Mars Science Poster 2

Candidate 1,2,#,* mud volcanoes in the Northern Plains of Mars 1,* 1,* 3,* Edwin S. Kite

, Niels Hovius , John K. Hillier , Jonathan Besserer

1. Department of Earth Sciences, Cambridge University; 2. now at Earth and Planetary Sciences, University of California, Berkeley; 3. Université de Nantes; # poster author ([email protected]); * abstract author.

Abstract

Figure 6. Håkon Mosby mud volcano. Barents Margin, Earth; bathymetry from multibeam sonar. Viewed from N. Long axis of volcano extends 1.5 km. Source: Beyer et al., 2005 [21]

We describe large moated domes from the Scandia region, near the Phoenix landing site, Mars. The Scandia Tholi are the only large moated domes on Mars, and have been previously interpreted as basaltic volcanoes [1], mud volcanoes [2], or glacial kames [3] Their morphology corresponds closely to that of terrestrial submarine mud volcanoes. However, dome spectra lack the signatures of phyllosilicates or evaporites. No hypothesis satisfies all our observations, but we interpret the domes as the probable result of either mud volcanism (MV) or, less likely, magma-ice/magma-water interactions. Either would require a stratigraphic column enriched in volatiles to substantial depths.

Figure 7. Dome 13 in our catalogue. Borealis back-basin, Mars; topography from laser altimetry. Viewed from S. Long axis of structure extends 50 km. Red is -4433m, white is -4863m. The straight ridge extending SE across the moat is an artifact.

Context

Results of morphological analysis Dome morphology is diverse, but an example showing many of the observed features is the 40x50 km Dome 13 [Fig. 7]. Maximum elevations (~420m above plains) are at the summit of a 7x5 km marginal peak. Just as the marginal peak is associated with a pit, the dome as a whole is encircled by a moat. Moat width (~2.5km) and floor elevation does not change with azimuth from the dome centre. The interior shows evidence for (formerly more complete?) marginal and interior annular ridges. Marginal slopes reach 5 - 8°. By drawing numerous profiles in gridded MOLA toporaphy from the dome centres to a hand-picked dome edge, and averaging the results between domes (n=29), we confirmed the visual impression of multiple ridges in many domes. Aspect ratio is (0.93±0.22)%. The best-fitting simple cone has aspect ratio (rise:run)~1:100 = 0.53° slope, confirming the remarkably low relief of the domes Relief is (418±90) m, diameter is (49±16) km. The amplitude of internal roughness on dome surfaces is greater than that on the background plains, making the domes resemble welts. Most domes have central peaks, which have 5°-9° flanks. Their extent is always 7x4 km or similar, and they are often accompanied by peakmarginal pits which resemble subsidence features. Concentric rings (indurated ring-faults or, less likely, flow features) are present in 6 domes and possibly present in an additional 6. Inner ridges are always circular, even though parent domes are always elliptical. Moat width and depth show no correlation with dome height, excluding flexural origin for the moats. Mean moat width is (2.3±0.37) km (n=12). This narrow range of values suggests that moat width is related to reservoir depth, and that the reservoir is of the same order as the moat width, that is, fairly shallow. However, moat width does increase with dome diameter, and the largest moats show box-profile, deep moats. These observations are most consistent with moat origin by collapse after removal of material from an underlying reservoir, as expected for mud volcanism. Slope-aspect relationships in the domes region show significant deviations from randomness. After detrending, the relative frequency of slopes facing S or N exceeds other aspects; this excess has local maxima at (S±5°) and (N±5°). Because N-facing slopes are more abundant than S-facing slopes, S-facing slopes must be, on average, steeper. A control region at the same latitude shows no comparable anomaly.

Olympia dome Figure 2. Location of catalogued domes. Topography has been detrended to remove basin slope. -0.4 0 0.2 (0.8-1.1)? (1-4)?? approx. depth below plains, km

The domes and associated mounds and ridge-bounded depressions are largely contained within a basin, the 'Borealis backbasin,' approximately 1500 km x 700 km and 400 m deep. A saddle at an elevation of -4850m near 78N, 240E connects the Borealis back-basin to the Borealis basin, which is the lowest region in the northern hemisphere of Mars [Fig. 3]. The northward extent of the basin is obscured by the Olympia dome, which is up to 600m thick. Studies of crater fill in the northern plains using a differential compaction model suggest a post-basement sedimentary cover thickness that is greatest in the Borealis back-basin and Scandia basin [4]. Lineations defined by multiple domes are subparallel to the continuation of extensional grabens associated with Late Hesperian / Early Amazonian diking at Alba Patera. Moreover, mound chains are radial to Alba Patera. The stratigraphy of the Northern Plains of Mars is not well known, and endmember stratigraphic hypotheses have very different implications for the history of water on Mars [5-6; Fig. 4]. Among the lacunae in our understanding are:- 1) what fills the gap (if there is one) between the top of the Early Noachian basement and the bottom of the Early Hesperian basaltic flood lavas [7] - igneous materials, or ocean sediments?; 2) how is ice distributed as a function of position and time? Here we address these problems through a geomorphological study supported by analysis of nearinfrared spectral data.

Figure 3. Hemispheric context. Stereographic projection centred on N Pole; prime meridian extends from bottom centre of figure. White: location of domes and ridge-bounded depressions. Orange: gypsum-bearing dunes. Green arrows: discharge direction from circum-Chryse channels. Yellow and red arrows: Spillover points at -4850m and -4350m respectively. Black contours at -5000m, -4800m, and 4600m.

Figure 4. Endmember stratigraphic hypotheses for the Borealis back-basin. Absolute depths very poorly constrained. Left: Thick layer of Noachian ocean sediments [5] as source for MV. Right: Volcanic materials extend to basement; domes are due to silicate volcanism. A periglaciallymodified mantle (blue) coats the terrain, underlain by the Vastitas Borealis Formation (light green), Late Hesperian and perhaps earlier basalts (light orange), and ancient basement breccias and sediments (purple/yellow).

Interpretation All bodies with silicate crusts and diameters >103 km in our solar system show widespread basaltic volcanism, but only Earth has confirmed MV. Therefore, when assessing an extraterrestrial construct, one should assign a high prior probability to igneous volcanism. However, (1) The suite of dome morphologies can be matched one-for-one with MV in the S Caspian Basin [e.g., 11] and Gulf of Cadiz [e.g, 12], although Martian domes have diameters ~5 times greater than their largest terrestrial counterparts. (2) Moat morphometry excludes flexural origin and suggests collapse during withdrawal of material from a subsurface reservoir. Collapse moats are found in association with major submarine MV on Earth. (3) The slope-aspect anomaly suggests that nearsurface dome material has been subject to incomplete insolation-driven processing, and that dome near-surface material was partially volatile. However, it is possible that dome near-surface material is compositionally distinct from material making up the bulk of the domes. (4) Few volcanic constructs are found closer to Alba Patera, inferred to have triggered dome construction. So there must be a major increase in the fusibility of materials overlying dykes in our study area. Granitoid rocks are extremely rare on Mars, but volatile-rich deposits satisfy this requirement. (5) Mars has thick sediment piles, permitting MV. Crater fill studies suggest that the Scandia and Borealis back- basins contain the greatest thickness of post-Late Noachian sedimentary cover in the entire Northern Plains [4], making these preferred MV locations. We infer that the domes were likely emplaced by MV.

Discussion: isostasy Isostatic considerations indicate that, if the domes formed subaerially, their source reservoir must be below the Early Hesperian flood basalts, because the thickness of the Vastitas Borealis Formation is insufficient to engender 400+ m of extrusional relief. If dome-forming material is mud with a density of 1.8 - 2.3 g/cm3 , and the overburden has the density of basalt, our results indicate a depth to the source layer of 535-1367 m. This brackets the inferred thickness of the Hesperian flood basalt [7], and tends to support the proposal that ancient ocean sediments are present at depth beneath the Northern Plains [5]. Alternatively, if the dome-forming material is silicate magma with a density of 2.65 - 2.9 g/cm3, our results indicate a depth to the source layer for an overburden with the density of basalt of ~3.5 km. This might represent a magma chamber at or near the top of the buriedcrater-bearing basement [16] .

Discussion: triggering Tectonic convergence aids MV on Earth by pressurizing shale piles, generating anticlines that localise the buoyant mud, and weakening faults that provide mobile mud with access to the surface [17]. Although wrinkle ridges, interpreted as blind thrusts, are present in the study area, both the orientations of individual dome long axes and the lineations defined by adjacent domes are oblique to the continuation of wrinkle wridge trends beneath our study area. Possible alternatives to tectonic convergence as triggers for MV include:- 1) Rapid (> 10 mm/yr) sedimentation: Rapid sedimentation following the most recent catastrophic-outflow event would led to transient overpressure of fluids in the sediment pile. We provisionally assign the mud source to sub-basalt, inferred Noachian – Early Hesperian, outflow or ocean deposits (Fairen et al., 2003). It is difficult to see how a fairly uniform outflow-deposit load could expel mud from such deep deposits. However, if the mud eruptions occurred into a transient flood deposit of water or ice, eruptions sourced from within the post-basalt cover could generate the observed relief. In that case, overpressure would be a possible trigger for mud volcanism. 2) Ice: Differential loading (107 Pa in the vertical) of a regionally connected mud reservoir by an advancing or retreating ice sheet would expel mud outboard of the ice-sheet margin through any pre-existing fractures. Increased reservoir temperature, following thermal insulation by overlying ice, would decrease mud viscosity and exsolve CO2. 3) Diking: The intersection of hot hydrothermal fluids circulating above dykes with a thick, ice-rich sediment column could have directly triggered MV in the planes above the dykes [18-19]. Alternatively, graben radiating from Alba Patera may have created planes of weakness later exploited by pressurized fluids.

Summary Although puzzles remain and work is ongoing, these results support the previous suggestion [2] that the Scandia Tholi were emplaced by MV. If confirmed, this may have implications for the ease of future drillrig access to ancient sedimentary deposits in the Northern Plains. There is also a tempting geographic link with young evaporates [20].

Results of spectral analysis OMEGA visible/near-IR spectra of the domes show no significant differences between dome and non-dome terrain. Both band depth methods and linear unmixing models [8] yield compositions dominated by ferric oxides and pyroxenes, everywhere in the region. Some structures are visible, e.g. in low-calcium pyroxene (LCP) unmixing ratio [Fig. 1], but there is no general correlation with topography. This spectral uniformity may be due to recent mantling. However, if it reflects the composition of the dome interiors, the observations favour but do not require dome origin by silicate volcanism. If the domes were emplaced by MV, the spectral data may indicate that they tap a source layer of similar composition to the adjacent plains, which are believed to have been reworked from Late Hesperian flood deposits [9-10].

Figure 1. Left: Hypersthene/albedo percentage using linear unmixing model. Right: MOLA topography (m) draped over shaded relief. OMEGA 3.9km/pixel track.

Figure 5. a) Terrestrial seismic impedance data for S. Caspian submarine mud volcano (from [22]. b) Regionally-detrended digital elevation model of Dome 5. c) Interpretation of DEM, with putative indurated ring-faults marked by yellow dashed lines.

Discussion: alternatives Could the domes result from magma-ice or magma-water interactions? At present, we disfavor magma-ice or magma-water interaction as a hypothesis for dome origin because:- most importantly, subsidence moats with the characteristics we describe are lacking in terrestrial tuyas; kilometrewavelength, hectometre-amplitude surface roughness would also be unexpected for tuyas; if dome extrusion was triggered by Alba Patera-radial dyking, it would be surprising that silicate volcanism only nucleated at great distances from the volcanic centre; the rise:run (1:100) of the Martian domes is very low compared to relevant terrestrial analogues; and, least importantly, the Scandia Tholi do not resemble extrusive structures in Cavi Angusti, previously interpreted as tuyas [13]. However, magma-water and magma-ice interactions can produce a very wide range of morphologies [e.g., 14], and their physics is incompletely understood [15]. Therefore, we continue to investigate this hypothesis.

References: [1]Garvin, J., et al., Icarus 145(2), 648-652, 2000. [2]Tanaka, K., Nature 437, 991-994, 2005. [3]Fishbaugh, K.E., and J.W. Head, LPSC 33, 1426, 2001.[4] Buckzcowski, D., J. Geophys. Res. 112, E09002, doi:10.1029/2006JE002836, 2007. [5] Fairen, A.G., et al., Icarus 165(1), 53-67, 2003. [6] McEwen, A.S., et al., Science 317, 1706-1709, 2007. [7] Head, J.W., et al, J. Geophys. Res. 107 (E1), doi: 10.1029/2000JE001445, 2002. [8] Combe, J.P., PhD thesis, U. Nantes, 2005. [9] Tanaka, K. et al., USGS Scientific Investigations Map 2888. [10] Carr, M., and J.W. Head, J. Geophys. Res., 108 (E5), 5042, doi:10.1029/2002JE001963.2, 2003. [11] Evans, R.J., et al. Basin Res., 19(1), 153-163, 2007. [12] Somoza, L., et al. , Mar. Geol. 195, 153-176, 2003. [13] Ghatan, G. J., et al., J. Geophys. Res., 108 (E5), 5045, doi:10.1029/2002JE001972. [14] Tuffen, H. , J. Geophys. Res., 112, B03203, doi:10.1029/2006JB004523. [15] Wohletz, K.H. and Zimanowski. B. , Terra Nostra 2000/6, 515-523, 2000. [16] Frey, H., et al., Geophys. Res. Lett., 29(10), 1384, doi:10.1029/2001GL013832., 2002. [17] Kopf, A.J., Rev. Geophys. 40(2), art. no. 1005., 2002. [18] Tanaka, K.L., Fourth Mars Polar Sci. Conf. #8024, 2006. [19] McKenzie, D. and F. Nimmo, Nature 397, 231-233. [20] Langevin, Y., et al., Science 307, 15841586, 2005. [21] Beyer, A., et al., Mar. Geophys. Res., 26(1), 61-75, 2005. [22] Stewart, S.A., and R.J. Davies, AAPG Bull. 90(5), 771-786, 2006. Acknowledgements: E.K. acknowledges financial support from Pembroke College, Cambridge, and from a Berkeley Fellowship. We thank Talfan Barnie and Andy Lea-Cox for assistance, and Michael Manga for comments.