow channels in proximity of Chryse Planitia, Kasei

Valles, Maja Valles, Shalbatana Valles, and Ares Vallis, respectively. Unfortunately, no good

MOLA profiles were available for the final part of Valles Marineris at Simud Valles, because they all were parallel to it, but at least a good one has been found at . The depth of the flat floors of the outflow channels ranged from the 400 m at Ares Vallis to the 1.8 km at

Shalbatana Valles with averages of 1 km at Kasei and 600 m at Maja Valles. Some significant images have been found to sample the Chryse Planitia off the outflow channels showing that lava flows have reached distances several hundreds of km beyond their outlets (Fig. 5.12).

197 The floor of Maja Valles is draped with lava flows embaying the mesas, carving channels inside them and finally modeling them with the characteristic tear-eye shape (Fig. 5.13) previously thought to be formed by water floods. Pyroxenes have been detected both on the terraces and the floor of Shalbatana and Ares Valles, where stratifications of lava flows are seen.

5.4.4 Valles Marineris

Lava flows also mantle the course of Valles Marineris and its nearby outflow channels.

Pyroxenes have been detected along the course of Ares Vallis at Tiu Valles, at the source of

Shalbatana and , along Ganges, Eos-Capri, and inter-chaos chasmata. Landslides have also been imaged on the inter-chaos chasmata along the Valles Marineris (Fig. 5.14) while the MOLA profiles show a gradual increase of the depth of the outflow channels floors,

1 km at Ares Vallis, 2 km at Tiu Valles, 2 km at and Ravi Vallis, and 3 km at

Eos-Capri Chasma. This trend continues in Coprates Chasma, although the width of the main channel narrows again, a straight and long channel just uphill the enlargements occupied by the chaos terrains in the lower course of Valles Marineris. The Coprates Chasma section of

Valles Marineris is particularly interesting because shows the continuity of the enlargement of the pit chains into fossae and then into chasmata. A direct evidence of this process can be seen at Coprates Catena where two out of three parallel fossae are visibly connected to Coprates

Chasma and the third, still connected underground, terminates into a pit chain. The first two fossae then merge into an enlargement that is connected again to Coprates Chasma further downhill, a situation symmetrical to the opposite western side, an area characterized by landslides (Fig. 5.16) where a few available CRISM observations revealed presence of pyroxenes under the usual cover of iron phyllosilicates. Going uphill again along Coprates

Chasma it is possible to see areas where lava flows cover landslides, erode the base of the chasm walls and the middle-channel ridge, all that remains of the paleoflows once separating

198 two nearby fossae (Fig. 5.17). A CRISM observation sampling a landslide on the wall of

Coprates Chasma and its floor shows pyroxenes on both, suggesting that they are made of the same igneous rock. The alternation of landslides and lava flows can be observed all along the course of Valles Marineris. A MOLA profile taken in the middle of Coprates Chasma shows the difference between the V-shaped sections of the fossae and the section of the chasm where the depth of the floor reaches 8.5 km. Another MOLA profile taken on the western Coprates

Chasma shows the difference between the main channel and its southern parallel. The main channel reaches 9 km, 1 km deeper than its parallel where widespread landslides accumulated on its floor. It is interesting to note how the difference of 1000 m between the two channels is likely due to the different amount of material distributed on their floors. Even the same main channel shows 500 m of difference between its uphill and downhill profiles where more lava could flow downhill. The uphill profile also shows remains of the middle ridge (Fig. 5.18) that has been nearly totally erased in the downhill profile. The middle ridge tends to disappear quicker than the walls of the channel because it is eroded on both sides (Fig. 5.16) by lava flows; its thicker portions are the ones that resist much longer and can still be seen today.

Similarities in fossa-chasm continuity can also be seen on east Candor Chasma where a

MOLA profile shows how its section is changed by the material accumulated on its floor by mass wasting processes.

The Melas-Candor-Ophir chasmata system is another lateral enlargement of the Valles

Marineris, after the ones observed along its lower course, and the same alternation of lava flows and landslides is also observed here (Fig. 5.19). A MOLA profile taken on the eastern

Melas Chasma shows how its floor section, as well as its depth, is modified by this alternation. Portions of pyroxenes and flat floor can also be seen on eastern Candor Chasma where it is not covered by landslide debris. The Coprates Chasma bottleneck might be the cause of the lava accumulation that formed the lateral enlargement of Melas Chasma and breached the ridges between Melas-Candor-Ophir (Fig. 5.20), perhaps also aided by the lava

199 supply going directly into Candor and Ophir from underground feeders and from Tithonium

Chasma. The MOLA profiles show how the landslides modify the section of the chasmata floors. The few CRISM observations on landslide-free floors of , Hebes

Chasma, and Ius Chasma, show pyroxenes. Sporadic LTDs superposed both on landslide material and the floor of Ius Chasma can be observed (Fig. 5.21).

The main Valles Marineris channel narrows again at Tithonium and Ius Chasma, when approaching its source located in eastern Labyrinthus Noctis, forming a similar lateral channel enlargement north of crater and parallel chasmata like the ones seen at Coprates

Chasma. Landslides still fill the channels floors because they are not covered by lava flows.

MOLA profiles along Tithonium and Ius Chasma, as well as Calydon Fossa, show how the channels are still flat floored although the sections tend to be more V-shaped, when compared to the median course of Valles Marineris, because less lava flooding occurred in this area. The depth of Tithonium and Ius Chasma floors decreased from 6.5 km to 6 km from east to west, a trend confirmed by another profile along the Oudemans crater and the channel enlargement north of it. The CTX/HiRISE images have shown LTDs again superposed on landslides (Fig.

5.22) making it difficult to understand why the chasmata floors are not completely covered as they should be in the case of formation in an aqueous environment. The MOLA profiles taken on eastern, central, and western Labyrinthus Noctis show a maximum channel depth of 4, 4.5, and 3.2 km, respectively, while samples of inter-channel floors show stratifications of lava flows with superposed LTDs.

200 5.5 Discussion

The overview of the last section provided detailed information to understand how pit chains evolve to fossae, outflow channels and then to mensae. The initial stage starts from lava both erupted on the surface and injected underground that erodes its own pathways to form lava tubes. This stage is still visible on Tharsis between Pavonis Mons and Labyrinthus

Noctis (Fig. 5.2 and Fig. 5.3). The thermal erosion of the floor of the tube can extend down into the pre-existing surface materials (paleoflows) to erode a very deep lava tube if the eruption goes on for long enough. If the lava flow on the surface (or within the lava tube) has a sustained eruption rate, it is aided by the gravity along a steep slope, and does not find topographic obstacles that would force it to deviate from its way, the pattern may also likely develop linearly. However, when fossae are straight despite the undulations of the topography, without curvilinear channels and breakouts, this could be the sign of the presence of gas cavities on top of a dike. If channels form by the collapse of the roofs of tubes that formed by roofing-over of lava flows they will tend to be sinuous and to develop breakouts just as flows are sinuous as they pass through pre-existing topography. But if channels form by the collapse of the roofs of gas cavities at the tops of dikes they are much more likely to be relatively straight and to ignore the topography - both in the sense of not being sinuous and in the sense of not being oriented down the local surface slopes (i.e. Acheron Catena). There are also cases in which lava channels open their way through mesas of paleoflows due to the combined thermal and mechanical erosion of lava: an example is Grjota Vallis (Fig. 5.10) and another is Maja Valles (Fig. 5.13). A wedge is carved in the obstacle, often made of paleoflows, while lava in excess surrounds the obstacle eroding its base and giving the classic tear-eye shape. Subsequent flows enlarge the wedge and open the channel. It will also be shown ahead how networks of these channels through the mesas characterize the chaos terrains. Depending on the local eruption rates, lava tubes can be a few tens of meters (Fig.

5.2) up to a km wide (Fig. 5.6) while the smallest fossae can be a few hundreds meters (Fig.

201 5.11) up to several km wide. Abrupt changes of slope and/or channel bottlenecks, which slow down the lava flows, favour the formation of lateral tube networks, breakouts, or channel enlargements. A network of lava tubes can be seen in Labyrinthus Noctis adjacent to the channel enlargement that can be seen in the basin north of Oudemans crater at the slope break between 9 and 6 km of height, in the basin formed by the Melas – Candor – Ophir Chasmata at the slope break between 6 and 3 km of height, in the Eos – Capri Chasma system at the slope break between 3 and 0 km of height, and in the at the slope break between 0 and -3 km of height. Topography also influences the direction of the lava channels forcing them to deviate or form breakouts (i.e. Fig. 5.9, Fig. 5.10, Fig. 5.11). This happens in the lower course of Valles Marineris, Ares and that turn to the north and also Kasei and Maja Valles that turn to the east almost parallel. When the eruption stops because of the depletion of the source, the pressure in the lava tubes drops to the collapse of the roofs thus forming the first pit chains. The collapse occurs all along the pipeline so the pit chain follows the pattern of the lava tube, thus also explaining the curvilinear shapes already seen at Labyrinthus Noctis (Fig. 5.1). Since the lava tubes are likely almost all connected, the collapse also occurs all along the network of tubes revealing the pattern that once was concealed underground. Labyrinthus Noctis may thus likely be such a network while the system Ius – may be considered the beginning of the Valles Marineris where lava carved its main channel. The axis Tithonium – Candor, as well as the axis Ius –

Melas – Coprates Chasma, offers further insight into the evolution of the pit chain into fossae and then chasmata. Just south of Ius and Tithonium Chasma lie parallel to them Calydon

Fossa and Tithonium Catena, respectively, where a regional context imageshows that the former enlarges into a wider channel merging the lower Ius Chasma while the latter ends up to Candor Chasma alternating as pit chain and fossa. Why such a difference when they both started from the same basin located north of Oudemans crater? An answer comes from the geomorphologic analysis of the terrains located north of Tithonium Chasma where there are

202 many surface channels, and likely many others underground, heading towards Echus, Hebes and even Ophir Chasma. The channel towards Ophir is mostly underground and its direction has been inferred from the direction of a pit chain located north of Candor Chasma, unless more evident collapse structures can be seen as in the case of the channel between Ganges

Chasma and Shalbatana – Ravi Vallis. The majority of these channels (and underground tubes) drained lava from the Tithonium – Candor axis mainly towards the Echus Chasma.

Although no good HiRISE coverage is available on the terrains north of the Tithonium –

Candor axis, a CTX image has enough resolution to show a pit chain at the end of a curvilinear drainage channel that starts from the north Oudemans basin. The lava drainage towards Echus – Hebes Chasma has likely prevented the axis Tithonium – Candor to be carved and developed like the axis Ius – Melas – Coprates. The lower (eastern) Candor

Chasma appears longer than its upper (western) side probably also due to the contribution of lava flows coming from Melas through the breach of their dividing ridge (Fig. 5.20). No lateral drainages are observed on the axis Ius – Melas, although some start to appear parallel to Coprates Chasma. Subsequent episodes of lava flooding have further eroded the pit chains and widened them into fossae as it can be seen at the eastern tips of Tithonium, Candor and

Ophir Chasma causing processes of mass wasting. Landslides commonly occur on their walls and those of Ius, Melas and Coprates Chasmata as well as on the chaos terrains located in the lower course of Valles Marineris (Fig. 5.14). Perhaps, if enough magma supply were available for some additional time, Hebes Chasma would have been included in the Melas –

Ophir channel. Even might have merged Candor Chasma and probably

Hebes Chasma in the same way in western Ophir Planum. This mechanism can be observed in a more advanced stage at Ganges Chasma, where only a thin ridge separates it from Capri

Chasma and where access is open to towards Aurorae Chaos. This is exactly the point from where the Valles Marineris system is characterized by chaos terrains and it is widened by the same degradational/erosional processes that formed islands of terrains typical

203 of the mensae, particularly visible between Simud and Tius Valles, where is located the main

Valles Marineris mouth into Chryse Planitia.

The role of lava is much simpler to understand, whereas that of water is quite difficult. The available images show LTDs, thought to be formed by sedimentation in an aqueous environment (Bibring et al., 2006; Weitz et al., 2010), scattered in topographic lows (Fig.

5.19) but also in topographic highs (Fig. 5.15) or superposed on intact landslides (Fig. 5.20 and Fig. 5.21) or as thin veneers on lava flows (Fig. 5.15). In some cases, there are topographic basins filled with LTDs and none nearby (Fig. 5.22). If outflow channels are carved by water as claimed so far, why is the presence of the LTDs not as ubiquitous as it should be? And why LTDs are not seen uniformly distributed on the channels floors or in sedimentary layers on their walls? If the water floods were so catastrophic to deeply carve the outflow channels, why do the LTD-topped landslides appear so intact? The diapiric formation proposed by Baioni (2013) does not provide a satisfactory explanation for the LTDs superposed on landslides; also the formation in a closed depositional system (Grindrod et al.,

2012) does not explain why some basins are filled and some nearby not. It is beyond the scope of this paper to provide a full satisfactory explanation for the LTDs formation.

However, both LTDs (and water, if associated) and (wind transported) dunes of iron phyllosilicates, postdate the volcanic activity that it is here proposed for the formation of the

Labyrinthus Noctis – Valles Marineris system. The role of water is also weakened by the ubiquitous presence of lava flows on the channel floors between the knobs and mesas of the chaos terrains. Lava flows barely appear from the iron phyllosilicate cover at Aromatum

Chaos but they are particularly evident in the chaos terrains located in the Eos – Capri

Chasma basin and at Hydraotes Chaos, where also occasional rootless cones resulting from lava flows and ground ice interaction are seen. The main channel between Aurorae and

Hydraotes Chaos is subject to the same mass wasting processes seen both in the upper and in the median course of the Valles Marineris (Fig. 5.15).

204

5.6 Conclusion

From all the observations collected so far the main scenario of formation for the

Labyrinthus Noctis and the Valles Marineris, including its system of perpendicular outflow channels, may be explained as an initial stage of lava tubes in the very large Tharsis paleoflows. CRISM observations have confirmed the same mineralogy both in the latest lava flows and in the paleoflows. Pit chains are the first stage of an erosional evolution that leads to fossae after the collapse of the lava tubes roofs. A tectonic origin can be definitely excluded after the analysis of the available imagery and by the absence of vertical displacements along both fossae and chasmata in MOLA profiles. The persistent availability of magma supply, evident from massive lava flooding, favoured subsequent eruptions that carved tubes into fossae first and chasmata later as far as the mouth to Chryse Planitia. The traces left by water are dubious and discontinuous and, in any case, postdate lava flows. It seems unlikely that the amount of water necessary to carve the ouflow channels has ever been concentrated or recycled only at their hydraulic heads without a comparable situation on other non-volcanic provinces of Mars. This lack of evidence has profound implications on the water inventory supposed to fill the lowlands with an ocean and on the whole hydrological cycle thought to move water to the equatorial regions of the planet. So the Valles Marineris and the other outflow channels are mainly the result of erosional processes caused by the passage of lava flows. The various stages of the erosional evolution are still visible from Labyrinthus

Noctis to the mouth of Valles Marineris, which can be defined as an outflow channel at an intermediate stage of evolution. The other outflow channels are at the extreme ends, Kasei

Valles is at the latest stage, while Maja, Shalbatana and Ravi Vallis are at the earliest stage.

Also the chaos terrains are likely the result of the erosional process that have been observed at a more advanced stage between Simud and Tiu Valles and at Nilokeras, the mouths of Valles

Marineris and Kasei Valles into Chryse Planitia, respectively. It is difficult to define the

205 erosional stage of Ares Vallis. Characterized by chaos terrains since its upper course, its main channel is wider than that of Maja Valles and there is no clear magmatic source other than that of a lateral enlargement of Valles Marineris leading to Holden crater through the

Erythraea Fossa. However, the rocks and the deposits imaged at the Pathfinder landing site have shown little compelling evidence of large water floods but rather the results of volcanic processes (Carr, 2006).

Acknowledgements

The author has been supported by the ETH Research Commission Grant ETH-03 10-1. The copyright for the CTX images is acknowledged to the Malin Space Science Systems, CRISM to Johns Hopkins University – Applied Physics Laboratory, HiRISE to , other copyrights are indicated directly on the images where applicable. The author thanks

Nick Lang and two anonymous reviewers for their helpful reviews.

206 Chapter 6

General overview and conclusions

207 6.1 General overview and conclusions

The results of the studies included in this thesis provide an alternative view, based on a consistent interpretation of the same data released by the various missions to Mars and profoundly different than the one currently and commonly accepted. Mars is not and was not the planet similar to Earth that has been described so far. Mars shared with the Earth only the hellish environmental conditions during (and typical of) the Hadean eon. Their paths diverged after the Hadean, Mars became an arid and cold world while the Earth became the watery and warm planet that we live today. Plate tectonics developed on Earth but never did on Mars because the presence of water may have made the difference. The fact that plate tectonics developed only on Earth, the only terrestrial planet with clear evidence of water, is the better proof that water was never able to form oceans on other planets, otherwise we would see plate tectonics elsewhere. This is another indirect proof that Mars never had water. However, it is not established yet that the lack of water is the only factor that causes differences between the tectonics of Earth and Mars and Venus, for example, lack of plate tectonics on Mars may additionally be due to its smaller size (O’Neill et al., 2007; Foley et al., 2012). Lack of plate tectonics on Venus may also be due to its higher surface temperature (Foley et al., 2012).

Volcanism is peculiar on Mars because of the SPGI otherwise it would probably have had a randomly distributed volcanism. Considering that water offered a good shelter to life on

Earth, it might be reasonable to think that life could not develop on the other terrestrial planets. The results and the implications of my research will be described in the next subchapters.

6.1.1 Formation of the Martian dichotomy

The alternative hypothesis to the formation of the Borealis basin, the SPGI model, is now supported by the geological evidence of the volcanic features on Mars, mainly distributed all

208 over the southern hemisphere according to twelve loxodromical alignments. Although there are still uncertainties regarding the upper end of the age of the giant impact, these are within the limit estimated by the geochemical studies of mantle-core differentiation and should not be above ~ 15 Ma. The lower end of 4 Ma is supported by the short life of the radiogenic elements Al26 and Fe60, before this age any trace of the Martian dichotomy would have been remelted again by the heat produced by these radiogenic elements. Two alternative impactors in size and composition, both viable from the astronomical point of view, might explain the formation of the Martian dichotomy: a siderite (80% of iron in radius) of 1600 km in radius and a meso-siderite (50% of iron in radius) of 2000 km in radius hitting the South Pole of

Mars at a speed of 5 km s-1. The velocity of migration of the plumes estimated along the alignments of volcanic features on Mars showed that the first might be the best candidate for the formation of the Martian dichotomy (Leone, 2016). In both cases, the core of the impactor reached the core of the planet and would have interrupted any existing magnetic field so it is reasonable to assume that the giant impact triggered the transient magnetic field of Mars. The decline time of such a magnetic field, as calculated by the SPGI model, is consistent with the ages estimated for the magnetic anomalies detected on Mars. The heat flux across the core- mantle boundary (CMB) showed a drop in the range 20-40 mW m-2, which is passed at around 4.0 Ga in all the simulations, and supported a declining volcanism until 3.5 Ga. This result is independent from the size of the impactor and it is essentially confirmed in all the simulations.

6.1.2 Volcanism in the history of Mars

Many new volcanoes were identified on the surface of Mars, mainly outside the known volcanic provinces along the way to the South Pole. The ages of the volcanism of Mars correspond to Pre-Noachian – Noachian in the geological history of Mars, when the eruptions took place on the surface. Noachian – Hesperian were instead the ages inferred from the crater

209 counts. However, the volcanic alignments observed on the surface of Mars, subsequent to the

SPGI, support the Pre-Noachian ages, the tracks of these plumes were found on the surface, starting from the observation of the largest volcanoes of Mars, and were formed by > 100 volcanic centres distributed along twelve hemispherical alignments. The thermal anomaly following the giant impact formed only minor volcanism in the lowlands, mainly around the northern polar regions, and it is then very difficult that the heat flux might have supported volcanism until the Amazonian. No alignment was observed in the northern hemisphere of

Mars, the sub-equatorial large volcanoes contributed to the formation and subsequent modification of the transition topography. Elysium now has to be considered part of the transition topography and not anymore located in the lowlands. The volcanoes have their maximum size along the sub-equatorial regions where the plumes had their peak and stationed for a few tens of Ma immediately after their formation. The size of the volcanic centres decreased along the track once the plumes started their migration towards the south pole.

The evidence is that today, when no ongoing volcanism is observed on the surface of

Mars, the heat flux estimated by various other studies has the same values of 3.5 Ga ago, while the time needed for a single plume to reach the south pole from the equator was between ~ 100 and ~ 300 Ma. On average, there are ~ 200 Ma of time difference between the largest volcanic centres located around the equator and those located near the south pole. This difference of emplacement time, also existing in the ages obtained from the crater counts available in the literature, supports the hypothesis of the impactor of 1600 km. The mismatch between the crater count ages and the SPGI ages is due to the uncertainties related to the crater counts methods. According to the SPGI model, the phase of the plume migration started

50 Ma after the giant impact (about 4.517 Ga ago) and lasted for 500 Ma with another 500 Ma of widespread volcanism of minor intensity and characterized by no plume migration, no volcanic centres of large dimensions were observed outside the alignments, ending in the

Hesperian (3.5 Ga ago) for a total of 1.0 Ga. According to the crater counts collected in the

210 literature, the volcanism lasted on average ~ 500-600 Ma, starting from the Noachian (4.1-4.0

Ga ago) and ending in the Hesperian (3.5 Ga ago), with an exception in some lava flows dated as Amazonian and mostly located in the lava fields of Tharsis and Elysium (3.0 – 0.2 Ga ago).

However, this interpretation is more related to the uncertainty of the crater counts rather than to a long eruptive history of the volcanic centres, the end of large impactors just before the emplacement of the latest flow altered the geologic clock. Even assuming an hot interior of

Mars, which might not be anymore the case, the crust could be so thick that no magma would make it to the surface of the planet to feed eruptive activity. The in situ K-Ar measurements of the Cumberland sample within the lava filled Gale crater, likely deposited by one of the eruptions of Tyrrhenum Mons, provided an absolute age of 4.21 Ga, a difference of 500-700

Ma with the crater count ages of 3.7-3.5 Ga, confirming that the SPGI ages of the volcanism are consistent with the K-Ar measurements.

6.1.3 The absence of Martian plate tectonics

The observation of the whole surface of the planet has neither shown any trace of recent nor past plate tectonics, including traces of proto-plate tectonics. There are neither subduction nor rift zones. Valles Marineris, once considered as a large rift zone of Mars (i.e. Sleep,

1994), is just the longest lava channel emplaced on the largest lava field of the largest volcanic province of the Solar System. Baltis Vallis on Venus is longer than Valles Marineris.

The observation of the surface at high resolution, integrated with useful MOLA profiles, has revealed that many “fractures” are lava channels formed by the continued erosion of collapsed lava tubes (Leone, 2014). The smallest of these channels still retain the original roof of the lava tube along their less eroded segments (Leone, 2016). Mars never had plate tectonics but a stagnant lid since the beginning of its history and this might be related to the extreme scarcity or absence of water on its surface (O’Rourke and Korenaga, 2012).

211 6.1.4 The extreme scarcity of water on Mars

Although the main part of the planetary community still argues otherwise with always less convincing arguments, the extreme scarcity of water on Mars is already known since the

Viking missions, the disappointment that followed the findings of both Vikings even stopped further missions of the Mars exploration programme for about two decades. Now this scarcity is also proved not only by the analysis of the Mars Odyssey spacecraft data but also by the

Mars Reconnaissance Orbiter and all the robotic landers gone to Mars so far. The widespread presence of unaltered olivine contained in the basaltic composition of the surface, the volcanic origin of the outflow channels and the fluvial networks, and the geomorphologic setting of the geologic units sampled on the floor of the craters support the view that water possibly never flowed, ponded, or even existed in significant amounts on Mars. A good example of how the geologic setting of Mars was misinterpreted so far is the lava filling of the Gale crater’s floor, very similar to that of many other craters observed on the surface of Mars, which includes the geologic units of the Yellowknife Bay formation thought as formed in aqueous environment.

In conclusion, this formation is just another exposed stratigraphic section of layers of basaltic composition likely formed by several episodes of deposition of lava flows coming from the

Tyrrhenum Mons via Farah Vallis and from the deposition of volcanic ashes transported by the wind. The stratigraphic section of the Yellowknife Bay formation is similar to the sections of layers observed on the walls of Valles Marineris, that is, a sequence of deposition of lava flows from periodic eruptive episodes that likely occurred in absence of water. The Point

Lake outcrop contains voids that, associated to the basaltic composition of the rock, suggests emplacement of vesicular lava flows. Blebs of light-toned deposits, similar to those observed within the lava-draped Valles Marineris, were analysed and determined as calcium sulphate.

Calcium sulphate is anhydrous, absorption of water would form gypsum, thus indicating lack of water. Major element compositions of the Glenelg member of the formation indicate no substantial weathering after the deposition, this implies a frigid and arid climate like the one

212 observed today. The Gillespie Lake member is formed by sand of basaltic composition.

Basaltic composition indicates a volcanic origin of the grains. Dunes and deposits of basaltic composition moved by the wind are quite common on Mars, the high degree of cementation of the sandstones does not necessarily mean fluvial transport and deposition as water would be incompatible with all the other observations and, most of all, the low pressure of the that would not allow liquid water to be stable on the surface of the planet.

Sandstones can form in arid desertic environments from compaction of sand grains by the pressure of overlying layers. The time required for the deposition of the layers in the

Yellowknife formation was estimated between 10k and 10M years (Grotzinger et al., 2015) while olivine is altered in serpentine between 100 and 10k years (Oze and Sharma, 2007). At last, the Sheepbed member located at the base of the stratigraphic column, containing ~ 20% of saponitic smectite clay minerals, is perfectly explainable with low-grade metamorphism of pre-existing lava flows, this process would be more consistent with the observation of superposed layers of volcanic origin. Again, craters filled by lava flows are very common on the surface of Mars, including the crater , where also the rover Opportunity found essentially a similar geologic setting observed at Gale. Jarosite, a hydrated mineral found nearby the Opportunity landing site (Meridiani Planum), on the highlands at the equator of

Mars, is unaltered and in close association with unaltered olivine; jarosite would rapidly decompose to produce ferric oxyhydroxides in humid climates otherwise (Madden et al.,

2004).

Other lines of evidence argue against the presence of water on Mars. Unaltered olivine is present both at the heads and along the course of the fluvial valleys. The heads of these valleys are located on the slopes of volcanic centres and are aligned with other highland volcanoes. Furthermore, the location of the valleys is not consistent with the models of past climate on Mars (Wordsworth et al., 2015). This is particularly evident along the Alignment

10, from Terra Sabaea to Sinus Margaritifer, starting from the Syrtis Major. Unaltered olivine

213 was also found within craters where water should have ponded for long time. This observation was already confirmed both within Holden, Gale, Meridiani Planum (Rogers and

Bandfield, 2009), and Gusev crater (McSween et al., 2006), coupled with the compelling evidence of lava flows mantling their floors. So the craters regarded as the cornerstones of water on Mars are now seriously questioned. This evidence also raises some other questions.

If water came before lava in the history of Mars, how is it possible that all the fluvial valleys are emplaced on volcanic fields? If water came after lava, how is it possible that olivine has been left unaltered? If many incisions ascribed to water are considered recent, how is it possible that Noachian olivine is unaltered? If does not rain on Mars and the main outflow channels are lava channels, how the putative ocean formed? Patches of smectites, thought to be formed in aqueous environment, were found in topographic highs of volcanic origin unreachable by groundwater and stratigraphically coincident with the Noachian – Hesperian age. A study revealed that phyllosilicates can be formed by volcanic processes (Schenato et al., 2003). The easiest answer to all these questions is that liquid water never flowed or even existed on the surface of Mars. Lava flows were found inside Valles Marineris and inside other circum-Chryse outflow channels, all carved on the very large lava fields of Tharsis, as far as Chryse and Acidalia Planitia, where unaltered olivine was also reported. After all these evidences already available in the literature, there are two possibilities remaining: a) the much claimed aqueous environment has a volcanic origin, that is, the few hydrothermal water contained in magma at average amounts of 0.67 – 1.0 wt% altered the olivine undeground where the pressure still allowed its stability, and b) the formation of the clays has to be ascribed to low grade metamorphism in cooling lava while gaseous water was lost to space and/or stored at the higher latitudes where it is more stable as ice in the or in the polar caps. Considering that the serpentine was observed in few locations on Mars, and that ground ice was not found at equatorial latitudes where the “fluvial networks” and the outflow channels are present, both hypotheses are likely. However, in both cases, the hypothesis that

214 there were flowing rivers of water on Mars should be inesorably abandoned. Liquid water is unstable on the surface of Mars and this is not only a problem of temperature but rather a problem of pressure to which volatiles like water are particularly sensitive. Even the addition of salts, much claimed in the literature as a proof of the presence of water, would not make water stable at all.

So the picture of Mars arising from these results and observations is profoundly different from the planet similar to the Earth that has been thought so far. Mars is a planet that had a peculiar volcanic history only because of the SPGI event otherwise it might have had a volcanic history completely different. The scarcity of water, the absence of plate tectonics, and a CO2 atmosphere are also points in common with Venus, another planet that notoriously did not host life.

215 References

Abramov, O., Wong, S.M. and Kring, D.A., 2012. Differential melt scaling for oblique

impacts on terrestrial planets. Icarus, 218(2): 906-916.

Ackiss, S.E., Seelos, K.D., and Buczkowski, D.L., 2014. Mineralogic mapping of Huygens

crater, Mars: A transect of the highlands crust and Hellas basin rim. VIII International

Conference on Mars. Abstract #1038.

Ackiss, S.E. and Wray, J., 2014. Occurrences of possible hydrated sulfates in the southern

high latitudes of Mars. Icarus, 243: 311-324.

Acocella, V., 2007. Understanding caldera structure and development: An overview of

analogue models compared to natural calderas. Earth-Sci Rev, 85(3-4): 125-160.

Acuna, M.H., Connerney, J.E.P., Ness, N.F., Lin, R.P., Mitchell, D., Carlson, C.W.,

McFadden, J., , K.A., Reme, H., Mazelle, C., Vignes, D., Wasilewski, P. and

Cloutier, P., 1999. Global distribution of crustal magnetization discovered by the Mars

Global Surveyor MAG/ER experiment. Science, 284(5415): 790-793.

Acuna, M.H., Connerney, J.E.P., Wasilewski, P., Lin, R.P., Mitchell, D., Anderson, K.A.,

Carlson, C.W., McFadden, J., Reme, H.R., Mazelle, C., Vignes, D., Bauer, S.J.,

Cloutier, P. and Ness, N.F., 2001. Magnetic field of Mars: Summary of results from

the aerobraking and mapping orbits. J Geophys Res-Planet, 106(E10): 23403-23417.

Adams, J.B., Gillespie, A.R., Jackson, M.P.A., Montgomery, D.R., Dooley, T.P., Combe, J.-

P., Schreiber, D.C., 2009. Salt tectonics and collapse of Hebes Chasma, Valles

Marineris, Mars. Geology, 37: 691-694.

Agnor, C.B., Canup, R.M. and Levison, H.F., 1999. On the character and consequences of

large impacts in the late stage of terrestrial planet formation. Icarus, 142(1): 219-237.

Allen, C.C., 1979. Volcano-Ice Interactions on Mars. J Geophys Res, 84: 8048-8059.

216 Amelin, Y., Krot, A.N., Hutcheon, I.D. and Ulyanov, A.A., 2002. Lead isotopic ages of

chondrules and calcium-aluminum-rich inclusions. Science, 297(5587): 1678-1683.

Andrews-Hanna, J.C. and Phillips, R.J., 2007. Hydrological modeling of outflow channels

and chaos regions on Mars. J Geophys Res-Planet, 112(E8).

Andrews-Hanna, J.C., Zuber, M.T., Arvidson, R.E. and Wiseman, S.M., 2010. Early Mars

hydrology: Meridiani playa deposits and the sedimentary record of Arabia Terra. J

Geophys Res-Planet, 115.

Andrews-Hanna, J.C., Zuber, M.T. and Banerdt, W.B., 2008. The Borealis basin and the

origin of the martian crustal dichotomy. Nature, 453(7199): 1212-U1227.

Anguita, F., Farelo, A.F., Lopez, V., Mas, C., Munoz-Espadas, M.J., Marquez, A. and Ruiz,

J., 2001. Tharsis dome, Mars: New evidence for Noachian-Hesperian thick-skin and

Amazonian thin-skin tectonics. J Geophys Res-Planet, 106(E4): 7577-7589.

Anguita, F., Fernandez, C., Cordero, G., Carrasquilla, S., Anguita, J., Nunez, A., Rodriguez,

S. and Garcia, J., 2006. Evidences for a Noachian-Hesperian orogeny in Mars. Icarus,

185(2): 331-357.

Baker, V.R., Strom, R.G., Gulick, V.C., Kargel, J.S., Komatsu, G. and Kale, V.S., 1991.

Ancient Oceans, Ice Sheets and the Hydrological Cycle on Mars. Nature, 352(6336):

589-594.

Baker, V.R., Maruyama, S., and Dohm, J.M., 2007. Tharsis superplume and the geological

evolution of early Mars. In: D.A. Yuen, S. Maruyama, S. Karato, B.F. Windley (Eds.),

Superplumes: Beyond Plate Tectonics, Springer, Doordrecht (The Netherlands), pp.

516.

Baker, V.R., 2015. Planetary geomorphology: Some historical/analytical perspectives.

Geomorphology, 240: 8-17.

217 Baker, V.R., Hamilton, C.W., Burr, D.M., Gulick, V.C., Komatsu, G., Luo, W., Rice Jr, J.W.,

and Rodriguez, J.A.P., 2015. Fluvial geomorphology on Earth-like planets: A review.

Geomorphology 245, 149-182.

Baioni, D., 2013. Morphology and geology of an interior layered deposit in the Western

Tithonium Chasma. Planetary and Space Science, 89: 140-150.

Banks, M.E., Lang, N.P., Kargel, J.S., McEwen, A.S., Baker, V.R., Grant, J.A., Pelletier, J.D.

and Strom, R.G., 2009. An analysis of sinuous ridges in the southern Argyre Planitia,

Mars using HiRISE and CTX images and MOLA data. J Geophys Res-Planet, 114.

Baratoux, D., Toplis, M.J., Monnereau, M. and Gasnault, O., 2011. Thermal history of Mars

inferred from orbital geochemistry of volcanic provinces (vol 472, pg 338, 2011).

Nature, 475(7355): 254-254.

Barr, A.C. and Canup, R.M., 2008. Constraints on gas giant satellite formation from the

interior states of partially differentiated satellites. Icarus, 198(1): 163-177.

Bernhardt, H., Hiesinger, H., Reiss, D., Ivanov, M. and Erkeling, G., 2013. Putative eskers

and new insights into glacio-fluvial depositional settings in southern Argyre Planitia,

Mars. Planetary and Space Science, 85: 261-278.

Bibring, J.P., Langevin, Y., Mustard, J.F., Poulet, F., Arvidson, R., Gendrin, A., Gondet, B.,

Mangold, N., Pinet, P., Forget, F. and Team, O., 2006. Global mineralogical and

aqueous mars history derived from OMEGA/ data. Science, 312(5772):

400-404.

Bishop, J.L., Loizeau, D., McKeown, N.K., Saper, L., Dyar, M.D., Des Marais, D.J., Parente,

M. and Murchie, S.L., 2013. What the ancient phyllosilicates at Mawrth Vallis can tell

us about possible habitability on early Mars. Planetary and Space Science, 86: 130-

149.

Bistacchi, N., Massironi, M. and Baggio, P., 2004. Large-scale fault kinematic analysis in

Noctis Labyrinthus (Mars). Planetary and Space Science, 52(1-3): 215-222.

218 Bleacher, J.E., Glaze, L.S., Greeley, R., Hauber, E., Baloga, S.M., Sakimoto, S.E.H.,

Williams, D.A. and Glotch, T.D., 2009. Spatial and alignment analyses for a field of

small volcanic vents south of Pavonis Mons and implications for the Tharsis province,

Mars. J Volcanol Geoth Res, 185(1-2): 96-102.

Bleacher, J.E., Hamilton, C.W., Scheidt, S.P., Garry, W.B., de Wet, A., Whelley, P., and

Williams, D.A., 2015. No erosion needed: development of streamlined islands during

lava channel construction. XXXXVI Lunar and Conference, The

Woodlands, TX, USA, #2182.

Borlina, C.S., Ehlmann, B.L., and Kite, E.S., 2015. Modeling the thermal and physical

evolution of Mt. Sharp’s sedimentary rocks, Gale crater, Mars: implications for

diagenesis on the MSL Curiosity rover traverse. J. Geophys. Res. DOI

10.1002/2015JE004799.

Bottke, W.F., Levison, H.F., Nesvorny, D. and Dones, L., 2007. Can planetesimals left over

from terrestrial planet formation produce the lunar Late Heavy Bombardment? Icarus,

190(1): 203-223.

Bouley, S., Craddock, R.A., Mangold, N. and Ansan, V., 2010. Characterization of fluvial

activity in Parana Valles using different age-dating techniques. Icarus, 207(2): 686-

698.

Breuer, D., and W. B. Moore (2007), Dynamics and thermal history of the terrestrial planets,

the moon, and Io, in Treatise on Geophysics, edited by T. Spohn, Elsevier B. V.,

Amsterdam.

Bridges, N.T., Banks, M.E., Beyer, R.A., Chuang, F.C., Dobrea, E.Z.N., Herkenhoff, K.E.,

Keszthelyi, L.P., Fishbaugh, K.E., McEwen, A.S., Michaels, T.I., Thomson, B.J. and

Wray, J.J., 2010. Aeolian bedforms, yardangs, and indurated surfaces in the Tharsis

Montes as seen by the HiRISE Camera: Evidence for dust aggregates. Icarus, 205(1):

165-182.

219 Bunch, T.E., Keil, K. and Huss, G.I., 1972. THE LANDES METEORITE. Meteoritics, 7(1):

31-38.

Bunte, M.K., Williams, D.A. and Greeley, R., 2008. Geologic mapping of the Zal region of

Io. Icarus, 197(1): 354-367.

Cabrol, N.A., Grin, E.A. and Landheim, R., 1998. Ma'adim Vallis evolution: Geometry and

models of discharge rate. Icarus, 132(2): 362-377.

Cabrol, N.A., Grin, E.A., Carr, M.H., Sutter, B., Moore, J.M., Farmer, J.D., Greeley, R.,

Kuzmin, R.O., DesMarais, D.J., Kramer, M.G., Newsom, H., Barber, C., Thorsos, I.,

Tanaka, K.L., Barlow, N.G., Fike, D.A., Urquhart, M.L., Grigsby, B., Grant, F.D. and

de Goursac, O., 2003. Exploring Gusev Crater with : Review of science

objectives and testable hypotheses. J Geophys Res-Planet, 108(E12).

Cailleau, B., Walter, T.R., Janle, P. and Hauber, E., 2005. Unveiling the origin of radial

grabens on Alba Patera volcano by finite element modelling. Icarus, 176(1): 44-56.

Calvari, S. and Pinkerton, H., 1999. Lava tube morphology on Etna and evidence for lava

flow emplacement mechanisms. J Volcanol Geoth Res, 90(3–4): 263-280.

Cantor, B., Malin, M. and Edgett, K.S., 2002. Multiyear (MOC)

observations of repeated Martian weather phenomena during the northern summer

season. J Geophys Res-Planet, 107(E3).

Capitan, R.D. and Van De Wiel, M., 2011. Landform hierarchy and evolution in Gorgonum

and Atlantis basins, Mars. Icarus, 211(1): 366-388.

Carr, M.H., Greeley, R., Blasius, K.R., Guest, J.E. and Murray, J.B., 1977. Some Martian

Volcanic Features as Viewed from Viking Orbiters. Eos T Am Geophys Un, 58(8):

829-829.

Carr, M.H., 1986. Mars - a Water-Rich Planet. Icarus, 68(2): 187-216.

Carr, M.H., 1996. Water on early Mars. Ciba F Symp, 202: 249-267.

220 Carr, M.H., 2001. Mars Global Surveyor observations of Martian . J Geophys

Res-Planet, 106(E10): 23571-23593.

Carr, M.H. and Head, J.W., 2003. Oceans on Mars: An assessment of the observational

evidence and possible fate. J Geophys Res-Planet, 108(E5).

Carr, M.H., 2006. The surface of Mars. Cambridge University Press, p.75.

Carr, M.H. and Head, J.W., 2010. Geologic history of Mars. Earth Planet Sc Lett, 294(3-4):

185-203.

Catalano, J.G., 2013. Thermodynamic and mass balance constraints on iron-bearing

phyllosilicate formation and alteration pathways on early Mars. J Geophys Res-Planet,

118(10): 2124-2136.

Chambers, J.E. and Wetherill, G.W., 1998. Making the terrestrial planets: N-body integrations

of planetary embryos in three dimensions. Icarus, 136(2): 304-327.

Chapman, M.G. and Tanaka, K.L., 2002. Related magma-ice interactions: Possible origins of

chasmata, chaos, and surface materials in Xanthe, Margaritifer, and Meridiani Terrae,

Mars. Icarus, 155(2): 324-339.

Chapman, M.G., Neukum, G., Dumke, A., Michael, G., van Gasselt, S., Kneissl, T.,

Zuschneid, W., Hauber, E., Ansan, V., Mangold, N. and Masson, P., 2010. Noachian-

Hesperian geologic history of the Echus Chasma and Kasei Valles system on Mars:

New data and interpretations. Earth Planet Sc Lett, 294(3-4): 256-271.

Christensen, P.R., 2006. Water at the poles and in permafrost regions of Mars. Elements, 2(3):

151-155.

Christensen, P.R., McSween, H.Y., Bandfield, J.L., Ruff, S.W., Rogers, A.D., Hamilton,

V.E., Gorelick, N., Wyatt, M.B., Jakosky, B.M., Kieffer, H.H., Malin, M.C. and

Moersch, J.E., 2005. Evidence for magmatic evolution and diversity on Mars from

infrared observations. Nature, 436(7050): 504-509.

221 Chuang, F.C. and Crown, D.A., 2005. Surface characteristics and degradational history of

debris aprons in the Tempe Terra Mareotis fossae region of Mars. Icarus, 179(1): 24-

42.

Clifford, S.M. and Parker, T.J., 2001. The evolution of the Martian hydrosphere: Implications

for the fate of a primordial ocean and the current state of the northern plains. Icarus,

154(1): 40-79.

Connolly, J.A.D. (2005) Computation of phase equilibria by linear programming: A tool for

geodynamic modeling and its application to subduction zone decarbonation. Earth

Planet. Sci. Lett., 236, 524–541.

Conway, S.J., Lamb, M.P., Balme, M.R., Towner, M.C. and Murray, J.B., 2011. Enhanced

runout and erosion by overland flow at low pressure and sub-freezing conditions:

Experiments and application to Mars. Icarus, 211(1): 443-457.

Cushing, G.E., Titus, T.N., Wynne, J.J. and Christensen, P.R., 2007. THEMIS observes

possible cave skylights on Mars. Geophys Res Lett, 34(17). de Koker, N., Steinle-Neumann, G. and Vlcek, V., 2012. Electrical resistivity and thermal

conductivity of liquid Fe alloys at high P and T, and heat flux in Earth's core. P Natl

Acad Sci USA, 109(11): 4070-4073. de Pablo, M.A. and Pacifici, A., 2008. Geomorphological evidence of water level changes in

Nepenthes Mensae, Mars. Icarus, 196(2): 667-671. de Pablo, M.A., Michael, G.G. and Centeno, J.D., 2013. Age and evolution of the lower NW

flank of the Hecates Tholus volcano, Mars, based on crater size-frequency distribution

on CTX images. Icarus, 226(1): 455-469.

Dehant, V., Banerdt, B., Lognonne, P., Grott, M., Asmar, S., Biele, J., Breuer, D., Forget, F.,

Jaumann, R., Johnson, C., Knapmeyer, M., Langlais, B., Le Feuvre, M., Mimoun, D.,

Mocquet, A., Read, P., Rivoldini, A., Romberg, O., Schubert, G., Smrekar, S., Spohn,

T., Tortora, P., Ulamec, S. and Vennerstrom, S., 2012. Future Mars geophysical

222 observatories for understanding its internal structure, rotation, and evolution. Planetary

and Space Science, 68(1): 123-145.

Di Achille, G. and Hynek, B.M., 2010. Ancient ocean on Mars supported by global

distribution of deltas and valleys. Nat Geosci, 3(7): 459-463.

Dohm, J.M., Ferris, J.C., Baker, V.R., Anderson, R.C., Hare, T.M., Strom, R.G., Barlow,

N.G., Tanaka, K.L., Klemaszewski, J.E. and Scott, D.H., 2001. Ancient drainage basin

of the Tharsis region, Mars: Potential source for outflow channel systems and putative

oceans or paleolakes. J Geophys Res-Planet, 106(E12): 32943-32958.

Dohm, J.M. and Tanaka, K.L., 1999. Geology of the Thaumasia region, Mars: plateau

development, valley origins, and magmatic evolution. Planetary and Space Science,

47(3-4): 411-431.

Dohm, J.M., Williams, J.P., Anderson, R.C., Ruiz, J., McGuire, P.C., Komatsu, G., Davila,

A.F., Ferris, J.C., Schulze-Makuch, D., Baker, V.R., Boynton, W.V., Fairen, A.G.,

Hare, T.M., Miyamoto, H., Tanaka, K.L. and Wheelock, S.J., 2009. New evidence for

a magmatic influence on the origin of Valles Marineris, Mars. J Volcanol Geoth Res,

185(1-2): 12-27.

Ehlmann, B.L., Mustard, J.F., Murchie, S.L., Poulet, F., Bishop, J.L., , A.J., Calvin,

W.M., Clark, R.N., Des Marais, D.J., Milliken, R.E., Roach, L.H., Roush, T.L.,

Swayze, G.A. and Wray, J.J., 2008. Orbital Identification of Carbonate-Bearing Rocks

on Mars. Science, 322(5909): 1828-1832.

Ehlmann, B.L., Mustard, J.F. and Murchie, S.L., 2010. Geologic setting of serpentine deposits

on Mars. Geophys Res Lett, 37.

Ehlmann, B.L., Mustard, J.F., Murchie, S.L., Bibring, J.P., Meunier, A., Fraeman, A.A. and

Langevin, Y., 2011. Subsurface water and clay mineral formation during the early

history of Mars. Nature, 479(7371): 53-60.

223 Ehlmann, B.L., Berger, G., Mangold, N., Michalski, J.R., Catling, D.C., Ruff, S.W.,

Chassefiere, E., Niles, P.B., Chevrier, V. and Poulet, F., 2013. Geochemical

Consequences of Widespread Clay Mineral Formation in Mars' Ancient Crust. Space

Sci Rev, 174(1-4): 329-364.

Erkeling, G., Reiss, D., Hiesinger, H., Poulet, F., Carter, J., Ivanov, M.A., Hauber, E. and

Jaumann, R., 2012. Valleys, paleolakes and possible shorelines at the Libya

Montes/Isidis boundary: Implications for the hydrologic evolution of Mars. Icarus,

219(1): 393-413.

Erkeling, G., Leone, G., and Kereszturi, Á., 2015. Paleoshoreline. In: Encyclopedia of

Planetary Landforms, Springer, New York. DOI 10.1007/978-1-4614-9213-9_650-I,

pp. 1501-1507.

Ernst, R.E., Grosfils, E.B. and Mege, D., 2001. Giant dike swarms: Earth, Venus, and Mars.

Annu Rev Earth Pl Sc, 29: 489-534.

Fagan, A.L., Sakimoto, S.E.H. and Hughes, S.S., 2010. Formation constraints on Martian

north polar volcanic edifices. J Geophys Res-Planet, 115.

Farrand, W.H., Gaddis, L.R. and Keszthelyi, L., 2005. Pitted cones and domes on Mars:

Observations in Acidalia Planitia and Mensae using MOC, THEMIS, and

TES data. J Geophys Res-Planet, 110(E5).

Farrand, W.H., Glotch, T.D., Rice, J.W., Hurowitz, J.A. and Swayze, G.A., 2009. Discovery

of jarosite within the Mawrth Vallis region of Mars: Implications for the geologic

history of the region. Icarus, 204(2): 478-488.

Fassett, C.I. and Head, J.W., 2008. Valley network-fed, open-basin :

Distribution and implications for Noachian surface and subsurface hydrology. Icarus,

198(1): 37-56.

224 Fenton, L.K. and Hayward, R.K., 2010. Southern high latitude fields on Mars:

Morphology, aeolian inactivity, and climate change. Geomorphology, 121(1-2): 98-

121.

Foley, B. J., D. Bercovici and W. Landuyt, 2012. The conditions for plate tectonics on super-

Earths: Inferences from convection models with damage. Earth and Planetary Science

Letters 331–332(0): 281-290.

Fornasier, S., Clark, B.E., Dotto, E., Migliorini, A., Ockert-Bell, M. and Barucci, M.A., 2010.

Spectroscopic survey of M-type asteroids. Icarus, 210(2): 655-673.

Frey, H., 2008. Ages of very large impact basins on Mars: Implications for the late heavy

bombardment in the inner solar system. Geophys Res Lett, 35(13).

Frey, H. and Jarosewich, M., 1982. Sub-Kilometer Martian - Properties and

Possible Terrestrial Analogs. J Geophys Res, 87(Nb12): 9867-9879.

Frey, H. and Schultz, R.A., 1988. Large Impact Basins and the Mega-Impact Origin for the

Crustal Dichotomy on Mars. Geophys Res Lett, 15(3): 229-232.

Frey, H.V., 2006. Impact constraints on, and a chronology for, major events in early Mars

history. J Geophys Res-Planet, 111(E8).

Garvin, J.B., Sakimoto, S.E.H., Frawley, J.J., Schnetzler, C.C. and , H.M., 2000.

Topographic evidence for geologically recent near-polar volcanism on Mars. Icarus,

145(2): 648-652.

Gerya, T.V. and Yuen, D.A., 2007. Robust characteristics method for modelling multiphase

visco-elasto-plastic thermo-mechanical problems. Phys Earth Planet In, 163(1-4): 83-

105.

Gerya, T. V. (2010), Introduction to Numerical Geodynamic Modelling, 345 pp., Cambridge

University Press, Cambridge.

225 Ghatan, G.J. and Head, J.W., 2002. Candidate subglacial volcanoes in the south polar region

of Mars: Morphology, morphometry, and eruption conditions. J Geophys Res-Planet,

107(E7).

Ghatan, G.J., Head, J.W. and Pratt, S., 2003. Cavi Angusti, Mars: Characterization and

assessment of possible formation mechanisms. J Geophys Res-Planet, 108(E5).

Ghatan, G.J. and Zimbelman, J.R., 2006. Paucity of candidate coastal constructional

landforms along proposed shorelines on Mars: Implications for a northern lowlands-

filling ocean. Icarus, 185(1): 171-196.

Gillmann, C., Lognonne, P. and Moreira, M., 2011. Volatiles in the atmosphere of Mars: The

effects of volcanism and escape constrained by isotopic data. Earth Planet Sc Lett,

303(3-4): 299-309.

Gilmour, J., 2002. The solar system's first clocks. Science, 297(5587): 1658-1659.

Glamoclija, M., Marinangeli, L. and Komatsu, G., 2011. Harmakhis Vallis Source Region,

Mars: Insights into the recent geothermal history based on geological mapping.

Planetary and Space Science, 59(11-12): 1179-1194.

Glaze, L.S., Anderson, S.W., Stofan, E.R., Baloga, S. and Smrekar, S.E., 2005. Statistical

distribution of tumuli on pahoehoe flow surfaces: Analysis of examples in and

Iceland and potential applications to lava flows on Mars. J Geophys Res-Sol Ea,

110(B8).

Gleason, G.C. and Tullis, J., 1995. A Flow Law for Dislocation Creep of Quartz Aggregates

Determined with the Molten-Salt Cell. Tectonophysics, 247(1-4): 1-23.

Golabek, G.J., Keller, T., Gerya, T.V., Zhu, G.Z., Tackley, P.J. and Connolly, J.A.D., 2011.

Origin of the martian dichotomy and Tharsis from a giant impact causing massive

magmatism. Icarus, 215(1): 346-357.

Goldstein, J.I., Scott, E.R.D. and Chabot, N.L., 2009. Iron meteorites: Crystallization, thermal

history, parent bodies, and origin. Chem Erde-Geochem, 69(4): 293-325.

226 Gomi, H., Ohta, K., Hirose, K., Labrosse, S., Caracas, R., Verstraete, M.J. and Hernlund,

J.W., 2013. The high conductivity of iron and thermal evolution of the Earth's core.

Phys Earth Planet In, 224: 88-103.

Grant, J.A., 2000. Valley formation in Margaritifer Sinus, Mars, by precipitation-recharged

ground-water sapping. Geology, 28(3): 223-226.

Grant, J.A. and Schultz, P.H., 1990. Gradational Epochs on Mars - Evidence from West-

Northwest of Isidis Basin and Electris. Icarus, 84(1): 166-195.

Grant, J.A. and Parker, T.J., 2002. Drainage evolution in the Margaritifer Sinus region, Mars.

J Geophys Res-Planet, 107(E9).

Grant, J.A., Irwin, R.P., Wilson, S.A., Buczkowski, D. and Siebach, K., 2011. A lake in

Uzboi Vallis and implications for Late Noachian-Early Hesperian climate on Mars.

Icarus, 212(1): 110-122.

Greeley, R., Fagents, S.A., Harris, R.S., Kadel, S.D., Williams, D.A. and Guest, J.E., 1998.

Erosion by flowing lava: Field evidence. J Geophys Res-Sol Ea, 103(B11): 27325-

27345.

Greeley, R., Kuzmin, R.O. and Haberle, R.M., 2001. Aeolian processes and their effects on

understanding the chronology of Mars. Space Sci Rev, 96(1-4): 393-404.

Greeley, R., Foing, B.H., McSween, H.Y., Neukum, G., Pinet, P., van Kan, M., Werner, S.C.,

Williams, D.A. and Zegers, T.E., 2005. Fluid lava flows in Gusev crater, Mars. J

Geophys Res-Planet, 110(E5).

Greeley, R. and Spudis, P.D., 1978. Volcanism in Cratered Terrain Hemisphere of Mars.

Geophys Res Lett, 5(6): 453-455.

Grindrod, P.M., West, M., Warner, N.H. and Gupta, S., 2012. Formation of an Hesperian-

aged sedimentary basin containing phyllosilicates in Coprates Catena, Mars. Icarus,

218(1): 178-195.

227 Grosse, P., , B.V., Euillades, P.A., Kervyn, M. and Petrinovic, I.A., 2012. Systematic

morphometric characterization of volcanic edifices using digital elevation models.

Geomorphology, 136(1): 114-131.

Grott, M., Hauber, E., Werner, S.C., Kronberg, P. and Neukum, G., 2005. High heat flux on

ancient Mars: Evidence from rift flank uplift at Coracis Fossae. Geophys Res Lett,

32(21).

Grotzinger, J.P., Sumner, D.Y., Kah, L.C., Stack, K., Gupta, S., Edgar, L., Rubin, D., Lewis,

K., Schieber, J., Mangold, N., Milliken, R., Conrad, P.G., DesMarais, D., Farmer, J.,

Siebach, K., Calef, F., Hurowitz, J., McLennan, S.M., Ming, D., Vaniman, D., Crisp,

J., Vasavada, A., Edgett, K.S., Malin, M., Blake, D., Gellert, R., Mahaffy, P., Wiens,

R.C., Maurice, S., Grant, J.A., Wilson, S., Anderson, R.C., Beegle, L., Arvidson, R.,

Hallet, B., Sletten, R.S., Rice, M., Bell, J., Griffes, J., Ehlmann, B., Anderson, R.B.,

Bristow, T.F., Dietrich, W.E., Dromart, G., Eigenbrode, J., Fraeman, A., Hardgrove,

C., Herkenhoff, K., Jandura, L., Kocurek, G., Lee, S., Leshin, L.A., Leveille, R.,

Limonadi, D., Maki, J., McCloskey, S., Meyer, M., Minitti, M., Newsom, H., Oehler,

D., Okon, A., Palucis, M., Parker, T., Rowland, S., , M., Squyres, S., Steele,

A., Stolper, E., Summons, R., Treiman, A., Williams, R., Yingst, A. and Team, M.S.,

2014. A Habitable Fluvio-Lacustrine Environment at Yellowknife Bay, Gale Crater,

Mars. Science, 343(6169).

Grotzinger, J.P., Gupta, S., Malin, M.C., Rubin, D.M., Schieber, J., Siebach, K., Sumner,

D.Y., Stack, K.M., Vasavada, A.R., Arvidson, R.E., Calef, F., Edgar, L., ,

W.F., Grant, J.A., Griffes, J., Kah, L.C., Lamb, M.P., Lewis, K.W., Mangold, N.,

Minitti, M.E., Palucis, M., Rice, M., Williams, R.M.E., Yingst, R.A., Blake, D.,

Blaney, D., Conrad, P., Crisp, J., Dietrich, W.E., Dromart, G., Edgett, K.S., Ewing,

R.C., Gellert, R., Hurowitz, J.A., Kocurek, G., Mahaffy, P., McBride, M.J.,

McLennan, S.M., Mischna, M., Ming, D., Milliken, R., Newsom, H., Oehler, D.,

228 Parker, T.J., Vaniman, D., Wiens, R.C. and Wilson, S.A., 2015. Deposition,

exhumation, and paleoclimate of an ancient lake deposit, Gale crater, Mars. Science,

350(6257): 177-U169.

Guest, J.E., Duncan, A.M., Stofan, E.R. and Anderson, S.W., 2012. Effect of slope on

development of pahoehoe flow fields: Evidence from Mount Etna. J Volcanol Geoth

Res, 219: 52-62.

Hand, E., 2012. Dreams of water on Mars evaporate. Nature, 484(7393): 153-153.

Hanna, J.C. and Phillips, R.J., 2005. Hydrological modeling of the Martian crust with

application to the pressurization of aquifers. J Geophys Res-Planet, 110(E1).

Hardersen, P.S., Gaffey, M.J. and Abell, P.A., 2005. Near-IR spectral evidence for the

presence of iron-poor orthopyroxenes on the surfaces of six M-type asteroids. Icarus,

175(1): 141-158.

Harrison, K.P. and Grimm, R.E., 2004. Tharsis recharge: A source of groundwater for

Martian outflow channels. Geophys Res Lett, 31(14).

Hauck, S.A. and Phillips, R.J., 2002. Thermal and crustal evolution of Mars. J Geophys Res-

Planet, 107(E7).

Head, J.W., Hiesinger, H., Ivanov, M.A., Kreslavsky, M.A., Pratt, S. and Thomson, B.J.,

1999. Possible ancient oceans on Mars: Evidence from Mars Orbiter Laser Altimeter

data. Science, 286(5447): 2134-2137.

Head, J.W. and Kreslavsky, M.A., 2002. Northern lowlands of Mars: Evidence for

widespread volcanic flooding and tectonic deformation in the Hesperian Period. J

Geophys Res-Planet, 107(E1).

Head, J.W. and Wilson, L., 2007. Heat transfer in volcano-ice interactions on Mars: synthesis

of environments and implications for processes and landforms. Ann Glaciol, 45: 1-13.

229 Head, J.W., Wilson, L., Dickson, J. and Neukum, G., 2006. The Huygens-Hellas giant dike

system on Mars: Implications for Late Noachian-Early Hesperian volcanic resurfacing

and climatic evolution. Geology, 34(4): 285-288.

Hernlund, J. W., and P. Tackley (2008), Modeling mantle convection in the spherical annulus,

Phys. Earth Planet. Int., 171(1-4), 48-54, doi:10.1016/j.pepi.2008.07.037.

Hiesinger, H. and Head, J.W., 2002. Topography and morphology of the Argyre Basin, Mars:

implications for its geologic and hydrologic history. Planetary and Space Science,

50(10-11): 939-981.

Hiesinger, H. and Head, J.W., 2004. The Syrtis Major volcanic province, Mars: Synthesis

from Mars Global Surveyor data. J Geophys Res-Planet, 109(E1).

Hobbs, S.W., Pain, C.F. and Clarke, J.D.A., 2011. Slope morphology of hills at the Mars

Pathfinder landing site. Icarus, 214(1): 258-264.

Hoke, M.R.T. and Hynek, B.M., 2009. Roaming zones of precipitation on ancient Mars as

recorded in valley networks. J Geophys Res-Planet, 114.

Hynek, B.M. and Phillips, R.J., 2008. The stratigraphy of Meridiani Planum, Mars, and

implications for the layered deposits' origin. Earth Planet Sc Lett, 274(1-2): 214-220.

Hynek, B.M., Robbins, S.J., Sramek, O. and Zhong, S.J., 2011. Geological evidence for a

migrating Tharsis plume on early Mars. Earth Planet Sc Lett, 310(3-4): 327-333.

Irwin, R.P., Howard, A.D. and Maxwell, T.A., 2004. Geomorphology of Ma'adim Vallis,

Mars, and associated paleolake basins. J Geophys Res-Planet, 109(E12).

Irwin, R.P., Watters, T.R., Howard, A.D. and Zimbelman, J.R., 2004. Sedimentary

resurfacing and fretted terrain development along the crustal dichotomy boundary,

Aeolis Mensae, Mars. J Geophys Res-Planet, 109(E9).

Irwin, R.P., Craddock, R.A., Howard, A.D. and Flemming, H.L., 2011. Topographic

influences on development of Martian valley networks. J Geophys Res-Planet, 116.

230 Isherwood, R.J., Jozwiak, L.M., Jansen, J.C. and Andrews-Hanna, J.C., 2013. The volcanic

history of Olympus Mons from paleo-topography and flexural modeling. Earth Planet

Sc Lett, 363: 88-96.

Ivanov, M.A. and Head, J.W., 2006. Alba Patera, Mars: Topography, structure, and evolution

of a unique late Hesperian-early Amazonian shield volcano. J Geophys Res-Planet,

111(E9).

Ivanov, M.A., Hiesinger, H., Erkeling, G., Hielscher, F.J., Reiss, D., 2012. Major episodes of

geologic history of Isidis Planitia on Mars. Icarus, 218: 24-46.

Jacobsen, S.B., 2005. The Hf-wisotopic system and the origin of the earth and moon. Annu

Rev Earth Pl Sc, 33: 531-570.

Jaeger, W.L., Keszthelyi, L.P., McEwen, A.S., Dundas, C.M. and , P.S., 2007.

Athabasca Valles, Mars: A lava-draped channel system. Science, 317(5845): 1709-

1711.

Jaeger, W.L., Keszthelyi, L.P., Skinner, J.A., Milazzo, M.P., McEwen, A.S., Titus, T.N.,

Rosiek, M.R., Galuszka, D.M., Howington-Kraus, E., Kirk, R.L. and Team, H., 2010.

Emplacement of the youngest flood lava on Mars: A short, turbulent story. Icarus,

205(1): 230-243.

Kageyama, A. and Sato, T., 2004. "Yin-Yang grid'': An overset grid in spherical geometry.

Geochem Geophy Geosy, 5.

Kargel, J.S. and Strom, R.G., 1992. Ancient Glaciation on Mars. Geology, 20(1), 3-7.

Karlsson, N.B., Schmidt, L.S., Hvidberg, C.S., 2015. Volume of Martian midlatitude glaciers

from radar observations and ice flow modeling. Geophys. Res. Lett. 42(8), 2627-2633.

Katayama, I., Karato, S.-I. (2008) Rheological structure and deformation of subducted slabs

in the mantle transition zone: implications for mantle circulation and deep

earthquakes. Phys. Earth Planet. Interiors, 168, 125–133.

231 Keller, T. and Tackley, P.J., 2009. Towards self-consistent modeling of the martian

dichotomy: The influence of one-ridge convection on crustal thickness distribution.

Icarus, 202(2): 429-443.

Kerber, L., Head, J.W., Madeleine, J.B., Forget, F. and Wilson, L., 2012. The dispersal of

pyroclasts from ancient explosive volcanoes on Mars: Implications for the friable

layered deposits. Icarus, 219(1): 358-381.

Kereszturi, A., 2010. Analysis of two sections of Shalbatana Vallis' tributary channels.

Planetary and Space Science, 58(14-15): 2008-2021.

Keszthelyi, L., Jaeger, W., McEwen, A., Tornabene, L., Beyer, R.A., Dundas, C. and

Milazzo, M., 2008. High Resolution Imaging Science Experiment (HiRISE) images of

volcanic terrains from the first 6 months of the Mars Reconnaissance Orbiter Primary

Science Phase. J Geophys Res-Planet, 113(E4).

Keszthelyi, L., McEwen, A.S. and Thordarson, T., 2000. Terrestrial analogs and thermal

models for Martian flood . J Geophys Res-Planet, 105(E6): 15027-15049.

Khan, A., Connolly, J.A.D. (2008) Constraining the composition and thermal state of Mars

from inversion of geophysical data. J. Geophys. Res., 113, E07003.

Kleine, T., Mezger, K., Munker, C., Palme, H. and Bischoff, A., 2004. Hf-182-W-182 isotope

systematics of chondrites, eucrites, and martian meteorites: Chronology of core

formation and early mantle differentiation in Vesta and Mars. Geochim Cosmochim

Ac, 68(13): 2935-2946.

Kleine, T., Munker, C., Mezger, K. and Palme, H., 2002. Rapid accretion and early core

formation on asteroids and the terrestrial planets from Hf-W chronometry. Nature,

418(6901): 952-955.

Kokubo, E. and Ida, S., 2000. Formation of protoplanets from planetesimals in the solar

nebula. Icarus, 143(1): 15-27.

232 Konopliv, A.S., Yoder, C.F., Standish, E.M., Yuan, D.N. and Sjogren, W.L., 2006. A global

solution for the Mars static and seasonal gravity, Mars orientation, and

Deimos masses, and Mars ephemeris. Icarus, 182(1): 23-50.

Kostama, V.P., Ivanov, M.A., Raitala, J., Tormanen, T., Korteniemi, J. and Neukum, G.,

2010. Evidence for multiple ice deposits on the northeastern rim of Hellas basin, Mars.

Earth Planet Sc Lett, 294(3-4): 321-331.

Lammer, H., Chassefiere, E., Karatekin, O., Morschhauser, A., Niles, P.B., Mousis, O., Odert,

P., Mostl, U.V., Breuer, D., Dehant, V., Grott, M., Groller, H., Hauber, E. and Le,

B.S.P., 2013. Outgassing History and Escape of the Martian Atmosphere and Water

Inventory. Space Sci Rev, 174(1-4): 113-154.

Lang, N.P., Tornabene, L.L., McSween, H.Y. and Christensen, P.R., 2009. Tharsis-sourced

relatively dust-free lavas and their possible relationship to Martian meteorites. J

Volcanol Geoth Res, 185(1-2): 103-115.

Langdon, T.G., 2005. An analysis of flow mechanisms in high temperature creep and

superplasticity. Mater Trans, 46(9): 1951-1956.

Laskar, J., Correia, A.C.M., Gastineau, M., Joutel, F., Levrard, B. and Robutel, P., 2004.

Long term evolution and chaotic diffusion of the insolation quantities of Mars. Icarus,

170(2): 343-364.

Le Feuvre, M. and Wieczorek, M.A., 2008. Nonuniform cratering of the terrestrial planets.

Icarus, 197(1): 291-306.

Leask, H.J., Wilson, L. and Mitchell, K.L., 2006. Formation of Aromatum Chaos, Mars:

Morphological development as a result of volcano-ice interactions. J Geophys Res-

Planet, 111(E8).

Leinhardt, Z.M. and Richardson, D.C., 2002. N-body simulations of planetesimal evolution:

Effect of varying impactor mass ratio. Icarus, 159(2): 306-313.

233 Leone, G., , A.G., Wilson, L., Williams, D.A., Keszthelyi, L.P., Jaeger, W.L. and

Turtle, E.P., 2009. Volcanic history, geologic analysis and map of the Prometheus

Patera region on Io. J Volcanol Geoth Res, 187(1-2): 93-105.

Leone, G., Tackley P. J., Gerya T. V., May D. A., Zhu, G., 2013. 3D numerical model for the

formation of the Martian dichotomy and the Tharsis and Elysium Rises. XXXXIV

Lunar and Planetary Science Conference, The Woodlands, TX, USA, #1089.

Leone, G., 2014. A network of lava tubes as the origin of Labyrinthus Noctis and Valles

Marineris on Mars. J Volcanol Geoth Res, 277: 1-8.

Leone, G., Tackley P. J., Gerya T. V., May D. A., Zhu, G., 2014. The Southern Polar Giant

Impact for the Formation of the Martian Dichotomy. Geophys. Res. Lett., 41 (24):

8736-8743.

Leone, G., 2015. Transition topography (Mars). In: Encyclopedia of Planetary Landforms,

Springer, New York. DOI 10.1007/978-1-4614-9213-9_650-I, pp. 2169-2173.

Leone, G., 2016. Alignments of Volcanic Features in the Southern Hemisphere of Mars

produced by Migrating Mantle Plumes. J. Volcanol. Geoth. Res. Doi:

10.1016/j.jvolgeores.2015.10.028

Leveille, R.J. and Datta, S., 2010. Lava tubes and basaltic caves as astrobiological targets on

Earth and Mars: A review. Planetary and Space Science, 58(4): 592-598.

Leverington, D.W., 2009. Reconciling channel formation processes with the nature of

elevated outflow systems at Ophir and Aurorae Plana, Mars. J Geophys Res-Planet,

114.

Leverington, D.W., 2011. A volcanic origin for the outflow channels of Mars: Key evidence

and major implications. Geomorphology, 132(3-4): 51-75.

Liao, J. and Gerya, T., 2015. From continental rifting to seafloor spreading: Insight from 3D

thermo-mechanical modeling. Gondwana Research, 28(4): 1329-1343.

234 Lillis, R.J., Frey, H.V., Manga, M., Mitchell, D.L., Lin, R.P., Acuna, M.H. and Bougher,

S.W., 2008. An improved crustal magnetic field map of Mars from electron

reflectometry: Highland volcano magmatic history and the end of the martian dynamo.

Icarus, 194(2): 575-596.

Lin, J.-R., T. V. Gerya, P. J. Tackley, D. A. Yuen, and G. J. Golabek (2009), Numerical

modeling of protocore destabilization during planetary accretion: Methodology and

results, Icarus (USA), 204(2), 732-748.

Loizeau, D., Werner, S.C., Mangold, N., Bibring, J.P. and Vago, J.L., 2012. Chronology of

deposition and alteration in the Mawrth Vallis region, Mars. Planetary and Space

Science, 72(1): 31-43.

Lucchitta, B.K., 1992. The canyon system of Mars. Mars: 453-492.

Lunine, J.I., Chambers, J., Morbidelli, A. and Leshin, L.A., 2003. The origin of water on

Mars. Icarus, 165(1): 1-8.

Luo, W., 2002. Hypsometric analysis of Margaritifer Sinus and origin of valley networks. J

Geophys Res-Planet, 107(E10).

Mandt, K.E., de Silva, S.L., Zimbelman, J.R. and Crown, D.A., 2008. Origin of the Medusae

Fossae Formation, Mars: Insights from a synoptic approach. J Geophys Res-Planet,

113(E12).

Mandt, K.E., and Leone, G., 2015. Yardang (Mars). In: Encyclopedia of Planetary

Landforms, Springer, New York. DOI 10.1007/978-1-4614-9213-9_650-I, pp. 2339-

2347.

Mangold, N., 2012. Fluvial landforms on fresh impact ejecta on Mars. Planetary and Space

Science, 62(1): 69-85.

Marchant, D.R. and Head, J.W., 2007. Antarctic dry valleys: Microclimate zonation, variable

geomorphic processes, and implications for assessing climate change on Mars. Icarus,

192(1): 187-222.

235 Marinova, M.M., Aharonson, O. and Asphaug, E., 2008. Mega-impact formation of the Mars

hemispheric dichotomy. Nature, 453(7199): 1216-1219.

Masson, P., 1977. Structure Pattern-Analysis of Noctis Labyrinthus-Valles Marineris Regions

of Mars. Icarus, 30(1): 49-62.

Maxwell, T.A. and Craddock, R.A., 1995. Age Relations of Martian Highland Drainage

Basins. J Geophys Res-Planet, 100(E6): 11765-11780.

Mccauley, J.F., Masursky, H., , D.J., Sharp, R.P., Wilhelms, D.E., Carr, M.H., Cutts,

J.A. and Hartmann, W.K., 1972. Preliminary Mariner 9 Report on .

Icarus, 17(2): 289-&.

McCoy, T.J., Walker, R.J., Goldstein, J.I., Yang, J., McDonough, W.F., Rumble, D., Chabot,

N.L., Ash, R.D., Corrigan, C.M., Michael, J.R. and Kotula, P.G., 2011. Group IVA

: New constraints on the crystallization and cooling history of an asteroidal core

with a complex history. Geochim Cosmochim Ac, 75(22): 6821-6843.

McCubbin, F.M., Smirnov, A., Nekvasil, H., Wang, J.H., Hauri, E. and Lindsley, D.H., 2010.

Hydrous magmatism on Mars: A source of water for the surface and subsurface during

the Amazonian. Earth Planet Sc Lett, 292(1-2): 132-138.

McEwen, A.S., Eliason, E.M., Bergstrom, J.W., Bridges, N.T., Hansen, C.J., Delamere, W.A.,

Grant, J.A., Gulick, V.C., Herkenhoff, K.E., Keszthelyi, L., Kirk, R.L., Mellon, M.T.,

Squyres, S.W., Thomas, N. and Weitz, C.M., 2007. Mars Reconnaissance Orbiter's

High Resolution Imaging Science Experiment (HiRISE). J Geophys Res-Planet,

112(E5).

Mcgill, G.E. and Dimitriou, A.M., 1990. Origin of the Martian Global Dichotomy by Crustal

Thinning in the Late Noachian or Early Hesperian. J Geophys Res-Solid, 95(B8):

12595-12605.

Mcgill, G.E. and Squyres, S.W., 1991. Origin of the Martian Crustal Dichotomy - Evaluating

Hypotheses. Icarus, 93(2): 386-393.

236 McKeown, N.K., Bishop, J.L., Dobrea, E.Z.N., Ehlmann, B.L., Parente, M., Mustard, J.F.,

Murchie, S.L., Swayze, G.A., Bibring, J.P. and Silver, E.A., 2009. Characterization of

phyllosilicates observed in the central Mawrth Vallis region, Mars, their potential

formational processes, and implications for past climate. J Geophys Res-Planet, 114.

McSween, H.Y., Wyatt, M.B., Gellert, R., Bell, J.F., Morris, R.V., Herkenhoff, K.E.,

Crumpler, L.S., Milam, K.A., Stockstill, K.R., Tornabene, L.L., Arvidson, R.E.,

Bartlett, P., Blaney, D., Cabrol, N.A., Christensen, P.R., Clark, B.C., Crisp, J.A., Des

Marais, D.J., Economou, T., Farmer, J.D., Farrand, W., Ghosh, A., Golombek, M.,

Gorevan, S., Greeley, R., Hamilton, V.E., Johnson, J.R., Joliff, B.L., Klingelhofer, G.,

Knudson, A.T., McLennan, S., Ming, D., Moersch, J.E., Rieder, R., Ruff, S.W.,

Schroder, C., de Souza, P.A., Squyres, S.W., Wanke, H., Wang, A., Yen, A. and

Zipfel, J., 2006. Characterization and petrologic interpretation of olivine-rich basalts at

Gusev Crater, Mars. J Geophys Res-Planet, 111(E2).

Mege, D. and Masson, P., 1996a. Amounts of crustal stretching in Valles Marineris, Mars.

Planetary and Space Science, 44(8): 749-+.

Mege, D. and Masson, P., 1996b. A plume tectonics model for the Tharsis province, Mars.

Planetary and Space Science, 44(12): 1499-1546.

Mehta, M., Renno, N.O., Marshall, J., Grover, M.R., Sengupta, A., Rusche, N.A., Kok, J.F.,

Arvidson, R.E., Markiewicz, W.J., Lemmon, M.T. and Smith, P.H., 2011. Explosive

erosion during the Phoenix landing exposes subsurface water on Mars. Icarus, 211(1):

172-194.

Meresse, S., Costard, F., Mangold, N., Masson, P., Neukum, G. and Team, H.C.-I., 2008.

Formation and evolution of the chaotic terrains by subsidence and magmatism:

Hydraotes Chaos, Mars. Icarus, 194(2): 487-500.

Michalski, J.R., Bibring, J.P., Poulet, F., Loizeau, D., Mangold, N., Dobrea, E.N., Bishop,

J.L., Wray, J.J., McKeown, N.K., Parente, M., Hauber, E., Altieri, F., Carrozzo, F.G.

237 and Niles, P.B., 2010. The Mawrth Vallis Region of Mars: A Potential Landing Site

for the (MSL) Mission. Astrobiology, 10(7): 687-703.

Michalski, J.R. and Bleacher, J.E., 2013. Supervolcanoes within an ancient volcanic province

in Arabia Terra, Mars. Nature, 502(7469): 47-+.

Michalski, J.R., Niles, P.B., Cuadros, J. and Baldridge, A.M., 2013. Multiple working

hypotheses for the formation of compositional stratigraphy on Mars: Insights from the

Mawrth Vallis region. Icarus, 226(1): 816-840.

Michaut, C., Baratoux, D. and Thorey, C., 2013. Magmatic intrusions and deglaciation at

mid-latitude in the northern plains of Mars. Icarus, 225(1): 602-613.

Monteux, J., Coltice, N., Dubuffet, F. and Ricard, Y., 2007. Thermo-mechanical adjustment

after impacts during planetary growth. Geophys Res Lett, 34(24).

Morgan, G.A., Head, J.W. and Marchant, D.R., 2009. (LVF) and lobate

debris aprons (LDA) in the Deuteronilus Mensae northern dichotomy boundary

region, Mars: Constraints on the extent, age and episodicity of Amazonian glacial

events. Icarus, 202(1): 22-38.

Morgan, G.A., Head, J.W. and Marchant, D.R., 2011. Preservation of Late Amazonian Mars

ice and water-related deposits in a unique crater environment in Noachis Terra: Age

relationships between lobate debris tongues and gullies. Icarus, 211(1): 347-365.

Morishima, R., Stadel, J. and Moore, B., 2010. From planetesimals to terrestrial planets: N-

body simulations including the effects of nebular gas and giant planets. Icarus, 207(2):

517-535.

Murray, J.B., de Vries, B.V., Marquez, A., Williams, D.A., Byrne, P., Muller, J.P. and Kim,

J.R., 2010. Late-stage water eruptions from Ascraeus Mons volcano, Mars:

Implications for its structure and history. Earth Planet Sc Lett, 294(3-4): 479-491.

Neukum, G. and Hiller, K., 1981. Martian Ages. J Geophys Res, 86(Nb4): 3097-3121.

238 Neukum, G., Jaumann, R., Hoffmann, H., Hauber, E., Head, J.W., Basilevsky, A.T., Ivanov,

B.A., Werner, S.C., van Gasselt, S., Murray, J.B., McCord, T. and Team, H.C.-i.,

2004. Recent and episodic volcanic and glacial activity on Mars revealed by the High

Resolution Stereo Camera. Nature, 432(7020): 971-979.

Neumann, G.A., Zuber, M.T., Wieczorek, M.A., McGovern, P.J., Lemoine, F.G. and Smith,

D.E., 2004. Crustal structure of Mars from gravity and topography. J Geophys Res-

Planet, 109(E8).

Newsom, H.E., Lanza, N.L., Ollila, A.M., Wiseman, S.M., Roush, T.L., Marzo, G.A.,

Tornabene, L.L., Okubo, C.H., Osterloo, M.M., Hamilton, V.E. and Crumpler, L.S.,

2010. Inverted channel deposits on the floor of Miyamoto crater, Mars. Icarus, 205(1):

64-72.

Nimmo, F. and Agnor, C.B., 2006. Isotopic outcomes of N-body accretion simulations:

Constraints on equilibration processes during large impacts from HfAV observations.

Earth Planet Sc Lett, 243(1-2): 26-43.

Nimmo, F., Hart, S.D., Korycansky, D.G. and Agnor, C.B., 2008. Implications of an impact

origin for the martian hemispheric dichotomy. Nature, 453(7199): 1220-U1232.

Nimmo, F. and Kleine, T., 2007. How rapidly did Mars accrete? Uncertainties in the Hf-W

timing of core formation. Icarus, 191(2): 497-504.

O'Neill, C., A. M. Jellinek and A. Lenardic, 2007. Conditions for the onset of plate tectonics

on terrestrial planets and . Earth and Planetary Science Letters 261(1-2): 20-32.

O'Rourke, J.G. and Korenaga, J., 2012. Terrestrial planet evolution in the stagnant-lid regime:

Size effects and the formation of self-destabilizing crust. Icarus, 221(2): 1043-1060.

Ockert-Bell, M.E., Clark, B.E., Shepard, M.K., Isaacs, R.A., Cloutis, E.A., Fornasier, S. and

Bus, S.J., 2010. The composition of M-type asteroids: Synthesis of spectroscopic and

radar observations. Icarus, 210(2): 674-692.

239 Oehler, D.Z. and Allen, C.C., 2010. Evidence for pervasive mud volcanism in Acidalia

Planitia, Mars. Icarus, 208(2): 636-657.

Ohman, T. and McGovern, P.J., 2014. Circumferential graben and the structural evolution of

Alba Mons, Mars. Icarus, 233: 114-125.

Okubo, C.H. and Martel, S.J., 1998. Pit crater formation on Kilauea volcano, Hawaii. J

Volcanol Geoth Res, 86(1-4): 1-18.

Oze, C. and Sharma, M., 2007. Serpentinization and the inorganic synthesis of H-2 in

planetary surfaces. Icarus, 186(2): 557-561.

Pacifici, A., Komatsu, G. and Pondrelli, M., 2009. Geological evolution of Ares Vallis on

Mars: Formation by multiple events of catastrophic flooding, glacial and periglacial

processes. Icarus, 202(1): 60-77.

Parker, T.J., Saunders, R.S. and Schneeberger, D.M., 1989. Transitional Morphology in West

Deuteronilus Mensae, Mars - Implications for Modification of the Lowland Upland

Boundary. Icarus, 82(1): 111-145.

Pasckert, J.H., Hiesinger, H. and Reiss, D., 2012. Rheologies and ages of lava flows on

Elysium Mons, Mars. Icarus, 219(1): 443-457.

Penido, J.C., Fassett, C.I. and Som, S.M., 2013. Scaling relationships and concavity of small

valley networks on Mars. Planetary and Space Science, 75: 105-116.

Peterson, D.W., and Swanson, D.A., 1974. Observed formation of lava tubes during 1970-

1971 at Kilauea Volcano, Hawaii. Stud. Speleol., 2: 209-222.

Pierazzo, E. and Melosh, H.J., 2000. Melt production in oblique impacts. Icarus, 145(1): 252-

261.

Pinkerton, H. and Wilson, L., 1994. Factors Controlling the Lengths of Channel-Fed Lava

Flows. B Volcanol, 56(2): 108-120.

240 Platz, T., Michael, G., Tanaka, K.L., Skinner, J.A. and Fortezzo, C.M., 2013. Crater-based

dating of geological units on Mars: Methods and application for the new global

geological map. Icarus, 225(1): 806-827.

Plescia, J.B., 1994. Geology of the Small Tharsis Volcanos - Jovis-Tholus, Ulysses-Patera,

Biblis-Patera, Mars. Icarus, 111(1): 246-269.

Plescia, J.B., 2004. Morphometric properties of Martian volcanoes. J Geophys Res-Planet,

109(E3).

Poulet, F., Gomez, C., Bibring, J.P., Langevin, Y., Gondet, B., Pinet, P., Belluci, G. and

Mustard, J., 2007. Martian surface mineralogy from Observatoire pour la Mineralogie,

l'Eau, les Glaces et l'Activite on board the Mars Express spacecraft (OMEGA/MEx):

Global mineral maps. J Geophys Res-Planet, 112(E8).

Pozzo, M., Davies, C., Gubbins, D. and Alfe, D., 2012. Thermal and electrical conductivity of

iron at Earth's core conditions. Nature, 485(7398): 355-U399.

Ranalli, G. (1995), Rheology of the Earth, 2nd ed., 413 pp., Chapman and Hall, London.

Reese, C.C., Orth, C.P. and Solomatov, V.S., 2011. Impact megadomes and the origin of the

martian crustal dichotomy. Icarus, 213(2): 433-442.

Reese, C.C. and Solomatov, V.S., 2006. Fluid dynamics of local martian magma oceans.

Icarus, 184(1): 102-120.

Reese, C.C. and Solomatov, V.S., 2010. Early martian dynamo generation due to giant

impacts. Icarus, 207(1): 82-97.

Richardson, J.A., Bleacher, J.E. and Glaze, L.S., 2013. The volcanic history of Syria Planum,

Mars. J Volcanol Geoth Res, 252: 1-13.

Righter, K. and Shearer, C.K., 2003. Magmatic fractionation of Hf and W: Constraints on the

timing of core formation and differentiation in the Moon and Mars. Geochim

Cosmochim Ac, 67(13): 2497-2507.

241 Robbins, S.J., Di Achille, G. and Hynek, B.M., 2011. The volcanic history of Mars: High-

resolution crater-based studies of the calderas of 20 volcanoes. Icarus, 211(2): 1179-

1203.

Robbins, S.J., Hynek, B.M., Lillis, R.J. and Bottke, W.F., 2013. Large impact crater histories

of Mars: The effect of different model crater age techniques. Icarus, 225(1): 173-184.

Roberts, J.H. and Zhong, S.J., 2006. Degree-1 convection in the Martian mantle and the origin

of the hemispheric dichotomy. J Geophys Res-Planet, 111(E6).

Roberts, J.H., Lillis, R.J. and Manga, M., 2009. Giant impacts on early Mars and the cessation

of the Martian dynamo. J Geophys Res-Planet, 114.

Rogers, A.D. and Bandfield, J.L., 2009. Mineralogical characterization of Mars Science

Laboratory candidate landing sites from THEMIS and TES data. Icarus 203, 437-453.

Rogers, A.D. and Nazarian, A.H., 2013. Evidence for Noachian flood volcanism in Noachis

Terra, Mars, and the possible role of Hellas impact basin tectonics. J Geophys Res-

Planet, 118(5): 1094-1113.

Ruiz, J., Lopez, V. and Dohm, J.M., 2010. The present-day thermal state of Mars. Icarus,

207(2): 631-637.

Russell, P.S. and Head, J.W., 2007. The Martian hydrologic system: Multiple recharge centers

at large volcanic provinces and the contribution of snowmelt to outflow channel

activity. Planetary and Space Science, 55(3): 315-332.

Salvatore, M.R., Mustard, J.F., Wyatt, M.B. and Murchie, S.L., 2010. Definitive evidence of

Hesperian basalt in Acidalia and Chryse planitiae. J Geophys Res-Planet, 115.

Scambos, T.A. and Jakosky, B.M., 1990. An Outgassing Release Factor for Nonradiogenic

Volatiles on Mars. J Geophys Res-Solid, 95(B9): 14779-14787.

Schenato, F., Formoso, M.L.L., Dudoignon, P., Meunier, A., , D. and Mas, A., 2003.

Alteration processes of a thick basaltic lava flow of the Parana Basin (Brazil):

petrographic and mineralogical studies. J S Am Earth Sci, 16(5): 423-444.

242 Schmeling, H., et al. (2008), A benchmark comparison of spontaneous subduction models--

Towards a free surface, Phys. Earth Planet. Inter. (Netherlands), 171(1-4), 198-223.

Schultz, P.H., 1976. Floor-Fractured . Moon Planets, 15(3-4): 241-273.

Schultz, P.H. and Glicken, H., 1979. Impact Crater and Basin Control of Igneous Processes

on Mars. J Geophys Res, 84: 8033-8047.

Schultz, R.A., 1991. Structural Development of Coprates Chasma and Western Ophir Planum,

Valles Marineris Rift, Mars. J Geophys Res-Planet, 96(E5): 22777-22792.

Scott, D. H., Carr, M. H., 1978. Geologic map of Mars. In: of Mars, 1:25,000,000

Geologic Series M25M I-1083. U.S. Geol. Surv., Reston, VA.

Scott, D.H. and Tanaka, K.L., 1982. Ignimbrites of Amazonis-Planitia Region of Mars. J

Geophys Res, 87(Nb2): 1179-1190.

Scott, D. H., Tanaka, K. L., 1986. Geologic map of the Western Equatorial region of Mars. In:

Atlas of Mars, 1:15,000,000 Geologic Series M15M I-1802-A. U.S. Geol. Surv.,

Reston, VA.

Scott, D. H.; Dohm, J.M.; Zimbelman, J.R., 1998. Geologic Maps of Pavonis Mons, Mars.

USGS, I-2561.

Scott, E.D. and Wilson, L., 2002. Plinian eruptions and passive collapse events as

mechanisms of formation for Martian pit chain craters. J Geophys Res-Planet,

107(E4).

Scott, E.D., Wilson, L. and Head, J.W., 2002. Emplacement of giant radial dikes in the

northern Tharsis region of Mars. J Geophys Res-Planet, 107(E4).

Senshu, H., Kuramoto, K. and Matsui, T., 2002. Thermal evolution of a growing Mars. J

Geophys Res-Planet, 107(E12).

Sharp, R.P., 1973. Mars - South Polar Pits and Etched Terrain. J Geophys Res, 78(20): 4222-

4230.

243 Shepard, M.K., Clark, B.E., Ockert-Bell, M., Nolan, M.C., Howell, E.S., Magri, C., Giorgini,

J.D., Benner, L.A.M., Ostro, S.J., Harris, A.W., Warner, B.D., Stephens, R.D. and

Mueller, M., 2010. A radar survey of M- and X-class asteroids II. Summary and

synthesis. Icarus, 208(1): 221-237.

Solomon, S.C., Aharonson, O., Aurnou, J.M., Banerdt, W.B., Carr, M.H., Dombard, A.J.,

Frey, H.V., Golombek, M.P., Hauck, S.A., Head, J.W., Jakosky, B.M., Johnson, C.L.,

McGovern, P.J., Neumann, G.A., Phillips, R.J., Smith, D.E. and Zuber, M.T., 2005.

New perspectives on ancient Mars. Science, 307(5713): 1214-1220.

Som, S.M., Montgomery, D.R. and Greenberg, H.M., 2009. Scaling relations for large

Martian valleys. J Geophys Res-Planet, 114.

Spagnuolo, M.G., Figueredo, P.H. and Ramos, V.A., 2008. Reinterpretation of Tractus Fossae

region as an asymmetric rift system on Mars. Icarus, 198(2): 318-330.

Spencer, J.R. and Fanale, F.P., 1990. New Models for the Origin of Valles-Marineris Closed

Depressions. J Geophys Res-Solid, 95(B9): 14301-14313.

Squyres, S.W., 1978. Martian Fretted Terrain - Flow of Erosional Debris. Icarus, 34(3): 600-

613.

Sramek, O., Milelli, L., Ricard, Y. and Labrosse, S., 2012. Thermal evolution and

differentiation of planetesimals and planetary embryos. Icarus, 217(1): 339-354.

Sramek, O., Ricard, Y. and Dubuffet, F., 2010. A multiphase model of core formation.

Geophys J Int, 181(1): 198-220.

Stevenson, D.J., 2001. Mars' core and magnetism. Nature, 412(6843): 214-219.

Stevenson, D.J., 2003. Mission to Earth's core - a modest proposal. Nature, 423(6937): 239-

240.

Stewart, E.M. and Head, J.W., 2001. Ancient Martian volcanoes in the Aeolis region: New

evidence from MOLA data. J Geophys Res-Planet, 106(E8): 17505-17513.

244 Stewart, S.T. and Leinhardt, Z.M., 2012. Collisions between Gravity-Dominated Bodies. Ii.

The Diversity of Impact Outcomes during the End Stage of Planet Formation.

Astrophys J, 751(1).

Tackley, P.J., 2008. Modelling compressible mantle convection with large viscosity contrasts

in a three-dimensional spherical shell using the yin-yang grid. Phys Earth Planet In,

171(1-4): 7-18.

Tanaka, K.L., 1997. Origin of Valles Marineris and Noctis Labyrinthus, Mars, by structurally

controlled collapse and erosion of crustal materials. XXVIII Lunar and Planetary

Science Conference, Houston, #1169.

Tanaka, K.L., 1997. Sedimentary history and mass flow structures of Chryse and Acidalia

Planitiae, Mars. J Geophys Res-Planet, 102(E2): 4131-4149.

Tanaka, K.L. and Kolb, E.J., 2001. Geologic history of the polar regions of Mars based on

Mars global surveyor data I. Noachian and Hesperian Periods. Icarus, 154(1): 3-21.

Tanaka, K.L., Robbins, S.J., Fortezzo, C.M., Skinner, J.A. and Hare, T.M., 2013. The digital

global geologic map of Mars: Chronostratigraphic ages, topographic and crater

morphologic characteristics, and updated resurfacing history. Planetary and Space

Science(0).

Tanaka, K.L., Robbins, S.J., Fortezzo, C.M., Skinner, J.A. and Hare, T.M., 2014. The digital

global geologic map of Mars: Chronostratigraphic ages, topographic and crater

morphologic characteristics, and updated resurfacing history. Planetary and Space

Science, 95: 11-24.

Tanaka, K.L., Rodriguez, J.A.P., Skinner, J.A., Bourke, M.C., Fortezzo, C.M., Herkenhoff,

K.E., Kolb, E.J. and Okubo, C.H., 2008. North polar region of Mars: Advances in

stratigraphy, structure, and erosional modification. Icarus, 196(2): 318-358.

Teiser, J., Kupper, M. and Wurm, G., 2011. Impact angle influence in high velocity dust

collisions during planetesimal formation. Icarus, 215(2): 596-598.

245 Terasaki, H., , D.J., Rubie, D.C. and Langenhorst, F., 2005. The effect of oxygen and

sulphur on the dihedral angle between Fe-O-S melt and silicate minerals at high

pressure: Implications for Martian core formation. Earth Planet Sc Lett, 232(3-4): 379-

392.

Thommes, E., Nagasawa, M. and Lin, D.N.C., 2008. Dynamical shake-up of planetary

systems. II. N-body simulations of solar system terrestrial planet formation induced by

secular resonance sweeping. Astrophys J, 676(1): 728-739.

Tonks, W.B. and Melosh, H.J., 1993. Magma Ocean Formation Due to Giant Impacts. J

Geophys Res-Planet, 98(E3): 5319-5333.

Villanueva, G.L., Mumma, M.J., Novak, R.E., Kaufl, H.U., Hartogh, P., Encrenaz, T.,

Tokunaga, A., Khayat, A. and Smith, M.D., 2015. Strong water isotopic anomalies in

the martian atmosphere: Probing current and ancient reservoirs. Science, 348(6231):

218-221.

Wadge, G., 1978. Effusion Rate and Shape of Aa Lava Flow-Fields on Mount Etna. Geology,

6(8): 503-506.

Walker, G.P.L., 1973. Lengths of Lava Flows. Philos T R Soc A, 274(1238): 107-118.

Ward, A.W., 1979. Yardangs on Mars - Evidence of Recent Wind Erosion. J Geophys Res,

84: 8147-8166.

Ward, A.W. and Greeley, R., 1984. Evolution of the Yardangs at Rogers Lake, .

Geol Soc Am Bull, 95(7): 829-837.

Watters, T.R., McGovern, P.J. and Irwin, R.P., 2007. Hemispheres apart: The crustal

dichotomy on Mars. Annu Rev Earth Pl Sc, 35: 621-652.

Webb, V.E., 2004. Putative shorelines in northern Arabia Terra, Mars. J Geophys Res-Planet,

109(E9).

Weitz, C.M., Milliken, R.E., Grant, J.A., McEwen, A.S., Williams, R.M.E., Bishop, J.L. and

Thomson, B.J., 2010. Mars Reconnaissance Orbiter observations of light-toned

246 layered deposits and associated fluvial landforms on the plateaus adjacent to Valles

Marineris. Icarus, 205(1): 73-102.

Werner, S.C., 2009. The global martian volcanic evolutionary history. Icarus, 201(1): 44-68.

Whaler, K.A. and Purucker, M.E., 2005. A spatially continuous magnetization model for

Mars. J Geophys Res-Planet, 110(E9).

Wilhelms, D.E., 1974. Comparison of Martian and Lunar Geologic Provinces. J Geophys Res,

79(26): 3933-3941.

Wilhelms, D.E. and Squyres, S.W., 1984. The Martian Hemispheric Dichotomy May Be Due

to a Giant Impact. Nature, 309(5964): 138-140.

Wilkins, S.J. and Schultz, R.A., 2003. Cross faults in extensional settings: Stress triggering,

displacement localization, and implications for the origin of blunt troughs at Valles

Marineris, Mars. J Geophys Res-Planet, 108(E6).

Williams, D.A., Greeley, R., Fergason, R.L., Kuzmin, R., McCord, T.B., Combe, J.P., Head,

J.W., Xiao, L., Manfredi, L., Poulet, F., Pinet, P., Baratoux, D., Plaut, J.J., Raitala, J.,

Neukum, G. and Team, H.C.-I., 2009. The Circum-Hellas Volcanic Province, Mars:

Overview. Planetary and Space Science, 57(8-9): 895-916.

Williams, J.P. and Nimmo, F., 2004. Thermal evolution of the Martian core: Implications for

an early dynamo. Geology, 32(2): 97-100.

Wilson, L. and Head, J.W., 1994. Mars - Review and Analysis of Volcanic-Eruption Theory

and Relationships to Observed Landforms. Rev Geophys, 32(3): 221-263.

Wilson, L. and Head, J.W., 2002. Tharsis-radial graben systems as the surface manifestation

of plume-related dike intrusion complexes: Models and implications - art. no. 5000. J

Geophys Res-Planet, 107(E8).

Wise, D.U., Golombek, M.P. and Mcgill, G.E., 1979. Tectonic Evolution of Mars. J Geophys

Res, 84: 7934-7939.

247 Wordsworth, R. D., L. Kerber, R. T. Pierrehumbert, F. Forget, and J. W. Head, 2015.

Comparison of “warm and wet” and “cold and icy” scenarios for early Mars in a 3-D

climate model, J. Geophys. Res. Planets, 120, 1201–1219.

Wray, J.J., Hansen, S.T., Dufek, J., Swayze, G.A., Murchie, S.L., Seelos, F.P., Skok, J.R.,

Irwin, R.P. and Ghiorso, M.S., 2013. Prolonged magmatic activity on Mars inferred

from the detection of felsic rocks. Nat Geosci, 6(12): 1013-1017.

Wyrick, D., Ferrill, D.A., Morris, A.P., Colton, S.L. and Sims, D.W., 2004. Distribution,

morphology, and origins of Martian pit crater chains. J Geophys Res-Planet, 109(E6).

Xiao, L., Huang, J., Christensen, P.R., Greeley, R., Williams, D.A., Zhao, J.N. and He, Q.,

2012. Ancient volcanism and its implication for thermal evolution of Mars. Earth

Planet Sc Lett, 323: 9-18.

Yin, Q.Z., Jacobsen, S.B., Yamashita, K., Blichert-Toft, J., Telouk, P. and Albarede, F., 2002.

A short timescale for terrestrial planet formation from Hf-W chronometry of

meteorites. Nature, 418(6901): 949-952.

Zharkov, V.N., Gudkova, T.V. and Molodensky, S.M., 2009. On models of Mars' interior and

amplitudes of forced nutations 1. The effects of deviation of Mars from its equilibrium

state on the flattening of the core-mantle boundary. Phys Earth Planet In, 172(3-4):

324-334.

Zhong, S.J., 2009. Migration of Tharsis volcanism on Mars caused by differential rotation of

the lithosphere. Nat Geosci, 2(1): 19-23.

Zhong, S.J. and Zuber, M.T., 2001. Degree-1 mantle convection and the crustal dichotomy on

Mars. Earth Planet Sc Lett, 189(1-2): 75-84.

Zimbelman, J.R. and Griffin, L.J., 2010. HiRISE images of yardangs and sinuous ridges in

the lower member of the Medusae Fossae Formation, Mars. Icarus, 205(1): 198-210.

248

ETH Zurich [email protected] Department of Earth Sciences Phone: +41 44 6328349 Sonneggstrasse 5 Mobile: +41 77 4059186 Zurich, Zurich CH-8092 Website: Switzerland http://www.geophysics.ethz.ch/people/all_people/leoneg Giovanni Leone https://www.researchgate.net/profile/Giovanni_Leone/

Education

Jun 2011 – Feb 2016 ETH Zurich PhD, Earth Sciences Zurich, Switzerland Continuum Mechanics, Basics and Principles of Radar Remote Sensing for Environmental Applications, Numerical Modelling in Fortran Sep 1996 – Aug Lancaster University 2007 PhD, Environmental (Planetary) Science Parental leave from Lancaster, United Kingdom Mar 1998 – Aug Planetary Dynamics, Remote Sensing. Transcripts were not issued, 2001 courses were not compulsory for the award of the PhD. Sep 1986 – Jul 1993 Università degli Studi di Palermo M.Sc., Geological Sciences. B.Sc included in M.Sc. in Italian universities. Palermo, Italy Elements of Mathematics, Chemistry, Experimental Physics, Mineralogy, Geology, Topography and Cartography, Petrography (Petrology),

249 Geological Survey, Palaeontology, Geography, Physical Geography, Geodynamics, Geotechnical Engineering, Applied Geophysics, Mineral Physics, Earth Physics. Additional training: - USGS, Astrogeology Center Flagstaff, Training Course with HiRISE imagery on ISIS – SOCET SET and ArcGIS software for Digital Terrain Models production, Sep 2013. - ETH Zurich, Training Course of Computational Magma Geodynamics, Feb 2014; Learning to Teach, May 2016.

Theses

Giovanni Leone: The Southern Polar Giant Impact hypothesis for the origin of the Martian dichotomy and the evolution of volcanism on Mars. 02/2016, ETH Zurich. Degree: PhD, Supervisor: Prof. Paul James Tackley Giovanni Leone: The Internal Structure of Io's Lithosphere and Its Influence on Volcanic Eruption Styles. 08/2007, Lancaster University. Degree: PhD, Supervisor: Prof. Lionel Wilson Giovanni Leone: Vertical Seismic Profiling in a Landslide Area of the Greek Temples’ Valley, Agrigento (Sicily). 07/1993, Universitá degli Studi di Palermo. Degree: M.Sc., Supervisor: Prof. Dario Luzio

Research Experience

Jun 2011 – Feb PhD Position 2016 Swiss Federal Institute of Technology, Department of Earth Sciences, Institute of Geophysics. Zurich, Switzerland Member of Prof. Paul Tackley’s Geophysical Fluid Dynamics group and member of the Geophysics Institute Teaching Committee at ETH Zurich. Worked at the origin of the Martian dichotomy through numerical modeling validated with a geologic and volcanologic study of the surface of Mars to investigate an alternative process of formation that involved a giant impact on the South Pole with a lunar sized impactor of 1600 km of radius with a 80% iron (by radius) fraction, or 2000 km radius with a 50% iron fraction, hitting at a speed of 5 km/s (the escape velocity of Mars), melting much of the interior and 1/2 of the planetary surface with the creation of a magma ocean that formed the highlands upon cooling and solidification. Using a combination of I3ELVIS (immediate post-impact and core formation) and STAGYY (long- term) C and Fortran thermo-mechanical codes, respectively. I studied the long-term consequences of such a giant impact: 1) thermal and compositional effect on core formation; 2) triggering of a transient magnetic field, traces of which have been detected on both the hemispheres; 3) migration of mantle plumes from the southern polar region to the equator following preferential paths, tracked through volcanic alignments on the surface; 4) identification of volcanic features at global scale on Mars. Sep 1996 – Aug PhD Position 2006 Lancaster University

250 Parental leave Lancaster, UK from Mar 1998 – Worked with Prof. Lionel Wilson, in collaboration with NASA Aug 2001 JPL, under the NASA and Geophysics Program, for the study of eruption processes on Io from Galileo NIMS, SSI and Photo-Polarimeter-Radiometer (PPR) data and estimates of material volume during the October 1999 episode at Prometheus which resulted in: 1) production of a detailed geological map of the Prometheus Patera Area with subsequent 2) determination of conduit geometry and magma chamber depth and size, 3) determination of the global average geothermal gradient through numerical modelling of the lithosphere using a finite- difference analysis. Contributed, as JGR referee, to review the detailed examination of the proposed Loki Patera “magma sea” and the modelling of lava flows and lakes after solidification with a finite-difference analysis model. Sep 2001 – Sep Assegno di Ricerca (Italian postdoctoral activity of research) 2002 Università del Salento, Department of Mathematics and Physics "Ennio De Giorgi" Lecce, Italy Analysed Martian putative paleolakes and estimated salts sedimentation preliminary to the A mission “Spirit” (see Grin et al., 2002 in the Conference Proceedings Section below). Mar 1997 – Mar Visiting Researcher 1998 Università degli Studi di Palermo, Institute of Physics, Department of Mathematical, Physical and Natural Sciences. Palermo, Italy Used SPH (Smoothed Particle Hydrodynamics) numerical modeling techniques for simulation of planetary systems origins.

University Committees

Nov 2012 – Nov Member of Teaching Committee 2014 Department of Earth Sciences, ETH Zurich Zurich, Switzerland

Teaching Experience

Sep 1996 – Jan 1997 Demonstrator, Lab Assistant Lancaster University Lancaster, UK Assisted Prof. Jennie in Geological Processes Lab for undergraduate and graduate students exercises and rock specimen identification. Sep 2007 – Jun 2008 Teacher High School “Duca degli Abruzzi” Palermo, Italy

251 Taught Planetary Geology and coordinated team of teachers, assisted students in identification of geological and volcanological features in images released by planetary missions and in development of their own research project. Sep 2008 – Jul 2010 Teacher Primary School “N. Botta” Cefalú (Palermo), Italy Taught Lab of Planetary Science, Natural and Environmental Sciences to K-10, coordinated team of teachers and assisted students in practice of basic scientific experiments of chemistry and physics.

Awards & Grants

Mar 2008 Grant: EU Fund PON 2007-2013 - Scientific Lab in School

Skills & Activities

Skills Geological Mapping, Satellite Image Analysis and Interpretation, Numerical Modeling (Fortran, Matlab, Finite Differences) of Volcanic Processes, Seismic and Electric Prospections, Sampling and Preparation of Thin Sections, Petrologic Analysis through Polarized Microscope. Languages English (fluent), Italian (mother tongue), German and French (beginner). Scientific AGU (1996 – 2000; 2014) - JGR and Icarus referee, IAVCEI (Lead Memberships and Convenor, 2010 – 2011), EGU, The Planetary Society (Regional activities Coordinator, 2007 - 2014), Societá Geologica Italiana – Sezione di Geologia Planetaria (2014 – today), Societá Geologica Italiana 88th Meeting Naples 2016 (Convenor, 2015 – 2016), NASA Postdoctoral Program Reviewer (2016 –). Six new names of volcanoes on Mars approved by the IAU WGPSN in year 2013: Aonia Mons, Aonia Tholus, Electris Mons, Eridania Mons, Sirenum Mons, Sirenum Tholus (Gazetteer of Planetary Nomenclature, 4 and 12 August 2013). Interests Identification of volcanic processes and geophysical modelling of planetary surfaces and interiors.

Numerical modeling of magma rise and internal structure of terrestrial planets and volcanic moons, cryovolcanism of icy satellites.

Geologic and thematic mapping of planetary surfaces.

Geomorphometric analysis of lava flows, volcanoes, and other geologic features such as fluvial valleys, paleolakes, and outflow channels.

Scientific outreach of terrestrial and planetary comparative volcanology, life-rock interactions, life in extreme environments (i.e. volcanic or glacial environments) and evaluation of potential exobiological landing sites.

252 Science Outreach

Oct 1986 Exhibit of planetary photographs taken through small telescopes High School Palermo, Italy Sep 1993 – Jul 1996 Co-author of planetary science TV programs TSK, Canale 21 Palermo, Italy Co-authored two TV programs entitled ”A come Astronomia” and “Nova” based on the hot topics in astronomy and planetary science. Been on air to answer questions of the audience, exobiology issues the most asked. Nov 2007 Public talk on the history of Solar System exploration City Hall Library Isola delle Femmine (Palermo), Italy Sep 2008 Invited scientific talk: Io’s geology, volcanology, and new perspectives for future exploration NASA JPL Pasadena, USA Sep 2010 Talk on the volcanism in the Solar System Department of Physics, Universitá degli Studi di Palermo Palermo, Italy Jul 2014 Scientific talk: Life on Mars? Department of Earth Sciences, ETH Zurich Zurich, Switzerland May 2014 Against the current with lava flows ETH News https://www.ethz.ch/en/news-and-events/eth-news/news/2014/05/mit- lavafluessen-gegen-den-strom.html Zurich, Switzerland May 2014 Researcher finds lava, not water, flowed on Mars Swissinfo.ch http://www.swissinfo.ch/eng/researcher-finds-lava--not-water--flowed- on-mars/38563356 Bern, Switzerland Jun 2014 Water didn’t carve Mars canyons, LAVA did: Liquid rock melted massive gorges on red planet’s surface, study reveals. Daily Mail http://www.dailymail.co.uk/sciencetech/article-2647295/Water-didnt- carve-Mars-canyons-LAVA-did-Liquid-rock-melted-massive-gorges- planets-surface-study-reveals.html London, UK Jul 2014 Zürcher Marsforscher schwimmt gegen den Strom (A Mars researcher from Zurich swims against the stream) Tages Anzeiger http://www.tagesanzeiger.ch/wissen/technik/Zuercher-Marsforscher- schwimmt-gegen-den-Strom/story/21672484

253 Zurich, Switzerland Jan 2015 The two faces of Mars ETH News https://www.ethz.ch/en/news-and-events/eth-news/news/2015/01/two- faces-of-mars.html Zurich, Switzerland Jan 2015 ETH präsentiert neue Theorie zum Mars Tages Anzeiger http://www.tagesanzeiger.ch/wissen/natur/ETH-praesentiert-neue- Theorie-zum-Mars/story/26942921 Zurich, Switzerland

Jan 2015 Einschlag auf den Südpol verformte den Mars Wiener Zeitung http://www.wienerzeitung.at/themen_channel/wissen/natur/731355_Einsc hlag-auf-den-Suedpol-verformte-den-Mars.html Vienna, Austria

Feb 2015 Giant Asteroid Collision May Have Radically Transformed Mars Scientific American http://www.scientificamerican.com/article/giant-asteroid-collision-may- have-radically-transformed-mars/ New York, USA Feb 2015 Mars verslond ooit een maan (Mars ever devoured a moon) De Standaard http://www.standaard.be/cnt/dmf20150201_01504972 Brussels, Belgium

Mar 2015 Einschlag liess Mars zur Haelfte schmelzen (Impact allowed to melt half Mars) Die Welt http://www.welt.de/wissenschaft/weltraum/article138455198/Einschlag- liess-Mars-zur-Haelfte-schmelzen.html Berlin, Germany Aug 2015 Der leblose Planet Süddeutsche Zeitung http://www.sueddeutsche.de/wissen/der-rote-planet-leblos-1.2596706-2 München, Germany Aug 2015 Gibt es Wasser auf dem Mars? Verband Schweizerischer Assistenz und Oberärtzinnen und – ärtze (VSAO) Journal, vol. 4, pag. 45-47. Bern, Switzerland Sep 2015 Water on Mars? NASA reveals briny flows on the surface – as it happened The Guardian (15:38, click on the view 23 more updates button) http://www.theguardian.com/science/across-the- universe/live/2015/sep/28/water-on-mars-buildup-to--mystery- solved-announcement-live#comments

254 London, UK Oct 2015 Invited scientific talk: Vulcanismo nel Sistema Solare Aula Strozzi, Museum of Natural History Florence, Italy

Publication Highlights

Giovanni Leone: Alignments of Volcanic Features in the Southern Hemisphere of Mars produced by Migrating Mantle Plumes. Journal of Volcanology and Geothermal Research, 01/2016: vol. 309, 78-95. Giovanni Leone, Paul James Tackley, Taras Gerya, Alexander May, Guizhi Zhu: Three-Dimensional Simulations of the Southern Polar Giant Impact Hypothesis for the Origin of the Martian Dichotomy. Geophysical Research Letters 12/2014: 41(24), 8736- 8743- doi:10.1002/2014GL062261. (Solicited paper) Giovanni Leone: A network of lava tubes as the origin of Labyrinthus Noctis and Valles Marineris on Mars. Journal of Volcanology and Geothermal Research 05/2014; 277, 1- 8. Giovanni Leone, Lionel Wilson, Ashley Gerard Davies: The geothermal gradient of Io: Consequences for lithosphere structure and volcanic eruptive activity. Icarus 01/2011; 211(1): 623-635. Giovanni Leone, Lionel Wilson Ashley Gerard Davies: Sulphur from below. Nature Geoscience 11/2010; 3:818. Giovanni Leone, Ashley Gerard Davies, Lionel Wilson, David A. Williams, Laszlo P. Keszthelyi, Windy L. Jaeger, Elizabeth P. Turtle: Volcanic history, geologic analysis and map of the Prometheus Patera region on Io. Journal of Volcanology and Geothermal Research 08/2009; 187(1-2): 93-105.

Book Chapters

Giovanni Leone: Transition Topography (Mars). Encyclopedia of Planetary Landforms, Springer Ed. 2015, pp. 2169-2173. Gino Erkeling, Giovanni Leone, Ákos Kereszturi: Paleoshorelines. Encyclopedia of Planetary Landforms, Springer Ed. 2015, pp. 1501-1507. Kathleen Mandt, Giovanni Leone: Yardang. Encyclopedia of Planetary Landforms, Springer Ed. 2015, pp. 2339-2347.

Peer Reviewed Journal Publications

Giovanni Leone: Mineralogical and volumetric evidence for the predominant role of lava and the scarcity of liquid water in erosional, depositional, and weathering processes on the surface of Mars since the pre-Noachian. (manuscript submitted to Precambrian Research).

255 Giovanni Leone: Mangala Valles, Mars: a reassessment of formation processes based on a new geomorphological and stratigraphic analysis of the geological units. (manuscript submitted to GSA Bulletin). Giovanni Leone: Alignments of Volcanic Features in the Southern Hemisphere of Mars produced by Migrating Mantle Plumes. Journal of Volcanology and Geothermal Research, 01/2016: vol. 309, 78-95. Giovanni Leone, Paul James Tackley, Taras Gerya, David Alexander May, Guizhi Zhu: Three-Dimensional Simulations of the Southern Polar Giant Impact Hypothesis for the Origin of the Martian Dichotomy. Geophysical Research Letters 2014: vol. 41(24), 8736-8743. (Solicited paper). Giovanni Leone: A network of lava tubes as the origin of Labyrinthus Noctis and Valles Marineris on Mars. Journal of Volcanology and Geothermal Research 05/2014: vol. 277, 1-8. Imke de Pater, Ashley Gerard Davies, Amistair McGregor, Chad Trujillo, Máté Ádámkovics, Glenn J. Veeder, Dennis L. Matson, Giovanni Leone and the Gemini Io Support Team: Global Near-IR Maps from Gemini-N and Keck in 2010, with a special focus on Janus Patera and Kanehekili Fluctus. Icarus – 11/2014: 242, DOI10.1016/j.icarus.2014.06.019. Giovanni Leone, Lionel Wilson, Ashley Gerard Davies: The geothermal gradient of Io: Consequences for lithosphere structure and volcanic eruptive activity. Icarus 01/2011; 211(1): 623-635. Giovanni Leone, Lionel Wilson Ashley Gerard Davies: Sulphur from below. Nature Geoscience 11/2010: vol. 3, p. 818. Giovanni Leone, Ashley Gerard Davies, Lionel Wilson, David A. Williams, Laszlo P. Keszthelyi, Windy L. Jaeger, Elizabeth P. Turtle: Volcanic history, geologic analysis and map of the Prometheus Patera region on Io. Journal of Volcanology and Geothermal Research 08/2009: vol. 187(1-2), 93-105. Ashley Gerard Davies, Lionel Wilson, Dennis Matson, Giovanni Leone, Laszlo Keszthelyi, Windy Jaeger: The heartbeat of the volcano: The discovery of episodic activity at Prometheus on Io. Icarus 01/2006; Giovanni Leone, Lionel Wilson: Density structure of Io and the migration of magma through its lithosphere. Journal of Geophysical Research 01/2001; 106:32983-32996. Giuseppe Lucido, Giovanni Leone, Vincenzo Cataldo: Mineralogical evidence for liquid immiscibility in magmatic rocks from Western Sicily. Geologica Carpathica 05/1995: vol. 46(3), 127-136.

Conference Proceedings

Giovanni Leone, Paul J. Tackley, Taras Gerya, David A. May, Guizhi Zhu: 3D Numerical Model of the Southern Polar Giant Impact for the Formation of the Martian Dichotomy. European Geosciences Union General Assembly 2013, Vienna, Austria; 04/2013 Giovanni Leone, Paul James Tackley, Taras Gerya, David Alexander May, Guizhi Zhu: 3D numerical model for the formation of the Martian dichotomy and the Tharsis and Elysium Rises. XXXXIV Lunar and Planetary Science Conference, The Woodlands, TX, USA; 03/2013

256 Giovanni Leone, Lionel Wilson, Ashley Gerard Davies: The geothermal gradient of Io: consequences for lithosphere structure and volcanic eruptive activity. XXXX Lunar and Planetary Science Conference, The Woodlands, TX, USA; 03/2010 Giovanni Leone, Lionel Wilson, Ashley Gerard Davies, Giuseppe Giunta, Vincenzo Cataldo: Links between volcanism and tectonism on Io: a comparative study of Monan Patera, Amirani and Prometheus. XXXVIII Lunar and Planetary Science Conference, League City, TX, USA; 03/2008 Giovanni Leone, Lionel Wilson, Ashley Gerard Davies, David Williams, Laszlo Keszthelyi: Evidence of tectonic fractures as magma conduits in the Prometheus area on Io. AGU Fall Meeting, San Francisco, California, USA; 12/2007 Ashley Gerard Davies, Dennis Matson, Giovanni Leone, Lionel Wilson, Laszlo Keszthelyi: “Active” and “passive” lava resurfacing processes on Io: A comparative study of Loki Patera and Prometheus. XXXV Lunar and Planetary Science Conference, Houston, TX, USA; 03/2004 Giovanni Leone, Lionel Wilson: Links between depths of magma reservoirs and volcanic eruption rates on Io.. XXXIV Lunar and Planetary Science Conference, Houston, TX, USA; 03/2003 Giovanni Leone, Vincenzo Orofino, Armando Blanco, Sergio Fonti: Estimates of carbonate sedimentation in Martian paleolakes. European Geophysical Society, XXVII General Assembly, Nice, France; 04/2002 Edmond A. Grin, Nathalie E. Cabrol, Giovanni Leone, Vincenzo Orofino: Strategy for the in situ search of evaporite and carbonate deposits in Gusev crater within the 2003 MER A landing ellipse. XXXIII Lunar and Planetary Science Conference, Houston, TX, USA; 03/2002 Giovanni Leone, Armando Blanco, Federica De Carlo, Sergio Fonti, Vincenzo Orofino: Ismenius Lacus, Mars: Morphometric analysis of paleolake basins and search for carbonates. European Geophysical Society, XXVI General Assembly, Nice, France; 04/2001 Vincenzo Cataldo, Lionel Wilson, Giovanni Leone: Ascent and eruption of magmas on Io: application to Pele. XXXI Lunar and Planetary Science Conference, Houston, TX, USA; 03/2001 Giovanni Leone, Lionel Wilson, Vincenzo Orofino, Sergio Fonti: Gorgonum Chaos: Are the seepage-runoff features really recent? XXXII Lunar and Planetary Science Conference, Houston; 03/2001 Giovanni Leone, Lionel Wilson, Vincenzo Cataldo: Lava flow fields and calderas on Io: their correlation to magma reservoir size after collapse events. XXXI Lunar and Planetary Science Conference, Houston, TX, USA; 03/2000 Giovanni Leone, Lionel Wilson: The geothermal gradient of Io. XXX Lunar and Planetary Science Conference, Houston, TX, USA; 03/1999 Giovanni Leone, Lionel Wilson: Ascent and emplacement of magmas on Io: some considerations on compositions and solubilities of volatiles and their roles in volcanic activity. XXIX Lunar and Planetary Science Conference, Houston, TX, USA; 03/1998 Giovanni Leone, Lionel Wilson: Magma densities, magma reservoirs, and volcanic eruption styles on Io. XXVIII Lunar and Planetary Science Conference, Houston, TX, USA; 03/1997

257

258