Motivation and hypothetical model for hydrocarbons on Mars

We discovered a correlation between the aligned (combed) gravity strike angles, based on the global gravity field models, and known oil and gas deposit locations in the Caspian Sea and in Ghawar, Saudi Arabia (see Klokočník and Kostelecký, Reference Klokočník and Kostelecký2014). Later, we extrapolated to possible oil and gas sources in diverse areas on the Earth (Klokočník et al., Reference Klokočník, Kostelecký, Bezděk, Cílek, Kletetschka and Staňková2020c, Reference Klokočník, Kostelecký, Bezděk and Cílek2021b) and to groundwater on the Moon (Kletetschka et al., Reference Kletetschka, Klokočník, Hasson, Kostelecký, Bezděk and Karimi2022). It is logical to extend this method to Mars because after the Earth and the Moon, we now have sufficient knowledge about the gravity field of Mars. Finding hydrocarbons on Mars would mean finding indication for a form of past life on Mars.

The idea of hydrocarbons on Mars is not a new one. Thiophenes, formed at the bottom of a hypothetical ancient Martian freshwater lake (Gale Crater), found by Curiosity, could be markers of life on Mars (Science 2013; Heinz and Schulze-Makuch, Reference Heinz and Schulze-Makuch2020). [Thiophenes contain four carbon atoms and one sulphur atom; they occur in kerogens, bitumen, coals, crude oil (petroleum), sediments (ibid., p. 552) and white truffles (mushrooms).] The rover of Perseverance found evidence for ancient hot magma, impact melt or volcanic deposit, variably altered to carbonate and water (e.g. Farley et al., Reference Farley, Stack, Shuster, Horgan, Hurowitz, Tarnas, Simon, Sun, Scheller, Moore, McLennan, Vasconcelos, Wiens, Treiman, Mayhew, Beyssac, Kizovski, Tosca, Williford, Crumpler, Beegle, Bell, Ehlmann, Liu, Maki, Schmidt, Allwood, Amundsen, Bhartia, Bosak, Brown, Clark, Cousin, Forni, Gabriel, Goreva, Gupta, Hamran, Herd, Hickman-Lewis, Johnson, Kah, Kelemen, Kinch, Mandon, Mangold, Quantin-Nataf, Rice, Russell, Sharma, Siljeström, Steele, Sullivan, Wadhwa, Weiss, Williams, Wogsland, Willis, Acosta-Maeda, Beck, Benzerara, Bernard, Burton, Cardarelli, Chide, Clavé, Cloutis, Cohen, Czaja, Debaille, Dehouck, Fairén, Flannery, Fleron, Fouchet, Frydenvang, Garczynski, Gibbons, Hausrath, Hayes, Henneke, Jorgensen, Kelly, Lasue, Le Mouélic, Madariaga, Maurice, Merusi, Meslin, Milkovich, Million, Moeller, Núnez, Ollila, Paar, Paige, Pedersen, Pilleri, Pilorget, Pinet, Rice, Royer, Sautter, Schulte, Sephton, Sharma, Sholes, Spanovich, St. Clair, Tate, Uckert, VanBommel, Yanchilina and Zorzano2022). River deltas are potentially preserving carbon-containing organic compounds. The Earth's river and fossil deltas, such as the Nile, Niger and Mississippi, represent some of the most suitable places for the origin of petroleum deposits, because the rivers are carrying biogenic elements (P, Fe, V and others), released from weathered rocks. The organic detritus is quickly buried by eroded sediments to the depths suitable for oil genesis (e.g. Hunt, Reference Hunt1996; Montmessin et al., Reference Montmessin, Smith, Langevin, Mellon and Fedorova2017).

These and other findings, independent of our novel approach (based on global gravity aspects) provided a clear motivation for us to investigate this topic, having already had some experience with the Earth. Here we present our first results.

We start with the following hypothesis. Just as life arose very quickly on the Earth after the cooling of the Earth's crust and the world's oceans (Mojzsis et al., Reference Mojzsis and Arrhenius1998; Altermann, Reference Altermann and Seckbach2007), an aquatic environment suitable for the evolution of life existed on Mars for a sufficiently long time (Baker, Reference Baker2001; Ehlmann et al., Reference Ehlmann, JF, SL, J-P, Meunier, AA and Langevin2011). However, the magnetic field of Mars was weaker than that of the Earth and the atmosphere lacked the protection of the ozone layer, so we consider submarine hot springs to be the most likely ‘oases’ where life originated (Abramov and Mojzsis, Reference Abramov and Mojzsis2009). These, as a result of strong volcanism and meteoritic impacts, probably existed for a sufficiently long time (Rossi et al., Reference Rossi, Neukum, Pondrelli, van Gasselt, Zegers, Hauber, Chicarro and Foing2008) for the evolution of life, e.g. in the form of biofilms and microbial coatings (Blumenberg, Reference Blumenberg, Kile and Tyler2010; Seckbach and Chela-Flores, Reference Seckbach and Chela-Flores2012).

The most likely type of life on Mars, like on the Earth, was anaerobic chemotrophs probably tied to vicinity of warm underwater springs (Westall et al., Reference Westall, Foucher, Bost, Bertrand, Loizeau, Vago, Kminek, Gaboyer, Campbell, Bréhéret, Gautret and Cockell2015). We hypothesize that these organic remains were redistributed by later processes such as pressure and temperature changes, resulting from sediment burial, and may appear as biosignatures or chemofossils in suitable locations that may be analogous to terrestrial oil traps.

The earliest morphological manifestations of life are tied to simple shapes that can arise both organically and inorganically, e.g. during the coralloid growth of crystalline aggregates in aqueous glass or silicate colloids. This is one of the reasons for the long discussion concerning, e.g. the findings of real or apparent microbes from the Australian Apex Chert and other sites (Brasier et al., Reference Brasier, Green, Lindsay, McLoughlin, Jephcoat, Kleppe, Steele and Stoakes2005; Schopf, Reference Schopf2006), despite the sites being relatively easy to reach and sample, at least in comparison with Mars.

In search of the earliest terrestrial life, we often focus on the geochemical remains of life – on chemofossils with characteristic isotope compositions mainly of carbon and sulphur, and the abundance of organic compounds and trace elements. But this yields ambiguous rather than probabilistic results (Blumenberg, Reference Blumenberg, Kile and Tyler2010). Terrestrial crude oil can be viewed as both biologically and catalytically pressure- and temperature-processed chemofossils (in the presence of iron and manganese oxides). Unlike many other remnants of organic life that remain more or less ‘in situ’ in the rock mass, petroleum usually has a long dynamic and migratory history. The basic mechanism of oil reservoir formation is the transport of oil droplets in a water emulsion, the gravitational separation of lighter liquid and gaseous hydrocarbons, and their deposition in oil traps. This is mostly an impermeable, anticlinal structure that morphologically resembles an umbrella (Hunt, Reference Hunt1996).

Extensive fluid migration occurred on Mars at the end of the Noachian (ca. 4.5–3.5 Ga) and during the Hesperian (ca. 3.5–2.5 Ga). During this long period, water on Mars either moved by gravity, retreated into drying basins or underwent hydrothermal circulation (Zurek, Reference Zurek, Tolson, Bougher, Lugo, Baird, Bell and Jakosky2017). Good evidence for the existence of large hydrothermal systems comprises, for example, ridges linked to linear tectonic structures (see the figure of Candor Chasma in Vogt, Reference Vogt2008, p. 36). They resemble silicified areas commonly found on the Earth along faults. We guess that on Mars, with its huge volcanoes and voluminous lava outpourings, we have to expect the existence of large hydrothermal systems and thus water flows, including possible hydrocarbons (for the Earth analogue, see Van Kranendonk, Reference Van Kranendonk2006). This is a situation commonly addressed by terrestrial petroleum geology.

The question of petroleum traps on Mars is highly problematic because most strata, as shown by images of the Martian landscape, are deposited horizontally or slightly tilted. However, the alternation of layers indicates, in some cases, regular environmental changes as well as the possibility of deposition of insulating clay layers of, e.g. aeolian origin (Kramer et al., Reference Kramer, Potter, Marais and Peterson2003; Kite et al., Reference Kite, Rafkin, Michaels, Dietrich and Manga2011). A similar situation between the arrangement of strike angles and the more frequent occurrence of oil (as a specific type of chemofossil) is expected on Mars.

It is important to note at the outset of this paper that the gravity aspects (and specifically the strike angles) provide three types of mutually overlapping information:

1. (1) Information about geological structure. In evaluating terrestrial analogies, large-scale, uniform structures little affected by fault tectonics turn out to be important. On Mars, large-scale deltas, outwash cones and possible lahars would probably best suit these conditions.

2. (2) Information on water circulation. The amount of water as a medium carrying petroleum emulsions is many times greater than the hydrocarbons themselves. Therefore, ‘ordered’ (aligned, combed) strike angle formations are more likely to give information about the groundwater flows than about oil/gas. Again, based on terrestrial analogies, we assume that they mainly indicate the areas with uniform linear flows.

3. (3) Occurrence of hydrocarbons. This case is mainly due to both of these factors, but also due to the existence of source rocks and their porosity.

In recent years, we have been analysing the gravity aspects (Kalvoda et al., Reference Kalvoda, Klokočník, Kostelecký and Bezděk2013; Klokočník and Kostelecký Reference Klokočník and Kostelecký2014), especially the strike angles, around major oil/gas fields such as Ghawar, the Caspian Sea, and the Middle East. For the fields, but also for the areas where groundwater reservoirs exist (additional to or in combination with hydrocarbons), a combed structure with a parallel course of strike angles is typical (Klokočník et al., Reference Klokočník, Kostelecký and Bezděk2017a, Reference Klokočník, Kostelecký, Cílek, Bezděk and Pešek2017b, Reference Klokočník, Kostelecký, Cílek, Bezděk and Pešek2018, Reference Klokočník, Kostelecký and Cílek2020a, Reference Klokočník, Kostelecký, Bezděk and Cílek2021b, Reference Klokočnik, Bezděk, Kostelecky, Foulger, Hamilton, Jurdy, Stein, Howard and Stein2022b). These structures on the Earth indicate places where underground reservoirs or oil traps are not disturbed by later folding or fault tectonics. The same methodology also succeeded in a search for groundwater for the Artemis mission near the southern pole of the Moon (Kletetschka et al., Reference Kletetschka, Klokočník, Hasson, Kostelecký, Bezděk and Karimi2022). Now we ‘expand’ to Mars, having at our disposal, not only wishful thinking, but also the data from quality satellite measurements.

People often transfer directly our knowledge valid for the Earth to the other, at least partly similar celestial bodies, but it may have some obstacles. On the one hand, the size of the body is important, and in turn, the local geology differs. For example, because the Moon is small, bending stresses are fully capable of supporting topographic loads up to long-wavelengths (expressed by low degree/order harmonics in the spherical expansion of the disturbing potential, see below) and this will naturally affect how the gravity field is expressed. For that reason, the geological features on the Earth, the Moon and Mars will have different form of gravitational expressions just due to planetary size. On the other, gravitation, as it is, is universal: crater/basin/fault/volcano/lake, etc. have generally common signal in terms of the gravity aspects everywhere (many examples are provided in our Supplements 2 and 3). We can and we will transfer but with caution and under control of geologists.

From the viewpoint of a geologist: what are the possibilities of comparing the gravimetric record on Earth, Moon and Mars, which have a different geological evolution? Some features are very similar to each other, as geomorphology shows – in particular volcanic formations, including secondary deformations such as slope movements and lahars. The steep walls of impact craters are very often subject to slope deformation and landslides, as we know from Earth. The very important analogy is the filling of linear depressions and valleys, where, for example, the old Nile valleys beneath the aeolian sediments of the Sahara have gravimetric characteristics similar to river valleys on Mars. The record of flat-lying sedimentary basins is analogous. In contrast, the record of impact processes is largely different. If mascons existed on Earth, they were reworked by plate tectonics, and likewise most large terrestrial impact craters were heavily altered by erosion, which denuded several kilometres of rock, especially in Precambrian terrains.

Our procedure is not able to precisely determine the locations of possible hydrocarbons on Mars, but could serve as a refinement criterion for further research; in other words, as an additional pre-drilling constraint, an indicator, ‘conditioner’ or information that the probability of the presence of oil/gas/water deposits in the area is high or low. This remote-sensing tool is cheap.

Gravity and topography data and reference ellipsoid

As an input to all of our analyses, we make use of the gravity field models in terms of spherical harmonic expansion (Stokes parameters). For Mars, we use model NASA JPL JGMRO_120F (Konopliv et al., Reference Konopliv, Park Ryan, Rivoldini, Baland, Le Maistre and Van Hoolst2020) to degrees and orders (d/o) = 80. The corresponding ground resolution is about 130 km or ~2.5°. This is much less compared to the Earth and the Moon (about 10 km in both cases) and means a partial hindrance for interpretations. One must be cautious. On the contrary, it is much better than our present knowledge of the gravity field of Venus or Mercury.

The surface topography data come from the altimeter MOLA on board the MGS (Lemoine et al., Reference Lemoine, Smith, Rowlands, Zuber, Neumann, Chinn and Pavlis2001; Smith et al., Reference Smith, Zuber, Frey, Garvin, Head, Muhleman, Pettengill, Phillips, Solomon, Zwally, Banerdt, Duxbury, Golombek, Lemoine, Neumann, Rowlands, Aharonson, Ford, Ivanov, Johnson, McGovern, Abshire, Afzal and Sun2001). The MOLA is linked to the gravity model, fixed areocentrically. The total elevation uncertainty is ‘at least 3 m’ (Lemoine et al., Reference Lemoine, Smith, Rowlands, Zuber, Neumann, Chinn and Pavlis2001); we take 10 m. The topography from the MOLA in this study is given in the ground (horizontal) resolution of 0.25°.

We work with planetographic (ellipsoidal, areographic, areodetic) latitude – analogous to geodetic (not geocentric) latitude on the Earth, related to the reference rotational ellipsoid; the adopted semimajor axis of the reference ellipsoid: a = 3396.19 km, the polar flattening ratio: 1/flattening =169.8. Smith et al. (Reference Smith, Zuber, Frey, Garvin, Head, Muhleman, Pettengill, Phillips, Solomon, Zwally, Banerdt, Duxbury, Golombek, Lemoine, Neumann, Rowlands, Aharonson, Ford, Ivanov, Johnson, McGovern, Abshire, Afzal and Sun2001) wrote that the MOLA team provided planetocentric (areocentric) coordinates, thus a transformation was needed, because the difference in both types of latitudes can be significant at small to medium scales. It is not, however, in our case – we have global and large-scale views.

Further on, we will speak simply about ‘latitudes’ θ and ‘longitudes’ λ. We count them from 0 to 180° to the east and from 0 to −180° to the west or from 0 to 360° to the east. The choice of zero longitude (prime meridian) has been made by astronomers and it is de facto arbitrary. Mariner 9 provided extensive imagery of Mars in 1972 and a small crater Airy-0 in Sinus Meridiani (Arabia Terra) was chosen for the definition of 0.00° longitude (by the IAU, the definition is still valid).

For more information about Data for Mars, see Section ‘Preparatory truncation tests’ (truncation tests). To perform the truncation tests for the Earth and the Moon, as a ‘tutorial’ for Mars, we used EIGEN 6C4 to d/o = 2190 (Förste et al., Reference Förste, Bruinsma, Abrikosov, Lemoine, Schaller, Goetze, Ebbing, Marty, Flechtner, Balmino and Biancale2014) and GRGM1200A to d/o = 600 (Lemoine et al., Reference Lemoine, Goossens, Sabaka, Nicholas, Mazarico, Rowlands, Loomis, Chinn, Neumann, Smith and Zuber2014), respectively.

Preparatory truncation tests

An important prerequisite for proper analyses is to estimate which part, if any, from the gravity signal, created by the strike angles, can be a real signal, and whether there may exist potentially dangerous and misleading artefacts.

Test 1

Let us assume that a gravity model is published to a degree and order (d/o) but recommended (by the authors of the model) to be used only to a lower maximum d/o_(reduced). The reason is the use of the Kaula rule in the big inversion from satellite (planetary orbiters) data to gravity data for the gravity field models for a high d/o to stabilize the system numerically. This is true for the Earth, the Moon, Mars and Venus as well as Mercury. In the case of Mars, the gravity field model JGMRO_120F (Konopliv et al., Reference Konopliv, Park Ryan, Rivoldini, Baland, Le Maistre and Van Hoolst2020) is available to d/o = 120 but recommended to be used only to d/o = 80.

Thus, we have performed the following testing: we used the full gravity model GRGM1200A, Lemoine et al. (Reference Lemoine, Goossens, Sabaka, Nicholas, Mazarico, Rowlands, Loomis, Chinn, Neumann, Smith and Zuber2014) for the Moon, to d/o = 1200, we cut it at d/o_(reduced) 600, 360 and 130 and mutually compared (Fig. 1(a)–(c) here, and others in Klokočník et al., Reference Klokočník, Kostelecký, Bezděk and Kletetschka2020b).

Figure 1. Moon's crater Copernicus, the strike angles, from the gravity model (Konopliv et al., Reference Konopliv, Park Ryan, Rivoldini, Baland, Le Maistre and Van Hoolst2020): (a) cut at d/o = 130 (it is too low, thus losing information); (b) cut at the recommended optimum d/o = 600 and (c) a devastating graining effect appears when d/o is kept up to 1200. Note that the blue and red colours relate to the quadrants of the strike angles. The θ [°] is expressed with respect to the local meridian; the red colour means its direction is to the north and blue is to the south of the east (for more information, see S2: 2).

Figure 1(a)–(c) compare the strike angles computed for I < 0.9 and to d/o = 130, 600 and 1200 for the Moon's crater Copernicus. A loose signal, its resolution, smoothing and generalization for d/o < 600 is evident, and trivial, but what happens for d/o = 1200 may be surprising for the reader (we have many such examples in Klokočník et al., Reference Klokočník, Kostelecký, Bezděk and Kletetschka2021a). The limit d/o = 1200 is too high, the data from the Moon's orbiters are not adequate for such a high limit, and the gravity signal in terms of θ degenerates, and disintegrates due to a (well-known) graining effect into small ‘islands’ (Fig. 1(c)); the halo around the crater then disappears. The consequence is that we only see artefacts, nothing real.

Our conclusions from test 1 for Mars are: (i) to use JGMRO_120F just to d/o = 80, and not higher. (ii) The θ values at the NMPO, being combed into big ‘plates’ (see Section ‘Individual zones/segments in the palaeo-ocean’) are not those small grained ‘islands’ – artefacts, as in Fig. 1(c), but might be the actual features on Mars. Further testing will be useful (tests 2 and 3).

Test 2

The gravity model JGMRO_120F is practically available to d/o = 80 and represents the top-most knowledge now; and it will be much better soon. JGMRO_120F yields the ground resolution of about 130 km on Mars. For the Earth, we also have – in addition to satellite data – terrestrial gravimetric data for effectively high(er) degrees and orders of harmonic coefficients (Stokes parameters) in the combined gravity models. No such analogy exists for Mars; here are ‘only’ the orbits of planetary orbiters and measurements from the on-board laser altimeters.

Let us imagine for a moment that our knowledge about the Earth's gravity field ends at a maximum d/o = 80 (corresponding to the ground resolution of about 260 km in this case). Let us compute and plot θ for this limit and compare the results with the full gravity model (2190). We used gravity EIGEN 6C4 (Förste et al., Reference Förste, Bruinsma, Abrikosov, Lemoine, Schaller, Goetze, Ebbing, Marty, Flechtner, Balmino and Biancale2014), complete to d/o = 2190 (the relevant ground resolution of 9 km). This can be called the ‘true’ gravity field model for this test. There is the question as to what degree the strikes θ will become smoothed, generalized and degenerated with d/o_(reduced) and also, whether some of the artefacts, which could be potentially misleading, would arise.

We reply to the questions with Figs. 2(a) and (b) and 3(a) and (b) and S2: 21–39. It becomes obvious that the higher ground resolution means smaller areas of alignment, and that some artefacts appear. Colour dashes (abscissae) show θ for d/o = 2190, and black dashes (abscissae) for 80. While the cut at 80 simplifies the picture too much and omits important features, as must be expected, the general trend of alignment is roughly the same for both d/o = 2190 and 80. This is an encouraging finding, but one has to be very careful to avoid misinterpretations with d/o = 80; there are some artefacts showing a rotation in the alignments for 80 (Fig. S2: 32–34) which the ‘real’, full gravity model, does not confirm.

Figure 2. Cut at a low d/o to simulate a lower resolution of the Martian gravity field models now available, by means of the gravity field EIGEN 6C4, valid for the Earth. The areas are (a) the eastern Mediterranean Sea, and (b) the southern part of the Caspian Sea. Coloured dashes show θ for d/o = 2190, and black dashes (abscissae) for d/o = 80. Our finding based on these tests is: we can rely, with some caution, upon the general trend of alignment because it is roughly maintained from the limit d/o = 2190 down to the cut at d/o = 80. Various disturbing and misleading artefacts are, however, possible, like directional changes of the alignment, false rotation-like features, etc.; see Supplement S2: Fig. S2: 23, 26 or 32–34. Red dots are for deposits of oil/gas (gathered from various sources), incl. L. Eppelbaum, Tel Aviv Univ. (private commun.).

Figure 3. (a) Strike angles for the Earth with the full model EIGEN 6C4 to maximum d/o = 2190. In the window in South America, the alignment of the strike angles is observed in a very limited narrow belt, going along the seashore (~NS direction). This is taken as ‘reality’, as a reference to be now confronted with the cut at a lower d/o in (b). (b) The insufficient knowledge of the gravity field, here modelled for the Earth, simulating EIGEN 6C4 only to maximum d/o = 80. It degrades the strike angles not only due to significantly lower resolution ( = neglecting details) but also due to the origin of unexpected artefacts – in the windows we can see red areas with highly combed strikes while in the ‘real’ field (in Fig. 3(a)) with the d/o to a maximum of 2190, they were reduced or missing. This is a warning example of the treacherous artefacts. (c) The strike angles θ [deg] for the east Mediterranean (EIGEN 6C4 to d/o = 2190), computed for the ratio I = 0.9. A black and white version of Fig. 2(a) without dots giving better view of those large ‘plates’ of aligned areas. Topography relief is derived from the MOLA data. For further figures of this type, see Supplement 2. The strike angles θ [deg] for the Persian Gulf, EIGEN 6C4 (I = 0.9) to (d) d/o = 2190 and (e) d/o = 80. Note the remarkable difference in resolution but a common trend of alignment. The topography relief is from MOLA. Observe closely and mutually compare Figs. 2–4. This test is very important.

Figure 3(a) detects areas on the Earth with the high comb factor – brown contour lines with Comb = 0.95 and also the highest comb factor areas (Comb ≥ 0.99) for d/o = 2190; we can see that they are rare. In Fig. 3(b), we have the same but for d/o = 80, we can see significantly more combed areas here than in Fig. 3(a). Now, we transfer brown contour lines for Comb = 0.95 from Fig. 3(a) into Fig. 3(b). The result is that the red values (for Comb ≥ 0.99), computed from the cut model at d/o = 80, correlate well with the brown contours to d/o = 2190.

The two areas delineated by rectangles in Fig. 3(a) and (b) compare two examples with different behaviour while the comb factor between the solutions to 2190 and 80 is the same. The example of the west border of South America with d/o = 2190 and Comb > 0.95 contains only a narrow belt along the seashore, Fig. 3(a). The result to d/o = 80 (and Comb > 0.95) shows a wider alignment area, reaching far from the seashore, Fig. 3(b). This is false information. There is another situation for the second rectangle in Fig. 3(a), and (b), in Tibet. The model to 80 shows a high alignment there, while the full solution to 2190 does not. This is misleading and can be dangerous for interpretation. In general, however, the trends of the plates with the combed strikes to d/o = 80 reproduce the ‘reality’, i.e. the alignments depicted by the full model 2190, well, ignoring details, over the whole world. We can only hope that the same conclusion will be valid for Mars.

The conclusion from tests 1 and 2 – for sufficiently large ‘plates’ of aligned θ, not affected by graining but also sufficiently extensive to be well above the resolution of JGMRO_120F (i.e. about 130 km ≈ 2.5°) – is that we can rely upon general trends of alignment because they are roughly maintained for d/o = 2190 and d/o = 80 as well. Such ‘plates’ are shown in Fig. 3(c) for the east Mediterranean (to be compared to Fig. 4(c) for Mars) and in S2.

Figure 4. Strike angles θ [deg] for Mars with JPL JGMRO_120F (Konopliv et al., Reference Konopliv, Park Ryan, Rivoldini, Baland, Le Maistre and Van Hoolst2020) to maximum d/o = 80, computed for the ratio I = 0.9. East/west longitude on the x-axis, latitude on the y-axis. (a) The strike angle is expressed with respect to the local meridian; its red colour means its direction to the north and blue is to the south of the east. (b) The Comb statistics – the red colour valid for the highest degree of alignment – Comb ≥ 0.99. (c) Martian dichotomy in the form of a different look of alignments of the gravity strike angles for a part of the Northern and Southern Hemispheres. This is an example of the strike angles in a part of the hypothetical NMPO. We show a contrast between the varied terrain of the highlands on the Southern Hemisphere and the plain, drab ‘sea bottom’ in the lowlands, the Northern Hemisphere. In the lowlands, the strike angles are intensively combed and form large ‘plates’ with one prevailing direction of alignment in the plate. Artefacts cannot be fully ruled out, but our tests (Section ‘Preparatory truncation tests’) show that reality is prevailing. Contours with the MOLA topography are added in the blue colour.

Strike angles globally

These are samples of our global views on Mars with the gravity model JPL JGMRO_120F (Konopliv et al., Reference Konopliv, Park Ryan, Rivoldini, Baland, Le Maistre and Van Hoolst2020) to a maximum d/o = 80 showing the strike angles without (Fig. 4(a)) and with the comb statistics – the red areas in Fig. 4(b) have Comb > 0.99. For gravity aspects other than the strike angles, see Supplement 3 (Fig. S3: 4–9).

It looks like (Fig. 4(a) and (c)) the large areas (looking like ‘plates’) in the northern lowlands are highly combed. Figure 4(b) shows clearly that the highest alignment exists more frequently over the lowlands (mostly plain, drab terrain) than on the highlands (with more varied terrain). The strike angles thus confirm a conspicuous difference between the northern and southern parts of the globe of Mars and not only in the topography (well known). While Δg, T_zz and the invariants support the fact of a gravitationally, relatively ‘silent’ area (dumped due to the effect of thick sediments?) for the northern part, but not everywhere, θ and vd disclose more: a difference in porosity (higher in the lowlands) and various stress tendencies (higher in the highlands).

Strike angles for the polar areas

The northern cap (MOLA topography relief in Fig. 5(a)) is bigger and mostly comprised of ice water (H₂O), while the southern is mostly CO₂ (Fig. 6(a)), but water is present there, too (e.g. Arnold et al., Reference Arnold, Butcher, Conway, Gallagher and Balme2022). Both have a higher porosity than the surrounding rocks. This should be reflected in the strike angles that would be aligned.

Figure 5. (a) MOLA topography relief (m) for the northern polar cap. ‘Dunes’ and ‘rotation’ due to Coriolis force are well visible. More about the interpretation to the series of Figs. 5 and 6 is provided in Discussion. (b) The strike angles (I < 0.9) in the form of Comb factors, in a combination with abscissae for the strikes. MOLA contour lines are added. (c) The strike angles (I < 0.9) in the form of Comb factors. Together with the MOLA height profile.

Figure 6. (a) MOLA topography relief for the southern polar cap. This cap is smaller than the northern cap, but the extent of latitudes on this scale differs from that in Fig. 5(a); in turn this cap may look more massive. (b) The strike angles (I < 0.9) in the form of Comb factors, in a combination with abscissae for the strikes. MOLA contour lines added. (c) The strike angles (I < 0.9) in the form of Comb factors. Together with the MOLA height profile.

Here, we show the topography and strike angles only. The other gravity aspects are provided in Supplement 3, see Fig. S3: 69–81. Figures 5(b) and (c) and 6(b) and (c) show the relevant strike angles for the northern and southern caps. The reader can see that the alignment is (i) conspicuous for both caps, that (ii) the alignment is not so extreme on the south (cf. Fig. 6(c) to Fig. 5(c), note that the colour scales of these figures are the same) and (iii) the northern cap is more aligned than the surrounding NMPO.

Figure 5(a), showing the surface morphology, has two basic features – smooth and rugged surfaces. Both can be interpreted as manifestations of wind activity. The strike angles in Fig. 5(b) complement this information. Figure 5(b) shows beneath the planet's surface. The relatively low density of the gravimetric signal means that the diagram captures the basic most prominent features of the mass structure of the polar region of Mars. Terrestrial analogies, verified, e.g. by drilling and geophysical work and measurements, have repeatedly shown that gravimetry is capable of capturing the geological structure at a range of at least 2–5 km below the surface.

The spiral structure corresponding to the morphology is quite visible in a rough outline in Fig. 5(b). This probably indicates that similar aeolian processes have been operating on Mars for a long time. The whole scheme can be decomposed into approximately concentric features, which probably correspond to aeolian processes caused by planetary rotation (Coriolis force), and radial structures, which are not visible in the morphology in Fig. 5(a). The seven most pronounced structures lie between 200 and 210°E, but less pronounced lines can also be found at other locations, e.g. between 110 and 120° (see Fig. 5(c) as well). Our explanation is that here we have buried valleys formed by melt water erosion.

The southern polar region already confirms the hypothesis of radial erosion valleys on the geomorphological map (Fig. 6(a)). Note, however, that the position of the strike angles (Fig. 6(b) and (c)), e.g. between 60 and 70°E or between 300 and 320°E, does not correspond to the present morphology of Mars. This may indicate that another, older and somewhat different set of erosional features lies beneath the present valley network.

We consider the test for the polar caps of Mars as a very useful ‘test-bed’. We receive what we expected. This means that we have checked that (i) the strike angles are highly combed (due to the higher porosity of material of the caps). Moreover, we verified that (ii) the northern, bigger and thicker cap is more intensively combed than the southern.

Similar situations with highly combed areas but not uniquely focused on the causative bodies, are known from the Earth (Section ‘Introduction’) for groundwater, subglacial features, canyons, palaeo-lakes, oil and gas deposit localities and also, in a circular form (as a halo), for impact craters and basins (and this is also true for the Moon, see Klokočník et al., Reference Klokočník, Kostelecký, Cílek, Kletetschka and Bezděk2022a).

Thus, we conclude that the concept of the strike angles ‘works well’, and that it ‘works as should work’, so we can do the next step of our analyses.

Extent of the palaeo-ocean

Discussion about a hypothetical ocean or oceans (Arabia, Deuteronilus, Oceanus Borealis or Utopia, its part) on Mars is old (e.g. Baker et al., Reference Baker, Strom, Gulick, Kargel and Komatsu1991; Parker et al., Reference Parker, Gorsline, Saunders, Pieri and Schneeberger1993; Head et al., Reference Head, Hiesinger, Ivanov, Kreslavsky, Pratt and Thomson1999; Clifford and Parker Reference Clifford and Parker2001; Ghatan and Zimbelman Reference Ghatan and Zimbelman2006; Forget et al., Reference Forget, Costard and Lognonné2008). The ocean would cover the vast topographically low plains of Vastitas Borealis, Chryse, Arcadia, Acidalia and Utopia Planitia. The topography and the gravity aspects do not in general correlate as we would expect for the Earth at least for the gravity anomalies; these low areas are not ‘silent’ everywhere in terms of the gravity signal (see series of Figs. 4 and S3: 4–9, 21–66).

Since the early discovery of Martian north-south topographic dissymmetry or dichotomy, it has been logical to put a putative palaeo-ocean or oceans, lakes and rivers mostly towards the topographically low northern plains, like the Vastitas Borealis formation – looking like abyssal plains of the Earth's oceans (e.g. Carr and Head, Reference Carr and Head2003; Irwin and Watters, Reference Irwin and Watters2010). The rugged frontier, the transition zone in between the lowlands and highlands, known as fretted terrain, may define a part of the seashore (Forget et al., Reference Forget, Costard and Lognonné2008). But the shore is highly topographically variable, so we have to refer to one equipotential level (Lemoine et al., Reference Lemoine, Smith, Rowlands, Zuber, Neumann, Chinn and Pavlis2001; Smith et al., Reference Smith, Zuber, Frey, Garvin, Head, Muhleman, Pettengill, Phillips, Solomon, Zwally, Banerdt, Duxbury, Golombek, Lemoine, Neumann, Rowlands, Aharonson, Ford, Ivanov, Johnson, McGovern, Abshire, Afzal and Sun2001; Zuber, Reference Zuber2018). It is important to recall that the MOLA MGS topography is related to the global Mars' best-fitting ellipsoid and, in turn, to the particular gravity field model and Mars' geoid-areoid, in other words, to an equipotential surface (Section ‘Gravity and topography data and reference ellipsoid’).

In this experiment, we ignore any knowledge in the first step about the geological features and history of Mars and all previous attempts to delineate the extent of the palaeo-ocean. Constraints on estimating its hypothetical seashore are derived with the aid of the global MOLA topography (Fig. 7(a)–(e)), going from the southern highlands, using knowledge of geographical locations of the fretted terrain via the northern lowlands to the northern polar cap. The real seashore was evidently time variable (due to various reasons) during the long history of Mars – before the NMPO had desiccated after all. A more fragmented potential shoreline can be expected.

Figure 7. (a) Global view on Mars with the MOLA topography (in (m)). The estimated ‘seashore’ of the NMPO is expected to be between −3000 and −4000 m. The red contour line corresponds to −3500 m. The border between the Northern and Southern Hemispheres contains in various places the fretted terrain, described in the literature. For us, it is important that the erosion that formed the fretted terrain was aided by water (e.g. Denton and Head, Reference Denton and Head2018). For comparison see, e.g. Kite et al. (Reference Kite, Rafkin, Michaels, Dietrich and Manga2011). (b) The northern polar cap on Mars by means of the MOLA topography (m). The tentative ‘seashores’ of the NMPO at various levels. (c) MOLA topography (m) for the Valles Marineris (VM). A possible connection to the hypothetical palaeo-ocean – its seashore is shown by the red contour line (for −3500 m). The numbers in the yellow colour localize black dots in (d). (d) The altitude profile of the VM from the west (Tharsis) to the east and north (NMPO), points 1–13 in (c). The distance on the x-axis is computed from point 1. Based on MOLA topography (m). (e) A compromise for the ‘seashore’ line found: the height may be ~−3500 m with respect to the reference ellipsoid, this is our model altitude for the hypothetical seashore of the hypothetical palaeo-ocean. Based on the MOLA topography (m).

Figure 7(a) yields −3000 to −3500 m for such a model seashore altitude, while the topography at the cap in Fig. 7(b) leads rather to −4000 m.

In the second step, we account for previous attempts to delineate the palaeo-ocean and for the generally accepted geological history of Mars (e.g. Irwin and Watters, Reference Irwin and Watters2010; Citron et al., Reference Citron, Manga and Hemingway2018a, Reference Citron, Manga and Tan2018b). The palaeo-ocean belongs to the Noachian era (from the planet's formation to 3.5 billion years ago, Gy), after some time desiccating. The cap should be newer – late Amazonian (~1 Gy). The fretted terrain is Noachian-Hesperian. We should not combine the polar cap and the lowland–highland border as a common condition for the height of the seashore. We should prefer to use the older features (Fig. 7(b)). Then, we cannot be surprised that, for example, at the height of −3000 m, the northern polar cap would be partly ‘flooded’ (Fig. 7(b)).

To compare our results, we used the estimate from Forget et al. (Reference Forget, Costard and Lognonné2008), based on the MOLA data, too. Their estimate is the seashore at −4350 m ‘above the zero level’ (which we hope is the same ellipsoid as in our case, Section ‘Gravity and topography data and reference ellipsoid’). The ocean might be 600 m deep.

As a check, Fig. 7(c) (a connection between the NMPO and the Valles (valleys) Marineris (VM) on a level at −3500 m) confirms that this choice is appropriate; compare Fig. 7(c) to (a), see Fig. S3: 11–19.

Figure 7(d) is an altitude profile of the VM from west to east and then to the north (follow points 1–13 in Fig. 7(c)). We can see (i) a general trend of topography from the top-down from west to east and to the flat ‘delta’ north, and (ii) we can speculate about a deeper lake in the VM between points 3 and 6 and hypothetical rapids or waterfalls or both between 2 and 3 or 9 and 10. There are terrestrial analogies where long valleys break up into sub-basins, in some periods. The headlands between the basins are either of tectonic origin or have been formed by filling with sediments from lateral tributaries or landslides.

After all the experimenting, our final estimate for the seashore height (giving its hypothetical long-term level) is about −3500 m.

Individual zones/segments in the palaeo-ocean

Six segments over the hypothetical NMPO are shown here, from west to east, based on the MOLA topography and the gravity strike angles from JPL JGMRO_120F (Konopliv et al., Reference Konopliv, Park Ryan, Rivoldini, Baland, Le Maistre and Van Hoolst2020), see Section ‘Gravity and topography data and reference ellipsoid’, to d/o = 80. The other gravity aspects (Δg, T_zz, I_j, I, θ and vd, and the MOLA topography) can be found in Supplement S3 as Fig. S3: 20–67.

For a better orientation on Mars, we recommend using the USGS (2003) topographic map of Mars. For a few comments on the strike angles of the individual segments, see in Section ‘Strike angles’ map documentation’.

This series of figures with the strike angles for the individual zones mostly for the northern lowlands clearly demonstrates highly aligned strikes. We will quantify the degree of their alignment by means of the comb factor and tentatively interpret these results in Section ‘Strike angles’ map documentation’.

Criteria

In an analogy to the Earth, we seek the largest and highest combed areas of θ giving a higher chance for hydrocarbon reservoirs. We suggest that promising oil areas could be located at an intersection of three basic factors:

1. (i) The ocean, as a larger body of water, which has a fairly significant thermal and chemical inertia. It is, therefore, not subjected to rapid changes that could threaten emerging life.

2. (ii) The existence of a thermal oasis containing substances that are capable of providing energy. These places are suitable for the evolution of chemotrophic bacteria. They can be detected as chemofossils containing a wide range of substances such as: hydrocarbons, iron oxides, sulphates, carbonates, etc. On the Earth, these sites are typically associated with fault structures related to mid-ocean ridges and other areas of increased volcanic activity.

3. (iii) A suitable place where chemofossils indicated by hydrocarbons could be preserved (oil traps).

We prefer areas in the NMPO, which are as large as possible, together with combed strike angles that are as high as possible. But such an inland sea like the Hellas basin probably also poses ground (ice) water. Within our first estimate, we focus on the NMPO. We know from Sections ‘Gravity and topography data and reference ellipsoid’ and ‘Preparatory truncation tests’ that the ground resolution, dictated by the gravity model JGMRO_120F (not by the method), is about 2.5°. Thus, we will ignore all (even very well combed) areas smaller than 10°, at least in one dimension (harder than the criterion ~3σ). Further, we select only the red zones of θ with Comb ≥ 0.99.

We know that for oil fields, but also for the areas where groundwater reservoirs exist, a combed structure with a parallel course of strike angles is typical. These structures on the Earth indicate places where underground reservoirs or oil traps are not disturbed by newer folding or fault tectonics. Thus, we prefer ‘gravitationally silent’ areas, meaning areas without sharp changes in Δg, T_zz or the invariants. This prefers flatter sedimentary areas, but proximity to volcanoes or trenches can also not be excluded, if Comb ≥ 0.99. The sediments, superimposed upon the pre-existing surface features, reduce Δg, T_zz (much more for T_zz and the invariants than for Δg) of those underlying structures, and affect the orientation of θ and vd.

The zones fulfilling our criteria are selected and presented with a commentary in the next section. The reader can perform her/his own selection according to her/his criteria from our results in Sections ‘Strike angles globally’, ‘Strike angles for the polar areas’ and ‘Individual zones/segments in the palaeo-ocean’.

Strike angles' map documentation

The results from Section ‘Individual zones/segments in the palaeo-ocean’ are now presented in the form of colour maps of the strike angles in Fig. 9(a)–(f) with the Comb factors (explained in Section ‘Method’), and in another projection. The selected features in Fig. 9(a)–(f) are discussed. The ovals indicate the largest areas with the highest Comb value (in red colour) in the respective segment. Names are according to the Topographic Map of Mars (USGS, 2003).

Particularly striking are the elongated anomalies, which usually have a red centre and a narrow blue rim. They are found especially at 50°N and −130°, 50° and −100° or in the southern part of the figure at 30°N and −150° to −120°. The continuation of the valley, oriented NE, at 50°N and −60°, which clearly represents the outflow area, opens up an interpretative possibility. Similarly, the elongated anomalies can be considered as the remains of broad streams or flow-through lakes.

The most striking feature of Fig. 9(b) is the large homogeneously arranged area around the VM. The analogous valley structure in the upper right corner (Sydonia Mensae) can also be used for its interpretation. We assume that it could be a fairly uniformly deposited sediment on tilted blocks of rift or tectonic blocks. Aeolian-transported or washed sediments can cause softening of the originally much more rugged relief. However, this depends on the prevailing wind direction, i.e. on the local climate, where the thicker sedimentary layers will form.

In the VM, one can surprisingly see the highest Comb values in the fault and in its wide surroundings along the fault, perhaps a real feature, not an artefact (Fig. 9(c)). One possible explanation is that there was a uniform deposition of aeolian and slope sediments on the long, inclined surfaces around the central rift. In the same figure, notice the ends of shallow valleys that merge into the sea basin (10°S–15°N, 50–20°W). This is a typical area where, in our opinion, we can look for occurrences of hydrocarbons.

From an interpretative point of view, it is appropriate to move from Figs. 9(c) to 3(c), which shows the Nile delta. The Nile Valley itself has an orientation of strike angles in the direction of the structure, but the delta grew gradually and was modelled by sea currents, so its orientation is perpendicular to the direction of the river. The direction of the combed θ follows the course of the long axis of the VM, as the reader expects.

It is not easy to interpret Fig. 9(c)–(f) either, but the most important features of the relief are neither striking craters like on the Moon nor linear structures like on the Earth; instead, we encounter more or less regular surfaces. Thus, there is no evidence of a clear river network manifested by well-defined valleys. Instead, broad, elongated anomalies appear in places. It is thought that this type of structure was mainly formed by aeolian processes, but on the whole the fine-grained sediments were periodically moved during wetter or more humid periods. This situation can be approximately compared to the present-day shallow Lake Chad Basin or to some Triassic formations of the US Midwest.

Summary of our observations and discussion

The strike angles in the northern lowlands of Mars are combed nearly everywhere and often more intensively than on the Southern Hemisphere (Figs. 4(a)–(c), 5(c) and 8(a), (b) and (d); S3: 7, 8, 73–75, 79–81). The strong alignment relates mainly to the NMPO. It corresponds to a higher porosity of the material, meaning the sedimentary layer, ground-ice (permafrost), groundwater sources, oil-type/hydrocarbon occurrences, lahars, palaeo-lakes, deltas of hypothetical rivers (e.g. Fig. 7(c)), fretted terrain (Figs. 8(c) and (d) and 9(c), (d) and (f)) or the relative vicinity to volcanoes and lahars (Fig. 8(a) and (f)).

Figure 8. (a) Topography (MOLA (m)) and the strike angles θ (for I < 0.9) for the selected zone. The feature at (φ = 40°N, λ = 110°W) is Alba Patera near Olympus Mons. 1° ~ 50 km. (b) Topography (MOLA (m)) and the strike angles θ (for I < 0.9) for the selected zone, mainly the Valles Marineris (VM). The feature at (7°N, 22°W) is the crater Aram Chaos. A halo-like shape of several features north of the VM resembles smaller volcanoes. (c) Topography (MOLA (m)) and the strike angles θ (for I < 0.9) for the selected zone, partly Arabia Terra (bottom right). The feature at (65°N, 10°W) is the crater Lomonosov in the Vastitas Borealis plain. Here, the reader can compare colour and black and white versions of the figure for the same zone. (d) Topography (MOLA (m)) and the strike angles θ (for I < 0.9) for the selected zone. The feature at (50°N, 30°E) is the crater Lyot, north of the Deuteronilus Mensae fretted terrain. (e) Topography (MOLA (m)) and the strike angles θ (for I < 0.9) for the selected zone. The feature, bottom left is Isidis Planitia, with an approximately hexagonal shape, the plains EN of it are a part of Utopia Planitia. The landing site of Perseverance is shown by the white star. (f) Topography (MOLA (m)) and the strike angles θ (for I < 0.9) for the selected zone. The feature at (25°N, 145°E) is Elysium Mons.

Figure 9. (a) Feature at (φ = 40°N, λ = 110°W) is Alba Patera near Olympus Mons (on the southern edge of this figure). The oval with the largest area with the highest Comb is located near a crater (with the misleading name of Erebus Montes) partly buried in sediments in Amazonis Planitia, NW of Olympus; see also the slides with other gravity aspects – Fig. S3: 20–31 in Supplement 3. The slant ‘rifts’ at (φ = 25–45°N, λ = 90–60°W) belong to Mareotis Fossae and Tempe Fossae. (b) The highest Comb values along the Valles Marineris (VM) and within its wide surroundings. This might, however, be an artefact like those shown in Section ‘Preparatory truncation tests’; compare (b) to Figs. 3(a) and (b); we cannot exclude this possibility. Tharsis would be on the left-hand side (outside). The crater at [φ = 7°N, λ = 22°W] in the ‘delta’ of the VM to the hypothetical NMPO (via Chryse Planitia) is Aram Chaos. The red belt at the upper right-hand corner belongs to Sydonia Mensae. (c) The ovals are (top-down) in: (1) Vastitas Borealis lowland [φ = 65–75°N, λ = 40–0°W]; the crater Lomonosov at (65°N, 10°W); (2) the lowland part at Acidalia Planitia [φ = 35–50°N, λ = 30–5°W]; (3) on the border between lowland and highland, Cydonia Mensae [φ = 25–45°N, λ = 20–0°W]. (d) The oval at [φ = 50–60°N, λ = 35–60°E] is in the lowlands, crater Lyot is at [φ = 50°N, λ = 30°E], Deuteronilus Mensae is at [φ ~ 40°N, λ = 10–60°E] and Arabia Terra is south of this. (e) Utopia Planitia [φ = 50–60°N, λ = 85–120°E]. Isidis Planitia [φ = 0–20°N, λ = 75–100°E]. (f) Elysium [φ = 20–30°N, λ = 125–140°E]; see also Fig. S3: 67. Gale crater at [φ = 5°S, λ = 138°E]; Apollinaris Pater volcano at [φ = 9°S, λ = 175°E].

The combed strike angles θ in the NMPO create features like ‘plates’ on the surface (large areas, zones and belts that can be 20° across) keeping a similar orientation of θ in the given plate (a series of Figs. 8(a)–(f); S3: 28, 29, 37–39, 46, 47, 55–57, 64, 65, 75, 81, 84). It appears that the plates do not in general correlate with the other gravity aspects, e.g. a high porosity does not automatically mean a low density, because it may be masked by various geological features beneath the sedimentary layer. The strikes react dominantly on the sediment and not on deeper layers.

We confirm that the polar caps are highly combed, the northern more than the southern (Fig. 5(b) and (c) versus 6(b) and (c), S3: 73–75, 79–81). The northern cap exhibits more intensive alignments than the NMPO as a whole.

The Isidis basin (Figs. 8(e), 9(e) and S3: 49–59, 86–88) would be a part of the edge of the NMPO. Around the Isidis basin, magnesium carbonate was discovered (by the MRO); this indicates non-acid water with favourable conditions for the evolution of life (Murchie et al., Reference Murchie, Mustard, Ehlmann, Milliken, Bishop, McKeown, Noe Dobrea, Seelos, Buczkowski, Wiseman, Arvidson, Wray, Swayze, Clark, Des Marais, McEwen and Bibring2009). Nearby, in Utopia Planitia, a part of the NMPO, groundwater was detected (by radar on the MRO, Section ‘Introduction’). The Zhurong rover of China's Tianwen-1 mission, which landed in southern Utopia Planitia, identified hydrated sulphate/silica materials on the Amazonian terrain at the landing site (Liu et al., Reference Liu, Wu, Zhao, Pan, Wang, Liu, Zhao, Zhou, Zhang, Wu, Wan and Zou2022). The locality fits well with the place with the highest Comb factor in Fig. 9(e). We should remember that when comparing such results with ours one has to be aware of the ground resolution accessible by us (dictated by the gravity field model used) and that it is about 130 km.

At the west-north of Tharsis (at φ = 35°N, λ = 170°S), there is a big crater nearly entirely buried under sediments, but well visible in the gravity aspects (Fig. 8(a) with θ but better in Fig. S3: 20–31).

In the Valles Marineris (VM), one can surprisingly see the highest Comb values in the fault and in the wide surroundings along the fault, perhaps a real feature, not an artefact (Fig. 9(c) and S3: 100–103). One possible explanation is that there was a uniform deposition of aeolian and slope sediments on the long, inclined surfaces around the central rift. In the same figure, notice the ends of shallow valleys that merge into the sea basin (10°S–15°N, 50–20°W). This is a typical area where, in our opinion, we can look for occurrences of hydrocarbons.

We investigated a possible extent of the NMPO (Section ‘Extent of the palaeo-ocean’). We used MOLA topography for a hypothetical seashore, the northern polar cap and the VM as a hypothetical river flowing to the VM from Tharsis Plato (series of Fig. 7). Figure 7(d) shows an altitude profile of the VM from west to east and north (with a hypothetical delta or deltas), confirming a general trend of topography from the top-down for running water, leaving space for speculations about deeper lakes, rapids or waterfalls along the trench. Here we recall Fig. 7(c) for the height profiles to the height profiles in Section ‘Extent of the palaeo-ocean’. Among many other places, the VM contains sulphates on its floor and walls (according to the Mars Express orbiter camera, combining a spectroscope and imaging device); sulphates require liquid water in their formation (Viviano-Beck et al., Reference Viviano-Beck, Seelos, Murchie, Kahn, Seelos, Taylor, Taylor, Ehlmann, Wisemann, Mustard and Morgan2014).

Potential hydrocarbons are more likely to be found in places with ordered combed structures, which probably represent here places with linear (uniform) water flow, as on Earth that would carry and concentrate hydrocarbon chemofossils in suitable places, e.g. under impermeable clay layers.

In interpreting areas with aligned strike angles, we so far mainly considered the overflow of fine-grained aeolian material, but there are other possibilities. These include dry granular flows, which are known on Earth and could be even more abundant on Mars given the height of the volcanoes and a lower gravity. Such terrain might be dangerous for future astronauts.

It turns out that the gravity aspects are sensitive not only to uniformly ordered sediments of shallow basins (the ‘plates’), but also to the supposed lahars (‘mud flows’), provided that the area covered by them is above the ground resolution limit of the gravity field model (in our case ~130 km). They were described on Mars more than 40 years ago and are further studied (Allen, Reference Allen1979; Christiansen, Reference Christiansen1989; Cuřín et al., Reference Cuřín, Brož, Hauber and Markonis2023) recently amended by laboratory experiments (Brož et al., Reference Brož, Krýza, Wilson, SJ, Hauber, Mazzini, Raack, MR, ME and MR2020).

Pedersen (2013 and a set of references in this paper) writes that ‘intrusions have been proposed as mechanisms for generating major lahars in various regions…’ [roughly for orientation at: φ…°N or S, λ…°E], such as Mangela Valles [10S, 210], Elysium Fessae/Fossae [25N, 140], Cerberus Fessae [10N, 165] and Athabasca Valles [10N, 155], ‘…either by melting a substantial amount of ground ice or by cracking the cryosphere allowing drainage of confined aquifers…’ and ‘…a plethora of volcano–ice interactions known from Earth have been suggested to occur on Mars, and the transition zone between the Elysium rise and Utopia Planitia has been commonly suggested to host a variety of these features…’; he studied these morphologically complicated features in detail (Pedersen, Reference Pedersen2013).

We have checked, using our figures in the main text and in S3, together with topography maps, that for example the huge Elysium lahar field [~30, ~135], situated approximately west of the Elysium volcano (Fig. 9(f); S3: 50–59), belongs to the areas with the highest Comb factor. These plates are large enough (~1000 km) to be well above our ground resolution (~100 km).