SECOND WORKSHOP ON MARS VALLEY NETWORKS
27
FLUVIAL VALLEYS AND SEDIMENTARY DEPOSITS IN XANTHE TERRA: IMPLICATIONS FOR
THE ANCIENT CLIMATE ON MARS. E. Hauber1, K. Gwinner1, M. Kleinhans2, D. Reiss3, G. Di Achille4,5, R.
Jaumann1, 1Institute of Planetary Research, DLR, Rutherfordstr. 2, 12489 Berlin, Germany, 2Departement of Physical Geography, University of Utrecht, PO-box 80115, 3508 TC Utrecht, The Netherlands, 3Institute of Planetology,
Westfälische Wilhelms-Universität, Wilhelm-Klemm-Str. 10, 48149 Münster, Germany, 4International Research
School of Planetary Science, Università d`Annunzio, Viale Pindaro, 65127 Pescara, Italy, 5now at: Laboratory for
Atmospheric and Space Physics, University of Colorado, Boulder, Colorado 80309, USA.
Summary: A variety of sedimentary deposits is
observed in Xanthe Terra, Mars, including Gilberttype deltas, fan deltas dominated by resedimentation
processes, and alluvial fans. Sediments were provided
through deeply incised valleys, which were probably
incised by both runoff and groundwater sapping. Mass
balances based on high-resolution HRSC digital terrain
models show that up to ~30% of the material that was
eroded in the valleys is present as deltas or alluvial fan
deposits. Stratigraphic relationships and crater counts
indicate an age of ~4.0 to ~3.8 Ga for the fluvial activity. Hydrologic modeling indicates that the deposits
were probably formed in geologically very short timescales. Our results indicate episodes of a warmer and
wetter climate on early Mars, followed by a long period of significantly reduced erosion rates.
Introduction: Two major questions in studies of
Martian valley networks concern the nature of the fluvial activity (surface runoff versus groundwater seepage) and the depositional environment (subaerial alluvial fans versus lacustrine deltas), a question that addresses the ability of Mars and its climate to host large
bodies of standing surface water [e.g., 1,2]. The interest in sedimentary deposits on Mars was recently increased by findings of the imaging spectrometers,
OMEGA and CRISM, which identified the spectral
signatures of phyllosilicates in association with sedimentary deposits in Holden and Jezero craters [3].
Research activities currently focus on a spectacular
deposit in the Eberswalde Crater, which is characterized by well-preserved layering and sinuous, meandering channels [e.g., 4-6], and on deposits in Holden
Crater [e.g., 7,8], both of which are considered as potential landing sites for the upcoming MSL mission.
However, there are many other sedimentary deposits
which have not been examined in detail using the new
high-resolution datasets from recent Mars missions.
Here we focus on sedimentary deposits in Xanthe
Terra, between the outflow channels Maja Valles to
the west and Shalbatana Vallis to the east (Fig. 1).
Geologic context: The study area is located in
Xanthe Terra between Maja and Shalbatana Valles
(Fig. 1). It is characterized by a high density of craters
with >10 km diameter and resurfacing in the late Noachian [9]. The study region is dissected by two types
Figure 1. Sketch map of the study area (the Subur Vallis delta
is located outside this base map at 11.72°N and 307.05°E). Black:
valleys; orange: sedimentary deposits; open circles: impact craters;
yellow: superposed craters and their ejecta, which formed after the
valleys onto which they are superposed; blue: outflow channel; rows
of red dots: ejecta and secondary crater chains radiating outward
from large impact crater (C); boxes: locations of Fig. 2 and 3.
of valleys. Valleys of the first type are narrow (few
km) and shallow (meters to tens of meters). They were
probably formed by fluvial processes that were not
very intense and did not lead to the development of
mature drainage patterns with high-order tributaries
and high drainage densities. A second type of valleys
is younger and and includes Nanedi, Hypanis, Sabrina,
Ochus, Drilon, and Subur Valles. Valleys of this type
are deeply entrenched and show characteristics like
SECOND WORKSHOP ON MARS VALLEY NETWORKS
amphitheater-shaped heads, few tributaries, low drainage density, and almost constant widths (see [10] for
terrestrial analogues on the Colorado Plateau). They
were interpreted to be indicative of an origin by
groundwater sapping rather than by precipitation
[11,12], although other processes such as late-stage
fluvial entrenchment, erosion through layered stratigraphy, and incision into a duricrust might also produce these morphological properties ([13]; see also
[14], who describe a terrestrial amphitheater-headed
canyon that was not eroded by sapping). A contribution from overland flow to sapping valley formation on
Earth seems to be likely from the analysis of terrestrial
analogues in Utah and Arizona [15]. The few very
high-resolution images taken by HiRISE do not show
unambiguous evidence for many small or shallow
tributary channels indicative of overland flow connecting with the large entrenched valleys. But at least one
of the images displays a faint pattern of very shallow
and small valleys (Fig. 2), which might have hosted
fluvial channels. Erosion and blanketing of the surface
with wide-spread eolian dunes may have obscured the
traces of others. Small interior channels are incised
into some valley floors [16] and suggest sustained flow
of water [17]. Valley formation might have begun in
the Noachian and continued into the Early Hesperian
[18; 9; 12]. Most deposits are situated where fluvial
channels breach the rims of strongly degraded impact
craters and debouch onto the crater floors. The region
is moderately dust-covered [19], and consequently no
spectral signatures of alteration minerals could be
identified [20,21; Poulet, pers. comm].
Figure 2. Morphologic evidence for surface runoff in Xanthe
Terra. (a) Portion of the Nanedi Valles system (HiRISE
PSP_003341_1855; see Fig. 1 for location), white boxes mark locations of panels b and c (not shown). (b) Faintly visible shallow valleys might have been fluvial channels localizing surficial runoff due
to precipitation. Similarly, shallow contributing streams have been
found on the Colorado plateau above theatre-headed valleys [15].
Inset shows details, contrast was enhanced for better visualization of
valley topography. The co-existence of both deeply incised valleys
28
and shallow upslope tributaries makes sapping as a sole cause for the
creation of the theatre-headed valleys unlikely (compare to Figs. 2b
and 2c of [17]).
Observations of deposits: Two of the deposits
identified in Xanthe Terra are described in detail here,
the others are covered elsewhere [22] (Fig. 1).
Nanedi delta. The smallest of the deposits is located near 8.5°N, 312°E at the end of the Nanedi
Valles (hereafter referred to as Nanedi Deposit). It is
located inside a degraded impact crater with 6.5 km
diameter (Fig. 3) and covers almost the entire crater
floor. The surface of the deposit covers an area of
~23 km2 and is dissected by a pattern of radially distributing channels. Where the deposit does not onlap
the inner crater wall, its distal parts are marked by a
steep front, along which fine layers are exposed. The
layers have thicknesses of only a few meters, and there
are no boulders or blocks that could be identified at the
scale of the HiRISE image (25 cm/pixel). The layers
can be traced for a few hundred meters and do not display evidence for faulting or unconformal contacts. It
must be noted, however, that the entire area is heavily
covered with wind-blown dunes and ripples which
obscure much of the underlying bedrock. The height of
the distal margin is about 50 m, according to the
HRSC DEM (Fig. 4a). Assuming a constant thickness
of 50 m over the entire area of the deposit, we obtain a
volume of about 1.15 km3. Since the thickness is
probably larger at the crater center, we consider this
value to be a lower limit of the total volume. The surface of the deposit has an average topographic gradient
of ~0.014, as measured in the HRSC DEM. The crater
has a small outlet channel at its eastern part. Outlet
channels from craters are strong indicators for openlake basins [23], and the morphology of the outlet
channel, including a small streamlined island, is suggestive of liquid flow. This and the small slope would
be in agreement with ponding of water in the crater,
formation of the deposit in a lake, and drainage of lake
water into the outlet channel. However, the available
data do not permit a definitive interpretation, and a
formation as alluvial fan, as suggested by [12] can not
be ruled out.
Subur delta: An interesting sedimentary deposit
can be observed in an unnamed 41 km-diameter impact crater centered at 11.97°N and 307.28°E (outside
area shown in Fig. 1). Where Subur Vallis breaches
the southeastern crater wall, a deposit (hereafter Subur
Deposit) extends for several kilometers onto the crater
floor (Fig. 4). This deposit is the most complex example described in this paper because, even from the plan
view images, it shows two distinct depositional subsystems. The upper part of the deposit consists of a flat
SECOND WORKSHOP ON MARS VALLEY NETWORKS
29
mented [26]. In most cases, the proximal part of the
delta complex is preserved and consists of a large Gilbert-type fan delta with foresets reaching the basin
floor and merging into a deep-sea resedimented system
[e.g., 27]. For example, some fan deltas in the Gulf of
Corinth (Greece) display a delta front-slope-fan apron
architecture, which strongly resembles the morphology
of the Subur Delta (compare Fig. 2a of [27] to Fig. 4).
Therefore, our tentative interpretation is that the Subur
Deposit might be composed of a proximal Gilbertdelta merging into deep-water resedimented facies.
Figure 3. Small delta in an unnamed 5 km-diameter impact crater at the terminus of Nanedi Valles (see Fig. 1 for location). (a) The
downstream part of Nanedi Valles cuts the southern crater wall, and
a little outlet emanates from the eastern crater wall (white arrows).
The small and relatively pristine appearing outlet curves into the
course of an abandoned, old and degraded channel (OC), and a small
streamlined island (SI) attests to the flow of a liquid medium (detail
of CTX P06_003407_1872; crater center at 8.62°N, 48.0°E; North is
up; white box shows location of panel b). (b) Distal portion of the
deposit, showing the ubiquitous presence of eolian ripples and dunes
(detail of HiRISE image PSP_006954_1885; white box shows location of panel c). (c) Blow-up of topographic scarp marking the distal
margin of the deposit. Layers with apparent thicknesses of a few
meters or less can be traced for hundreds of meters without evidence
for unconformal contacts.
plain with a steep front, like it is observed in Gilberttype deltas with flat topsets and steep foresets [24].
The lack of significant fluvial dissection of the deposit
suggests that fluvial activity sharply declined after
deposition, as already noted by [25]. According to
MOLA data, the proximal part slopes very gently with
an angle of ~1.1°, while the steep front displays slopes
of up to 10°. Since small-scale topography is smoothed
in MOLA data, this steep slope value is probably even
an underestimation. At the distal part of this deposit
and in a topographically deeper position there are low,
concentric terraces. These could be erosional scarps
from wave action. On the other hand, they might constitute strata which are not directly linked to a delta
front because they appear to be subhorizontal. On
Earth, a number of fan deltas show large resedimentation processes that form complex deep-sea deposits,
and in some cases almost all of the delta is resedi-
Figure 4. Gilbert-type delta in unnamed crater at the termination of Subur Vallis (at 11.72°N and 307.05°E). (a) High-resolution
view of the downstream part of Subur Vallis and the delta (mosaic of
MOC images; white scale bar is 2 km, illumination from the left). (b)
morphological sketch map of the delta. Note the well expressed
features that are typical for Gilbert-type deltas (flat topset and steep
foresets) and the possibly resedimented distal layers (black scale bar
is 2 km). (c) enlarged detail, marked by white box in panel a. Arrows
point to possible depositional lobes, although it can not be excluded
that they represent an erosional signature. Indentations of the delta
front might have been left by mass wasting processes.
Time scales of sedimentation: Flow discharge
and sediment transport computations (see [28,22] for
details) indicate minimum time scales for fluvial activity in the order of ~101 to ~104 years (Tab. 1). Such a
formation time is consistent with the study of [29] and
[30], who both give very short time-scales for deposit
formation. The lack of late-stage incision into sedimentary deposits indicates a relatively abrupt termination of the fluvial activity, in agreement with similar
observations by [25] and [30].
SECOND WORKSHOP ON MARS VALLEY NETWORKS
Table 1. Morphometry of valleys and deltas, sediment transport
rates, and minimum time scales for delta formation (years are Earthyears; see text for details). Note that the channel dimensions are
large when compared to most terrestrial rivers.
Nanedi
Valley volume (km3)
Delta volume (km3)
Channel width (m)
Channel depth (m)
Channel slope
Flow discharge
(106 m3/s)
Sediment transport
rate (km3/yr)
Minimum formation
time (yr)
assuming bankfull
depth†
Hypanis
3.74
1.15
744
45
0.0037
0.18
Sabrina
29
9.4
575
25
0.0082
0.073
0.56
0.47
0.82
2.1
20.2
183
~850
~150
750
28
0.0098
0.12
†
See [22] for time scales assuming 50% and 10% of bankfull depth.
To correct for intermittency, divide by the assumed intermittency.
Chronology: The valley systems of Nanedi, Hypanis, Sabrina, and Subur Valles, with which the deposits described here are associated, are all situated in
Noachian and Early Hesperian terrain [9]. Crater
counts on the floor of the Nanedi Valles yield a Late
Noachian age [31]. At several locations, the valleys are
superposed by impact craters (Fig. 1) [12]. Crater
counts on the ejecta of some of the craters that superpose the valleys provide absolute model ages of
~3.8 Ga (31), which would require that fluvial processes operated in the valleys before that time. These
absolute model ages [31] are based on crater diameters
between ~600 m and more than one kilometer.
Conclusions:
1. Different types of sedimentary deposits in
Xanthe Terra display morphologic similarities of some
deposits to fan deltas (e.g., Nanedi and Subur), while
others resemble alluvial fans (e.g., Hypanis [22]). They
are associated with deeply entrenched valleys, which
may have formed by a combination of runoff and sapping. Chronologic and stratigraphic relationships indicate a formation of valleys and deposits in the Late
Noachian to Early Hesperian.
2. The formation of the sedimentary deposits
occurred probably over relatively short time-scales of
several years to ~104 years, which is in agreement with
previous studies [e.g., 29,30]. The general lack of incision of the sedimentary bodies is indicative of an
abrupt end of flow, and would be in agreement with an
intense terminal period of fluvial activity that ended
abruptly [13,25].
30
3. The relatively well preserved nature of some
small deposits suggests that the fluvial activity ended
abruptly and that erosion rates after thereafter were
generally very low.
References:
[1] Irwin, R.P. et al. (2002) Science, 296, 22092212. [2] Wood, L.J. (2006) Geol. Soc. Am. Bull., 118,
557-566. [3] Ehlmann, B.L. et al. (2008) Nature Geosci., 1, 355-358. [4] Malin, M.C., Edgett, K.S. (2003)
Science, 302, 1931-1934. [5] Moore, J.M. et al. (2003)
Geophys.
Res.
Lett.,
30,
2292,
doi:10.1029/2003GL019002. [6] Pondrelli, M. et al.
(2008) Icarus, in press (available online).
[7] Pondrelli, M. et al. (2005) J. Geophys. Res., 110,
CiteID
E04016,
doi:10.1029/2004JE002335.
[8] Grant, J.A. et al. (2008) Geology, 36, 195-198.
[9] Rotto, S., Tanaka, K.L. (1995) U.S. Geol. Surv.
Misc. Invest. Ser. Map I-2441. [10] Laity, J.E., Malin,
M.C. (1985) Geol. Soc. Amer. Bull., 96, 203-217.
[11] Sharp, R.P., Malin, M.C. (1975) Geol. Soc. Amer.
Bull., 86, 593-609. [12] Harrison, K.P., Grimm, R.E.
(2005) J. Geophys. Res., 110, CiteID E12S16,
doi:10.1029/2005JE002455. [13] Howard, A.D. et al.
(2005) J. Geophys. Res., 110, CiteID E12S14, doi:
10.1029/2005JE002459. [14] Lamb, M.P. et al. (2008)
Science, 320, 1067-1070. [15] Irwin, R.P. et al. (2006)
LPS XXXVII, abstract 1912. [16] Carr, M.H., Malin,
M.C. (2000) Icarus, 146, 366-386. [17] Irwin, R.P. et
al. (2005) Geology, 33, 489-492. [18] Masursky, H. et
al. (1980) NASA Tech. Mem. 82385, 184-187.
[19] Ruff, S.W., Christensen, P.R. (2002) J. Geophys.
Res., 107, 5127, doi: 10.1029/2001JE001580.
[20] Poulet, F. et al. (2007) J. Geophys. Res., 112, CiteID
E08S02,
doi:
10.1029/2006JE002840.
[21] Gondet, B. et al. (2007) OMEGA/MEx Global
Mineralogical Maps. 7th Int. Conference on Mars, LPI
Contr. No. 1353, abstract 3185. [22] Hauber, E. et al.
(2008) Planet. Space Sci., in press (available online).
[23] Fassett, C.I., Head, J.W. (2008) LPS XXXIX,
abstract 1139. [24] Gilbert, D.K. (1885) Ann. Rep.
U.S. Geol. Surv., 5, 69-123. [25] Irwin, R.P. (2005) J.
Geophys. Res., 110, CiteID E12S15, doi:
10.1029/2005JE002460. [26] Postma, G. (1990) in:
Colella, A., Prior, D. B. (Eds.), Coarse-Grained Deltas, Int. Ass. Sedimentol., Spec. Publ. 10, Blackwell,
Oxford, pp. 13-27. [27] Ferentinos, G. et al. (1988)
Marine Geol., 83, 43-61. [28] Kleinhans, M.G. (2005)
J. Geophys. Res., 110, CiteID E12003, doi:
10.1029/2005JE002521. [29] Jerolmack, D.J. et al.
(2004) Geophys. Res. Lett., 31, L21701,
doi:10.1029/2004GL021326. [30] Kraal, E.R. et al.
(2008) Nature, 451, 973-976. [31] Reiss, D. Et al.
(2005) Meteor. Planet. Sci., 41, 1437-1452.