Morphology, evolution and tectonics of Valles Marineris wallslopes (Mars)

Abstract

Hillslopes up to 11 km in height can be found along the walls of the Valles Marineris troughs. The widest and deepest troughs are grabens, in which tectonics probably exerted the primary control on the wall morphology. Geographical variations in the wall morphology and profiles show that they result from complex, persistent tectonic influences, and that significant changes in erosional processes occurred during this evolution, from late Hesperian to late Amazonian. Preliminary calculations suggest that about 85–95% of the fault-controlled wall relief probably formed in an “ancient” stage prior to this transitional period. A study of the volatile content of the wall rocks, based upon the morphology and distribution of impact craters on the surrounding plateaus, shows that extreme erosional widening of the Central Valles Marineris troughs occurred during the “ancient” stage of high ground ice content. During the subsequent “recent” stage of tectonic and morphological evolution, the wall materials were partly desiccated.

Full text

Ž.
Geomorphology 37 2001 329–352 www.elsevier.nlrlocatergeomorph
Morphology, evolution and tectonics of Valles Marineris
ž/
wallslopes Mars
Jean-Pierre Peulvast a,), Daniel Mege b, Jan Chiciak c, Franc¸ois Costard a,
`
Philippe L. Masson a
aUMR 8616-Orsay Terre, UniÕersite Paris-Sud, Bat. 509, 91405 Orsay Cedex, France
´ˆ
bLaboratoire de Tectonique, case 129 ESA CNRS 7072, UniÕersite Pierre et Marie Curie, 4 Place Jussieu, 75252 Paris Cedex 05, France
´
cUniÕersite de Paris-Sorbonne, UFR de Geographic, 191 Rue Saint-Jacques, 75005 Paris, France
´´
Received 1 July 1994; received in revised form 1 August 1998; accepted 14 January 1999
Abstract
Hillslopes up to 11 km in height can be found along the walls of the Valles Marineris troughs. The widest and deepest
troughs are grabens, in which tectonics probably exerted the primary control on the wall morphology. Geographical
variations in the wall morphology and profiles show that they result from complex, persistent tectonic influences, and that
significant changes in erosional processes occurred during this evolution, from late Hesperian to late Amazonian.
Preliminary calculations suggest that about 85–95% of the fault-controlled wall relief probably formed in an AancientBstage
prior to this transitional period. A study of the volatile content of the wall rocks, based upon the morphology and distribution
of impact craters on the surrounding plateaus, shows that extreme erosional widening of the Central Valles Marineris troughs
occurred during the AancientBstage of high ground ice content. During the subsequent ArecentBstage of tectonic and
morphological evolution, the wall materials were partly desiccated. q2001 Elsevier Science B.V. All rights reserved.
Keywords: Morphology; Valles Marineris; Mars
1. Introduction
For Mars and Earth, very steep hillslopes with
Ž.
high relief several kilometers are related to special
Ž.
conditions, including 1 powerful geodynamic pro-
cesses, such as volcanism with induced gravitational
)Corresponding author.
Ž.
E-mail address: peulvast@geol.u-psud.fr J.-P. Peulvast .
Ž.
tectonics Cipa, 1994 or mountain tectonic uplift;
Ž. Ž.
2 exceptionally strong slope materials; and 3 deep
erosion in the uplifted areas. The longevity and
preservation of steep, high hillslopes also depend on
the age of the tectonics and on the efficiency of the
erosional processes that lower high-relief areas.
Moreover, the high potential energy related to great
Ž.
slope heights Embleton and Whalley, 1979 , the
possible effects of unloading related to rapid erosion,
and vibrations related to the seismic–tectonic activ-
ity result in the frequent occurrence of mass move-
0169-555Xr01r$ - see front matter q2001 Elsevier Science B.V. All rights reserved.
Ž.
PII: S0169-555X 00 00085-4

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352330
Ž.
Fig. 1. Simplified geomorphological map of Valles Marineris modified from Peulvast and Masson, 1993a . 1: Trough edge, ridge; 2: lower escarpment, bluff; 3: impact crater;
4: impact crater with lobate ejecta; 5: closed depression; 6: chaotic terrain; 7: landslide; 8: fault; 9: wrinkle ridge.

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 331
Ž
ments Brunsden, 1979; Lucchitta, 1979; Fort, 1993;
.
Fort and Peulvast, 1995 .
The Valles Marineris walls in the Tharsis region
of Mars have a relief up to 11 km in the central parts
of a 4000-km-long system of troughs that lie just
Ž.
south of the martian equator Fig. 1 . The remarkable
scale of this relief arose in part Abecause the absence
of recent pluvial and fluvial activities on Mars pre-
Ž
vented rapid infilling of the depressionsBLucchitta
.
et al., 1992, p. 478 . The very high hillslopes of
Valles Marineris form deep natural sections through
the Tharsis Plateau. The morphology reflects proper-
ties of the wall materials, thereby constraining mod-
els of structure, composition, and evolution of the
Ž.
upper layers of the martian crust Carr, 1981 .
Fault-related morphology is clearly evident along
Ž
many wall segments Carr, 1981; Mege and Masson,
`
1996; Schultz, 1991; Witbeck et al., 1991; Peulvast
.
and Masson, 1993a . Relationships among these
structures, and the erosional landforms of the walls
and wallfoot deposits show the sequences of events
Ž
that shaped the Valles Marineris troughs Lucchitta,
1987; Lucchitta and Bertolini, 1989; Peulvast and
.
Masson, 1993b . Thus, a geomorphological study of
these landforms is critical to understanding the very
high wallslopes of Valles Marineris in regard to
regional tectonics.
2. Morphological relationships
This study will focus on the central part of Valles
Ž.
Marineris Fig. 1 , which consists of a series of
parallel troughs or AchasmataB. Most of these troughs
are interconnected, except for Echus, Hebes, and
Juventae Chasmata, which form a discontinuous par-
allel set a few hundred kilometers north of the main
system; these chasmata form the deepest depressions
within the whole Valles Marineris. The troughs are
Ž.Ž.
of two types Lucchitta et al., 1990 : 1 scalloped
Ž
troughs and pit crater chains Tithonium and south
.Ž.
Coprates Chasmata and 2 straight and well-aligned
Ž.
troughs Ius and Coprates Chasmata , which widen
between 648and 77830Xlongitude to form a 700-
Ž.
km-wide system of isolated Hebes Chasma or in-
Ž
terconnected troughs Ophir, Candor and Melas
.
Chasmata . The latter comprise the ACentral Valles
Marineris troughsB, and are 120 to 160 km wide.
Ž.
Type 2 troughs are between 6 and 8 km deep, and
Ž
up to 9 to 11 km deep in Melas Chasma U.S.
.
Geological Survey, 1986a,b . The very high trough
Ž.
walls overlook either: 1 wide and relatively flat
Ž.
parts of the trough floors, or 2 moats that separate
them from interior benches or tablelands of layered
Ž.
deposits e.g. Hebes Chasma: Peterson, 1981 . In
both cases, excepting local landslide areas, only
minor volumes of waste deposits occur along the
walls. In several places, the pristine morphology of
downfaulted segments of upland plateau can be rec-
ognized on the trough floor, indicating the tectonic
Ž
origin Blasius et al., 1977; Schultz, 1991; Peulvast
.
and Masson, 1993a,b .
The Valles Marineris structures comprise a rift
Ž
system Hartmann, 1973; Sharp, 1973; Frey, 1979;
.
Masson, 1977, 1985; Witbeck et al., 1991 , belong-
Ž
ing to the Tharsis radial pattern of fractures Carr,
.
1981 , whose trend is expressed in many shallow
grabens in the surrounding plateaus, and in structures
inside the troughs. These fractures are perpendicular
to the trend of wrinkle ridges on the plateaus, which
Ž
are also related to Tharsis stress patterns Chicarro et
.
al., 1985; Tanaka et al., 1991; Watters, 1993 . The
boundary faults that control long and rectilinear sec-
tions of the chasmata are some of the most conspicu-
Ž.
ous among the regional structures Schultz, 1991 .
Though rifting alone cannot explain the box-like
terminations of Candor Chasma, or the transitions
between pit crater chains, scalloped troughs, and
grabens, good evidence exists for floor downfaulting
and probably asymmetric rifting.
Ž.
Lucchitta 1978 attributes the present configura-
tions of the Valles Marineris walls to erosional scarp
Ž.
retreat, recognizing three major types of walls: 1
the spur-and-gully type, which is most common along
Ž
straight and well-aligned troughs e.g. Coprates and
.Ž.
Ius Chasmata, interior ridges ; 2 walls dissected by
Ž.Ž.
tributary canyons south Ius Chasma ; and 3 land-
slide scars forming broad curved or straight recesses
in the chasma walls. This classification was later
Ž
simplified Bousquet et al., 1987; Lucchitta et al.,
.
1992 to distinguish only spur-and-gully morphol-
ogy, landslide scars, and small-scale talus slopes.
Ž.
Our classification Fig. 2 emphasizes the systematic
relationships among morphology, tectonics, and other
structural elements in Central Valles Marineris
Ž.
Peulvast and Masson, 1993a .

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352332
Fig. 2. Wall types of Valles Marineris. 1: Smooth talus slope; 2: spurs and gullies slope; 3: landslide deposits with hummocky material;4: landslide deposits with hummocky and
fan-shaped materials; 5: rotational slump; 6: aeolian flutes; 7: chaotic terrain; 8: slope gradient; 9: crest.

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 333
High, fault-controlled walls are well displayed
Ž.
along Ius and Coprates Chasmata Figs. 1 and 3 and
Ž
on interior ridges paralleling the chasma walls for
.
example, in the straight and well-aligned troughs .
They are also found in north Melas Chasma and in
the northern walls of Ophir and Hebes Chasmata
Ž.
Lucchitta et al., 1992 . Overall, rectilinear outlines
and long rows of triangular facets cut into the lower
Ž.
spurs northern wall of Coprates Chasma allow a
precise mapping of the main normal faults along the
Ž.
fault scarps Schultz, 1991 . These walls generally
show spur-and-gully morphology with widths that
range from 15 to 30 km and slopes in the head walls
up to 308. Dissection is locally controlled by minor
Ž
oblique sets of synthetic and antithetic faults Peul-
.
vast and Masson, 1993a . Though in many places the
gullies merge with the troughs at the level of the
floors, the presence of transverse fault steps and
continuous rows of triangular facets and fault scarps
at the base of the intervening spurs shows that
faulting locally continued after the end of the gully-
ing process. Narrow interior ridges, such as Geryon
Montes in Ius Chasma, are probably horsts or half
Ž.
horsts Mege and Masson, 1996; Schultz, 1991 . The
`
dissected ridge flanks intersect in sharp crests, where
local splitting into two parallel crests suggests that
slow mass movements contributed to the wasting
Ž.
process Flageollet, 1988 .
Tectonic control is also obvious along wall seg-
ments dissected by tributary canyons with blunt heads
Ž.
and V-shaped cross profiles Lucchitta, 1978 . For
example, the 30- to 130-km-long Louros Valles,
South of Ius Chasma, dissect a straight fault scarp,
and appear to be passively controlled by intersecting
oblique arrays of fractures. Northwest of Melas
Chasma and in parts of Tithonium Chasma, solitary
canyons are clearly incised along shallow oblique
grabens intersected by the main troughs. These
canyons have average gradients between 28and 58;
the canyons merge with the trough floors and are
considered to have been formed by structurally con-
Ž
trolled sapping processes Sharp, 1973; Lucchitta et
.
al., 1992 .
Most walls that lack obvious structural control
were probably formed by erosional retreat along
Ž
fault scarps for example, the main fault lines are
obliterated by erosion andror deposition processes;
.
Fig. 4 . Moreover, the headwalls of the erosional
landforms, which intersect the grabens and the wrin-
kle ridges of the surrounding plateaus, are only
locally controlled by these structures, as in the case
of some landslide scars. Most of these walls are
located in Central Valles Marineris, where extreme
widening and interconnection of parallel troughs
Ž.
Ophir, Candor and Melas Chasmata are at least
Ž
partly related to erosional wall retreat Peulvast and
.
Masson, 1993a,b . Some of these walls display spur-
and-gully morphology, and they generally overlook
interior plateaus of layered deposits or the moats that
separate these plateaus from the walls. Embayments
of these deposits commonly occur inside the ero-
sional re-entrants, and even inside the mouth of some
gullies.
Landslide scars form broad curved or straight
segments of chasma wall up to 100 km long and
locally recessed 5 to 10 km from the adjoining walls
Ž.
Lucchitta, 1978 . The scars are the only erosional
landforms whose deposits are recognizable on the
chasma floors, with slump blocks at the head and
vast aprons of longitudinally ridged or smooth mate-
Ž.
rial Fig. 5 . The scars occur mainly in central Valles
Marineris, between segments of dissected walls with-
out clear structural control; however, they are also
found in rectilinear segments of Ius, Hebes, Ophir,
and Melas Chasmata, and in Coprates Chasma, and
in scalloped troughs of western Valles Marineris
Ž.
Tithonium Chasma . Whereas some landslide de-
posits occur on the chasma floors, especially in the
straight and well-aligned troughs, many partly embay
or bury the interior layered deposits, in the moats,
below the level of the tablelands or benches.
The basal fault scarps, which are restricted to
rectilinear segments of the main troughs, and espe-
Ž.
cially to the northern sides Schultz, 1991 , suggest
that faulting continued after the end of the gullying
process. Gullying probably implies some kind of
vertical erosion and longitudinal waste transport by
fluids or viscous interstitial material, probably ice
Ž.
Lucchitta, 1978 , related to the widening of the
Central Valles Marineris troughs during the late Hes-
Ž.
perian Lucchitta et al., 1992 , and to the emplace-
ment of interior layered deposits. Embayment rela-
tions inside the mouth of some gullies shows that
most of the erosional work was completed before the
end of deposition. In all cases, gullying predates the
landslides, whose deposits generally obliterate and

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352334
destroy spur-and-gully topographies. The formation
Ž
of these relatively young features middle to late
.
Amazonian was followed in places by faulting
Ž.
Mege and Masson, 1996 . Because the morphology
`
of some landslides seems to be related to a certain
Ž.
content of groundwater or ice Lucchitta, 1987 , the

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 335
Ž
Fig. 4. Southern wall in eastern Candor Chasma. Note the morphological contrast of spur-and-gully topography between the northern Fig.
.Ž. Ž
6i and southern walls m , and the lack of basal scarps at the bottom of the southern wall. The spurs associated with a tectonic activity Fig.
.
6 have more orderly branching and shorter crests than other walls. The debris from spur-and-gully erosion may have been incorporated into
Ž.
interior deposits n or is simply removed by unknown processes. Small, ArecentBmass wasting processes, expressed on walls by talus
Ž. Ž. Ž.
slopes o , and fine grooves on cornices and talus slopes p , are responsible for debris tongues q at the base of the Valle Marineris walls.
Ž.
Viking picture 912 A 13 49 mrpixel .
AwetBconditions of erosion and trough widening
associated with landsliding must have been followed
by AdryBconditions. The present mean annual tem-
perature exceeds the frost point by 20 K in the study
Ž.
area Clifford, 1993 , and other current conditions
Ž
imply a lack of ground volatiles Squyres et al.,
.
1992 .
The distribution of the various types of very high
Ž.
wallslopes Fig. 2 shows that the highest slopes are
mainly restricted to the aligned troughs, which are
Fig. 3. Western Ius Chasma, displaying two grabens, located on the both sides of Geryon Montes. North of Geryon Montes: The southern
Ž. Ž.
wall displays spurs and gullies only a , and the northern wall displays spurs and gullies associated with AancientBfaceted spurs b and
Ž. Ž.
ArecentBbasal structural scarp c . All these features are cut by a landslide d on the left of the image. South of Geryon Montes: The whole
Ž.
northern wall displays a continuous fault scarp 1 km high e , suggesting a ArecentBorigin. The southern wall displays a spur-and-gully
Ž. Ž . Ž .
morphology f and tributary canyons g . Geryon Montes crests are locally split into two parts h , suggesting the occurrence of slow mass
movements. The white broken line underlines a 1-km topographic step on the trough floor. The two sets of small parallel arrows indicate the
Ž. Ž . Ž .
ends of two ArecentBscarps about 800 m left and 500 m right high from U.S. Geological Survey, 1986a,b corresponding to 11% and
Ž.
7%, respectively, of the total wall relief. Viking pictures 65 A 09, 12 75 mrpixel .

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352336
Ž.
Fig. 5. Landslide in Ius Chasma described by Lucchitta 1979 . Arrows: fissure-related volcanic flow. Viking pictures 919 A 15, 16, 17, 18
Ž.
46 mrpixel .
true grabens. The lesser depths of the highly scal-
loped troughs and pit crater chains are probably
Ž
related to a different structural origin Schultz, 1989;
.
Lucchitta et al., 1990 . Wall heights are also related
to differences between the altitude of the Tharsis
plateau and the altitude of the floors, which tend to
decrease from west to east, from 8 km in parts of the
Central Valles Marineris, to 3 km in East Coprates
Chasma. In Central Valles Marineris and in Coprates
Chasma, the northern walls are generally the highest.
This is partly related to the location of the grabens
on the southern side of the topographic crest that
forms the eastern continuation of the Tharsis rise,
and it may also be related to asymmetric faulting
ŽSchultz, 1991; Peulvast and Masson, 1993b; Mege
`
.
and Masson, 1996 .
The morphological contrast between the widened
Central Valles Marineris and the narrow troughs of
the western and eastern parts of the system cannot be
explained by structural and tectonic factors, because
wall retreat without clear tectonic control appears to
be the main mechanism of widening. No significant

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 337
wall retreat and deposition is related to the deep
dissection by tributary canyons of Ius and Tithonium
Chasmata, whereas gullying of the walls and deposi-
tion of thick layered deposits are chronologically, if
not genetically, linked in the widened troughs. Such
relationships suggest complex spatial heterogeneity
in the properties of the wall materials, including
composition, structure, shear strength, volatile con-
tent, and time-related changes in the erosionalrde-
positional processes and in the structural conditions
of trough formation. The long-lived resurfacing his-
Ž.
tory of Mars Tanaka et al., 1992 suggests that the
morphological diversity of the Valles Marineris walls
involves the coexistence of landforms shaped during
several stages of development, which may be re-
vealed by geomorphological study.
3. Structural interpretation
3.1. Valles Marineris and the origin of Tharsis
The Ius, Melas, Coprates, Candor, Ophir, and
Hebes Chasmata have a structural origin that is
likely to be linked to the Tharsis tectonics. Using
Valles Marineris as a constraint on Tharsis stress
Ž.
modeling, Sleep and Phillips 1979 showed that
isostasy is consistent with the observed martian grav-
ity field. The structural patterns expected from this
mechanism, however, are not fully consistent with
the location and orientation of Valles Marineris. The
latter may be explained by modeling a succession of
tectonic stages, first isostatic, followed by flexural
ŽBanerdt et al., 1982; Willeman and Turcotte, 1982;
Solomon and Head, 1982; Sleep and Phillips, 1985;
see reviews in Banerdt et al., 1992; Mege and Mas-
`
.Ž.
son, 1996 . Alternatively, Tanaka et al. 1991 and
Ž.
Banerdt and Golombek 1992 hypothesize a syn-
chronous mechanism based upon a different litho-
spheric structure beneath the internal and external
parts of the Tharsis surface. Convective forces may
have been important in the earliest history of Tharsis
Ž
prior to later flexural relaxation Kiefer and Hager,
.
1989; Schubert et al., 1990 . Because the most ap-
propriate model has to be consistent with structural
data, inferred from study of morphological features,
especially along the walls of Valles Marineris, the
morphotectonic interpretation of these walls is criti-
cal to the geophysical understanding of Tharsis.
A body of arguments suggests a major change in
morphogenetic processes on Mars at the Hesperianr
Ž
Amazonian boundary e.g. Lucchitta, 1984; Baker
.
and Strom, 1992; Parker et al., 1993 . This distinc-
tion between successive morphogenetic conditions
will be referred to as AancientBand ArecentBin the
discussion that follows.
One of two main extensional patterns for Valles
Marineris consists of triangular faceted spurs, devel-
oped on fault scarps below the faceted spur-and-gully
topography. Terrestrial faceted spurs indicate normal
Ž.
faulting e.g. Hamblin, 1976; Wallace, 1978 , and
morphological comparisons between the terrestrial
Ž.
and martian features are striking Figs. 3, 6, and 7 .
The Valles Marineris faceted spurs probably repre-
sent erosion of normal fault scarps under the Aan-
cientBconditions. Faceted spurs in regions with es-
Ž
pecially clear tectonic control e.g. eastern Candor
.
Chasma display spur-and-gully topography with
more orderly branching than occurs on the walls,
which are not fault-controlled. They also have shorter
Ž.
crests Figs. 4 and 6 , which can be compared to
Ž
normal fault scarps subject to erosion on Earth Fig.
.
7.
A second extensional pattern is expressed as nor-
mal fault scarps at the wall bases. These scarps
evolved during conditions of reduced erosion, proba-
bly by wind, and they probably formed under current
Ž.
ArecentBconditions. For the prolonged and continu-
ing tectonic activity of Valles Marineris, AancientB
features are expected to be observed on the upper
parts of the walls, whereas ArecentBfeatures are
expected to be restricted to the lowest parts.
3.2. The Ius Chasma example
A detailed description of Ius Chasma illustrates
the morphological distinction of AancientBpatterns
Ž. Ž.
Fig. 8 from ArecentBpatterns Fig. 9 . We studied
about 150 pairs of Viking stereo images, with resolu-
tion mostly ranging from 40 to 75 mrpixel. East of
Calydon Fossa and west of 838W, the southern
Geryon Montes wall is entirely composed of a con-
Ž.
tinuous, 1 km high fault scarp Figs. 3 and 9 .
Lacking faceted spurs, this scarp probably only expe-

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352338
Fig. 6. Northern wall in eastern Candor Chasma, showing clear tectonic control, expressed by AancientBfaceted spurs beneath spurs and
Ž. Ž. Ž .
gullies i , and a ArecentBbasal scarp j . k : Small, ArecentBlandslide cutting across the AancientBspur-and-gully morphology, and
Ž. Ž
probably across ArecentBbasal scarp. 1 Landslide material. Small parallel arrows indicate a ArecentBscarp about 800 m high U.S.
.
Geological Survey, 1986a,b , corresponding to about 10% of the whole wall relief. The ArecentBj scarp is about 400 m high and
corresponds to about 5% of the wall relief. Viking picture 911 A 12.
rienced ArecentBmorphogenetic conditions. If the
scarp were older, its height would likely have al-
lowed spur-and-gully morphology to develop. Al-
though this scarp has been well preserved, the wall
to the south is intensely eroded down to its base,
indicating an older age and the possible occurrence
of a volatile-rich wall material that facilitated erosion
during trough formation.
The most common case encountered is the occur-
rence of spur-and-gully topography, and the triangu-
lar facets on the higher portions of the trough walls.
ARecentBfault scarps are located on the lower por-
Ž.
tions Fig. 3 . Sometimes, however, triangular facets
are located at the basal part of the slope, and Are-
Ž
centBscarps are located inside the trough for exam-
.
ple, inside the AancientBtrough geometry; Fig. 9B .
This ArecentBdeformation probably occurred on
faults different than AancientBboundary faults, and it
locally defines narrower grabens.
The most important process to occur after the
formation of basal scarps is landsliding. Gullying
and sapping generally occurred before landsliding
Žduring early Amazonian for sapping, according to
.
Witbeck et al., 1991 , perhaps just at the turning
point of the erosional condition change. Although
sapping is often difficult to interpret except for the
Ž
dendritic network area in Ius Chasma Louros
.
Valles , some observations indicated that sapping

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 339
Ž.
Fig. 7. Diagrams showing the evolution of faceted spurs produced by periods of movement separated by periods of stability on Earth. A
Ž. Ž.
Undissected fault scarp; B development of faceted spurs by streams cutting across scarp; C period of stability with slope retreat and
Ž. Ž.
development of a narrow pediment; D recurrent movement; E dissection of new segment of scarp by major streams and by those
Ž.
developed on the face of faceted spurs formed in B; F new period of stability with slope retreat and development of another narrow
Ž. Ž.
pediment at base of mountain front upthrown block; G recurrent movement; H dissection of scarp formed in G, resulting in a line of
Ž
small faceted spurs at base of mountain front. Remnants of narrow pediments are preserved at apices of each set of faceted spurs redrawn
.
from Hamblin, 1976 . Compare the final morphology shown at stage H with the i and j sites of Fig. 6. Moreover, note the similarities
between stage G and the profile defined by b and c in Fig. 3. Whereas, stage G corresponds to fault reworking before fault scarp erosion on
Earth, this stage corresponds in Valles Marineris to the ArecentBconditions, with a reduced erosional activity. Stage H is not further
expected as far as the current morphogenic conditions exist on Mars.
might have locally lasted until the ArecentBperiod.
For example, on the southern side of Geryon Montes,
some ArecentBsapping might have cut across both
spur-and-gully topography and basal scarps. Locally,
some tributary canyons wholly or partially cut across
basal scarps, and others hang above them. The
chronology between the last tectonic activity and
mass-wasting processes is thus variable, perhaps be-
cause both are contemporaneous at the geological
scale.
Although the AancientBfracture set was partly
destroyed by the later geomorphological events or
covered by deposits, it is considered to be responsi-
ble for the major part of the extensional deformation.
Vertical offsets from the ArecentBperiod are usually
limited to a few hundred meters and only exception-
ally exceed 1 km. Two main orientations are recog-
Ž.Ž.
nized in the fault patterns Figs. 8 and 9 : 1 east-
trending, following the main walls of Ius Chasma,
Ž. Ž.
and 2 oblique to the trend 1 .
The east-trending fault orientation extends from
the boundary between Noctis Labyrinthus and Ius
Chasma to 838W. Except for the ArecentBwalls, all
of the walls of Ius Chasma have a structural and

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352340
AancientBorigin. Deformations under the ArecentB
conditions are shared among boundary faults and
other faults of various directions. East of 838W, the
Ius troughs follow the N758W direction, up to Melas
Chasma. All the trough walls north and south of
Geryon Montes have an AancientBgenesis. East of
Geryon Montes, the ArecentBtrough does not display
a clear graben geometry, for example, boundary
faults cannot be continuously followed.
The oblique fault set includes northeast, north-
west, east, and north orientations. Both the northeast
and northwest directions are strongly expressed from
the boundary with Noctis Labyrinthus to 838W. Far-
ther east, the northwest trend becomes more and
more predominant. West of 838W, the east trend is
the boundary fault orientation, whereas this trend
links right-stepping N758W trending faults at the
base of the northern wall between 738W and 768W.
Locally, several faults in the southernmost part of
Melas Chasma follow the north trend.
North of Geryon Montes, widespread landslides
complicate recognition of the AancientBstructural
geometry. Nevertheless, left- and right-stepping faults
developed AancientlyBbelow the northern wall from
738to 828W. The ArecentBtectonic patterns form a
more complex fracture system. The landslides gener-
ate numerous northeast- and northwest-trending fea-
tures, whereas inter-landslide areas preserve the
dominantly trough-trending boundary faults. The
landslide faults may not have any link with gravity-
controlled movements, which would be oriented
preferentially perpendicular to the sliding direction
Ž.
e.g. Jibson and Keefer, 1993 . Both northeast- and
northwest-trending faults are aligned with tributary
canyons to the north and south, which align along
Ž
fault sets that predate Sinai Planum, for example,
.
Noachian in age; Scott and Tanaka, 1986 . These
may result from faulting, whose offset is concealed
by the block morphology, or from the collapse of
wet landslide material, preferentially accenting old,
Noachian weakness zones. Some northeast- and
northwest-trending faults were active during the Aan-
cientBperiod, and these may also display inherited
Noachian trends.
South of Geryon Montes, Ius Chasma may be
Ž.
divided into four major graben G1 through G4 . The
Ž. Ž.
Fig. 8. Location of AancientBstructural features in eastern a and western b Ius Chasma from morphoclimatic arguments. Morphology:
Ž. Ž. Ž . Ž .
1 Cornice with smooth talus slope; 2 cornice with spurs and gullies; 3a crest with spurs and gullies on both sides; 3b crest with spurs
Ž. Ž. Ž. Ž.
and gullies on one side and smooth talus slope on the other side. Structure: 4a Fault; 4b conjectured fault; 5a normal fault; 5b
conjectured normal fault. G2–G4: Graben subdivisions south of Geryon Montes, discussed in the text.

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 341
Ž.
Fig. 8 continued .
westernmost graben partly formed under ArecentB
Ž.
conditions Fig. 9: G1 , and its floor is cut by a
Ž.
northwest-trending topographic step, Fig. 3 , 1000
Ž.
m high U.S. Geological Survey, 1986a,b . The sec-
Ž.
ond graben Figs. 8 and 9: G2 is bordered by faults
stepping southward, has a poorly constrained wall
geometry, and is shifted northward compared to the
Ž.
third graben Figs. 8 and 9: G3 . Considering the
AancientBfaults responsible for the main offsets, the
transition zone between G2 and G3 displays north-
west-trending extensional faults, with triangular
faceted spurs. The third graben is connected with the
Ž.
fourth one Figs. 8 and 9: G4 by two east–north-
east-trending extensional faults. The graben orienta-
tion reverts back to a north trend, south of Geryon
Montes.
East of Geryon Montes, the N758W boundary
fault direction continues into West Melas Chasma up
to 73.58W, but this trend is underlined by few fault
traces. Geryon’s gently east-dipping topography close
to Melas Chasma may indicate that the horst approx-
imately ends at its current topographic terminus,
implying lesser extension eastward than westward in
Ius Chasma. Conversely, this horst may have under-
gone increasing extension eastward, leading to com-
plete burial under more recent sediments at its cur-
rent topographic terminus. The latter hypothesis is
supported by observations that the Geryon Montes
boundary fault set may well prolongate further East,
with fragmented segments illustrated in Fig. 9.
Both the AancientBand ArecentBfault sets show
some northwest control, which becomes especially
strong between 728W and 758W. This trend also
Ž
occurs in Melas Chasma Peulvast and Masson,
.
1993a,b . The ArecentBtectonic activity appears to
be responsible for very small offsets. The north-

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352342
trending faults in southwest Melas Chasma may be a
structural expression of sediment packing, together
with undulations observed by Peulvast and Masson
Ž.
1993a,b .
3.3. AAncientBand ArecentBtectonics in Valles
Marineris
A striking morphological feature is the clear
prevalence of ArecentBdeformation at the base of the
northern walls rather than southern walls. This is
especially true for Candor and Ophir Chasmata, and
in parts of Ius, Coprates, and Tithonium Chasmata.
Spurs and gullies are often located on the upper
Ž
slopes and lie above continuous basal scarps Figs. 3,
.
6, and 10 . Landslides have locally destroyed the
Ž.
previous spur-and-gully morphology Fig. 5 , and
they postdate basal scarps. Triangular faceted spurs
seem to be widespread on the northern slopes above
basal scarps. In contrast, the southern walls display
few traces of either AancientBor ArecentBtectonic
activity. The lack of scarplets suggests that ArecentB
tectonics is reduced in comparison to the northern
wall tectonics. AAncientBfaulting may have oc-
curred, however, and been obliterated by further
erosion. For instance, a part of the western Candor
south wall has sapping morphological characteristics
Ž.
Sharp, 1973; Lucchitta, 1978; Kochel et al., 1982
Ž
or collapse features Baskerville, 1982; Tanaka and
.
Golombek, 1989 . This morphology is further devel-
oped at the western border of Candor Chasma and
Ž.
linked to sapping channels Fig. 11 . In these re-
gions, the current morphology is that of a fault scarp
which underwent significant retreat, probably during
AancientBas well as ArecentBstages of wall develop-
ment. The original boundary fault may coincide with
the pit chain joining Candor and Tithonium Chas-
mata.
The current rectilinear morphology of some inte-
rior deposit boundaries within East Candor Chasma
suggests that the AancientBCandor development
might have been influenced by transverse structures
Ž.
Fig. 10 . The borders of the western branch of
Candor Chasma are not exactly aligned with the
borders of the eastern branch. The central Candor
Chasma area might mask a buried transition zone,
perhaps responsible for the shifting of the Candor
Ž. Ž.
Fig. 9. Location of ArecentBstructural features in eastern a and western b Ius Chasma from morphoclimatic arguments. Symbols as in
Ž. Ž .
Fig. 8; 6 second-order drag fault Mege and Masson, in preparation . G1–G4: Graben subdivisions south of Geryon Montes, discussed in
`
Ž. Ž.
the text. A Faults possibly connecting Geryon Montes boundary faults and Melas Chasma internal structures. B Examples of ArecentB
extensional movements in grabens on faults, which are different than AancientBboundary faults. The two easternmost A faults may be
Ž.
correlated with a 2- to 3-km high topographic scarp U.S. Geological Survey, 1986a,b .

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 343
Ž.
Fig. 9 continued .
Ž
west and east grabens, as in Melas Chasma Peulvast
.
and Masson, 1993b .
Coprates Chasma has a complex structure of alter-
nating horsts and grabens. Several horsts occur close
to Melas Chasma and close to Eos and Capri Chas-
Ž
mata. Five major grabens may be observed Fig.
.Ž.
10b: G5–G9 , including one G6 whose original
geometry was destroyed by landsliding on the south-
ern wall. In contrast to Ius, the oblique trends be-
tween these grabens cannot be correlated to tributary
Ž
canyon directions for example, ancient buried struc-
.
tures . AAncientBand ArecentBfaulting is widespread,
except in the pit chains. ARecentBmovements are
reduced at their proximity.
Despite the local occurrence of spur-and-gully
topography, little clear evidence exists for faceted
spurs in Tithonium Chasma. Close to Noctis
Labyrinthus, Tithonium looks like Ius Chasma, a
trough whose morphology is primarily controlled by
Ž
tectonics and further enlarged by erosion Figs. 1, 3,
.
10, and 12 . The southern wall of this part of Titho-
nium Chasma displays spurs and gullies and some
possible faceted spurs, whereas the northern wall
displays basal scarps under a very degraded upper
slope with spur-and-gully remnants. This part of
Tithonium is clearly a graben, in which ArecentB
tectonic activity occurred on the northern wall only.
The morphology of Tithonium in the other parts of
the trough seems to be primarily related to erosional
processes and secondarily to tectonics. Some other
ArecentBfaulting is reflected in two discontinuous
trends following the trough direction at the bottom of
Ž.
both walls Fig. 10b .
The structural control of Echus Chasma is not
clear. AAncientBeast-trending tectonic structures may
have existed, like those in the Hebes trough. Some

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352344
Ž.
Fig. 10. a Generalized Valles Marineris tectonic patterns formed under AancientBmorphogenic conditions. b: Generalized Valles Marineris
tectonic patterns formed later, under ArecentBmorphogenic conditions. Symbols as in Fig. 8. G5–G9: Graben subdivisions in Coprates
Ž. Ž.
Chasma mentioned in the text. F1, F2: Faceted spurs on the Tithonium Chasma southern wall F1 , and northern scarp on F2 in Tithonium
Chasma, discussed in the text.
evidence of slight ArecentBevolution of Echus
Chasma occurs on the southern wall. The further evolution of Echus Chasma may have included the
same events as in Eos and Capri Chasmata, such as

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 345
Ž.
Fig. 11. Southwestern and western walls of Candor Chasma, displaying a collapse-like geomorphology r . The southern wall might have
Ž.
originally followed the pit-crater chain s trend at the West, and developed a spur-and-gully morphology, further destroyed by another
Ž.
process, such as collapse. Viking picture 66 A 18 75 mrpixel .
catastrophic flood discharge, which formed Kasei
Valles and modified Chryse Planitia.
3.4. RelatiÕe magnitudes of ArecentBand Aan-
cientBtectonism
Because central Valles Marineris troughs are likely
to have formed originally as grabens, it is possible to
estimate the relative relief generated by ArecentB
tectonic movements, in contrast to that attributed to
AancientBtectonics. For instance, in the western Ius
Ž.
area Fig. 3 , ArecentBscarp height measurements
Ž.
U.S. Geological Survey, 1986a,b give 800 and 500
m at two different sites, corresponding to 11% and
7% of the entire wall relief, respectively, assuming
Ž
that deposits are thin near these walls which is
likely because of the layered deposits, e.g. Schultz,
.
1991 . The AancientBmovements are, thus, expected
to be responsible for 89% to 93% of the tectonic
relief at these sites. The following values for the
ArecentBperiod are approximate maximum values in
Ž. Ž .
other central Valles Marineris sites: a 300 m 6% ,
Ž. Ž . Ž. Ž . Ž.
b 400 m 5% , c 800 m 10% , d 1000 m
Ž . Ž. Ž . Ž.
14% , and e 800 m 15% , respectively, for a
Ž.Ž.Ž. Ž.Ž.
Tithonium Fig. 12 , b and c Candor Fig. 6 , d
Ž. Ž
Ophir, and e Hebes Chasmata northern walls The
Hebes southern wall might have undergone move-
ments comparable to those on the Hebes northern
.
wall . The remaining percentages correspond to what
results from AancientBmovements. The amounts of
ArecentBmovement tend to increase northward, from
Ius to Hebes Chasmata, but this trend needs to be
confirmed by further study.
3.5. Volatile distribution and wall deÕelopment
The following features have been used to infer the
present or past occurrence of volatiles in Valles
Ž.
Marineris region: 1 rampart craters in the surround-
Ž.Ž.
ings of the troughs Costard, 1990 , 2 the tributary
Ž
canyons, most probably formed by sapping Kochel
.Ž.
and Piper, 1986 , 3 landslide deposits in Ophir–

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352346
Ž. Ž .
Fig. 12. Eroded faceted spurs t and basal scarps u in Tithonium Chasma near Noctis Labyrinthus. F1, F2: See Fig. 10 and the text. The u
Ž.
scarp is about 300 m high U.S. Geological Survey, 1986a,b and corresponds to about 6% of the whole slope gradient, assuming that the
Ž.
northern wall originally reached the same height as the southern wall. Viking picture 63 A 63 75 mrpixel .
Candor Chasmata may have been emplaced as
Ž.Ž.
Agigantic wet debris flowsBLucchitta, 1987 , 4
the layered deposits, which may have been emplaced
Ž.Ž.
in ice-covered lakes Nedell et al., 1987 , and 5 the
morphology of certain gully-mouth lobate deposits,
which may be related to the presence of interstitial
Ž.
ice Peulvast and Masson, 1993b . All of these fea-
tures are most abundant in the central Valles
Marineris troughs. Although AdryBprocesses have
also been imposed for explaining some of these
Ž.
features McEwen, 1989 , the general view is that
volatiles or ground ice were involved in the forma-
Ž
tion and erosional widening of the canyons McCau-
.Ž.
ley, 1978 . Fanale 1976 estimates that the 3- to
7-km wall heights coincide with the depth of the
permafrost, such that landslides would involve
volatile-rich layers present at depth or would corre-
spond to the emergence of underground aquifers
Ž. Ž.
Carr, 1979 . Battistini 1985 proposed a strong
association between the ground ice and the morphol-
ogy of Valles Marineris, linking collapse depres-
sions, aligned along grabens, to the presence of
volatile rich materials.
An objective sampling of the volatile distribution
around the Valles Marineris may be accomplished by
studying the distribution and morphological charac-
teristics of rampart craters, which have lobate ejecta
Ž
terminating distally in ramparts Mouginis-Mark,
.
1979; Horner and Greeley, 1987 . This morphology
implies emplacement by flows around craters over
the surface just after the impact event through the
Ž
melting of volatiles Carr et al., 1977; Mouginis-
.
Mark, 1979 . Thus, rampart craters are considered to
Ž
be excellent ground-ice indicators Squyres et al.,
.
1992 . In our study, all rampart craters were sur-
veyed within "58of the Valles Marineris. Data were
collected for 62 rampart craters in the size range
5–40 km located on the plateau surfaces above the
troughs. We did not survey the small number of
impacts on the trough floor. As a first hypothesis, we

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 347
infer that lobate ejecta located AnearBthe troughs
Ž.
less than 50 km reflect the nature of terrains that
form the Tharsis plateau around Valles Marineris
and the wall rocks.
We assume that the volatile content of the exca-
vated material directly influences the ejecta mobility
at the time of the impact event, as shown experimen-
Ž
tally Gault and Greeley, 1978; Wohletz and Sheri-
.
dan, 1983 . Because the extension of ejecta lobes is
Ž
theoretically proportional to the volatile content for
.
appropriate material properties, porosity, etc. , ejecta
Ž.
mobility EM can be expressed by the ratio of the
maximum diameter of ejecta deposits normalized by
Ž
the diameter of the parent crater Mouginis-Mark,
.
1979; Battistini, 1984; Kuzmin et al., 1988 . The EM
ratio indicates the volatile content of ground-ice at
the time of the impact event. EM ratios of 2 to 3
indicate low-mobility ejecta, while EM ratios of 5 to
7 indicate high-mobility ejecta.
Fig. 13 shows the characteristics of individual
rampart craters around Valles Marineris. Fig. 14
shows variations of the EM ratio for 62 rampart
craters, according to their location in longitude along
the troughs. The two regression lines in Fig. 14,
showing the EM variation with longitude, stop be-
tween 708W and 758W because of insufficient data
in that region. The scatter in the data may result from
target heterogeneity, variations in the total ice con-
tent of the excavated material, the relatively sparse
data set, andror related variations in the content of
ground ice. Fig. 13 shows a clear relationship be-
Ž
tween ejecta mobility located at 58on each side of
.
the troughs , and the distribution of the widened
parts of the troughs with the relative abundance of
volatile-related erosional landforms. Rampart craters
Ž
with high ejecta mobility deposits EM values close
.
to 4.5 predominate, especially south of Melas
Ž.
Chasma the widest trough , but the number of mo-
bile-ejecta craters decreases both to the west and east
of the central Valles Marineris, except in the Louros
Valles area. This same trend is reflected in the
regression lines of the EM ratios in Fig. 14.
Changes in ejecta mobility appear to be best
explained by a progressive enrichment in volatile
materials within the Tharsis region around the cen-
Ž.
tral Valles Marineris Costard, 1990 . The widening
of the chasmata in this region supports this interpre-
tation. The inferred high volatile content would pro-
Fig. 13. Distribution of rampart craters within "58of the Valles Marineris. Ejecta mobility is calculated using EM ratio of the maximum
range of ejecta lobes normalized to the diameter of the parent crater.

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352348
Ž. Ž.
Fig. 14. Variations of the EM ratio y-axis for 62 rampart craters according to their location on the canyon x-axis . Both regression lines
Ž.
exhibit a clear general rise in the EM ratio ejecta mobility towards the central part of the canyon. Note the enrichment in volatile materials
from the margins to the Central Valles Marineris. This concentration of volatiles may have contributed to the widening of the chasma.
vide a plausible mechanism for several erosional
processes including the very large wet debris flows
that seemed to occur until middle to late Amazonian
Ž.
Lucchitta and Bertolini, 1989 . The reduction of
wall retreat outside the central Valles Marineris re-
gion may be explained by the presence of a volatile
poor permafrost andror be the presence of more
coherent and less porous wall materials.
4. Discussion and conclusions
Valles Marineris exposes some of the deepest
vertical sections in the crust of any planetary body,
providing a unique morphotectonic record for under-
standing the structural development of the upper
martian crust. Although the range of slope processes
is not yet fully understood because of uncertainties
concerning the nature and evolution of the ground
volatiles, gullying and widening processes, and the
role of water or ice in mass movements, the regional
evolution fundamentally underwent different stages
from Hesperian to the present. A rather stable open-
ing strain regime persisted in Valles Marineris for a
Ž.
long time at least from Hesperian to present , which
has major implications for the 5000-km Tharsis
bulge. Moreover, some important differences occur
between terrestrial rifts and Valles Marineris, includ-
Ž
ing the lack of triple junctions a major rifting
.
process on Earth , transfer faults, and shifting major
boundary faults. Like terrestrial rifts, however,
prominent asymmetric wall topography occurs in
Ž.
Valles Marineris Schultz, 1991 . Finally, the major
change from AancientBto ArecentBwall develop-
ment, together with the change from possibly AwetB
Ž
Noachian climatic conditions e.g. Pollack et al.,
.
1987 to the drier Hesperian and Amazonian condi-
tions, are necessarily linked with other processes of
major importance for the history of Mars.
The detailed study of a particular graben, Ius
Chasma, reveals a long-lasting history and a complex
geometry, with several imbricated grabens and horsts.
Previous oblique structural trends, possibly Noachian,
are included in this geometry.
The northern slopes of some tectonically con-
Ž
trolled parts of Valles Marineris Candor, Ophir, and
.
Hebes Chasmata underwent intense faulting during
the Hesperian and Amazonian. In contrast, the south-
ern slopes mostly have a poorly constrained Aan-
cientBtectonic history, and the ArecentBhistory is
only characterized by rather weak erosion. The
southern wall of Coprates Chasma and part of Hebes
Chasma were most deformed, whereas the southern
wall of Candor Chasma was less deformed. Similar
Ž.
to the conclusions of Schultz 1991 for Coprates
Chasma, the morphotectonic study of the central
Valles Marineris and Ius Chasma indicates that the
former geometry of Valles Marineris may have been
that of an asymmetric parallel graben set. This asym-
metric rift evolved with varying active geomorpho-
logical processes, possibly related to a thermal
anomaly centered under the Melas and central Can-
Ž.
dor Chasmata Peulvast and Masson, 1993b .

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 349
Preliminary calculations suggest that, assuming
thin floor deposits near the walls, the maximum
ArecentBextensional movements are about 5% to
15% of the wall relief, leaving 95% to 98% of the
relief separated by tectonic movement. Although
these values may not be representative of the whole
trough system, the lowest values generally corre-
spond to the southern troughs, and the highest, to the
northern ones.
Ius and Coprates Chasmata, the simplest and
most-aligned grabens, appear to have recorded
slightly different deformational histories than the
above-mentioned troughs. The southern walls clearly
underwent AancientBand ArecentBnormal faulting,
and asymmetry is less apparent than in the other
troughs. The western part of Tithonium Chasma also
displays graben-like structures, but the evolution of
its eastern part was dominated by geomorphological
processes similar to those responsible for the scal-
Ž.
loped troughs Tanaka and Golombek, 1989 .
The distribution of mobile ejecta in relation to the
erosional landforms of wall retreat or dissection
strongly suggests that water contributed to the
widening of Central Valles Marineris troughs by
reducing the shear strength of the wall rock. This
implies localized enrichment of ice at depth, or the
Ž
presence of groundwater below the permafrost Bat-
.
tistini, 1985 . Water saturating the pore space of the
rock may have produced regolith mass movements at
Ž
the base of some high walls for example, in western
.
Ophir Chasma , though the porosity is reduced at
Ž
depth by the effect of lithostatic stresses McKinnon
.
and Tanaka, 1989 . Groundwater may also have
induced sapping along discontinuities in more coher-
Ž.
ent rocks Louros Valles . Data on EM suggest that
the central chasmata area may be relatively more
porous, and formerly ice-rich than in peripheral ar-
eas, though the existence of very high walls implies
a cohesion of the wall material that would derive
Ž
from limited amounts of interstitial ice Spencer and
.Ž.
Croft, 1986; Croft, 1989 . Lucchitta et al. 1992
consider the wall material to be either porous re-
Ž.
golith impact breccia or volcanic materials such as
Ž.
tuffs with volatiles trapped in their pores , probably
capped with more coherent rocks less than 1-km
thick. These coherent rocks, which may partly corre-
spond to a Noachian basement, seem to be thicker in
the western and eastern parts of the trough system.
The regional concentration of volatiles appearing
at central Valles Marineris may result from accumu-
lation of underground water from neighboring areas
or underlying magma in porous rocks or regoliths
Ž.
Battistini, 1985 . The widening of the chasmata by
sapping and poorly characterized mass movements,
which was largely completed before the occurrence
of most visible landslides, may have involved release
Ž
of water from a confined aquifer Carr, 1979; Hig-
.
gins, 1982; Kochel and Piper, 1986 . This process
would also be genetically and chronologically con-
sistent with a lacustrine origin for the layered de-
posits that were emplaced in central Valles Marineris
Ž.
during the same period Nedell et al., 1987 . Despite
the probable dessication of the wall materials that
followed these events, the effects of an active
groundwater system on the reduction of shear strength
may be one of the causes of the later stage of
landsliding that occurred during middle- to late-
Ž.
Amazonian Lucchitta et al., 1992 . This study indi-
cates, however, that a late tectonic stage, accompa-
nied by wall heightening and oversteepening and by
seismic activity, probably triggered these mass
movements, and this stage was not followed by any
erosional process which could be compared with the
gullying of the previous stage.
Multi-stage faulting along wide grabens is the
Ž
primary cause of the formation of high walls Peul-
.
vast and Masson, 1993b , though the subsidence of
the basins that prefigured the troughs during the first
stage was locally accompanied by the emplacement
of thick sedimentary and volcanic deposits. During
the AancientBstage, tectonic and probably volcanic
activity was associated with a great efficiency of
erosion, especially in central Valles Marineris, which
resulted in wide re-entrants, and an unknown amount
of wall retreat from fault lines or other down-warp-
ing structures in Melas Chasma. Gullying and even
Ž.
dissected landforms Louros Valles probably in-
volved the action of water or ice released by wall
rocks, but no large-scale mass movements devel-
oped. The end of this stage seems to be related to the
end of active subsidence of the basin floors.
The last stages of faulting probably occurred when
the wall materials were desiccated. This ArecentB
tectonic stage, which resulted in interconnection of
basins along the main or annex grabens and in
deepening to 7 or 9 km in the main troughs, seems to

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352350
be related to regional up-doming and renewed mag-
Ž.
matic activity of Tharsis Geissler et al., 1990 .
Spectacular erosional processes, mainly landslides
and talus formation, occurred, but the eroded vol-
umes remained moderate, and the localized distribu-
tion of the corresponding landforms allowed the
preservation of older wall topographies, even along
segments of renewed tectonic activity. This second
stage also gave rise to a local asymmetry, with mass
movements mostly related to the reactivated northern
boundary faults.
Acknowledgements
This work was supported by the Programme Na-
tional de Planetologie from Institut National des
´Ž.
Sciences de l’ Univers CNRS, France . The data
used in this study were made available by the Re-
gional Planetary Image Facility at University of Paris
Ž.
XI Orsay, France under the auspices of the NASA
Planetary Geology and Geophysics Program. The
figures were redrawn by Mrs. Genevieve Roche
´
ŽUnite C.N.R.S. de Geophysique et Geodynamique
´´´
.
Interne, Universite Paris-Sud .
´
References
Baker, V.R., Strom, R.G., 1992. Ancient Martian oceans. In:
Zuber, M.T., James, O.B., Lunine, J.I., McPherson, G.J.,
Ž.
Phillips, R.J. Eds. , Planetary Geosciences 1989–1990. NASA
SP, vol. 508, pp. 1–3.
Banerdt, W.B., Golombek, M.P., 1992. Constraints on the struc-
ture and evolution of the Tharsis region of Mars. In: Zuber,
M.T., James, O.B., Lunine, J.I., McPherson, G.J., Phillips,
Ž.
R.J. Eds. , Planetary Geosciences 1989–1990. NASA SP, vol.
508, pp. 4–6.
Banerdt, W.B., Phillips, R.J., Sleep, N.H., Saunders, R.S., 1982.
Thick shell tectonics on one-plate planets: application to Mars.
J. Geophys. Res., 87, 9723–9733.
Banerdt, W.B., Golombek, M.P., Tanaka, K.L., 1992. Stress and
tectonics on Mars. In: Kiefer, H.H., Jakowsky, B.M., Snyder,
Ž.
C.W., Matthews, M.S. Eds. , Mars. Univ. of Arizona Press,
Tucson, pp. 249–297.
Baskerville, C.A., 1982. Collapse: a mechanism for Martian scarp
retreat. NASA Tech. Memo., 85127, 244–252.
Battistini, R., 1984. Morphology and origin of ridges in low-lati-
tude areas of Mars. Earth, Moon, Planets, 31, 49–61.
Battistini, R., 1985. Hydrolithosphere and problems of subsurface
ice in the equatorial zone of Mars. In: Klinger, J., Bebest, D.,
Ž.
Dollfus, A., Smoluchowski, R. Eds. , Ices in the Solar Sys-
tem. Reidel, Boston, pp. 607–617.
Blasius, K.R., Cutts, J.A., Guest, J.E., Masursky, H., 1977. Geol-
ogy of the Valles Marineris: first analysis of imaging from the
Viking I Orbiter Primary Mission. J. Geophys. Res., 82,
4067–4091.
Bousquet, B., Jourdain-Bodart, J., Touzeau, J.F., 1987. Relief et
modele des versants de Valles Marineris: leur signification
´
pour Mars. Bull. Soc. Geol. Fr., 3, 43–48.
Brunsden, D., 1979. Mass movements. In: Embleton, C., Phornes,
Ž.
J. Eds. , Processes in Geomorphology. Arnold, London, pp.
131–186.
Carr, M.H., 1979. Formation of Martian flood features by release
of water from confined aquifer. J. Geophys. Res., 84, 2995–
3007.
Carr, M.H., 1981. The Surface of Mars. Yale Univ. Press, New
Haven, 232 pp.
Carr, M.H., Crumpler, L.S., Cutts, J.A., Greeley, R., Guest, J.E.,
Mazursky, H., 1977. Martian impact craters and emplacement
of ejecta by surface flow. J. Geophys. Res., 82, 4055–4065.
Chicarro, A.F., Schultz, P.H., Masson, P.H., 1985. Global and
regional ridge patterns on Mars. Icarus, 63, 153–174.
Cipa, A., 1994. Le volcan Olympus Mons et ses environs, Mars.
These de doctorat, Universite Paris Sud, Orsay, 243 pp.
´´
Clifford, S.M., 1993. A model for the hydrologic and climatic
behavior of water on Mars. J. Geophys. Res., 98, 10973–
11016.
Costard, F., 1990. Distribution et caracteristique du pergelisol
´´
martien: Son influence sur certains traits de la geomorpholo-
´
gie. These de doctorat, Universite Paris IV, 327 pp.
´´
Croft, S.K., 1989. Mars canyon systems: problems of origin.
Abstr. 4th Int. Conf. Mars, Tucson, pp. 88–89.
Embleton, C., Whalley, B., 1979. Energy forces resistance and
Ž.
responses. In: Embleton, C., Phornes, J. Eds. , Processes in
Geomorphology. Arnold, London, pp. 11–38.
Fanale, F.R., 1976. Martian volatiles: the degassing history and
geochemical fate. Icarus, 28, 179–202.
Flageollet, J.C., 1988. Les Mouvements de Terrain et Leur
Prevention. Masson, Paris, 224 pp.
´
Fort, M., 1993. Etude d’une chaıne de collision intracontinentale
ˆ
Ž.
Himalaya du Nepal, transversale des Annapurnas . These de
´´
doctorat d’Etat, Universite Paris VII, 702 pp.
´
Fort, M., Peulvast, J.P., 1995. Catastrophic mass-movements and
morphogenesis in the peri-Tibetan ranges: examples from
Ž.
West Kunlan, Pamir, and Ladakh. In: Slaymaker, O. Ed. ,
Steepland Geomorphology. Wiley, Chichester, pp. 171–198.
Frey, H., 1979. Martian canyons and African rifts: structural
comparisons and implications. Icarus, 37, 142–155.
Gault, D.E., Greeley, R., 1978. Exploratory experiments of impact
craters formed in viscous-liquid targets, analog for rampart
craters. Icarus, 34, 486–495.
Geissler, P.E., Singer, R.B., Lucchitta, B.K., 1990. Dark materials
in Valles Marineris: indications of the style of volcanism and
magmatism on Mars. J. Geophys. Res., 95, 14399–14413.
Hamblin, W.K., 1976. Patterns of displacement along the Wasatch
fault. Geology, 4, 619–622.

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352 351
Hartmann, W.K., 1973. Martian surface and crust: review and
synthesis. Icarus, 19, 550–575.
Higgins, C.G., 1982. Drainage systems developed by sapping on
Earth and Mars. Geology, 10, 147–152.
Horner, V.M., Greeley, R., 1987. Effects of elevation and ridged
plains thickness on Martian crater ejecta morphology. J. Geo-
phys. Res., 92, E561–E569.
Jibson, R.W., Keefer, D.K., 1993. Analysis of the seismic origin
of landslides: examples from the New Madrid seismic zone.
Geol. Soc. Am. Bull., 105, 521–536.
Kiefer, W.S., Hager, B.H., 1989. The role of mantle convection in
the origin of the Tharsis and Elysium provinces of Mars. In:
Ž.
Frey, H.V. Ed. , MEVTV Workshop on Early Tectonic Fea-
tures on Mars. LPI Technical Rept. 89-04. Lunar and Plane-
tary Institute, Houston, TX, pp. 48–50.
Kochel, R.C., Piper, J., 1986. Morphology of large valleys on
Hawaii: evidence for ground water sapping and comparisons
with Martian valleys. J. Geophys. Res., 91, E171–E192.
Kochel, R.C., Baker, V.R., Simmons, D.W., Lis, C.J., 1982.
Surface channel networks developed by groundwater sapping
in a laboratory model: analogy to sapping channels on Mars.
Prog., NASA Tech. Memo., 85127, 227–229.
Kuzmin, R.O., Bobina, N.N., Zabalueva, E.V., Shashkina, V.P.,
1988. Mars: estimation of the relative ice content in upper
layers of the permafrost. Lunar Planet. Sci., vol. XIX, Lunar
and Planetary Institute, Houston, TX, pp. 657–658.
Lucchitta, B.K., 1978. Morphology of chasma walls, Mars. J. Res.
U.S. Geol. Surv., 6, 651–662.
Lucchitta, B.K., 1979. Landslides in Valles Marineris, Mars. J.
Geophys. Res., 84, 8097–8113.
Lucchitta, B.K., 1984. A late climatic change on Mars. Lunar
Planet. Sci., vol. XV, Lunar and Planetary Institute, Houston,
TX, pp. 493–494.
Lucchitta, B.K., 1987. Valles Marineris: wet debris flows and
ground ice. Icarus, 72, 411–429.
Lucchitta, B.K., Bertolini, L.M., 1989. Mars: interior structure of
Valles Marineris, Mars. Lunar Planet. Sci., vol. XX, Lunar
and Planetary Institute, Houston, TX, pp. 590–591.
Lucchitta, B.K., Balser, R.A., Bertolini, L.M., 1990. Valles
Marineris, Mars: are pit chains formed by erosion and troughs
by tectonism? Lunar Planet. Sci., vol. XXI, Lunar and Plane-
tary Institute, Houston, TX, pp. 722–723.
Lucchitta, B.K., McEwen, A.S., Clow, G.D., Geissler, P.E., Singer,
R.B., Schultz, R.A., Squyres, S.W., 1992. The canyon system
on Mars. In: Kiefer, H.H., Jakosky, B.M., Snyder, C.W.,
Ž.
Matthews, M.S. Eds. , Mars. Univ. of Arizona Press, Tucson,
pp. 453–492.
Masson, Ph.L., 1977. Structure pattern analysis of the Noctis
Labyrinthus, Valles Marineris region of Mars. Icarus, 30,
49–62.
Masson, Ph.L., 1985. Origin and evolution of the Valles Marineris
region of Mars. Adv. Space Res., 5, 83–92.
McCauley, J.F., 1978. Geologic map of the Coprates Quadrangle
of Mars, scale 1:5,000,000. U.S. Geol. Survey Misc. Inv.
Series Map I-897.
McEwen, A.S., 1989. Mobility of large rock avalanches: evidence
from Valles Marineris, Mars. Geology, 17, 1111–1114.
McKinnon, D.J., Tanaka, K.L., 1989. The impacted Martian crust:
structure, hydrology, and some geologic implications. J. Geo-
phys. Res., 94, 17359–17370.
Mege, D., Masson, Ph.L., 1996. Stress models for Tharsis forma-
`Ž.
tion. Mars. Planet. Space Sci., 44, 12 , 1471–1497.
Mouginis-mark, P.J., 1979. Martian fluidized crater morphology:
variations with crater size, latitude, altitude, and target mate-
rial. J. Geophys. Res., 84, 8011–8022.
Nedell, S.S., Squyres, S.W., Anderson, D.W., 1987. Origin and
evolution of the layered deposits in the Valles Marineris,
Mars. Icarus, 70, 409–441.
Parker, T.J., Gorsline, D.S., Saunders, R.S., Pieri, D.C., Schnee-
berger, D.M., 1993. Coastal geomorphology of the Martian
northern plains. J. Geophys. Res., 98, 11061–11078.
Peterson, C., 1981. A secondary origin for the central plateau of
Hebes Chasma. Proc. 12th Lunar Planet Sci. Conf. Geochim.
Cosmochim. Acta, pp. 1459–1471.
Peulvast, J.P., Masson, Ph.L., 1993a. Melas Chasma: morphology
Ž.
and tectonic patterns in Central Valles Marineris Mars . Earth,
Moon, Planets, 61, 219–248.
Peulvast, J.P., Masson, Ph.L., 1993b. Erosion and tectonics in
Ž.
Central Valles Marineris Mars : a new morphostructural
model. Earth, Moon, Planets, 61, 191–217.
Pollack, J.B., Kastings, J.F., Richardson, S.M., Poliakoff, K.,
1987. The case for a wet, warm climate on early Mars. Icarus,
71, 203–224.
Schubert, G., Bercovici, D., Glatzmaier, G.A., 1990. Mantle
dynamics in Mars and Venus: influence on an immobile
lithosphere on three-dimensional mantle convection. J. Geo-
phys. Res., 95, 14105–14129.
Schultz, R.A., 1989. Do pit-crater chain grow up to be Valles
Marineris canyons? NASA Tech. Memo., 4130, 539–540.
Schultz, R.A., 1991. Structural development of Coprates Chasma
and western Ophir Planum, Valles Marineris Rift, Mars. J.
Geophys. Res., 96, 22777–22792.
Scott, D.H., Tanaka, K.L., 1986. Geologic map of the western
equatorial region of Mars, scale 1:15,000,000. U.S. Geological
Survey Misc. Inv. Series Map I-1802-A.
Sharp, R.P., 1973. Mars: troughed terrain. J. Geophys. Res., 78,
4063–4072.
Sleep, N.H., Phillips, R.J., 1979. An isostatic model for the
Tharsis province, Mars. Geophys. Res. Lett., 6, 803–806.
Sleep, N.H., Phillips, R.J., 1985. Gravity and lithospheric stress
on the terrestrial planets with reference to the Tharsis region
of Mars. J. Geophys. Res., 90, 4469–4489.
Solomon, S.C., Head, J.W., 1982. Evolution of the Tharsis
province of Mars: the importance of heterogeneous litho-
spheric thickness and volcanic construction. Proc. 3rd Int.
Colloquium on Mars. J. Geophys. Res., vol. 87, pp. 9755–
9774.
Spencer, J.R., Croft, S.K., 1986. Valles Marineris as karst: feasi-
bility and implications for Martian atmospheric evolution.
NASA Tech. Memo., 88383, 193–195.
Squyres, S.W., Clifford, S.M., Kuzmin, R.O., Zimbelman, J.R.,
Costard, F.M., 1992. Ice in the Martian regolith. In: Kiefer,
Ž.
H.H., Jakosky, B.M., Snyder, C.W., Matthews, M.S. Eds. ,
Mars. Univ. of Arizona Press, Tucson, pp. 523–554.

()
J.-P. PeulÕast et al.rGeomorphology 37 2001 329–352352
Tanaka, K.L., Golombek, M.P., 1989. Martian tension fractures
and the formation of grabens and collapse features at Valles
Marineris. Proc. 19th Lunar Planet. Sci. Conf., pp. 383–396.
Tanaka, K.L., Golombek, M.P., Banerdt, W.B., 1991. Reconcilia-
tion of stress and structural histories of the Tharsis region of
Mars. J. Geophys. Res., 96, 15617–15633.
Tanaka, K.L., Scott, D.H., Greeley, R., 1992. Global stratigraphy.
In: Kiefer, H.H., Jakosky, B.M., Snyder, C.W., Matthews,
Ž.
M.S. Eds. , Mars. Univ. of Arizona Press, Tucson, pp. 345–
382.
U.S. Geological Survey, 1986. Topographic orthophoto mosaic of
the Tithonium Chasma region of Mars. U.S. Geol. Survey
Misc. Inv. Series Map I-1294.
U.S. Geological Survey, 1986. Topographic map of the Coprates
Northwest quadrangle of Mars. U.S. Geol. Survey Misc. Inv.
Series Map I-1712.
Wallace, R.E., 1978. Geometry and rates of change of fault-gener-
ated range fronts, North-Central Nevada and some tectonic
implications. J. Res. U.S. Geol. Surv., 6, 637–650.
Watters, T.R., 1993. Compressional tectonism on Mars. J. Geo-
phys. Res., 98, 17049–17060.
Willeman, R.J., Turcotte, D.L., 1982. The role of lithospheric
stress in the support of the Tharsis rise. Proc. 3rd Int. Collo-
quium on Mars. J. Geophys. Res., vol. 87, pp. 9793–9801.
Witbeck, N.E., Tanaka, K.L., Scott, D.H., 1991. Geologic map of
the Valles Marineris region, Mars, scale 1:2,000,000. U.S.
Geol. Survey Misc. Inv. Series Map I-2010.
Wohletz, K.M., Sheridan, M.F., 1983. Martian rampart crater
ejecta: experiments and analysis of melt–water interaction.
Icarus, 56, 15–37.