1 Pacing Early Mars river activity: Embedded craters in the Aeolis Dorsa 2 region imply river activity spanned ≳ (1-­‐20) Myr 3 Edwin S. Kitea (ekite@caltech.edu), Antoine Lucasa, Caleb I. Fassettb 4 a Caltech, Division of Geological and Planetary Sciences, Pasadena, CA 91125 5 b Mount Holyoke College, Department of Astronomy, South Hadley, MA 01075 6 7 Abstract: We find net sedimentation rate ≲(13-­‐200) μm/yr in the Aeolis Dorsa region, Mars, 8 using the frequency of crater-­‐river interactions. This sets a lower bound of 1-­‐20 Myr on the 9 total interval spanned by fluvial activity, which we correlate to the Noachian-­‐Hesperian 10 transition. The main uncertainty is the impact flux at the time of deposition. This result rules 11 out basin-­‐filling by a single catastrophic episode, such as a single impact-­‐induced water-­‐vapor 12 greenhouse. 13 14 1. Introduction: On Mars, many craters are embedded within sedimentary sequences, leading 15 to the recognition that the planet’s geological history is recorded in “cratered volumes,” rather 16 than just cratered surfaces (Edgett and Malin, 2002). For a given impact flux, the density of 17 craters interbedded within a geologic unit is inversely proportional to the deposition rate of 18 that geologic unit (Smith et al. 2008). To use embedded-­‐crater statistics to constrain deposition 19 rate, it is necessary to distinguish interbedded craters from craters formed during and after 20 exhumation. However, on Mars, erosion can exhume intact impact craters complete with ejecta 21 blankets from beneath hundreds of meters of overlying sediment (e.g., Edgett, 2005). 22 Variations in burial, exhumation, erosion, and (for icy targets) the presence or absence of 23 viscous relaxation produce varying crater-­‐preservation styles. Therefore it is difficult to 24 determine which craters are syndepositional. 1 25 Regionally-­‐integrated highland valley networks formed during a period of enhanced 26 precipitation-­‐fed runoff around the Noachian-­‐Hesperian boundary ~3.6-­‐3.7 Gyr ago (Irwin et 27 al., 2005). This activity ceased planetwide at roughly the same time, perhaps even 28 synchronously (Fassett et al., 2011), although limited and/or localized fluvial activity 29 continued afterward (e.g., Grant & Wilson, 2011; Mangold et al., 2012). Lake-­‐basin hydrology 30 disfavors valley-­‐network formation by a single deluge (Barnhart et al., 2009; Matsubara et al., 31 2011). Improved constraints on the pace and persistence of fluvial activity during the 32 Noachian-­‐Hesperian transition are required to understand early Mars climate and habitability. 33 It remains unclear whether the fluvial geomorphology of the Noachian-­‐Hesperian transition is 34 a palimpsest of transient events triggered by volcanic eruptions or impacts (Segura et al., 2008; 35 Toon et al., 2010; Kite et al., 2011; Wordsworth et al., 2013), or alternatively records a 36 sustained (>>1 Kyr) wet interval(s) caused by unusual orbital conditions (Kite et al., 2013a), an 37 early greenhouse (Sagan & Mullen, 1972), or an impact-­‐triggered excursion to a warm stable 38 state (Segura et al., 2012). 39 Here we use interactions between craters and fluvial deposits to determine sedimentation 40 rate within a candidate Late Noachian/Early Hesperian sedimentary basin containing 41 numerous fluvial deposits. A river flowing over a heavily cratered sedimentary landscape will 42 be frequently diverted by crater rims and/or deposit sediment in pools corresponding to the 43 crater interiors (Figure 1). These craters, and craters partly overlain by river deposits, are 44 readily identified as being syndepositional. 45 46 2. River-­‐crater interactions: We search for river-­‐crater interactions within exceptionally 47 numerous and exceptionally well-­‐preserved fluvial channel deposits exhumed by erosion in 48 the Aeolis Dorsa region, formerly termed Aeolis-­‐Zephyria Planum (Burr et al., 2009, 2010; note 49 that we use Aeolis Dorsa to refer to the formally defined region, not just the ridges within that 2 50 region). The area of our search (~6S, 152E; Figure 2) is ~500m below a surface with an Early 51 Hesperian, minimum model age of 3.69(+0.05/-­‐0.07) Gyr (Zimbelman & Scheidt, 2012, using 52 the Ivanov et al., 2001 production function and the Hartmann & Neukum, 2001 chronology 53 function). It forms part of an eroded deposit that is thought to be Late Noachian or Early 54 Hesperian (Irwin et al., 2004). These dates correlate the Aeolis rivers to the Noachian-­‐
55 Hesperian transition and to the lower Gale Crater mound (Fassett & Head, 2008; Kerber & 56 Head, 2010; Thomson et al, 2011; Zimbelman & Scheidt, 2012). 57 The Martian rivers are qualitatively similar to meandering-­‐river deposits on Earth. 58 Interactions with craters are, therefore, easily recognized as anomalous. Impact craters are 59 identified by random distribution across the landscape, obliteration of older geological 60 structures within the crater, and especially circularity and upturned rims. Interfluve material 61 would be interpreted as overbank material (fine-­‐grained deposits from levee-­‐overtopping 62 floods) if this were a basin on Earth. However, in the absence of grain-­‐size data or diagnostic 63 levee-­‐breach features, wind-­‐blown dust and silt cannot be ruled out (Haberlah et al., 2010). 64 Locally the range in elevation spanned by deposits containing fluvial channels is > 100m. 65 Because the dips of the river-­‐deposit bounding surfaces are modest, the corresponding 66 stratigraphic thickness (Δz) is similarly > 100m. Regionally, channel deposits are exposed over 67 a ~1km elevation range, but postdepositional modification complicates reconstruction of Δz 68 (Lefort et al., 2012). A likely lower bound on Δz is the difference between the modern surface 69 and a surface interpolated inward from low points surrounding the fluvial region (using a 70 inverse-­‐distance-­‐weighted grid constrained to Mars Orbiter Laser Altimeter – MOLA – 71 Precision Experimental Data Record points). This gives Δz ≳ 300m. 72 17 exhumed craters are found (Supplementary Table) at multiple stratigraphic levels 73 within 2100 km2 of fluvial deposits surveyed with CTX images (5-­‐6 m/pixel). Identifications 74 were checked where possible with higher-­‐resolution images (HiRISE or MOC). In addition to 3 75 these features that are definitely identified as embedded within the stratigraphy and having 76 definite impact crater morphology, an additional 43 candidates were found. Crater diameters 77 were obtained by fitting a circle to the visible arc of the crater edge. The embedded-­‐crater size-­‐
78 frequency distribution for D >250m (D is diameter) has a cumulative power-­‐law slope slightly 79 shallower than –2, ~1 less than the production-­‐function slope of –3 in this range. This is 80 expected for a crater population embedded within a volume, as previously noted for 81 subsurface fault populations (Yielding et al., 1996), because the fraction of embedded craters 82 that are exposed at the ~horizontal deflation surface is proportional to the absolute difference 83 between their rim and floor elevations, which is greater for large than for smaller craters. (A 84 pristine crater population on a geologically stable surface would parallel the production 85 function.) Craters D ≲ 250 m are still further underrepresented, which we interpret as the 86 result of survey incompleteness or poor exhumed-­‐crater preservation at small sizes. 87 88 3. Constraints on sedimentation rate and fluvial timescales: We make the following initial 89 assumptions – 90 (i) Cratering is a Poisson process with an initial crater depth, d ≈ 0.2D for D < 1km (Melosh, 91 1989). (HiRISE DTM measurements support this d/D for fresh Mars craters; Wes Watters, 92 via e-­‐mail). 93 (ii) Erosion does not preferentially expose craters for counting (this would increase the 94 effective d/D). 95 (iii) During deflation of the deposit an embedded crater is invisible until the deflation 96 surface reaches the embedded-­‐crater rim. It is then visible at its original diameter until the 97 deflation surface reaches the level of the bottom of the crater. 98 We discuss possible violations of these assumptions later. 4 99 The flux of impact craters at the time the deposit was forming, f (D>Di), is obtained 100 using the crater-­‐production and crater-­‐chronology functions recommended by Werner & 101 Tanaka (2011), with one exception.1 Differences between the Hartmann and Ivanov/Neukum-­‐
102 Hartmann functions lead to <30% disagreement in resurfacing rate for the size-­‐range of 103 craters used here (Figure 3), which is unimportant compared to other uncertainties. The 104 expected number of embedded craters that outcrop at the modern surface, Ncr, is given by 105 Ncr (D>Di)= f (D>Di) (d/D) D a / S 106 where a is count area, and S is accumulation rate. Excluding D < 290m craters and assuming an 107 age in the range 3.7 – 3.9 Gyr, least-­‐squares fitting of accumulation rates to the data (Figure 3) 108 gives S = 50-­‐200 μm/yr (or 13-­‐50 μm/yr including all candidates). Therefore, S ≈ 13-­‐200 109 μm/yr for the range of likely ages. 110 The most important uncertainty is f. The Aeolis Dorsa river deposits were emplaced during 111 a period of higher crater flux f when rapid changes in f are also possible, potentially associated 112 with the Late Heavy Bombardment. Therefore, small changes in the age of the deposit may lead 113 to large changes in S (Werner & Tanaka, 2011). Crater-­‐chronology functions are defined based 114 on the lunar sample collection, whose interpretation is somewhat model-­‐dependent, and the 115 translation of these data to Mars is challenging (Ivanov, 2001; Robbins, 2013). 116 Other uncertainties tend to lead to an underestimate of exhumed-­‐crater frequency. (a) If 117 channel belts aggraded faster than interfluves and are erosionally resistant, and erosion is by 118 vertical downcutting uncorrelated with laterally adjacent terrain, then at any given time the 119 modern deflation surface will preferentially expose the erosion-­‐resistant, crater-­‐deficient 120 units. In this case our procedure would underestimate the frequency of craters per unit volume 121 averaged over the basin. On Earth, channel belts always aggrade faster than their floodplains 122 on interannual timescales, but levee breaches and avulsions maintain constant aggradation 1 The chronology function given for Hartmann (2005) by Werner and Tanaka (2011) is in error. Instead, we use N(1)=3.79 x 10-­‐14 (exp(6.93t)-­‐1)+5.84 x 10-­‐4 t (here, t is in Gyr). 5 123 rate across the floodplain averaged over depths greater than ~1 channel depth (Mohrig et al., 124 2000). Channel-­‐width measurements (Burr et al., 2010), standard fluvial scaling relations, and 125 our HiRISE DTM measurements of negative-­‐relief channel depths all indicate that Aeolis 126 channel depths should be small compared to the original depth of the craters in our count (i.e. 127 channel depth << 30-­‐60m). Therefore lateral gradients in sedimentation rate are unlikely to be 128 important if the interfluves are floodplain deposits. (b) Cliff-­‐forming units that are very 129 resistant to vertical abrasion will be removed by lateral mass wasting as surrounding weaker 130 material is eroded. In this case, no embedded craters within those cliff-­‐forming units will be 131 included in the count: craters on top of the cliff-­‐forming units will be hard to distinguish from 132 relatively recent synerosional craters, and craters within those units will be blanketed by talus 133 throughout the erosion process. Only craters on the substrate that have been partially 134 exhumed from beneath the cliff-­‐forming unit will be visible, and so the effective count area 135 then scales with the perimeter of the cliff-­‐forming unit, rather than with its area. (c) S would 136 halve in the extreme case that all craters initially have a secondary-­‐like d/D (i.e., ~0.1). d/D is 137 currently <0.1 in our HiRISE DTMs, but this could be due to incomplete erosion or recent 138 infilling. (d) Supposing small craters were all erosionally resistant and formed mesas as tall as 139 their diameter, asumption (iii) would be violated and a surface count would overestimate the 140 true crater density. However, our CTX and HiRISE DTMs show that small exhumed craters are 141 not locally highstanding in this region, although they do tend to be preserved with rims intact, 142 so (iii) is probably a good approximation. (e) We assume craters are either present at full 143 diameter or eliminated completely, but the apparent diameter of an eroding crater may be less 144 (or more) than its initial diameter. (f) Finally, we neglect possible erosion of craters during the 145 period of net accumulation. 146 Because these errors tend to lead to an underestimate of the number and size of craters 147 that formed during the interval of fluvial deposition, we interpret our data as a lower limit on 6 148 time for accumulation. Assuming Δz ≳ 300m, the range of minimum deposition timescale is 149 Δz/S ~1-­‐20 Myr for the range of likely ages (or 0.6-­‐40 Myr assuming a wider range of age 150 uncertainty, from 3.6-­‐4.0 Gyr).2 151 152 4. Discussion: The simplest interpretation of these data is fluvial aggradation at rates 153 comparable to Earth (fluvial aggradation rates ranging from 50-­‐600 μm/yr are compiled from 154 Earth data by Miall, 2012). Aeolis Dorsa sedimentation cannot be distinguished from later 155 sedimentation on the basis of sedimentation rate alone. Putzig et al. (2009) correlate radar 156 reflectors within the North Polar Layered Deposits to obliquity cycles over the last 4 Ma and 157 obtain S ~1 mm water ice/yr. Similarly, Lewis et al. (2008) correlate bed:bundle ratios in 158 Becquerel Crater at 22°N to Milankovitch beats and obtain S ~30 μm/yr. Typical equatorial 159 mound rhythmic-­‐layer thicknesses of 3-­‐20m (Lewis et al., 2010) imply accumulation at 20-­‐200 160 μm/yr if forced by obliquity cycles (0.12 Myr), or 100-­‐800 μm/yr if forced by precession 161 (0.025 Myr effective period at the equator). Zabrusky et al. (2012) find S > 3 μm/yr in 162 Meridiani using pedestral craters. Rhythmites have been hypothesized to be relatively young 163 (Grotzinger & Milliken, 2012). Modern gross sedimentation rate on Mars is 10-­‐100 μm/year 164 from dust storms (e.g., Drube et al., 2010). 165 On Earth, mean sedimentation rate frequently decreases with increasing measurement 166 duration (Jerolmack & Sadler, 2007) as a result of power-­‐law fluctuations of the boundary 167 between erosion and deposition (Schumer & Jerolmack, 2009). This “Sadler effect” is 168 ubiquitous at short timescales near coasts but is less relevant on the longest timescales, or 169 where erosion is unimportant (Jerolmack & Sadler, 2007). It is not clear whether the Sadler 170 effect should apply to Mars sediments. Martian weather is remarkably predictable on both 2 Our constraint is slightly tighter than that given in our preprint arXiv:1207.6726v1, because we have used an improved fitting procedure, used denser grids for interpolation, and corrected an error in the Werner & Tanaka (2011) chronology functions. 7 171 synoptic and interannual timescales, probably because of the lack of a large energy capacitor 172 analogous to Earth’s ocean (Read & Lewis, 2004). Therefore, we might speculate that Mars’ 173 sedimentary record, as the imprint of the atmosphere on rocks, is less chaotic than its Earth 174 counterpart. Consistent with this, quasi-­‐periodic bedding is common on Mars (Lewis et al., 175 2010) and angular unconformities are rare. CTX images suggest that exhumed craters in this 176 region are concentrated at a few stratigraphic levels, consistent with omission surfaces. 177 The sediment source for the Aeolis Dorsa deposits is uncertain, and so it is not clear 178 whether the accumulation of the deposits was rate-­‐limited by fluvial transport or by aeolian 179 transport. One option is that sediment was fluvially transported from highlands to the south. 180 The volume of valley networks draining toward Aeolis Dorsa appears to be much smaller than 181 the volume of the clastic wedge, and no complete transport pathways are visible. (However, 182 the volume of valley networks is only a lower bound on the total fluvially-­‐eroded volume). It is 183 conceivable that the Aeolis Dorsa deposits are fluvially reworked ancient highlands crust, 184 which would require that the dichotomy boundary was once further to the north. But crater-­‐
185 floor tilts suggest that the dichotomy boundary was in place near its current location prior to 186 the late Noachian (Watters et al., 2007). An attractive possibility is that the fluvial deposits are 187 reworked from relatively weak aeolian or niveoaeolian deposits accumulating at the highland-­‐
188 lowland boundary (Irwin et al., 2004). 189 The relatively low embedded-­‐crater frequency demonstrates that sedimentary processes 190 on Mars could outcompete cratering even when the impact flux was higher early in Solar 191 System history (Howard et al., 2007). This is in line with low Platinum Group Element 192 concentrations in 3.8 Ga metasediments from Earth (Anbar et al., 2001). 193 194 5. Implications for ancient climate and MSL’s mission to Gale Crater: A lower bound of 1-­‐
195 20 Myr for the total interval spanned by fluvial deposition rules out basin-­‐filling by a single 8 196 catastrophic episode, and is consistent with hydrologic and total-­‐erosion estimates for fluvial 197 activity around the Noachian-­‐Hesperian boundary (e.g., Barnhart et al., 2009; Hoke et al., 198 2011). Possible climate regimes include multiple climate transients (Segura et al. 2008, 199 Wordsworth et al. 2012), or intermittent precipitation-­‐fed runoff over ≥105 yr (Barnhart et al., 200 2009). The sedimentary rock record of Mars appears to record a small fraction of Mars history, 201 perhaps because surface liquid water was necessary for lithification and was only available 202 intermittently (Moore, 1990, Knoll et al., 2008, Andrews-­‐Hanna & Lewis, 2011, Kite et al., 203 2013a). 204 Based on observed and candidate embedded-­‐crater frequency in Aeolis Dorsa, we predict 205 an Aeolis Mons sedimentation rate of ≲13-­‐200 μm/yr, including nondepositional intervals, 206 averaged over the ~2km-­‐thick lower mound. This assumes that sedimentation rates across the 207 broader Aeolis region were fairly uniform. Our prediction is testable with MSL measurements 208 of cosmogenic noble gases (e.g., Shuster et al., 2012), meteoritic Ni (Yen et al., 2006), meteoritic 209 organic matter (OM), and small embedded craters. Mars Hand Lens Imager’s (MAHLI’s) 14μm-­‐
210 per-­‐pixel resolution permits identification of ≥40μm-­‐thick varves, if they exist. 211 River deposits exist at the Gale mound, but those clearly identifiable from orbit appear to 212 postdate the accumulation of the lower unit of Aeolis Mons. It has been argued that the 213 primary sediment source for the Gale mound is atmospherically-­‐transported sediment (Pelkey 214 et al., 2004; Thomson et al., 2011; Kite et al., 2013b). If this is true, then the extrapolation of the 215 Aeolis Dorsa sedimentation rate to Aeolis Mons depends on the assumption that river 216 sedimentation in the Aeolis Dorsa region was supply-­‐limited, with sediment introduced to the 217 region by aeolian processes (e.g., Irwin et al., 2004; Haberlah, 2010). 218 Sediment accumulation rate affects the hunt for OM on Mars. Fast sedimentation dilutes 219 OM delivered by meteoritic infall – thus reducing the “meteoritic background level” (Summons 220 et al., 2011). Preservation of OM introduced at the surface is affected by time-­‐to-­‐burial to a 9 221 depth of order 1m. During this time, OM is vulnerable to degradation by radiation, atmospheric 222 oxidants, and/or UV, to an extent that depends on the unknown redox state, composition and 223 thickness of Early Mars’ atmosphere (Pavlov et al., 2012). On Earth, overbank environments 224 and oxbow lakes in meander belts are favored among subaerial environments for preserving 225 organic matter (Summons et al., 2011). 226 Future work might use MSL data and further orbital analyses to assess the relation 227 between exposure time, crater frequency, OM preservation, and atmospheric paleopressure. 228 The existence of >3.7 Gyr-­‐old, ≤102m-­‐diameter craters should place an upper limit on ancient 229 atmospheric pressure, because small impactors do not form hypervelocity craters beneath 230 thick atmospheres (Popova et al., 2003); we are pursuing this quantitatively. 231 232 Acknowledgements: We thank Mike Lamb, Ken Farley, the participants in Caltech’s Mars 233 Fluvial Geomorphology Reading Group, Devon Burr, Alexandra Lefort, Robert Jacobsen, Wes 234 Watters, and especially Oded Aharonson, Mark Allen, Woody Fischer, Kevin Lewis, and Nick 235 Warner, for discussions. We additionally thank Kevin Lewis and Oded Aharonson for sharing 236 their preprint on cyclic bedding. We thank two anonymous reviewers for their comments, and 237 Jeff Moersch for editorial handling. We thank the HiRISE team for maintaining a responsive 238 public target request program, HiWish, which was useful for this work. DTMs produced for this 239 work, 240 ESP_024497_1745 (@2.5m/pixel) are available for unrestricted further use from the 241 corresponding author. 242 243 References: 244 Anbar, A.D., et al., 2001, Extraterrestrial iridium, sediment accumulation and the habitability of 245 the early Earth's surface. J. Geophys. Res., 106, 3219-­‐3236. ESP_17548_1740/ESP_019104_1740 10 (@1m/pixel) and PSP_007474_1745/ 246 Andrews-­‐Hanna, J.C. & Lewis, K.W., 2011, Early Mars hydrology: 2. Hydrological evolution in 247 the Noachian and Hesperian epochs. J. Geophys. Res. 116, E02007. 248 Barnhart, C.J. et al., 2009, Long-­‐term precipitation and late-­‐stage valley network formation: 249 Landform simulations of Parana Basin, Mars. J. Geophys. Res., 114, E01003. 250 Burr, D.M., et al., 2009, Pervasive aqueous paleoflow features in the Aeolis/Zephyria Plana 251 region, Mars. Icarus, 200, 52-­‐76. 252 Burr, D. M., et al., 2010, Inverted fluvial features in the Aeolis/Zephyria Plana region, Mars: 253 Formation mechanism and initial paleodischarge estimates. J. Geophys. Res., 115, E07011. 254 Drube, L., et al., 2010, Magnetic and optical properties of airborne dust and settling rates of 255 dust at the Phoenix landing site. J. Geophys. Res., 115, E00E23. 256 Edgett, K.S., 2005, The sedimentary rocks of Sinus Meridiani, Mars Journal, 1, 5-­‐58. 257 Edgett, K. S., & Malin, M. C., 2002. Martian sedimentary rock stratigraphy, Geophys. Res. Lett., 258 29, 2179. 259 Fassett, C.I., & Head, J. W. , 2008, The timing of martian valley network activity: Constraints 260 from buffered crater counting. Icarus, 195, 61-­‐89. 261 Fassett, C. I., & Head., J. W., 2011, Sequence and timing of conditions on early Mars. Icarus, 211, 262 1204-­‐1214. 263 Grant, J. A., & Wilson, S. A., 2011, Late alluvial fan formation in southern Margaritifer Terra, 264 Mars. Geophys. Res. Lett., 38, L08201. 265 Grotzinger, J.P., & Milliken, R.E., 2012, The sedimentary rock record of Mars, p. 1-­‐48 in 266 Grotzinger, J.P., & Milliken, R.E. (Eds.), Sedimentary Geology of Mars, SEPM Special Publication 267 no. 102. 268 Haberlah, D., et al., 2010, Loess and floods: High-­‐resolution multi-­‐proxy data of Last Glacial 269 Maximum (LGM) slackwater deposition in the Flinders Ranges, semi-­‐arid South Australia. Quat. 270 Sci. Rev., 29, 2673-­‐2693. 11 271 Hartmann, W.K., 2005, Martian cratering 8: Isochron refinement and the chronology of Mars, 272 Icarus, 175, 294-­‐320. 273 Hartmann, W.K. & Neukum, G., 2001, Cratering Chronology and the Evolution of Mars. Space 274 Sci. Rev., 96, 165-­‐194. 275 Hoke, M.R.T., et al. 2011, Formation timescales of large Martian valley networks. Earth Planet. 276 Sci. Lett., 312, 1-­‐12. 277 Howard, A., 2007, Simulating the development of martian highland landscapes through the 278 interaction of impact cratering, fluvial erosion, and variable hydrologic forcing, 279 Geomorphology 91, 332–363. 280 Irwin, R.P., et al., J.R., 2004, Sedimentary resurfacing and fretted terrain development along the 281 crustal dichotomy boundary, Aeolis Mensae, Mars. J. Geophys. Res., 109, E09011. 282 Irwin, R.P. III, et al., 2005, An intense terminal epoch of widespread fluvial activity on early 283 Mars: 2. Increased runoff and paleolake development. J. Geophys. Res. 110, E12S15. 284 Ivanov, B.A., 2001, Mars/Moon Cratering Rate Ratio Estimates. Space Sci. Rev., 96, 87. 285 Jerolmack, D.J. & Sadler, P., 2007, Transience and persistence in the depositional record of 286 continental margins. J. Geophys. Res., 112, F03S13. 287 Kerber, L. & Head, J.W., 2010, The age of the Medusae Fossae Formation, Icarus, 206, 669. 288 Kite, E.S., et al. 2011, Localized precipitation and runoff on Mars. J. Geophys. Res. 116, E07002. 289 Kite, E.S., et al., 2013a, Seasonal melting and the formation of sedimentary rocks on Mars, with 290 predictions for the Gale Crater mound, Icarus 223, 181–210. 291 Kite, E.S., et al., 2013b, Growth and form of the mound in Gale Crater, Mars: Slope-­‐wind 292 enhanced erosion and transport, Geology, doi:10.1130/G33909.1. 293 Knoll, A.H., et al., 2008, Veneers, rinds, and fracture fills: Relatively late alteration of 294 sedimentary rocks at Meridiani Planum, Mars. J. Geophys. Res., 113(E6), E06S16. 295 Lewis, K.W., et al., 2008, Quasi-­‐Periodic Bedding in the Sedimentary Rock Record of Mars. 12 296 Science, 322, 1532. 297 Lewis, K.W., et al., 2010, Lunar Planet Sci. Conf. 41, abstract # 2648. 298 Matsubara, Y. et al. 2011, Hydrology of early Mars: Lake basins. J. Geophys. Res., 116, E04001. 299 Lefort, A., et al., 2012, Inverted fluvial features in the Aeolis-­‐Zephyria Plana, western Medusae 300 Fossae Formation, Mars. J. Geophys. Res., 117, E03007. 301 Mangold, N., et al., 2012, The origin and timing of fluvial activity at Eberswalde crater, Mars. 302 Icarus, 220, 231-­‐251. 303 Melosh, H. J. 1989. Impact cratering: a geologic process. New York: Oxford University Press. 304 Miall, A.D., 2012, A new uniformitarianism: stratigraphy as just a set of “frozen accidents,” 305 preprint, http://www.geology.utoronto.ca/Members/miall/miall_bib/pdfs/Miall2012.pdf/. 306 Mohrig, D., et al., 2000, Interpreting avulsion process from ancient alluvial sequences: 307 Guadalope-­‐Matarranya system (northern Spain) and Wasatch formation (western Colorado). 308 Geol. Soc. Am. Bull., 112, 1787 – 1803. 309 Moore, J.M., 1990, Nature of the mantling deposit in the heavily cratered terrain of 310 northeastern Arabia, Mars. J. Geophys. Res., 95, 14279-14289.
311 Pelkey, S.M., et al., 2004 Surficial properties in Gale Crater, Mars, from Mars Odyssey THEMIS 312 data. Icarus, 167, 244-­‐270. 313 Popova, O., et al., 2003, Bolides in the present and past Martian atmosphere and effects on 314 cratering processes. Meteoritics and Planetary Science, 38, 905-­‐925. 315 Putzig, N., et al., 2009, Subsurface structure of Planum Boreum from Mars Reconnaissance 316 Orbiter Shallow Radar soundings, Icarus, 204, 443–457. 317 Read, P.L., & S.R. Lewis, 2004, The Martian Climate Revisited: Atmosphere and Environment of 318 a Desert Planet, Springer-­‐Praxis. 319 Robbins, S.J., 2013, Revised Lunar Cratering Chronology for Planetary Geological Histories, 320 LPSC 44, abstract no. 1619. 13 321 322 Sagan, C., & Mullen, G., 1972, Earth and Mars: Evolution of atmospheres and surface 323 temperatures. Science, 177, 52-­‐56. 324 Schumer, R., & Jerolmack, J.D., 2009, Real and apparent changes in sediment deposition rates 325 through time, J. Geophys. Res. 114, F00A06. 326 Segura, T.L., et al., 2008, Modeling the environmental effects of moderate-­‐sized impacts on 327 Mars. J. Geophys. Res., 113, E11007. 328 Segura, T.L., et al., 2012, An impact-­‐induced, stable, runaway climate on Mars. Icarus, 220, 144. 329 Shuster, D.L., et al., 2012, Cosmogenic 3He in hematite and goethite from Brazilian “canga” 330 duricrust demonstrates the extreme stability of these surfaces. Earth Planet. Sci. Lett., 329, 41. 331 Sivia, D.S., 2006, Data analysis: A Bayesian tutorial (2nd edition), Oxford University Press. 332 Smith, M.R., et al., 2008, Effect of obliteration on crater-­‐count chronologies for Martian 333 surfaces, Geophys. Res. Lett., 35, L10202. 334 Summons, R.E., et al., 2011, Preservation of Martian Organic and Environmental Records. 335 Astrobiology, 11, 157-­‐181. 336 Thomson, B.J., et al., 2011, Constraints on the origin and evolution of the layered mound in Gale 337 Crater, Mars using Mars Reconnaissance Orbiter data. Icarus, 214, 413–432. 338 Toon, O.B., Segura, T., & Zahnle, K., 2010, The formation of Martian river valleys by impacts, 339 Ann. Rev. Earth Planet. Sci. , 38, 303-­‐322. 340 Watters, T.R., et al., 2007, Hemispheres apart: The crustal dichotomy on Mars. Ann. Rev. Earth 341 Planet. Sci., 35, 621-­‐652. 342 Werner, S.C. & Tanaka, K.L., 2011, Redefinition of the crater-­‐density and absolute-­‐age 343 boundaries for the chronostratigraphic system of Mars. Icarus, 215, 603-­‐607. 344 Wordsworth, R., et al. 2013, Global modelling of the early martian climate under a denser CO2 345 atmosphere: Water cycle and ice evolution, Icarus, 222, 1-­‐19. 14 346 Yen, A.S., et al., 2006, Nickel on Mars: Constraints on meteoritic material at the surface. J. 347 Geophys. Res., 111, E12S11. 348 Yielding, G., et al., 1996, Sampling of fault populations using sub-­‐surface data: a review, J. 349 Struct. Geol., 18, 135-­‐146. 350 Zabrusky, K., Andrews-­‐Hanna, J.C., and Wiseman, S., Reconstructing the distribution and 351 depositional history of the sedimentary deposits of Arabia Terra, Mars, Icarus 220 (2012) 311–
352 330. 353 Zimbelman, J.R., & Scheidt, S.P., 2012, Hesperian age for Western Medusae Fossae Formation, 354 Mars, Science 336, 1683. 355 15 356 357 Figure 1. Gallery of ancient Martian craters. a) Crater being exhumed from beneath an 358 unconformity within Gale Crater’s mound (Mt. Sharp, Aeolis Mons), ESP_019988_1750. b) 359 Craters with fresh-­‐appearing ejecta being exhumed from beneath meander belts, Aeolis Dorsa, 360 P15_006973_1742_XI _05S205W; Burr et al., 2010. c) Crater being exhumed from beneath 361 fluvial channel deposit, Aeolis Dorsa. #6 in Supplementary Table, 238 m diameter. d) Crater 362 draped by fluvial channel deposit, Aeolis Dorsa, ESP_019104_1740. #3 in Supplementary Table, 363 141 m diameter. e) Crater from (c), but with 1m elevation contours from 1m DTM. DTM is 364 composed of ESP_019104_1740 and ESP_017548_1740. f) Crater from (d), with 1m contours 365 from the same DTM. 366 16 367 368 17 369 Figure 2. Count region (black polygons outlining bright pink areas) in relation to Gale Crater 370 and to areas with Crater-­‐Retention Ages (CRAs) determined by Zimbelman & Scheidt (2012). 371 Color scale corresponds to MOLA elevation, is linear, and saturates at -­‐1000m (red) and 372 -­‐3500m (white). Red-­‐outlined area is typically 500m higher than the fluvial-­‐channel deposits. 373 Gale’s lower mound accumulated at the same time, within error (Thomson et al., 2011). Inset 374 shows zoom on count region. Within the count region, bright pink areas correspond to counted 375 fluvial-­‐channel deposits, black circles to embedded craters associated with those fluvial 376 deposits, dark gray circles to possible embedded craters, and white areas to “holes” (areas not 377 counted due to poor preservation) within fluvial-­‐channel deposits. Circles representing 378 craters are not to scale. Only a small subset of fluvial-­‐channel deposits was used for this count 379 (Burr et al., 2009). 380 18 2
10
Assumed Age = 3.7 Gyr
0
Ncr (>D) / km2
10
−2
10−6 m/yr
10
10−5 m/yr
−4
10
−4
m/yr
−3
m/yr
−2
m/yr
−1
m/yr
10
10
−6
10
10
10
−8
100 m/yr
10
−1
10
381 a) 0
1
10
D (km)
10
2
10
Assumed Age = 3.9 Gyr
0
Dcr (>D) / km2
10
10−6 m/yr
−2
10
10−5 m/yr
10−4 m/yr
−4
10
−3
10
m/yr
10−2 m/yr
−6
10
10−1 m/yr
100 m/yr
−8
10
−1
10
382 383 b) 0
10
D (km)
19 1
10
384 Figure 3. Aeolis Dorsa crater-­‐river interaction frequency: crater counts plotted against 385 sedimentation rate curves for (a) assumed age 3.7 Gyr and (b) assumed age 3.9 Gyr. Red lines 386 with symbols correspond to counts of definite river-­‐crater interactions (Supplementary Table); 387 black lines with symbols correspond to counts including all candidates. Curves correspond to 388 lines of equal sedimentation rate using crater production functions and chronology functions 389 given by Werner & Tanaka (2011). Thin magenta lines employ Ivanov (2001) crater 390 production function and Hartmann & Neukum (2001) chronology function. Thin black lines 391 employ Hartmann (2005) crater production function and Hartmann (2005) chronology 392 function (see footnote 1). 1σ-­‐error bars assume Poisson statistics (Sivia, 2006, p.121). Thick 393 blue lines correspond to best-­‐fitting accumulation rates. 20