Amazonian geologic history of the Echus Chasma and Kasei Valles system on Mars

18 Pages·2010·2.43 MB·English

EarthandPlanetaryScienceLetters294(2010)238–255
ContentslistsavailableatScienceDirect
Earth and Planetary Science Letters
journal homepage: www.elsevier.com/locate/epsl
Amazonian geologic history of the Echus Chasma and Kasei Valles system on Mars:
New data and interpretations
Mary G. Chapmana,⁎, Gerhard Neukumb, Alexander Dumkeb, Greg Michaelb, Stephan van Gasseltb,
Thomas Kneisslb, Wilhelm Zuschneidb, Ernst Hauberc, Nicolas Mangoldd
aPlanetaryScienceInstitute,Tucson,Arizona,USA
bInstituteofGeosciences,FreieUniversitaetBerlin,Germany
cGermanAerospaceCenter(DLR),Berlin,Germany
dLPGN,CNRS,UniversitéNantes,France
a r t i c l e i n f o a b s t r a c t
Articlehistory: Newhigh-resolutiondatasetshavepromptedamapping-basedstudyofthe2500-km-longEchusChasma
Accepted19November2009 and Kasei Valles system that utilizes geomorphic details, stratigraphic relations, and cratering statistics
Availableonline29December2009 derivedfromthenewdatasets.OurresultssuggestthatbetweentheHesperianandAmazonianEpochson
Mars(3.7GatoRecent),thestudyareawasaffectedbyatleast4episodesofwidespreadvolcanicactivity
Keywords: and4periodsofepisodicfluvio-glacialactivity.ThispaperdiscussestheAmazonian(b1.8Ga)historyofthe
chasma
studyarea,duringwhichtimethelastofthefourvolcanicepisodesoccurredbetweenthelasttwoepisodes
valles
of fluvio-glacial activity. Highlights of our new findings from this time period include (1) evidence that
anastomosing
dendritic suggestsglaciersandnear-surfaceicemayhavepersistedthroughAmazoniantimeinlocalareasoverthe
permafrost entirelength ofKasei Valles; (2)anewwidespread platy-flowsurface material thatisinterpreted tobe
2100-km-runoutfloodlavassourcedfromEchusChasma;and(3)afractureinEchusChasma,identifiedto
havesourcedatleastonelate-stageflood,thatmayhavebeentheoriginfortheplaty-flowmaterialand
youngnorth-trendingKaseifloodwater.
©2009ElsevierB.V.Allrightsreserved.
1.Introduction (b)theKaseichannelfloorwasmappedasHesperianplainHchpor
olderchannelfloormaterialHcch;(c)EchusMontesandChaoswere
Thelargestofthecircum-ChryseoutflowchannelsisKaseiValles. eithermesasofwallrockmaterialHNuorchaoticmaterialunitHcth;
Kasei Valles heads at one of the Valles Marineris troughs: Echus and (d) Tharsis Montes Formation lava members At4 and At5 of
Chasma.EchusChasmaisa175-km-wide,100-km-long,open-ended, Amazonian age covered large parts of the channel, with older,
flat-floored depression centered at latitude 1°S, longitude 80°W. underlyingmemberHt2embayingTempeTerranorthofthismassive
North from Echus Chasma, Kasei Valles is nearly 2500km long, valley(ScottandTanaka,1986).[TharsisMontesFormationmember
extending about 1600 to latitude 20°N, longitude 75°W (near the Ht3doesnotoccurwithinthestudyarea.]ObservationsfromViking
locationofthe“enigmatic”UraniusDorsumridge,whoseoriginhas datasetsindicatedthatbothlavaunitsAt4andAt5postdatedflooding
longbeenuncertain,seeChapmanandScott(1989))wherethevalley fromKaseiValles(RottoandTanaka,1995).Flowsfromtheolderunit
turnsabruptlyeastandcontinuesforanother900kmtodebouchinto At4wereshowntobesourcedeastfromTharsisandmappedtocover
ChrysePlanitia(Fig.1). mostofthenorthernhalfofKaseiValles,andincludedtheenigmatic
NumerousmappingstudieshaveincludedpartoralloftheKasei northeast-trending ridge of Uranius Dorsum. Lava flows from the
Valles area (Milton, 1974; Scott and Carr, 1978; Scott and Tanaka, younger unit At5 flow were also shown sourced from Tharsis and
1986; Chapman and Scott, 1989; Jöns, 1990, 1991; Chapman et al., mappedtocovermostofthesouthernhalfofKaseiVallestoEchus
1991; Tanaka and Chapman, 1992, Rotto and Tanaka, 1995, Tanaka Chasma. In order to avoid lengthy redundancy, for references to
et al., 2005). Previous mapping indicated that (a) the chasma and previousmappingresultsthroughoutthetextthereaderisreferredto
vallessystemcutintoHesperianplainsmaterialunitsHrandHsmof thosementionedinthisparagraph.
theLunaePlanum,TempeTerraandSyriaPlanumboundingplateaus; Ourstudyofthissystemwaspromptedbytheavailabilityofnew
high-resolution data from the Mars Express High Resolution Stereo
Camera (HRSC) images and derived digital terrain models (DTMs),
Mars Odyssey Thermal Emission Imaging System (THEMIS), Mars
⁎ Correspondingauthor.PlanetaryScienceInstitute,1700EFortLowell,Suite106,
OrbiterCamera (MOC), Mars Orbiter LaserAltimeter (MOLA),Mars
Tucson,Arizona,85719-2395,USA.Tel.:[phone];fax:[phone].
E-mailaddress:[email protected](M.G.Chapman). Reconnaissance Orbiter Context Imager (CTX; Fig. 14), and Mars
0012-821X/$–seefrontmatter©2009ElsevierB.V.Allrightsreserved.
doi:10.1016/j.epsl.2009.11.034
M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255 239
(Fig. 2). So within north Kasei, the once low-lying Hesperian
channelfloorlavaswereerodedtoforminvertedtopographynow
located on the highest terrace of north Kasei Valles. The flooding
episodesfromEchuswereeithersimultaneouswithorfollowedby
a prolonged period of slow erosion that formed theater-headed
channels, narrow sinuous channels, and gullies on alluvial fans.
Capping off the major depositional processes in the valley was a
prolongedperiodofvolcaniceruptionsintheLateAmazonianthat
produced a massive output of lava from both Tharsis Montes and
EchusChasma.Usingthenewdata,wehavefurtherdelineatedand
redrawn the contact boundaries of the Tharsis lava units and
mapped a new material: Amazonian platy-flow unit Apf. This
materialisinterpretedtoconsistofveryfluid,highlymobileflood
lavasthatweredepositedinKaseiVallesandcanbeobservedfrom
within Echus Chasma to south of Sharonov impact crater in
northeast Kasei Valles (Fig. 1). Unit Apf is overlain by unit At5
fromTharsisMontes.Finally,wehavefoundevidencewithinEchus
Chasmathatafloorfracturemayhaveemittedlate-stagefluidflows
which ponded within the chasma. Details of these units, the
mapping, our new findings and interpretations follow in the rest
ofthispaper.ThispaperisasummaryofallAmazonianunitsinthe
studyarea,eachofwhichcouldmeritadetailedpaper.
2.Methods
Toavoiddatagaps,themappingwasdigitallycompiledonahigh-
resolutionshadedMOLAaltimetrybase.Complexareasweremapped
in greater detail on individual HRSC images and compiled THEMIS
mosaicsandthislineworkwastransferredtothealtimetrybase.Some
mapunitscorrespondorarepartlyequivalenttounitsfromprevious
mapping efforts, and the formal terminology of geologic units
proposed by Scott and Tanaka (1986) and revised by Rotto and
Tanaka(1995)wasfollowedwherefeasible.However,inmanyplaces
interpretationandcontactshavebeenrevisedtoreflectinformation
visible on the new higher resolution datasets. The depths of some
channels,thicknessesofsomeunits,andheightsofsomestreamlined
islands were obtained in places using HRSC and MOLA altimetry
profiles.Therelativeagesofgeologicunitsandgeomorphicfeatures
were established by observed stratigraphic relations and supported
bynewcratercountsofsurfacematerialsusingnewhigh-resolution
data.Theagesofunitsweredeterminedbymorethan200countsof
allcraterswithdiametersN1kmonareasof15HRSCorbitalimages
(resolution approximately 17m/pixel or m/p), 6 MOC images
(resolution typically about 1.5 to 3m/p), and 11 THEMIS images
(nominal resolution of IR=100m/p and VIS=18m/p). Amazonian
dataareshowntabulatedinthistext(Tables1–4).Forthesecounts
weusedtheproductionfunctioncoefficientsofIvanov(2001)andthe
Fig. 1. (A) Viking MDIM image showing physiographic features of study area
(EC=EchusChasma,EM=EchusMontes,ECh=EchusChaos,UD=UraniusDorsum crateringmodelofHartmanandNeukum(2001)toderiveabsolute
ridge, SM=Sacra Mensa, LM=Labeatis Mensa, 1 = Sharonov impact crater, ages;someageswerealsotakenfromWerner(2006)whousedthe
2=Fesenkovimpactcrater,3=Sulakimpactcraterandstreamlinedisland;narrow samemethods.Thegeologicunitsyield2typesofcratercountdata:a
innergorgesofnorthKaseiVallesinblack;northattop);(B)MOLAdigitalelevation unitagethatreflectstheageofdepositionandaresurfacingagethat
modelofstudyarea;notethatnarrowinnergorgesofnortheastKaseiValleslieatsame
elevationasChrysePlanitia. markstheageoferosion.
The cratering age of erosional units represent the time of their
modification relative to other units whose statistical age and
ReconnaissanceOrbiterHighResolutionImagingScienceExperiment stratigraphicpositioninthesequencereflecttheirtimeofemplace-
(HiRISE; Figs. 8B and 9B). In order to understand the relations ment. Erosional units may thus consist of older material that is in
betweensourceandchannel,weundertookanintegratedexamina- placebuthasbeenhighlymodifiedbylaterprocesses(Milton,1974).
tion of thisspecific chasma-channel systemusing thenewdatasets Forexample,floodunitsmaynotbedepositionalmaterialsbutmostly
and have produced a post-Viking revised map of the entire Echus erosional units that date from periods of resurfacing. As with the
Chasma–KaseiVallessystem(Fig.2). PleistoceneGlacialLakeMissoulafloods,eachmappedfloodunitmay
Theresultsofourmappinginthisreportontheb1.8Gahistory betheresultofnumerousindividualfloodscloselyspacedintime(Alt,
of the area indicate that during the early Amazonian, Kasei Valles 2001; Komatsu and Baker, 2007). In fact, several floods were
was inundated by north-directed flooding from Echus Chasma. suggestedtohaveformedamaterialontheKaseiVallesfloorproper,
WithinnorthKasei,olderHesperianlavasareerodedontheirsouth based on cratering statistics and geomorphic erosional details
edges and locally (eastward) by this last flooding episode that observedonVikingdata(NeukumandHiller,1981;Chapmanetal.,
carvedthelowereast-trendingterracelevelintonorthKaseiValles 1991).Onafinalnote,consistentwithnumerousstudies(forexample
240 M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255
Fig.2.Geologicmapofthestudyarea.
Table1
CratercountdatafordegradedunitHt4,whereerodedbynarrowchannels.
Unit Imagery Designation Area N(1) Age Errorup Errordown Errorfitrange N(1) Age Errorup Errordown Errorfitrange
(km2)
Unitage Resurfacingage
Channel-erodedHt4 HRSC 3261.000 90.909 0.000 0.576 0.148 0.155 0.15/0.38 0.000 0.090 0.027 0.030 0.07/0.13
HRSC 3261.000 146.847 0.001 1.116 0.302 0.332 0.2/0.4 0.000 0.100 0.021 0.022 0.07/0.11
HRSC 3261.000 75.305 0.001 2.070 0.565 0.593 0.2/0.36 0.000 0.098 0.074 0.083 0.09/0.16
HRSC 3261.000 145.631 0.001 1.050 0.284 0.299 0.18/0.39 0.000 0.168 0.023 0.024 0.06/0.14
HRSC 3261.000 229.412 0.001 1.056 0.242 0.257 0.2/0.47 0.000 0.117 0.015 0.016 0.06/0.12
HRSC 3261.000 134.752 0.001 2.097 0.458 0.478 0.2/0.44 0.000 0.112 0.022 0.023 0.06/0.12
HRSC 3261.000 93.656 0.001 1.330 0.377 0.384 0.2/0.3 0.000 0.093 0.026 0.036 0.06/0.15
HRSC 3261.000 112.038 0.001 1.237 0.826 1.030 0.35/0.45 0.000 0.191 0.037 0.038 0.08/0.15
HRSC 3261.000 568.707 0.001 2.225 0.585 0.614 0.35/0.6 0.000 0.207 0.026 0.027 0.1/0.2
HRSC 3261.000 121.495 0.001 1.328 0.504 0.540 0.25/0.4 0.000 0.646 0.164 0.173 0.15/0.3
Average 1.41 0.18
M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255 241
Table2
CratercountdataforunitAch.
Unit Imagery Designation Area N(1) Age Error Error Errorfit N(1) Age Error Error Errorfit
(km2) up down range up down range
Ach Unitage Resurfacingage
(NEofUraniusDorsum) HRSC 3283.000 636.036 0.001 1.435 0.282 0.296 0.25/0.6 0.000 0.026 0.026 0.026 0.09/0.17
MOC E0400068 18.467 0.000 0.813 0.185 0.193 0.09/0.2 5.192 0.101 0.005 0.005 0.02/0.04
MOC M0904101 5.523 0.001 1.189 0.294 0.294 0.09/0.2 3.443 0.067 0.003 0.003 0.01/0.04
HRSC 3283.000 196.687 0.001 1.315 0.253 0.258 0.2/0.33
HRSC 3272.000 403.863 0.001 1.428 0.155 0.155 0.2/0.6 0.000 0.584 0.029 0.029 0.1/0.2
Average 1.236 0.195
(SEofUraniusDorsum) MOC M0904101 4.682 0.000 0.303 0.090 0.090 0.07/0.11 5.836 0.113 0.007 0.007 0.02/0.06
MOC M0904101 7.507 nofit 0.000 0.813 0.169 0.169 0.07/0.15
MOC M0904101 1.944 0.000 0.401 0.281 0.281 0.11/0.07 2.281 0.038 0.009 0.009 0.02/0.05
MOC M1102254 333.227 0.001 0.916 0.172 0.172 0.12/0.2 0.000 0.363 0.024 0.024 0.05/0.08
HRSC 3283.000 341.905 nofit 0.000 0.683 0.105 0.118 0.15/0.4
HRSC 3283.000 447.343 0.001 1.467 0.315 0.331 0.25/0.5 0.000 0.267 0.017 0.017 0.07/0.11
HRSC 3283.000 62.276 0.000 0.752 0.132 0.132 0.13/0.25
HRSC 3283.000 73.791 0.000 0.620 0.355 0.355 0.25/0.3 8.316 0.148 0.023 0.023 0.08/0.14
HRSC 3272.000 283.420 0.001 1.027 0.112 0.112 0.17/0.5 0.000 0.425 0.033 0.033 0.1/0.17
HRSC 1235.000 2490.000 0.001 1.370 0.130 0.130 0.3/0.9
Average 0.784 0.468
(NWofSharonov) HRSC 5239.000 170.400 0.002 1.370 0.170 0.170 0.17/0.45
Average Total 1.00 0.385
Christensenetal.(1998)),volcanicflowsintheareaareallconsidered channelsthatcanbetracedtounitHt4moatsaroundLabeatisMensa
tobeapproximatelybasalticincomposition. andaroundtheadjacentTempeTerraplateau.Theselavamoatsare
hypothesizedtohavebeensitesofice-richlobatedebrisapronsthat
3.Geologichistory have been subsequently removed (Lucchitta and Chapman, 1988;
van Gasselt, 2007; Hauber et al., 2008). The dendritic feeder
3.1.YoungKaseiVallesmaterials channelscanalsobetracedtoroughandheavilydegradedareasof
unit Ht4 near Tempe Terra and to smooth surfaces of Ht4 directly
TheonsetoftheUpperAmazonianinthestudyareaismarkedby northofUraniusDorsum.Perhapstheanastomosingnarrowvalley
thepartialdegradationofTharsisMontesFormationmember4(unit and dendritic feeder channels were fed by melting of the local
Ht4), a Hesperian-age flood lava in the north Kasei area. This remnants of near-surface and surface ice. Therefore formation of
degraded area has an absolute crater model age of 1.41Ga±0.6 these fluvial features may mark a climate change on Mars (van
(Table1).Thedegradationisduetolocalerosionoftheunitbyan Gasselt, 2007; Hauber et al., 2008). This period of climate change
internal anastomosing narrow valley cut into unit Ht4 north of wasthebeginningofthethirdandlastfluvio-glacialcyclewithinthe
Uranius Dorsum and later disturbance by two 10-km-diameter study area (cycles 1 and 2 occurred during the Hesperian; see
impactcraters(Fig.3).Thevalleyisfedbynumeroussmalldendritic Chapmanetal.,thisvolume).
Table3
CratercountdataforunitApf.
Unit Imagery Designation Area N(1) Age Error Error Errorfit N(1) Age Error Error Errorfit
(km2) up down range up down range
Apf Unitage Resurfacingage
(EchusChasma) HRSC 2182.000 89.670 0.000 0.054 0.012 0.012 0.08/0.13
HRSC 2182.000 50.93 0.000 0.219 0.097 0.097 0.15/0.25
THEMIS V11686009 57.820 0.000 0.092 0.021 0.021 0.08/0.15
THEMIS V11686009 106.9 0.000 0.523 0.130 0.130 0.17/0.35 0.000 0.188 0.037 0.037 0.1/0.15
THEMIS V18425016 254.500 0.000 0.088 0.014 0.014 0.1/0.2
THEMIS V18425016 127.200 0.000 0.155 0.032 0.032 0.11/0.2
THEMIS V09815001 183.700 0.000 0.303 0.060 0.060 0.15/0.2
Average 0.205 0.188
(NearEchusMontes) HRSC 3283.000 356.200 0.000 0.673 0.300 0.300 0.35/0.4 0.000 0.164 0.018 0.018 0.11/0.3
HRSC 3283.000 313.300 0.000 0.228 0.061 0.061 0.16/0.25
HRSC 3283.000 412.100 0.000 0.160 0.026 0.026 0.12/0.2
HRSC 3283.000 324.300 0.000 0.147 0.067 0.067 0.2/0.3
HRSC 3283.000 316.200 0.000 0.198 0.049 0.049 0.17/0.33
Average 0.281 0.164
(Mid-channel) HRSC 3272.000 99.73 0.000 0.254 0.047 0.047 0.11/0.16 0.000 0.07 0.011 0.011 0.06/0.11
(Southgorge) HRSC 1202.000 1153.000 0.000 0.155 0.047 0.047 0.25/0.5 0.000 0.076 0.007 0.007 0.11/0.25
(Northgorge) HRSC 1224.000 402.600 0.000 0.131 0.022 0.022 0.13/0.25
HRSC 3261.000 155.300 0.000 0.119 0.028 0.028 0.12/0.2 0.000 0.086 0.024 0.024 0.12/0.35
HRSC 3261.000 125.3 0.000 0.150 0.026 0.026 0.1/0.2 0.000 0.126 0.024 0.024 0.1/0.2
Average 0.133 0.094
(SouthofSharonov) HRSC 5239.000 215.6 0.000 0.141 0.035 0.035 0.140.35
HRSC 5239.000 243.200 0.000 0.114 0.023 0.023 0.12/0.25
Average 0.128
242 M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255
Table4
CratercountdataforunitAt5.
Unit Imagery Designation Area N(1) Age Errorup Errordown Errorfitrange N(1) Age Errorup Errordown Errorfitrange
(km2)
At5 Unitage Resurfacingage
(SEofFesenkov) THEMIS V14232008 792.602 0.000 0.626 0.293 0.297 0.4/0.5 0.000 0.083 0.017 0.018 0.13/0.25
THEMIS V14232008 904.977 0.000 0.114 0.018 0.018 0.13/0.3
THEMIS V14232008 735.874 0.002 1.986 1.238 1.794 0.7/1 0.000 0.105 0.020 0.021 0.14/0.25
THEMIS V14232008 630.408 0.001 0.816 0.431 0.432 0.45/0.5 0.000 0.085 0.023 0.024 0.15/0.24
THEMIS V17639017 175.123 0.000 0.450 0.449 0.446 0.3/0.4 0.000 0.118 0.021 0.021 0.09/0.13
THEMIS V17639017 353.941 0.000 0.116 0.017 0.017 0.09/0.12
THEMIS V13608010 809.524 0.000 0.108 0.015 0.016 0.11/0.45
THEMIS V13608010 885.714 0.001 0.775 0.309 0.312 0.4/0.5 0.000 0.106 0.014 0.014 0.12/0.2
THEMIS V13608010 1294.120 0.000 0.163 0.052 0.053 0.250.35 0.000 0.057 0.006 0.006 0.09/0.17
Average 0.803 0.099
(WestofKaseimidpoint) HRSC 0097.000 235.3 0.000 0.172 0.036 0.036 0.14/0.35
HRSC 0097.000 122.1 0.000 0.116 0.022 0.022 0.09/0.13
HRSC 0097.000 215.2 0.000 0.188 0.028 0.028 0.12/0.35
HRSC 0097.000 151.9 0.000 0.206 0.038 0.038 0.1/0.3
HRSC 2957.000 141.0 0.000 0.217 0.026 0.026 0.09/0.15
HRSC 2957.000 161.4 0.000 0.207 0.032 0.032 0.11/0.25
HRSC 2957.000 93.81 0.000 0.196 0.034 0.034 0.1/0.2
Average 0.186
(MiddleKasei) HRSC 3272.000 215.2 0.000 0.147 0.021 0.021 0.1/0.2
HRSC 3272.000 171.2 0.000 0.118 0.023 0.023 0.09/0.17
HRSC 3272.000 110.4 0.00 0.12 0.017 0.017 0.08/0.2
Average 0.133 0.12
(NWofEchus) HRSC 2957.000 187.0 0.000 0.095 0.014 0.014 0.09/0.2
HRSC 2957.000 145.3 0.000 0.095 0.015 0.015 0.09/0.15
HRSC 0920.000 140.4 0.000 0.111 0.022 0.022 0.10/0.25
HRSC 0920.000 44.72 0.000 0.089 0.028 0.028 0.09/0.2
HRSC 0920.000 202.7 0.000 0.083 0.032 0.032 0.14/0.2
Average 0.089 0.10
This suggestedclimate change wasfollowed by the well-known The longitudinal grooves and ridges on the floor of Kasei Valles
catastrophicfloodingand/orglacialerosionfromEchusChasmathat were suggested to have formed by fluvial, mudflow, or glacial
formedthescouredsurfaceoftheKaseichannelfloor(Milton,1974; processes (Fig. 19A; Baker and Kochel, 1978; Thompson, 1979;
Masurskyetal.,1977;ScottandCarr,1978;BakerandKochel,1978, Lucchitta, 1982; Komar, 1983). Similar-appearing features occur
1979;Carr,1981;Baker,1982;Lucchitta,1982;ChapmanandScott, within catastrophically flooded and glacially eroded areas on Earth
1989;RobinsonandTanaka,1990).Weagreewiththecatastrophic (BakerandMilton,1974;Kehew,1982;Lucchitta,1982).Thecurrent
floodinterpretation,butaddmoredetailedmappingandaccurateage (and probably past) frigid climate of Mars suggests that the most
datingforthiserosionevent,thatwenowfindtohaveoccurredinthe likely explanation for the origin of Kasei Valles floor features is a
Amazonian (Fig. 4). For example, grooved Kasei Valles floors from combination of catastrophic flooding and glacial erosion (Chapman
UraniusDorsumtonearSharonovimpactcrater(Fig.5)producean and Scott, 1989). Water and even brines (Brass, 1980; Burt and
average crater retention age of 1.00Ga±0.22, allowing us to now Knauth,2003;ChevierandAltheide,2008)eruptedattypicalMartian
determine that the majority of the channel floor erosion was temperatureswouldquicklydevelopfrazilice(Baker,1979),which
Amazonian in age, rather than Hesperian (Table 2). Similar to would eventually consolidate and move downstream like glaciers
terrestrial analogs (see Method section, above), the channel floor (Lucchitta, 1982). Grooves and ridges can be observed eroded into
material unit Ach, a mostly erosional unit that occurred in fluvio- impact crater ejecta on the Kasei Valles floor with no observable
glacialepisode3,showsterracesthatmayhaveformedbynumerous damage to the impact crater rims and pits (Fig. 4A); perhaps
periodsoffloodingcloselyspacedintime.[Thestudyareaalsoshows suggesting infilling ice that protected the crater pit and rim from
evidence of 2 earlier episodes of east-trending Tharsis-sourced erosion(ChapmanandScott,1989).UnitAchalsoincludesrectilinear
Hesperian floods that occurred only in north Kasei Valles (see crevicesperpendiculartothetrendofthegrooves,someofwhichmay
Chapmanetal.,thisvolume).]Fig.5showshowwecannowproduce havebeenenlargedbylaterspringsappingprocesses(Fig.4A).These
betteragebracketsforthiserosion,asnewimagesclearlyshowhigh- crevices, which presumably formed by hydraulic plucking via flood
standingpreservedHesperianimpactcraterSharonovejecta(Hc)ona processes(Carr,1981)onlyoccurwithinthemappedareaofUnitAch
mesainnorthKaseiValleswithitsejectaawayerodedbythelater and not on older units that were also eroded by Amazonian floods
Amazonianfloods(unitAch;seeTable2)toproducedgroovedterrain (unitsHch2andHt4;Fig.2).
thatissuperposedbyyoungerAmazoniancraterejecta(Ac).Dating Another interesting observation to note is that although our
better-resolved materials with clear stratigraphic hierarchies intro- mapping now indicates the existence of additional streamlined
duces a level of accuracy not available for earlier Viking-based islands associated with unit Ach, none of these islands occur south
calculations. These late-stage floods from Echus cut the north- ofabout13°N(Fig.4B),stronglysuggesting thatfloodingprocesses
trending portion of Kasei Valles, formed large north-trending wereconfinedtothechannelfloorareathatliesnorthofthislatitude.
streamlinedislandsinthissamearea(coloredredonFig.4),cutthe Streamlinedislands(allstandinghigherthan100mabovetheKasei
grooves and ridges in the Kasei Valles floor, turned sharply east at floor),unlikelow-lyingridges,groovesandcrevices,aretoohightobe
latitude 20°N, longitude 75°W to follow the older Hesperian- buriedbylaterlavas.Areasofthechannelfloorsouthofabout13°N
established east-trend of Kasei, and eroded and removed parts of latitude are the location of topographically high-standing plateaus,
lavaunitHt4andmanyareasofHesperianchannelunit2(unitHch2) mesas,andhillsofEchusMontesandEchusChaos(Fig.4).Someareas
on a final path to Chryse Planitia (Fig. 2; see Chapman et al., this ofthesechaoticterrainshavelowerelevationsthanthestreamlined
volume). islands farther north, yet these locales are unburied by lavas,
M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255 243
Fig.3.FluvialerosionofTharsisMontesFormationmember4(At4);(A)GeologicmapshowingnarrowfluvialvalleyanddendriticfeederchannelscarvedintounitHt4(location
markedbystaronFig.13ofChapmanetal.,thisvolume),boxdenoteslocationof‘B’and‘C’;(B)enlargementshowingnarrowvalley;(C)THEMISphotomosaicbasefor‘B’.
indicating that topographically-higher streamlined islands are not collapsefeaturesarelocatedwithinaflat-flooredhorseshoe-shaped
likely located buried south of 13°N latitude. Previously mapping depressionboundbyconcave-inwardscarpsontheeastandwestside
indicatedthatEchusMontesandChaosweresurfacedbyHesperian and streamlined islands and grooved material on the north side
chaoticmaterialsthatformedbyLunaePlanumcollapse.Nocollapse (Fig.4B).Thehorseshoe-shapeddepressionmaymarkthenorthern-
features are located north of approximately 13°N latitude. New mostextentofLunaePlanumcollapseandperhapsthelocusofrelease
mappingindicatesthatthesehillsarefairlycoherentblocksofLunae for some catastrophic floods (and source of subsequent glacial ice)
Planummaterialthathavebeenfaulted,downdropped,partlyeroded thatcarvedsouthernKaseiValles.Youngerlavasnowburymostofthe
byflooding,andembayedbylavaflows.Theresurfacedareaswere hypotheticaldepressionsourceareaforAch.
mappedasAmazonianchaoticmaterialunitAct(Fig.6).Locallythe Theater-headedchannels(markedinblueonFig.7)arecutinto
mesas exhibit some high-standing levels of relatively undisturbed unitAch,andincisewallrockoftheplateausorunitHNunearEchus
plateau material mapped as Hesperian ridged plains (unit Hr), Chasma(seeFig.17AinChapmanetal.,thisvolume).Thesechannels
underlying layered material mapped as Hesperian–Noachian undi- are erosional features that expose much older materials (like unit
vided material (wallrock unit HNu), and younger embaying lavas HNu).Locally,drainagefromthesechannelsformedgulliesonalluvial
(Fig. 6). MOLA topography indicates that Echus Montes and Chaos fansthatextendoutfromplateauremnantsandoverlieunitAchin
244 M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255
Fig.4.GeologicmapshowingchannelfloorunitAchinblue,streamlinedislandsfromthiserosioninred,chaoticmaterialunitActingreen,olderunitslightlycolored,SC=Sharonov
crater,blackarrowmarksinset‘A’location,andboxesdenoteFigs.5and6locations;(A)THEMISVISV26599030(18kmwide)showingnortheast-trendinglongitudinalgroovesand
ridgesonKaseifloor(whitearrowmarked‘g’),nearlyperpendicularrectilinearcrevices(blackarrowmarked‘rc’),andpre-existingimpactcraters(marked‘c’)witherodedejecta
andalmostpristinecraterrimsandfloors,bluepolygonoutlinespresumedspringsappingenlargementofrectilinearcrevices;(B)stretchedversionsofMOLAdigitalelevationmodel
ofsouthKaseiValles,dottedredovalmarkstopographicdepressionboundbyconcave-inwardscarpsoneastandwestsides,containingchaoticmaterial,lackingstreamlinedislands,
andpartlyfilledbyyounglavas.
northgorgeofKaseiValles(seeFig.7inChapmanandTanaka,1996). andMalin,1985;KochelandPiper,1986).Inconsolidatedrocks,10to
Thereforethesesmallchannelsappeartopostdate(i.e.areyounger 100 times more water than volume of eroded material must be
than)theAmazonianfloodingevent.CutintounitAch,thesechannels dischargedtocreatespringsappingchannelsviasolutionofcement,
alsooccurattheheadsofthenarrownorthKaseigorges(Fig.7inset), salt fretting, ice melt, or freeze thaw (Laity and Malin, 1985). This
and in the north Kasei Valles floor as enlargement of rectilinear implies a fairly slow process compared with precipitation-induced
crevices(Fig.4A)andassmallisolatedsinuouschannels.Mostofthe channelerosion,andastheterrestrialtheater-headedchannelstake
channels cut into wallrock occur to the south and surround Echus severalhundredsofyearstoformonEarth,onMarstheywouldlikely
Chasma(Fig.7).Inmostcasestheserelativelyyoungchannelsfollow requireheatingofgroundwaterfromexpansivecontributingupland
pre-existingerosionalcrevicesorstructuralfractures.Ontheplateau areasoveralongperiodoftime(LaityandMalin,1985).Alternatively,
eastofEchusChasma,theyexposeunitHNuandfollowthetrendof onEarthmanysimilarfeaturesareformedbyvalleyglaciers,another
olderHesperiantrunkvalleys(fedbyoldnarrowdendriticchannels) source of a relatively slow erosional process, but one that requires
thatwerecutintounitHsm(seeFig.11inChapmanetal.,thisvolume, surfaceice(Lucchittaetal., 1991;Lucchitta,1982).Similartheater-
where HNu is shown in purple and Hesperian trunk and dendritic headedchannelsonthesouthrimofVallesMarinerishavebeennoted
channelsareshowninblue).Thetheater-headedchannelshavelong tobeanalogoustoglaciallyerodedtheater-headedchannelsonDevon
mainvalleyswithwidthsthatremainnearlyconstantandshortside Island(Lee,2000).Bothspringsappingandglacialerosionmechan-
tributaries. Alternatively, all of these features taken together are ismsoforiginforthetheater-headedchannelsrequireaclimatevery
consistentwithformationbygroundwaterorspringsapping(Laity different from the present conditions on Mars. However, a glacial
M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255 245
Fig.5.HRSCimageh5239_0000 showsremainsof100-km-diameterimpactcrater
Sharonovejecta(Hc)onmesatop(craterrimoutsideimage;imagelocationshownon
Fig.4), groovedAmazonian channelfloor unitAch,and overlyingyounger20-km-
diameter impact crater rim with ejecta (Ac) on channel floor; image width
approximately35km.
erosionoriginwouldbeconsistentwithgroovedchannelfloorsand
would imply a large ice mass in Kasei Valles south of the theater-
headed channels, feeding into the narrow curvilinear east-trending
gorgesofnorthKaseiValles.
Theflat-floorsofthecurvilinearnarrowgorgesofnorthKaseiare
topographically as low as Chryse Planitia (Fig. 1B). Many have
suggested that water from channels appears to have pooled in the
northern lowlands, forming an ocean (Witbeck and Underwood,
1983; Lucchitta et al., 1986; Parker et al., 1989, 1993; Jöns, 1990;
Fig.6.GeologicmapofHRSCh3283_001showingEchusMontes(locationonFigs.4and1A)
Baker et al., 1991; Scott et al., 1992; Lucchitta, 1993). Parker et al. andassociatedmaterialunits:Amazonianchaoticmaterial(Act),Hesperianridgeplain
(1989,1993)and Jöns(1991) mappedwhattheyinterpretedtobe material(Hr),andHesperian–Noachianundividedmaterial(HNu)allembayedbyyounger
shorelines of this ocean in the northern plains of Mars, showing a lavas(At5andApf),ancienterodedcraterrimsinbrown,youngererodedcratersinyellow;
large bay in Chryse Planitia. MOLA data are consistent with olderflowsofAt5shownindarkershade,arrowsmarkembaymentoferodedoutcropsofunit
At5byApfflows(bestseenonTHEMISVISV17177015andV26936031);boxdenoteslocation
hypotheses that suggest an ocean in the northern lowlands in the
ofupperleftinsetshowingmaterialunitsonTHEMISVISV09814013,insetlowerrightshows
past history of Mars, because a putative shoreline (mostly labeled HRSCtopographyofEchusMontes(blue=low).
contact 2 in Parker et al. (1989)) closely matches a global
equipotential surface with a mean elevation of −3760±560m
(Headetal.,1999).IfChrysePlanitiawereabayofanorthernocean
onMars,thentheeast-trendinggorgesmaynotbeduetoflooding,
but could be remnant fjords with glacially-carved theater-headed designatedasplaty(Chapmanetal.,2007;Fig.8BandC).Although
valleysattheirheads,similartoterrestrialfjords.Ifalargemassof high-resolution imagery shows a platy-textured surface, this unit
Kasei Valles floor ice fed into Chryse Planitia, then perhaps the differsfromthemember5lavaflowsinthatonthescalesofoldlow-
groovedfloorofKasei,thetheater-headedchannels,andthenarrow resolutiondatasetsitappearedextremelysmoothcomparedwiththe
gorges all formed at the same time, rather than the floor material rough-surfaced,steepflowterminationsoftheTharsisflows.Hence
beingformedfirstandsubsequentlybeingcutbylatertheater-headed thematerial'splatyappearancewasundetectedinVikingandMariner
channels. The erosional period that formed the theater-headed 9imageryoriginallyusedtomapKaseiValles,andtheaerialextentof
channels marks the end of the third fluvio-glacial cycle within the its smooth, relatively uncratered surface was thought to be young,
studyarea. mantled parts of the Kasei Valles floor or Tharsis member 5 lavas
Thefourthvolcaniccycleinthestudyareacapsoffallthemajor (RottoandTanaka,1995).UnitApfbearsstrikingsimilaritiestothe
depositional processes in Kasei Valles. This cycle was a prolonged distinctive “platy-ridged” surface morphology Cerberus flows in
periodofvolcaniceruptionsintheLateAmazonianthatproduceda ElysiumPlanitia(Keszthelyietal.,2000).Likelavaflowmaterialthe
massive output of lava from Tharsis Montes and possibly Echus unitlocallydisplayslobateterminations(seedarklobemarked‘B’on
Chasma. Previously, this volcanic material was mapped as Tharsis Fig. 17A in Chapman et al., this volume) and scalloped margins
Montes Formation member 5 (Scott and Tanaka, 1986; Rotto and (Fig. 11C). Unit Apf can be traced 2100km from Echus Chasma to
Tanaka,1995).Wehaverevisedthepreviouscontactboundariesof withinthesouthernmostnarrownorthKaseigorge,southofSharonov
member 5 lavas and have delineated and mapped a new material: impactcrater(Fig.8).Outcropsofthisunitcanbebestobserved(1)in
Amazonian platy-flow unit Apf. Large areas of this unit display EchusChasma,(2)onthefloorsoftheater-headedchannelswestand
relatively flat, featureless angular to subangular plates of material northofEchus,(3)directlysouthandeastofEchusMontes(chaotic
surrounded by topographically-higher rough areas. This texture is terrain in south Kasei Valles), (4) within amphitheater-headed
246 M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255
We cannot determine with certainty whether the platy-flow
material was derived from the Tharsis Montes area or from Echus
Chasma.Howeverstreamlinedislandsandgroovessuggestthatthe
last catastrophic flood was sourced by Echus Chasma, and unit Apf
directlyoverliesthischannelmaterialandistheprimarymaterialon
thefloorofEchusChasma.Inaddition,platy-flowsurfaceshavenot
beenobservedonTharsismaterialunitselsewhere.Wesuggestthat
theserelationssupportthestrongpossibilitythatEchusChasmawas
the source of the platy-flow material. Same-source lava floods
followingcatastrophicfluidflowisnotunique,assourcesforseveral
other outflow channelson Mars appearto have emitted lavas after
floodwatersuchasMangalaValles(Basilevskietal,thisvolume)and
AthabascaValles(Jaegeretal.,2007).
Emplacement of unit Apf took tens to a hundred million years.
Numerouscratercountsfrom6separateareasoftheplaty-flowmaterial
indicateaverageabsoluteagesfromsouthtonorthof205Ma±0.05(in
Echus Chasma), 281Ma±0.1 (near Echus Montes), 254Ma±0.13
(channel midpoint), 155Ma±0.05 (south narrow gorge), 133Ma±
0.03 (north narrow gorge), and 128Ma±0.03 (south of crater
Sharonov;Fig.10andTable3).Theagedatesformaterialonthefloor
ofEchusChasmarangefrom155Mato523Ma(Table3).Thesevariable
ages make sense when one takes into account the characteristics of
terrestrialfloodlavas(andthepossibilityofanothermaterial onthe
floorofEchus,suchasunitApec;seebelow).Themajorityofterrestrial
floodlavasareinflatedpahoehoeflowstransported100sofkmfrom
ventsunderneathaninsulatingcrust(Selfetal.,1997;Thordarsonand
Self,1998).Thelocalplaty-ridgesurfaceindicatesmultiplesurgesoflava
emplacement.Platesformina2-stageprocess:athickstablecrustforms
on moltenlava;latera surgein thelava fluxdisrupts thiscrust and
transportslargepiecesasraftsonmoltenlava(KeszthelyiandMcEwen
2007).
UnitApfisoverlainbytheyoungestflowssourcedfromtheTharsis
rise:TharsisMontesFormationmember5(unitAt5;Fig.10).Material
of member 5 forms relatively thick lobes of material, whose
terminations were clearly observed on moderate resolution Viking
imagery. Emplacement of unit At5 took hundreds of million years.
Crater counts from 4 separate areas of member 5 material indicate
average absolute ages from north to south of 803Ma (southeast of
Fesenkov crater), 186Ma (west of the Kasei midpoint), 133Ma
(withinKaseiValles),and89Ma(northwestofEchusChasma;Fig.10
and Table 4). The oldest (northmost) flows at 803Ma may have
Fig.7.GeologicmapshowingAmazoniantheater-headedchannelfloorunitAchtin occurredduringthelaststagesofEchusChasmaflooding,butasthese
blue,olderunitslightlycolored,andboxdenotinginsetlocation;(inset)MOLA-shaded
reliefshowingtheater-headedchannels(arrows)cutintoKaseiVallesfloorandfeeding older lavas cannot be traced to the floor of Kasei Valles, the
intonorthKaseiVallesgorges. stratigraphicrelationcannotbedetermined(Fig.10).However,itis
possible that an older flow may have entered the channel, as thick
lobes of older lobate material can be observed overlying Echus
Montes,erodedbyflooding,andembayedbyyoungerflowsofunit
Apf and member At5 (Fig. 6). The surface appearance of the Echus
channels of north Kasei Valles, and (5) within the east-trending Montesflowsissosimilartomember5thatwedesignatedittobethe
gorgesonthefloorofnorthKaseiValles(Fig.8).Inthesouthernmost samematerial,butolderthanlocalEchusMontes-embayingpartsof
gorge of north Kasei Valles, these putative flood lavas can be seen memberAt5.Lavasthatdatefrom186Maalsodidnotreachthefloor
enteringandexitingnarrowchannels(Fig.9).Thesenarrowchannels ofKaseiValles,butthelateryoungerflowscoveredlargepartsofthe
mayhavebeeninplacebeforetheflowsandonlybeenutilizedbythe southKaseiVallesfloor.OnlytheyoungestflowsintheKaseiValles
flowmaterial,buttheirassociationwithunitApfsuggeststhatthey channeloverlieplaty-flowmaterial(Fig.11).Thedecreasingflowage
wereformedbythisunit'sheat-inducederosionofunderlyingfriable fromnorthtosouthmaymarkashiftinthelocusofTharsismagma
orice-richmaterials.Thesesmallnarrowchannelshavenotributaries chamberswithtime.
ordistributaryfansandappearverysimilartolava-formedrills(Guest There may have been some subsurface ice present during the
andGreeley,1977,p.62–68).UnitApfwasinterpretedtobepossible emplacement of member 5 lava flows, as a small sinuous and
mudflowdeposits(WilliamsandMalin,2004)orfloatingicedeposits anastomosing channel can be observed to arise from a shallow
(Woodworth-Lynas and Guigne, 2004). However, we interpret the depressionwithintheunit'soutcropareajustsouthofimpactcrater
unittohavebeenveryfluid,highlymobile,floodlavasbasedonthe Sulak at about lat. 18.6 and long. −78.63 (Fig. 10A). Removal of
platy-flow texture, rough surface with a material strength that subsurface ice may also be responsible for tilting (subsidence and
supports crisp crater rims (Fig. 11C), large areal extent, and uplift)oflava-covered,relativelythinandcoherentplatesofsurface
association with possible rills. Within Echus Chasma this material material (marked ‘p’ on Figs. 10A and 12). Alternatively, uplift
may have formed a ponded lava lake, and apparently it filled and resulting from rise of magma may have also formed the coherent
traversedtheshallowtopographicdepressionoutlinedinFig.4. plates. These plates may have broken away from the surface along
M.G.Chapmanetal./EarthandPlanetaryScienceLetters294(2010)238–255 247
Fig.8.GeologicmapshowingoutcropsofAmazonianplaty-flowunit(Apf)inbrightorange,otherunitsinlightcolorsandnotlabeled,lettersdenoteinset‘A–D’locations;Fig.9
locationsnumbered;(A)THEMISVISV21220003(7kmwide)showingunitApfinbetween(overlying)northKaseiVallesfloorgrooves;(B–C)denoteextentsofplaty-flowtextures:
(B)HiRISEPSP_006625_2075(2kmwide);(C)THEMISVISV02162006(3kmwide);(D)THEMISVISV18425016(3kmwide).
enlargedpre-existingzonesofweaknesssuchasjoints,lavachannels, beingemittedafterdepositionofmemberAt5.Geomorphicrelations
andcoolingcrackswithinlavadeposits.Weinterpretsomeofthese that support late-stage outflow can be observed within the south-
fracturestohaveoriginatedaspossiblefracturingalongunit-confined ernmost,younglavalobeofTharsisMontesFormationmemberAt5
zonesofweakness,aslocalfracturescanbeobservedineitherunits thatflowedintoEchusChasma(Figs.13Aand14A).Thislavalobeis
Apf and At5 that do not continue onto the surface of the adjacent eroded and dissected by shallow channels, leaving a ring of rubble
unit or disrupt plateau wallrock (Fig. 12B). This type of fracture is marking the former lobate borders of the flow (Fig. 13A). Similar-
another indication of the age difference between units Apf and At5 appearing terrestrial analogs of this eroded morphology can be
(Fig.12B).Otherfloorfractures,innorthEchusChasma(Fig.13A)and observed in Iceland, where lava flows have been eroded by being
directlynortheastofthechasma(onplatemarked‘p’inFig.12),form “washed” by catastrophic flood waters to leave behind shallow
theboundaryofsuchlargecoherentblocksthatthesefracturescould channels, blocky terminus areas, and piles of internal rubble
bebetterinterpretedasboundingfracturesondowndroppedblocks surrounded by mud-covered or mantled terrain (Fig. 13B). [As
ofplateauthathavebeensubsequentlycoveredbyyoungerlavas.If notedintheIntroduction,thisisasummarypaperofallAmazonian
thisisthecase,laterupliftandsubsidencemusthaveoccurredalong unitsinthestudyarea,eachofwhichcouldmeritadetailedpaper,
theblockedges,asoneofthesefeaturesinnorthEchusChasmacanbe suchasonedevotedtoEchusChasmafloormaterialsandadditional
observedtooffsetapromontoryofwallrock(Fig.14E)—suggesting detailsregardingIcelandicflood-erodedbasaltanalogs.]
continuedfaultingoccurredalongthiszoneofweakness. Theshallowchannelingcanbetracedbacktoafracture(fossa)on
Outflowofwater(andpossiblylava)mayhavealsooccurredalong thefloorofEchusChasma(Figs.13A and 14A, andD).Presumably,
thesetypesoffracturesinEchusChasma,withlimitedfluidamounts water forced upward from a buried aquifer would behave like a