Reports                                                                                                                   ...
The channels identified in R3 align with the truncation of the L1R        Scabland of the northwestern United States (24)....
Mars: Key evidence and major implications. Geomorphology 132, 51   (2011). doi:10.1016/j.geomorph.2011.05.02221. B. A. Cam...
Downloaded from on March 7, 2013Fig. 1. SHARAD Radargrams display time delay of the echo signal against...
Downloaded from on March 7, 2013                                           Fig. 2. Spatial distribution...
Fig. 3. 3D visualization of the R3 reflector [corrected                                                   for depth (13)] ...
Upcoming SlideShare
Loading in …5

3 d reconstruction_of_the_source_and_scale_of_buried_young_flood_channels_on_mars


Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

3 d reconstruction_of_the_source_and_scale_of_buried_young_flood_channels_on_mars

  1. 1. Reports  Radargrams reveal that the northern and southern termini of the R3 reflector3D Reconstruction of the Source and dip upwards and reconnect with the sur- face, delineating a discrete facies bound-Scale of Buried Young Flood Channels ary (Fig. 1A and fig. S1). The R3 reflector is located exclusively below a mapped unit of young volcanics, ACyon Mars [(5), also mapped as AEc3 by (14)] (Fig. 2C and fig. S2C). This unit is interpreted to be formed of voluminous lava flows <Gareth. A. Morgan,1* Bruce. A. Campbell,1 Lynn. M. Carter,2 230 Ma old (5, 14), suggesting that R3Jeffrey. J. Plaut,3 Roger. J. Phillips4 represents the base of a distinct surficial1 2 flow. The bases of young lava flows Center for Earth and Planetary Studies, Smithsonian Institution, Washington, DC, USA. NASA Goddard 3Space Flight Center, Greenbelt, MD 20771, USA. Jet Propulsion Laboratory, Pasadena, CA, USA. have also been identified by SHARAD4 Planetary Science Directorate, Southwest Research Institute, Boulder, CO, USA. west of Ascraeus Mons (15). The north- ern boundary of R3 shows strong spatial*Corresponding author. E-mail: correlation with the boundary between ACy and the older unit ACo [>500 MaOutflow channels on Mars are interpreted as the product of gigantic floods due to (5)] (Fig. 2), implying that the lava em- Downloaded from on March 7, 2013the catastrophic eruption of groundwater that may also have initiated episodes of bayed the preexisting ACo surface southclimate change. Marte Vallis, the largest of the young martian outflow channels of Cerberus Fossae and flowed toward(<500 Ma), is embayed by lava flows that hinder detailed studies and comparisons the northeast (dominant slope directionwith older channel systems. Understanding Marte Vallis is essential to our of the present surface). The northernassessment of recent Mars hydrologic activity during a period otherwise considered portion of the R3 reflector exhibitsto be cold and dry. Using the SHARAD sounding radar we present a three- prominent depressions, delineating sub-dimensional (3D) reconstruction of buried channels on Mars and provide estimates surface channels (Fig. 1A). These chan-of paleo-hydrologic parameters. Our work shows that Cerberus Fossae provided the nel features are ~20 km wide and extendwaters that carved Marte Vallis, and extended an additional 180 kilometers to the for at least 50 km in a northeast direc-east prior to the emplacement of the younger lava flows. We identify two stages of tion. Seen in plan form, the channelchannel incision and determine that channel depths were more than twice that of features begin abruptly adjacent to oneprevious estimates. another along an orientation trending northwest to southeast (Fig. 2C and fig. The majority of outflow channels on Mars are attributed to megafloods S3). caused by the catastrophic release of groundwater. The most prominent We interpret these features to be the highest elevated channels of outflow channels – located around the Chryse basin - are > 1000 km Marte Vallis (Fig. 3), implying that the lava flow whose base is defined long and are estimated to be Hesperian (~3.7 – 3.1 Ga) in age (1–3). by R3 infilled the channels as the lavas flowed to the northeast. This Marte Vallis in Elysium Planitia is the largest of the young (late Amazo- indicates that the erosion of the outflow channel cut into the original nian: ~0.5 Ga to present) outflow channels on Mars. The channel system underlying surface of unit ACo prior to the emplacement of the younger extends over ~1000 km in length and ~100 km in width, making Marte ACy lavas (fig. S4). This sequence of events confirms the young age of Vallis comparable in scale to the Chryse basin channel systems. Young Marte Vallis and places the channel formation between the emplacement lava flows have fully embayed the most elevated portions of Marte of units ACo and ACy (10 – 500 Ma), in agreement with (5). Vallis, and as a consequence the fundamental characteristics of the The L1R and L2R reflectors are found extensively across the study channels, including their source, depth and morphology are less well region, suggesting that they represent regional boundaries between three understood than the Hesperian channels (4), despite being over 2.6 bil- bedrock facies (Figs. 1 and 2). Further inspection of the radargrams re- lion years younger (5). veals that both reflectors are punctuated by incisions of varying width Two possible sources have been proposed for Marte Vallis: Water (Fig. 1, B and C). These incisions are not random, but spatially align to flowing from the Athabasca Valles outflow channel in the west (4, 6, 7), reveal complex networks. We interpret the spatial location of the inci- possibly forming bodies of water such as the putative frozen central sions to represent where channels have been eroded through either one Cerberus sea (8); and water flowing from a now-buried section of Cer- or both of the two bedrock boundaries delineated by L1R and L2R. A berus Fossae (5, 6, 9). It is impossible to resolve which of the above similar methodology has been applied to map out buried flood channels hypotheses are correct from investigations of the surface geology alone. on Earth through the use of seismic profiles (16). No incisions are ob- Using the Shallow Radar (SHARAD) sounder (10, 11) on the Mars Re- servable in L1R and L2R below unit ACo and most SHARAD tracks connaissance Orbiter, we present a tomographic visualization of the reveal reflector incisions that spatially correlate with the boundary of buried Marte Vallis channels (12). ACo and ACy (Fig. 2). This indicates that the facies boundaries are sys- All 58 SHARAD tracks covering the uppermost reaches of Marte tematically correlated with ACo. Our interpretation is further corroborat- Vallis (as identifiable in Mars Orbiter Laser Altimeter - MOLA gridded ed by the surface morphology exhibiting streamlined features (which data) display multiple reflecting horizons (Fig. 1 and fig. S1). From represent bedrock ‘islands’ between the channels) that are spatially cor- mapping of the spatial distribution of SHARAD subsurface returns related with isolated patches of the bedrock reflectors (Fig. 1, B and C, (Figs. 1 and 2), three distinct reflectors have been identified. Two of and Fig. 2). We argue that the teardrop-shaped hills and associated re- these reflectors are found extensively across the study area and occupy flectors are remnant sections of the older ACo plains isolated by the different depth ranges (13), referred to here as L1R (the shallower reflec- erosional formation of the Marte Vallis channels (such streamlined fea- tor) and L2R (the deeper reflector). The third reflector, R3, is located tures characterize most outflow channels on Mars) prior to infilling of only in the southern portion of the region (Fig. 1C). the channels by the young ACy lavas. / / 7 March 2013 / Page 1 / 10.1126/science.1234787
  2. 2. The channels identified in R3 align with the truncation of the L1R Scabland of the northwestern United States (24). However, the Marteand L2R reflectors (fig. S3) implying that they are part of the same Vallis estimate represents a minimum value, as the L2R reflector haschannel network. Many martian outflow channels are sourced in chaos been fully bisected by this channel. The depths of the channels are atterrain, interpreted to be the result of subsidence from the rapid evacua- least double the previous maximum estimates for Marte Vallis of 40 mtion of groundwater (17). The abrupt opening of the channels as seen in (5), demonstrating that the scale of the floods has been underestimated.R3 and the lack of any evidence for depressions at their source argueagainst the existence of chaos terrain at the head of Marte Vallis. Instead References and Notesthe channel alignment at the source matches the orientation of the Cer- 1. D. H. Scott K. L. Tanaka, Geological map of the western equatorialberus Fossae graben system to the west (Fig. 3). This, and the absence of region of Mars. U.S. Geological survey Misc. Map I-1802-A (1986).observable subsurface channels in the southern portions of R3, strongly 2. K. L. Tanaka, The stratigraphy of Mars. J. Geophys. Res. 91, (B13), E139 (1986). doi:10.1029/JB091iB13p0E139supports the hypothesis that Cerberus Fossae was the source of the Marte 3. S. Rotto, K. L. Tanaka, Geologic/Geomorphic map of the ChryseVallis floods. This allows the position of this now-buried portion of Planitia region of Mars. U.S. Geological survey Misc. Inv. Map I-Cerberus Fossae to be inferred, and suggests that the fissure previously 2441 (1995)extended at least ~180 km to the east of its present surface expression 4. D. C. Berman, W. K. Hartmann, Recent fluvial, volcanic and tectonic(Fig. 3 and fig. S3). activity on the Cerberus Plains of Mars. Icarus 159, 1 (2002). The bed of the channel features cannot be mapped ‘downstream’ be- doi:10.1006/icar.2002.6920yond the areal extent of the R3 reflector. The sounder signal is increas- 5. J. Vaucher et al., The volcanic history of central Elysium Planitia,ingly attenuated as the thickness of lava above the buried channel bed implications for martian magmatism. Icarus 204, 418 (2009).increases, so we postulate that the R3 reflector may extend over a larger doi:10.1016/j.icarus.2009.06.032 6. D. M. Burr, J. A. Grier, A. S. McEwan, L. P. Keszthelyi, Reapeted Downloaded from on March 7, 2013part of the study region at depths beyond the SHARAD signal- aqueous flooding from the Cerberus Fossae: Evidence for very recentpenetration limit. Quantifying the channel erosion north of Cerberus extant, deep groundwater on Mars. Icarus 159, 53 (2002).Fossae can be achieved by measuring the extent of the incisions into the doi:10.1006/icar.2002.6921L1R and L2R reflectors. The horizontal and vertical position of the re- 7. E. R. Fuller, J. W. Head, Amazonis Planitia: The role of geologicallyflectors on each side of an incision was recorded and combined with the recent volcanism and sedimentation in the formation of the smoothestelevations of the streamlined forms in the MOLA data. Connecting these plains on Mars. J. Geophys. Res. 107, (E10), 5081 (2002).erosional control points yields an approximation of the buried channel doi:10.1029/2002JE001842morphology for each radargram. Due to the density of the SHARAD 8. J. B. Murray et al.; HRSC Co-Investigator Team, Evidence from thecoverage it was possible to interpolate between neighboring tracks to Mars Express High Resolution Stereo Camera for a frozen sea closeproduce a three-dimensional model of the buried Marte Vallis channels to Mars’ equator. Nature 434, 352 (2005). doi:10.1038/nature03379 Medline(Fig. 4). 9. J. B. Plescia, Cerberus Fossae, Elysium, Mars: A source for lava and The SHARAD data reveal a complex channel system consisting of a water. Icarus 164, 79 (2003).broad ~40 km wide main channel that is adjacent to a raised bench, 120 10. R. Seu et al., SHARAD: The MRO 2005 shallow radar. in width and incised by anastomosing channels formed around four Space Sci. 52, 157 (2004). doi:10.1016/j.pss.2003.08.024streamlined islands. Such morphology is consistent with the majority of 11. R. Seu et al., SHARAD sounding radar on the Mars Reconnaissancemartian outflow channels and the scale is comparable to that of the main Orbiter. J. Geophys. Res. 112, (E5), E05S05 (2007).channel and associated perched tributaries of the prominent Ares Vallis doi:10.1029/2006JE002745outflow channel system (Hesperian in age) that flows into the Chryse 12. Materials and methods are available as supplementary materials onBasin (18). The geomorphic configuration of Marte Vallis implies the Science Online. 13. The time delay offset between a reflector and the surface echo withinsystem experienced two phases of erosion: 1) to erode the islands and 2) a radargram can be converted to depth. This depth estimation isto cut the main channel below the level of the perched channels. This inversely proportional to the square root of the assumed permittivityalso implies that the initial stages of channel formation consisted of (ε′) of the material above the reflector. Figures 2 to 4 assume an ε′ ofsmall-scale, anastomosing patterns before flow was concentrated in a 8, consistent with dense, dry terrestrial lava flows. Applying a lowerdeeper, wider channel, leaving the central islands and smaller channels value for ε′ would provide a greater depth estimate for each of theas perched remnants (Fig. 4, fig. S4, and supplementary text). Recent citing the turbulent nature of the young Elysium Planitia lavas 14. K. L. Tanaka, J. A. Skinner, T. M. Hare, Geologic map of thesuggests that lava flows may have partially or fully eroded the outflow northern plains of Mars, U. S. Geol. Surv. Sci. Invest. Map, 2888channels (19, 20). While they are filled with later lavas, our tomographic (2005).models show that the Marte Vallis channels are morphologically similar 15. L. M. Carter et al., Dielectric properties of lava flows west of Ascraeus Mons, Mars. Geophys. Res. Lett. 36, L23204 (2009).to the circum-Chryse outflow channels, and thus could reasonably have doi:10.1029/2009GL041234been carved entirely by water. 16. C. S. Fulthorpe, J. A. Austin, Jr., Shallowly buried, enigmatic Estimates of the depth of the channels from time-delay offsets to the seismic stratigraphy on the New Jersey outer shelf: Evidence forsurface echo is based upon an assumed range of values for the permittiv- latest Pleistocene catastrophic erosion? Geology 32, 1013 (2004).ity of the material overlying the L1R and L2R reflectors (21). The per- doi:10.1130/G20822.1mittivity of these materials is uncertain, but the typical range for dry 17. M. Carr, Formation of matian flood features by release of water fromgeological materials is 3 to 8 (22, 23), with lower values leading to confined aquifers. J. Geophys. Res. 84, (B6), 2995 (1979).greater depths below the surface for any given time delay. For example, doi:10.1029/JB084iB06p02995the depth of the perched channel between islands ii and iii is constrained 18. N. H. Warner, S. Gupta, J.-P. Muller, J.-R. Kim, S.-Y. Lin, A refined chronology of catastrophic outflow events in Ares Vallis, Mars.between the incised L1R and the continuous L2R reflectors. Applying a Earth Planet. Sci. Lett. 288, 58 (2009).permittivity range of 8-3 yields a depth range of 35 – 56 m (L1R) to 67 - doi:10.1016/j.epsl.2009.09.008110m (L2R), respectively. With regards to the main channel, the depth 19. W. L. Jaeger et al., Emplacement of the youngest flood lava onestimate (between island i and the south bank) is 69 – 113 m. This is Mars: A short, turbulent story. Icarus 205, 230 (2010).comparable with the depth of incision of the largest known megaflood doi:10.1016/j.icarus.2009.09.011on Earth, the Missoula floods, responsible for carving the Channeled 20. D. W. Leverington, A volcanic origin for the outflow channels of / / 7 March 2013 / Page 2 / 10.1126/science.1234787 2
  3. 3. Mars: Key evidence and major implications. Geomorphology 132, 51 (2011). doi:10.1016/j.geomorph.2011.05.02221. B. A. Campbell et al., SHARAD radar sounding of the Vastitas Borealis Formation in Amazonis Planitia. J. Geophys. Res. 113, (E12), E12010 (2008). doi:10.1029/2008JE00317722. F. T. Ulaby et al., Microwave dielectric spectrum of rocks, Rep. 23817-1-T, Univ. of Mich. Radiat. Lab., Ann Arbor (1988).23. W. D. Carrier, G. R. Ohloeft, W. Mendell, Physical properties of the lunar surface, in G. H. Heiken, D. T. Vaniman, B. M. French Eds., Lunar Sourcebook, pp. 457–567, (Cambridge Univ. Press, New York, 1991).24. V. R. Baker, D. Nummedal Eds., The Channeled Scabland; a Guide to the Geomorphology of the Columbia Basin (NASA, Washington, DC, 1978).25. B. A. Campbell, N. E. Putzig, L. M. Carter, R. J. Phillips, Autofocus correction of phase distortion effects on SHARAD echoes. IEEE Geosci. Remote Sens. Lett. 8, 939 (2011). doi:10.1109/LGRS.2011.2143692Acknowledgments: T. Watters provided a useful review of an earlier version of the paper. Comments and suggestions by four anonymous referees were extremely helpful. Funding for this work was provided Downloaded from on March 7, 2013 by the NASA MRO Project. The SHARAD data and MOLA altimetry are available through NASA’s Planetary Data system.Supplementary and MethodsSupplementary TextFigs. S1 to S4Table S1Reference (25)4 January 2013; accepted 18 February 2013Published online 7 March 2013;10.1126/science.1234787 / / 7 March 2013 / Page 3 / 10.1126/science.1234787 3
  4. 4. Downloaded from on March 7, 2013Fig. 1. SHARAD Radargrams display time delay of the echo signal against along-trackdistance. Subsurface contrasts in permittivity reflect the transmitted signal, producingthe observed reflectors. Spatial coordinates of the radargram sections presented: (A)376102000 (8.42°N, 175.61°E – 4.07° N, 175.08° E), (B) 504001000 (10.13° N,177.81° E – 5.78° N, 177.28° E) (C) 724203000 (9.45° N, 176.79° E – 5.1° N, 176.26°E). The locations of the SHARAD tracks are presented in Fig. 2. / / 7 March 2013 / Page 4/ 10.1126/science.1234787
  5. 5. Downloaded from on March 7, 2013 Fig. 2. Spatial distribution of L1R (A), L2R (B) and R3 (C) reflectors corrected for depth (13). Background shows the (5) geologic map above a hillshade image of MOLA gridded data (200 × vertical exaggeration). [Geologic map adapted from (5) with permission from Elsevier]/ / 7 March 2013 / Page 5 / 10.1126/science.1234787 5
  6. 6. Fig. 3. 3D visualization of the R3 reflector [corrected for depth (13)] below the MOLA surface overlain with geologic units (5). Note the surface has been elevated, and scaled by a factor of 1/150 for clarity. The Marte Vallis channels are visible as prominent depressions in the R3 reflector and begin abruptly along an orientation that aligns with the subsurface vertical projection of Cerberus Fossae. Blue arrow highlights direction of flow in channels. [Geologic map adapted from (5) with permission from Elsevier] Downloaded from on March 7, 2013 Fig. 4. (A) Tomographic model of the buried Marte Valles channels and associated NW-SE topographic and tomographic profile (13). The model shows a ~40 km wide main channel and adjacent teardrop- shaped islands. (B) and (C) show perspective view of eastern Elysium Planitia at present (B) and before embayment by young lavas (C). White arrows represent the direction of flow./ / 7 March 2013 / Page 6 / 10.1126/science.1234787 6