Structure, stratigraphy, and origin of Husband Hill,
Columbia Hills, Gusev Crater, Mars
T. J. McCoy,1 M. Sims,2 M. E. Schmidt,1 L. Edwards,2 L. L. Tornabene,3 L. S. Crumpler,4
B. A. Cohen,5 L. A. Soderblom,6 D. L. Blaney,7 S. W. Squyres,8 R. E. Arvidson,9
J. W. Rice Jr.,10 E. Tre?guier,11 C. d?Uston,11 J. A. Grant,12 H. Y. McSween Jr.,13
M. P. Golombek,7 A. F. C. Haldemann,14 and P. A. de Souza Jr.15,16
Received 15 November 2007; revised 26 February 2008; accepted 7 April 2008; published 17 June 2008.
[1] The strike and dip of lithologic units imaged in stereo by the Spirit rover in the
Columbia Hills using three-dimensional imaging software shows that measured dips (15?
32
) for bedding on the main edifice of the Columbia Hill are steeper than local
topography (
8?10
). Outcrops measured on West Spur are conformable in strike with
shallower dips (7?15
) than observed on Husband Hill. Dips are consistent with
observed strata draping the Columbia Hills. Initial uplift was likely related either to the
formation of the Gusev Crater central peak or ring or through mutual interference of
overlapping crater rims. Uplift was followed by subsequent draping by a series of
impact and volcaniclastic materials that experienced temporally and spatially variable
aqueous infiltration, cementation, and alteration episodically during or after deposition.
West Spur likely represents a spatially isolated depositional event. Erosion by a variety of
processes, including mass wasting, removed tens of meters of materials and formed the
Tennessee Valley primarily after deposition. This was followed by eruption of the
Adirondack-class plains basalt lava flows which embayed the Columbia Hills. Minor
erosion, impact, and aeolian processes have subsequently modified the Columbia Hills.
Citation: McCoy, T. J., et al. (2008), Structure, stratigraphy, and origin of Husband Hill, Columbia Hills, Gusev Crater, Mars,
J. Geophys. Res., 113, E06S03, doi:10.1029/2007JE003041.
1. Introduction
[2] For more than 1450 Martian days, or ??sols,?? the Mars
Exploration Rover Spirit has explored the interior of the
Gusev Crater, Mars (Figure 1) [Crisp et al., 2003; Squyres
et al., 2003, 2004, 2006, 2007; Arvidson et al., 2006]. After
Spirit spent the first 155 sols exploring the basaltic cratered
plains on which it landed, the majority of its time has been
spent exploring the Columbia Hills, including a traverse
over Husband Hill and into the Inner Basin, a valley in the
center of the Hills. On sol 780, Spirit?s right front wheel
failed, preventing Spirit from climbing McCool Hill or any
other of the Columbia Hills constructs. As a result, this is an
opportune time to both revisit and summarize our knowl-
edge of the structure and stratigraphy of the Columbia Hills
and to discuss their implications for the origin of these hills.
[3] Studies of individual rocks on Husband Hill have
revealed a rich record of volcanic, impact and aqueous
processes in the formation and/or modification of these
rocks [Squyres et al., 2006]. The instruments that Spirit
and its twin Opportunity use to investigate these rocks (the
Athena payload) were designed to examine the geology,
geochemistry, mineralogy, and physical properties of the
rocks and soils [Squyres et al., 2003]. The Athena payload
includes the Panoramic camera (Pancam) [Bell et al., 2003],
three engineering cameras [Maki et al., 2003], and the
Miniature Thermal Emission Spectrometer (Mini-TES)
[Christensen et al., 2003]. Other instruments are on the
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, E06S03, doi:10.1029/2007JE003041, 2008
Click
Here
for
Full
Article
1Department of Mineral Sciences, National Museum of Natural History,
Smithsonian Institution, Washington, D. C., USA.
2Intelligent Systems Division, NASA Ames, Moffett Field, California,
USA.
3Lunar and Planetary Laboratory, University of Arizona, Tucson,
Arizona, USA.
4New Mexico Museum of Natural History and Science, Albuquerque,
New Mexico, USA.
5Institute of Meteoritics, University of New Mexico, Albuquerque, New
Mexico, USA.
6U.S. Geological Survey, Flagstaff, Arizona, USA.
7Jet Propulsion Laboratory, California Institute of Technology,
Pasadena, California, USA.
Copyright 2008 by the American Geophysical Union.
0148-0227/08/2007JE003041$09.00
E06S03
8Department of Astronomy, Cornell University, Ithaca, New York, USA.
9Department of Earth and Planetary Sciences, Washington University,
St. Louis, Missouri, USA.
10Department of Geological Sciences, Arizona State University, Tempe,
Arizona, USA.
11CESR/OMP, Toulouse, France.
12Center for Earth and Planetary Studies, National Air and Space
Museum, Smithsonian Institution, Washington, D. C., USA.
13Department of Earth and Planetary Sciences, University of Tennessee,
Knoxville, Tennessee, USA.
14ESA/ESTEC HME-ME, Noordwijk, Netherlands.
15Vallourec Research Center, Aulnoye-Aymeries, France.
16Now at Tasmanian ICT Centre, Hobart, Tasmania, Australia.
1 of 14
rover?s arm, or instrument deployment device, including the
Alpha Particle X-ray Spectrometer (APXS) [Rieder et al.,
2003], which analyzes major and some minor element
concentrations in rocks and soils; the Mo?ssbauer Spec-
trometer [Klingelhofer et al., 2003], which analyzes for
iron-bearing mineralogy; the Microscopic Imager camera
[Herkenhoff et al., 2003]; and the Rock Abrasion Tool
[Gorevan et al., 2003].
[4] It has proven difficult to link the petrologic properties
of individual rocks and outcrops along the traverse of
Husband Hill to understand the formation of the Columbia
Hills. In addition to a restricted data set along a narrow
traverse, information on the structure and stratigraphy (key
elements to understanding the petrogenesis of geologic
landforms on Earth) are more difficult to infer for Mars.
Structural measurements that are common on Earth are
difficult on Mars by conventional methods (i.e., with a
compass) because they require a magnetic field and com-
plex rover movements. Finally, this is, to our knowledge,
the first effort at in situ structural geology on Mars. The
combination of the limited data set and experience leads to
an element of uncertainty, recently termed ??conceptual
uncertainty?? [Bond et al., 2007]. In this paper, we first
describe the geologic setting and the lithologies that make
up Husband Hill. We then utilized three-dimensional visual-
ization models of outrcrops created from stereo imaging pairs
to determine the strikes and dips along the Husband Hill
traverse. Finally, we discuss the likely origin of Husband Hill
and the Columbia Hills as uplifted by impact processes and
later draped by volcanic and impact-derived debris. Sub-
sequent modification occurred by erosion, mass wasting,
aeolian reworking, and alteration processes.
2. Geologic Setting
[5] Gusev Crater (Figure 1) is a 
160 km diameter crater
close to the highland-lowland boundary south of Elysium
[Golombek et al., 2003; Cabrol et al., 2003]. The flat floor
of the crater is thought to be significantly younger results
from later infilling. The floor consists of a diversity of both
thermophysical and morphologic units [Milam et al., 2003].
Kuzmin et al. [2000] mapped Gusev Crater as Noachian
with Thyra crater as Late Noachian to Early Hesperian.
Crater floor material becomes progressively younger from
east to west, with Early to Late Hesperian material flooring
the eastern side of the crater, Late Hesperian in the central
portion and Early Amazonian in the western portion. The
Late Hesperian central portion of the crater corresponds to a
viscous flow that emanates from Ma?adim Vallis, an 800 km
long channel that drains the southern highlands [Golombek
et al., 2003]. The Columbia Hills are mapped as an outcrop
of the Early to Late Hesperian (AHbm2) unit surrounded by
younger Late Hesperian (AHgf1) material. A digital elevation
model (DEM) (Figure 2) reveals a few circular impact
craters of 5?20 on the floor of Gusev that were later buried
by Hesperian lava flows. The 
21 km diameter Thyra crater
is the youngest and least buried of these craters. Greeley et
al. [2005] argued that the density of impact craters on the
floor of Gusev points to an ancient age for the crater floor
(
3.65 Ga), which would make the Columbia Hills even
older and thus likely at least Late Noachian. Many of the
features are orthogonal and, given the complex infilling
history of Gusev, may not be impact craters and their origin
is highly uncertain. The position of Gusev Crater at the
northern terminus of Ma?adim Vallis, suggests that it may
have once served as a lacustrine depocenter for that drainage
system. Spirit found that the cratered plains she traversed
that surround the Columbia Hills are volcanic basalt flows
[e.g., Golombek et al., 2006a] and she has found no
evidence of aqueous activity associated with Ma?adim
drainage [e.g., Squyres et al., 2004]. Extending from the
mouth of Ma?adim Vallis is a broad N-S trough that contains
the Columbia Hills. About 250 km north of Gusev Crater is
the large Hesperian volcano Apollinaris Patera, which might
be the source for crater-filling volcanics [Martinez-Alonso et
al., 2005], or the basalts erupted from fissures now beneath
the flows. Features in and around Gusev Crater suggest a
complex history that may have extended from the late
Noachian through the Hesperian-Amazonian and likely
included lacustrine, fluvial, glacial, volcanic, impact, aeolian
and subsurface hydrothermal processes.
[6] In map view, the Columbia Hills (Figure 3) are
roughly the shape of 
8.4 by 
4.5 km north pointing
isosceles triangle that has sharp margins with the surround-
ing basaltic plains. The north trending margins of the hills
are linear to slightly concave outward, while the southern
margin consists of two equal-length, concave-outward seg-
ments. First viewed from the Spirit landing site, the Co-
lumbia Hills consist of seven distinct peaks informally
named the Columbia Hills in honor of the space shuttle
Columbia astronauts. From orbit, the Columbia Hills are a
complex raised landform consisting of multiple peaks,
ridges, and depressions. The centralmost low area has been
informally called the Inner Basin. Orbital studies, supple-
mented by new images from High Resolution Imaging
Science Experiment (HiRISE), provide our best overview
of the large-scale structure of the Columbia Hills and remain
key data to understanding their structure and origin.
[7] Husband Hill, one of the highest summits in the
range, rises to an elevation of about 90 m above the plains
(more than 110 m above the landing site) and is located at
175.54
E, 14.59
S [Squyres and the Athena Science Team,
2005; Squyres et al., 2006]. Among the more prominent
features of Husband Hill is an amphitheater-shaped valley
that bisects it northeast edge that has been named the
Tennessee Valley (Figure 4). The West Spur abuts from
the main edifice of Husband Hill to the northwest and rises
to 
45 m above the surrounding plains. Basaltic plains of
Adirondack lithology wrap around the northern edge of the
West Spur and embay the area between West Spur and the
northern edge of Husband Hill. Figure 4 is a HiRISE image
of Husband Hill overlain on a DEM produced by the U.S.
Geological Survey from a stereo pair of HiRISE images
pointing out the locations of the major features.
3. Lithologies and Distribution
[8] Here we briefly review the character and origin of the
rocks that compose the Columbia Hills [Squyres et al.,
2006]. Rock properties and petrogenesis have been derived
from a combination of spectral, geochemical, mineralogic
and physical measurements performed with the Athena
science payload aboard the Spirit rover. Rock classes that
connect often texturally diverse rocks are, by convention,
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
2 of 14
E06S03
determined on the basis of APXS analyses. Rock distribu-
tions, which are a key component for deciphering the extent
and relationship of lithologic units, are derived largely from
Mini-TES observations [Blaney et al., 2005; Ruff et al.,
2006]. Note that place names used in this paper for land-
forms, rocks, and soils encountered during Spirit?s mission
have not been approved by the International Astronomical
Union and are meant to be informal and convenient ways to
remember features (e.g., an outcrop).
[9] The West Spur is dominated by Clovis class rocks,
which occur as both outcrop and float. This rock class has
not been observed elsewhere on Husband Hill [Squyres et
al., 2006]. Petrogenetically, the Clovis class rocks are
poorly sorted clastic rocks of basaltic bulk composition.
These rocks respond to grinding similar to that of weak
sedimentary rocks, suggesting that that they are pervasively
altered and susceptible to erosion, and some show cavernous
weathering. They are compositionally similar to plains
(Adirondack) basalts for some elements (Si, Ti, Al and
Fe), although significantly more variable and with elevated
Ni, S, Cl and Br. Formation as either volcanogenic or
impact deposits have been postulated [Squyres et al.,
2006; Ming et al., 2006]. Their elevated Ni concentration
compared to Adirondack basalts as well as the dominant
basaltic glass component determined from Mini-TES [Ruff
et al., 2006] favors an origin as impact into ancient crustal
material. Indeed, the Ni concentrations are comparable to
that of soils, which have a small but significant Ni enrich-
ment [Yen et al., 2005]. Enrichments in S, Cl, and Br, as
well as the oxidized nature of Clovis Class (high Fe3+/FeTot)
point to later alteration, likely by aqueous fluids.
[10] The main edifice of Husband Hill is much more
diverse petrologically. The northwest flank is dominated by
Wishstone class rocks [Blaney et al., 2005; Ruff et al.,
2006]. Although dominant as float in this area, no outcrops
of Wishstone class were observed. These rocks are weakly
altered clastic rocks that are rich in plagioclase with lesser
mafic silicates and minor iron oxides and oxyhyrdroxides
Figure 1. A normalized THEMIS daytime thermal infrared brightness temperature mosaic draped on
the Mars Orbiter Laser Altimeter digital elevation model with a 5X vertical exaggeration. Black arrow
points to the Columbia Hills, which rises above the surrounding basaltic cratered plains near the center of
Gusev Crater. The inset at the bottom right is approximately 
3X and has a width of 
8 km. This inset
has been linearly stretched and sharpened to bring out the thermally distinct Columbia Hills.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
3 of 14
E06S03
[Squyres et al., 2006]. A distinctive feature of Wishstone
class rocks is their high P and Ti and low Cr concentrations.
These rocks are composed of poorly sorted, sometimes
highly angular clasts of 1?2 mm in size in a fine-grained
matrix. No layering is observed in any Wishstone class
rocks. The textural similarity to ash flow tuffs, coupled with
low Ni concentrations, suggests that these rocks probably
formed by pyroclastic volcanism. The occurrence of an
unconsolidated in situ sulfate-rich soil, the Paso Robles
class, within the region of Wishstone float rocks is note-
worthy. The mineralogy, variability, association with local
compositions (particularly P enrichment), and geologic
setting of the deposits suggest that Paso Robles class soils
likely formed as hydrothermal and fumarolic condensates
derived from magma degassing and/or oxidative alteration
of crustal iron sulfide deposits [Yen et al., 2008].
[11] Within the region dominated by Wishstone class
float rocks are two in situ outcrops of Peace Class rocks.
These rocks mineralogically are similar to a lightly altered
lherzolite with the majority of the iron in olivine and
pyroxene. Compositionally, Peace class rocks exhibit strik-
ingly low Al, Na, and K and high SO3 (up to 12.9 wt %).
Peace class outcrops exhibit fine-scale layering, with layers
of a few millimeters in thickness. Although clasts up to a few
millimeters in size are observed, the rock appears dominated
by grains of a few hundred micrometers in size. These are the
weakest rocks encountered by Spirit, explaining the lack of
prominent outcrops. Correlations between S, Ca and Mg
suggest the presence of 16?17%Mg and Ca sulfates, perhaps
as a hydrated phase [Ruff et al., 2006]. These rocks, despite
the evidence for a cementing sulfate agent, exhibit relatively
little alteration of the mafic component, with no evidence of
hematite, goethite or Fe sulfates. The most plausible origin
is cementation of modestly altered ultramafic, magnetite-
bearing sands by Mg and Ca sulfates.
[12] On the northwest and southeast sides of the Tennessee
Valley, Watchtower class rocks are observed as rugged and
prominent outcrops. Watchtower class rocks are less resis-
tant to grinding than Adirondack class basalts, but substan-
tially more durable than Peace and Clovis class rocks.
Watchtower Class rocks are similar in Al, Ti, Ca, Na, and
P to Wishstone Class but enriched in Mg, Zn, S, Br, and Cl.
This enrichment likely resulted by interaction between a
Wishstone-like protolith and a fluid phase enriched in the
latter elements [Hurowitz et al., 2006]. The extent of
alteration is highly variable [Ruff et al., 2006] over length
scales of meters to tens of meters laterally and meters
stratigraphically [Squyres et al., 2006].
[13] At the head of the Tennessee Valley amphitheater
Independence class (including the Assemblee subclass) and
Descartes class rocks are found. These rocks are clastic and
dominated by pyroxene and nanophase oxides. Indepen-
dence class rocks are exceptionally low in iron, high in SiO2
(
53%), exhibit a high Al/Si ratio, and a high Ni concen-
tration. Clark et al. [2007] argue that extensive aqueous
alteration has produced the chemical equivalent of mont-
morillonite in these rocks. Descartes class rocks are only
modestly depleted in Fe compared to plains basalts. Tex-
turally, Descartes class rocks are breccias with angular clasts
that can reach a centimeter in size. Both classes share a
common signature of Ti and P enrichments with Cr deple-
tions, a signature which Clark et al. [2007] use to link these
rocks in a superclass that includes Wishstone and Watch-
tower. The presence of goethite and hematite in Descartes
class rocks, coupled with a moderate Fe3+/FeTot point to
extensive alteration. The origin of these rocks is uncertain,
but aqueous alteration of either a pyroclastic volcanic or
impact breccia appears likely. Clark et al. [2007] suggest
that the protolith may resemble Wishstone class rocks.
[14] Rimming the Tennessee Valley, unaltered to modestly
altered, massive, aphanitic, alkaline basalts of the Irvine and
Backstay classes are found as exotics [McSween et al.,
2006a; Ruff et al., 2006], as are Adirondack class basalts
that were prevalent on the plains [McSween et al., 2004,
2006b]. Irvine class basalts also dominate the Inner Basin of
the Columbia Hills to the south of Husband Hill, where they
are often vesicular to scoriaceous [Crumpler et al., 2007].
[15] Spirit descended the southeast flank of Husband Hill
along Haskin Ridge, where Algonquin class rocks domi-
nate. These rocks are clastic, olivine-rich rocks that make up
an apparent olivine fractionation sequence, increasing in
MgO and decreasing CaO and Al2O3 as Spirit descended
the SE flank of Husband Hill [Mittlefehldt et al., 2006]. The
origin of these rocks is likely igneous, perhaps by pyro-
clastic volcanism.
4. Measuring Strikes and Dips
[16] Extensive mass wasting and erosion have softened
the edifice of Husband Hill. As a consequence, outcrops are
rare. Further, most outcrops exhibit no obvious lamination
or bedding and, thus, are unsuitable for measurements of
attitude. We have examined seven locations for which
strikes and dips can be measured. These outcrops include
the Clovis class rocks Tetl and Palenque, the Peace class
Figure 2. Mars Orbiter Laser Altimeter digital elevation
model derived by subtracting the original DEM from a
terrain smoothed on an 8 km scale to reveal small-scale
topographic features on the crater floor. Brightness scale is
+50 m (white) to 50 m (black). A number of circular
features corresponding to craters of 20?50 km diameter are
observed on the crater floor, including Thyra crater. An
arcuate trough extending from the mouth of Ma?adim Vallis
is also observed. The Columbia Hills (white arrow) lie
within this broad trough on the extension of a linear ridge.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
4 of 14
E06S03
F
ig
u
re
3.
(l
ef
t)
H
iR
IS
E
im
ag
e
P
S
P
_0
03
84
_1
65
0
an
d
(r
ig
ht
)
co
lo
r-
co
de
d
D
E
M
m
ap
de
ri
ve
d
fr
om
H
iR
IS
E
st
er
eo
-p
ai
re
d
im
ag
es
P
S
P
_0
01
71
7_
16
50
an
d
P
S
P
_0
01
51
3_
16
55
of
th
e
C
ol
um
bi
a
H
il
ls
,
w
hi
ch
in
m
ap
pl
an
ar
e
th
e
sh
ap
e
of
a


8.
4
by


4.
5
km
no
rt
h
po
in
ti
ng
is
os
ce
le
s
tr
ia
ng
le
.T
he
no
rt
h-
so
ut
h
tr
en
di
ng
m
ar
gi
ns
of
th
e
hi
ll
s
ar
e
li
ne
ar
to
sl
ig
ht
ly
co
nc
av
e,
w
hi
le
th
e
so
ut
he
rn
m
ar
gi
n
co
ns
is
ts
of
tw
o
eq
ua
l-
le
ng
th
,c
on
ca
ve
se
gm
en
ts
.T
he
sh
ap
e
of
th
e
hi
ll
s
co
ul
d
be
in
te
rp
re
te
d
as
th
e
re
su
lt
of
bo
un
di
ng
by
fo
ur
im
pa
ct
cr
at
er
ri
m
s.
H
iR
IS
E
D
E
M
co
ur
te
sy
of
R
.
K
ir
k
an
d
hi
s
U
S
G
S
ba
se
d
te
am
.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
5 of 14
E06S03
outcrop Alligator, the Watchtower class rocks Larry?s
Lookout and Summit (Hillary), the Descartes class outcrop
Voltaire, and the type outcrop for the Algonquin class rocks.
Information about the sols imaged, image sequence numbers,
and derived strike and dip are given in Table 1. Pancam
false color images of outcrops from which strikes and dips
are derived are illustrated in Figure 6. Note that the image
sequences illustrated in Figure 6 are not necessarily those
used to derive the strikes and dips (Table 1) because color
images required multispectral Pancam imaging, while the
strike and dip models require stereo imaging pairs.
[17] Traditionally, the orientation of geologic bedding is
determined via a classic three point problem where three
points in x, y, and z space are used to define a plane. In this
study, a three-dimensional computer visualization model
was created for each outcrop from stereo image pairs from
either the Pancam or the Navcam cameras. This three-
dimensional computer visualization method of fitting a
plane to bedding has distinct advantages over the more
traditional method of measuring points and then fitting a
plane. First, it gives immediate feedback to the quality of the
fitting across the entire fitted surface which does not occur
with a post hoc calculation of planes. Second, beddings
planes are most easily distinguished when they intersect the
surface along a flat plane and this is exactly the time when
the accuracy of the determination of the bedding plane is
most imprecise. There is a tendency for features lying
outside that plane to be differentially degraded relative to
the most obvious bedding surface. What seems an obvious
extension of bedding in one view may look completely
unconvincing in another perspective. The virtual reality
plane fitting used in this work allows one to experiment
with the fitting of planes and then examine the quality of
that fit for the entire extended region as viewed from various
perspectives. Once models of the sites have been built the
fitting of planes is intuitive and straightforward with results
immediately comparable from site to site.
[18] The models were created using NASA Ames? Stereo
Pipeline [Edwards and Broxton, 2006] and viewed using
their Viz visualization program [Edwards et al., 2005]. The
created models can be fully rotated, viewed, and measured
in three dimensions. To measure strike and dip, the program
allows introduction of a plane which is itself fully rotatable
and viewable in three dimensions. This plane was fit to the
bedding attitude of the outcrop. An example of such a fit of
the plane to the bedding attitude is shown in Figure 5 for
Larry?s Lookout. As expected from an ancient terrain, local
variations in strike and dip exist across the surface of the
bedding and the derived strikes and dips represent an overall
fit to the attitude. Each outcrop model was built in Mars
Exploration Rovers? (MER) site coordinate system which
maintains a fixed vertical vector and northerly zero angular
orientation across all models and thereby allows compar-
isons of strikes and dips from different sites. Strike and dip
were measured from the yaw and roll of the plane, respec-
tively, with pitch fixed at 0
. Realistic uncertainties are ?5

for strike and ?3
 for dip based on multiple fits to the same
bedding plane.
5. Bedding Attitudes
[19] The Clovis class outcrops Palenque (Figure 6a) and
Tetl (Figure 6b) were illustrated by Squyres et al. [2006],
who noted that they exhibit fine-scale plane-parallel fabric
Figure 4. HiRISE image of the Columbia Hills overlain on the stereo-derived DEM from Figure 3. The
view is southward across the Columbia Hills. In the center of the image is the 
90 m Husband Hill
edifice. West Spur extends to the right in this image. A prominent feature of Husband Hill is the
Tennessee Valley, which cuts Husband Hill from the peak extending northward and exposing outcrops
along its margins.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
6 of 14
E06S03
with individual planar elements of approximately centimeter
thickness. Like most of the units examined here, Palenque is
relatively small with a surface outcrop of 
1 m by 
30 cm,
with lower fine-scale units overlain by a more massive unit.
Tetl, located some 7 m down and slightly across slope, is
smaller (
10 by 30 cm) and exhibits a slightly undulating
layering across the face of the outcrop. Measured strikes
diverge by 10
 and dips by 8
. The latter is slightly outside the
uncertainty quoted for errors in measuring the respective dips
and probably reflects postformational slumping or distur-
bance by small impact events on this ancient surface. Despite
their small size, the coherence of strikes and dips on these two
units suggests that they likely represent an in-place unit.
[20] The Peace class outcrop Alligator (Figure 6c)
measures 
50 by 20 cm, but is fractured into several
fragments. Our model (Table 1) gives us a dip of 32
 to
the northwest and is consistent with observations made by
Squyres et al. [2006]. Alligator appears to exhibit shallower
dips in the upper portion, perhaps suggesting that the lower
portion, from which we measured the dip, might have
experienced some slumping. In this context, we note that
the dip of Alligator is the steepest of all those we measured.
[21] The Watchtower class outcrops Larry?s Lookout
(Figure 6d) and Summit (Figure 6f) form large, rugged
outcrops that measure more than a meter in width and tens
of centimeters in length. Given the size of these outcrops,
they are almost certainly in place, as also suggested by
Squyres et al. [2006] for the Larry?s Lookout outcrop. These
two units flank the Tennessee Valley. Larry?s Lookout, to
the northwest of the valley, dips to the northwest at 20
 and
Summit, which lies to the southeast, dips 18
 to the
southeast.
[22] The Descartes class outcrop Voltaire (Figure 6e) is

5 m wide and exhibits a stair step pattern. The outcrop
covers 
1.5 m, but the individual stair steps exhibit compa-
rable protrusions above the surface in the northwesterly and
Table 1. Strikes and Dips Derived From the West Spur and Husband Hill
Rock Class Sol
Image
Sequence
Image
Type Strike Dip
Palenque Clovis 263 P1902 Navcam N85
W 15
N
Tetl Clovis 264 P2598 Pancam N85
E 7
N
Alligator Peace 386 P1919 Navcam N65
E 32
N
Larry?s Lookout Watchtower 454 P2402 Pancam N45
E 20
N
Voltaire Descartes 549 P2352 Pancam N12
E 17
N
Summit Watchtower 612 P0770 Navcam N15
E 18
S
Alqonquin Algonquin 682 P2408 Pancam N53
E 15
S
Figure 5. Three-dimensional model produced from stereo imaging of Pancam sequence P1907 of the
outcrop Larry?s Lookout. This image represents one view in a rotatable, three-dimensional model. The
red plane is matched to the bedding attitude, and strike and dip are measured as yaw and roll of the plane,
respectively.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
7 of 14
E06S03
Figure 6. False color (L257) Pancam images of outcrops from which strikes and dips were derived.
(a) Palenque, sol 269, seq ID P2534, (b) Tetl, sol 264, seq ID P2598, (c) Alligator, sol 386, seq ID P2546,
(d) Larry?s Lookout, sol 456, seq ID P2279, (e) Voltaire, sol 550, seq ID P2353, (f) Summit, sol 608, seq
ID P2582, (g) Alqonquin, sol 682, seq ID P2408.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
8 of 14
E06S03
southeasterly direction. We have derived a northwesterly
dip of 17
, implicitly assuming that the northwesterly
dipping portion of the sttif represents the top of the unit.
One of us (REA) has argued that the southeasterly dipping
face is, in fact, the top of the unit, suggesting a dip to the
southeast. Given the extent and style of outcrop, we argue
that this material is in situ.
[23] The Algonquin outcrop is the type locality for
Algonquin class materials (Figure 6g) and exhibits a series
of prominent layered units with a dip to the southeast of
15
. The outcrop occurs over a lateral distance of a few
meters, with individual exposures of 
1 m. Layers within
these exposures are typically tens of centimeters in thick-
ness. Again, the extent and size of the outcrops suggests that
these rocks were sampled in situ.
6. Structure
[24] Our measured strikes and dips can be used to infer
the structure of units on Husband Hill. Before doing so, it is
appropriate to ask whether bedding attitudes from such
widely spaced, small outcrops can be used to infer subsur-
face structure. With the exception of Alligator, most of the
bedding attitudes were measured from relatively large out-
crops or those that projected from the local topography.
Further, at least two of the outcrops, Larry?s Lookout and
Summit, satisfied both of these criteria and essentially
anchor the structure and stratigraphy. The consistency of
all of the other strikes and dips with these two outcrops
gives us confidence that our bedding attitudes can be used
to infer subsurface structure. We note that many features of
such a structure (e.g., exact dip, unit thickness) must be
viewed with caution, as it is derived from topography which
has been substantially modified by postformation impact
and erosion and for which no unit contacts were observed.
[25] Measured dips on Husband Hill (15?32
) are steeper
than local topography (
8?10
). As Spirit traversed from
northwest to southeast, outcrops on the northwest side of
Husband Hill dipped to the northwest while those on the
southeast side dipped to the southeast (Figure 7a). The
change in dips occurred near the location of Voltaire just
up slope from central axis of the Tennessee Valley. Rocks
dip to the north on West Spur and are shallower than those
observed on Husband Hill (7?15
). A schematic cross
section through Husband Hill and West Spur with segment
A-A0 bisecting West Spur and segment B-B0 bisecting
Husband Hill shows strata dipping consistently away from
a NNE-SSW axis cutting the Tennessee Valley (Figure 7b).
Crumpler et al. [2006] argued that this was an anticline
based on units directly outcropping the rim of the Tennessee
Valley. While the inference of an anticline is possible, a
more likely and conservative interpretation is that beds
drape a preexisting structure.
7. Stratigraphy
[26] At no point within the Columbia Hills can a com-
plete stratigraphic section be observed and Spirit did not
sample any of lithologic contacts between units. As a result,
stratigraphy must be inferred from a combination of lithol-
ogies and structure, rather than measured directly. The
antiformal structure at Husband Hill suggests that the oldest
rocks should be found at its center. This interpretation
depends on whether units can be correlated across the axis
of the apparent fold at the Tennessee Valley. The core of the
antiform and stratigraphically lowest rocks are the Wish-
stone, Watchtower, Independence and Descartes classes.
Overlying this core are Peace class rocks to the northwest
and Algonquin class rocks to the southeast.
[27] We suggest that the oldest materials within the
Husband Hill are a superclass of rocks including Wishstone,
Watchtower, Independence and Descartes. As noted earlier,
these classes share a P-Ti-Cr signature [Clark et al., 2007].
The exact relationship between these classes is unclear. The
principal components analysis of E. Tre?guier et al. (Over-
view of Mars surface geochemical diversity through APXS
data multidimensional analysis: First attempt at modeling
rock alteration, submitted to Journal of Geophysical
Researh, 2008) noted the similarity between Wishstone,
Watchtower and Independence class rocks, while distin-
guishing Descartes class rocks. Hurowitz et al. [2006] have
argued that Wishstone class rocks represent a geochemical
end-member that was mixed with a fluid enriched in MgO,
Zn, S, Br and Cl to produce Watchtower class rocks. Clark
et al. [2007] have argued for formation of Independence
class rocks by extensive aqueous alteration of Wishstone
class rocks. It is unclear if the same process occurred in the
formation of Descartes class rocks, but we tentatively
suggest that these rocks may have formed from a common
precursor of pyroclastic volcanic or impact origin which
experienced a wide degree of alteration which varied at
length scales of meters to tens of meters. Alternatively,
Independence, Descartes and Watchtower class rocks may
have formed in a similar manner but from slightly different
protoliths, although an explosive history of pyroclastic
volcanism or impact is still indicated.
[28] Stratigraphically above this basal unit, Peace and
Algonquin class rocks appear to form another unit with a
similar history. Both are olivine-rich ultramafic clastic
rocks. The grains that make up the Peace class rocks are
cemented by Mg-Ca sulfate, indicating that they have
experienced a different alteration history. Principal compo-
nents analyses support the similarity between the mafic
component of Peace class and the Algonquin class rocks
(Tre?guier et al., submitted manuscript, 2008). This overly-
ing unit is a widespread olivine-rich pyroclastic unit, which
experienced local inundation of and cementation by sulfate-
rich fluids. It is important to note that this unit appears to
thicken to the southeast, where Algonquin appears much
thicker than the corresponding outcrops of Peace to the
northwest.
[29] The similarity between the Wishstone-Watchtower-
Descartes-Independence rocks and the overlying Peace-
Algonquin rocks may have had a deep-seated origin.
McSween et al. [2006a] hypothesized that all of the igneous
rocks at the Spirit landing site, including the late stage,
embaying Adirondack basalts, are related through polybaric
fractionation of some common mildly alkaline basaltic
magma. The relatively ancient age of the Columbia Hills
units precludes direction derivation from the Adirondack
flows themselves, but protracted volcanism eventually led
to a package of volcanic rocks observed by Spirit.
[30] Given the conformable nature of dips on West Spur
to those of the main Husband Hill edifice, it would be
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
9 of 14
E06S03
tempting to infer a capping unit of Clovis class rocks, likely
formed as an impact deposit. However, given the lack of
Clovis class materials on the Husband Hill edifice or in the
stratigraphically high Inner Basin to the south of Husband
Hill, we suggest that West Spur and the Clovis class rocks are
a localized deposit and did not occur as a widespread unit.
8. Formation of the Columbia Hills Edifice
[31] The same processes that have acted on Mars through-
out its history, namely impact, volcanism, erosion, and
aqueous alteration and deposition have likely contributed
to the formation of the Columbia Hills. Indeed, each of
these processes have been recognized in the geomorphology
and in individual rocks analyzed on Husband Hill [Squyres
et al., 2006]. A variety of hypotheses have been proposed to
explain the origin of the Columbia Hills including formation
as an erosional remnant of layered material, volcanic
construct, crater rim interference, central peak uplift due
to impact, and tectonic wrinkle ridge [Rice, 2005]. The
important questions we attempt to address in this section is
which process(es) dominated the formation of the Columbia
Hills and in what sequence did they occur?
[32] The first class of processes are constructional and
propose that the Columbia Hills were constructed over a
significant period of time through deposition of impact,
eolian or volcanic materials and their expression above the
surrounding plains results either from those depositional
processes or from later erosion of flat-lying sediments. In
considering a purely constructional model, it is apparent
from the consistent outwardly dipping beds of Husband Hill
that Spirit is not sampling an eroded, flat-lying stratigraphic
section. While this does not preclude a constructional
origin, it would have to provide a mechanism to explain
the dipping strata, such as through large-scale slumping or
construction of a localized edifice. We suggest instead that
units sampled by Spirit were draped on preexisting topog-
raphy that resulted from deformation.
[33] The second class of models, deformational, may
include those related to impact or tectonism. In the class
of tectonic models, compression, most commonly thought
Figure 7. (a) Mars Orbital Camera image of Husband Hill showing rover traverse and locations where
bedding attitudes were determined. The strikes and dips suggest layers that drape a preexisting uplifted
structure. (b) Cross section from A-A0-B-B0 with 2X vertical exaggeration. Note that a substantial offset
in the section occurs at the break between A0 and B. Dips are shown by short black line segments within
the individual units and correspond to locations in Figure 7a.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
10 of 14
E06S03
of as wrinkle ridges, is a viable model for linear uplifts.
There is little direct evidence to confirm or refute such a
model, although the triangular shape of the Columbia Hills
and their truncation to the south appears less consistent with
a wrinkle ridge origin. Further, the source of compression
that would have produced such a linear uplift is unknown.
[34] A more likely source of deformation providing an
underlying uplift for the Columbia Hills is by impact
processes, given the Noachian age of Gusev and the
importance of impact both during and after its formation.
Within the class of impact models, two hypotheses are
possible. The Columbia Hills lie near the geographic center
of Gusev Crater and rise some 110 m above the surrounding
plains. Gusev Crater, at 
160 km in diameter, should have
formed as a complex crater with a prominent central uplift
complex. Depending on the crater diameter, the uplift
complex can be a single peak or expand to an interior
mountain ring. The size of the central peak complex can be
estimated from empirical relationships: the height of the
central peak generally increases linearly with crater diam-
eter but levels off at 
3 km for craters >80 km [Melosh,
1989]; the width of the central peak or ring Dcp also
increases linearly with crater diameter, Dcp = 0.17D +
1.97 [Hale and Head, 1980]. Using these relationships,
the Gusev central uplift complex should be 
29 km across
and reach a height of 
3 km from the flat floor of the crater.
On the basis of observed final depth/diameter relationships
for Mars [Smith et al., 2001], the 160 km diameter Gusev
Crater should be 4.2?5.6 km deep, but it is currently only

1.9 km deep.
[35] Materials of various types have clearly filled the
crater since its formation, including the observed cap of
basaltic lava flows and possibly also including other lava
flows, aeolian deposits, fluvial and lacustrine materials
[Cabrol et al., 1998], ashfall from Appolinaris Patera
[Milam et al., 2003], and ejecta material from nearby craters
[Cohen, 2006]. Airborne sedimentation processes such as
ejecta and ashfall would have draped Gusev?s central peak
complex, creating distinct rock units conformable with
surface slopes. For example, the cumulative thickness of
ballistically emplaced impact ejecta material in the interior
of Gusev Crater is estimated at 
45 m based on models of
ejecta thickness from the likely history of impacts [Cohen,
2006]. Nearby craters Thira and Zutphen each are expected
to have created ejecta blankets 
5m deep over the Columbia
Hills, each with a distinctly different chemical composition
[Cohen, 2006]. Given the ancient (Noachian) age of Gusev
Crater, both depositional and erosional processes probably
acted on the central peak complex, so its current height above
the plains has to be merely a coincidence. However, the
location and size of the Gusev central uplift complex make it
the most likely mechanism for providing the base under the
Columbia Hills, upon which the observed rocks were
emplaced.
Figure 8. Overlapping impacts on Earth and the Moon. (a) A view of the Henbury craters, Northern
Territory, Australia, from 1 km altitude (Google Earth). Diameter of largest crater is 
150 m.
Overlapping rim (white arrow) has an additional 
50% uplift with respect to the nonoverlapping rim of
the largest crater [Hodge, 1994]. White dotted circles approximate the rim to rim diameters of the impact
craters. (b) Three ancient overlapping craters as seen in the Apollo image ATLAS AS16-M-1419. The
overlapping rims of these craters have additional uplift, which can be observed as the triangular structure
with the notably long shadow. Illumination is from the bottom left.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
11 of 14
E06S03
[36] Alternatively, impacts could have uplifted the
Columbia Hills as the result of mutually overlapping crater
rims, either at a modest scale or large scale. The additional
uplift of shared rims of multiple impacts can be observed on
the Earth (Figure 8a), the Moon (Figure 8b), and elsewhere
on Mars (Figure 9a) and even in Gusev Crater where two
small craters overlap. The arcuate boundaries of the
Columbia Hills could approximate the rims of four bound-
ing craters (c.f., Figures 9b and 9c), with those to the east
and west 
10 km in diameter each and the two craters
bounding the southern edge of the Hills at 2?3 km in
diameter. A problem with this model as the final formative
event in the origin of the Columbia Hills is that the raised
rims of such craters are expected to have upturned strata
and, thus, overlapping rims might be expected to produce
inward dipping strata. In contrast, stratigraphic units on
Husband Hill appear to dip outward on a SE?NW profile
that approximates a cross section between the two, large N?S
bounding arcs of the Columbia Hills. It seems more plausible
that overlapping crater rims played a role in constructing the
Columbia Hills edifice, although later draping occurred.
[37] In either of these impact scenarios, we can ask
whether any of the materials examined by Spirit might date
from preimpact uplifted materials. As noted by Squyres et
al. [2006], some Columbia Hills materials cannot be easily
explained as being exposed uplifted materials. One such
example is the occurrence of in situ sulfate-rich soils like
Paso Robles, which are completely unconsolidated and
would not survive intact during uplift. Further, Paso Robles
is spatially associated with the Watchtower-Wishstone-
Independence-Descartes superclass of rocks that outcrop
at the rim of the Tennessee Valley. It is possible to
speculate that West Spur represents the primary material
that was uplifted and tilted during the impact process that
formed the underlying structure of Husband Hill, but we
find it difficult to explain why such a block would have
not been draped by later materials. We suggest that all of
the rocks observed by Spirit simply drape underlying
topography produced by impact-induced uplift, probably
as the result of smaller craters that formed on the floor of
Gusev well after its formation and perhaps even after signif-
icant infilling of Gusev by sediments from Ma?adim Vallis.
[38] If all of the units observed on the edifice of Husband
Hill were formed by later draping, what was the sequence of
impact, aeolian and volcaniclastic deposition, aqueous
inundation or alteration, and erosion? We suggest that
deposition and aqueous inundation or alteration were epi-
sodic events. A variety of scenarios can be envisioned for
the timing of alteration, including impact reworking of
previously aqueously altered materials, postdepositional or
Figure 9. Additional terrain uplift from overlapping rims in Terra Sabaea compared to the Columbia
Hills. (a) MOLA gridded DEM shaded relief map overlain on the MOLA DEM of heavily cratered terrain
(32.2
E, 17.9
S) in Terra Sabaea. The largest crater in the scene is Denning crater (upper right; D =
165 km). The white outline highlights the area of significant rim overlap between multiple craters. This
white line also represents a topographic contour. (b) Is the same 3-D view from Figure 9c except the base
level of the DEM has been changed to the level of the white outline in Figure 9a. This is a simple
simulation of local deposition either by volcanism, or sedimentation. As we can see here most craters are
buried beyond recognition with the most prominent feature remaining being the triangular-shaped
overlapping rims of the now buried craters. (c) The same HiRISE image of the Columbia Hills overlain
on the stereo-derived DEM from Figure 4 (zoomed out and looking NNW) shown here for comparison
with Figure 9b There is an unmistakable resemblance between the example in Figure 9b with the
Columbia Hills.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
12 of 14
E06S03
syndepositional alteration, or spatially variable alteration.
Similarly, erosion may have been variable both temporally
and spatially. Grant et al. [2006] and Golombek et al.
[2006a] concluded, on the basis of the surficial geology
and measurement of rock populations, that erosion of the
basaltic plains since the Hesperian (
3.5 Ga) was limited to
the order of 10 cm and that of the Husband Hill to meters.
Differential erosion of the Spirit landing site is apparent in
the distribution of craters, which are abundant on the plains
and on the tops of peaks within the Columbia Hills and rare
in the Inner Basin and on crater slopes.
[39] The full extent of erosion is poorly known, but the
Tennessee Valley may provide some insights. Although
Spirit did not explore the Tennessee Valley, the Peace-
Algonquin layer does not appear to the drape the valley.
Instead, we suggest that the Tennessee Valley is an erosional
feature formed largely after deposition. If the draping unit
now sampled by the Watchtower class outcrops at Larry?s
Lookout and Summit were once continuous across the now
eroded Tennessee Valley, erosion on the order of tens of
meters since deposition is inferred.
[40] Taken together, we can constrain the timing of
aqueous alteration of the Columbia Hills. All indicators
suggest that the bulk of both erosion and weathering of the
Columbia Hills occurred prior to deposition of the plains
basalts. These Adirondack class basalts have experienced
minimal interaction with water [Yen et al., 2005]. Further,
estimates of erosion are similar in the cratered plains
[Golombek et al., 2006a, 2006b] and Husband Hill [Grant
et al., 2006], suggesting a minimal role for water since the
deposition of the basalts of the cratered plains [Golombek et
al., 2006b]. There is no direct evidence that the aqueous
alteration of the Columbia Hills is related to flooding by
Ma?adim Vallis, while the bulk of alteration occurred prior
to the deposition of the plains basalts in the Late Hesperian.
[41] The final constructional phase of the Columbia Hills
was the eruption and embayment of the Adirondack class
basaltic plains [McSween et al., 2004, 2006b; Greeley et al.,
2005]. Subsequent to this event, limited erosion [Grant et
al., 2006], impact reworking [Haldemann et al., 2006], and
eolian reworking [Greeley et al., 2006] have caused minor
modification of the Columbia Hills.
9. Conclusions
[42] The initial stage of the formation of the Columbia
Hills was probably uplift related to either the formation of
the Gusev Crater central peak or ring or through mutual
interference of overlapping crater rims. This raised platform
of low hills was subsequently draped by a series of impact
and volcaniclastic materials that experienced temporally and
spatially variable aqueous infiltration, cementation and
alteration episodically during or after deposition. West Spur
likely represents a spatially isolated depositional event, as
Clovis class rocks are not observed elsewhere on Husband
Hill. Erosion by a variety of processes, including mass
wasting, removed tens of meters of materials and formed the
Tennessee Valley primarily after deposition. This was
followed by eruption of the Adirondack class plains basalt
lava flows which embayed the Columbia Hills. Minor
erosion, impact, and aeolian processes have subsequently
modified the Columbia Hills.
[43] This history is consistent with the lithologic, physio-
graphic, structural and stratigraphic data collected by the
Spirit Rover in the Columbia Hills and orbital data collected
by a number of missions and instruments. While it is
impossible to uniquely define the sequence of events that
produced the Columbia Hills, these processes are widely
invoked to explain a diverse array of terrains on Mars. In this
sense, the Columbia Hills are not extraordinary, although the
range of rock types they have produced are outside the range
observed elsewhere on Mars at orbital scales. Indeed, the
processes are very ordinary, dominating the geologic history
of Mars. What is extraordinary about the Columbia Hills is
that we have been able to use multiple scales of observation,
from microscopic to orbital, and types of data, from mor-
phology to geochemistry, to decipher their origin.
[44] Acknowledgments. This work was supported by NASA through
funding to the Mars Exploration Rover mission. We thank the entire
engineering and science team of the MER mission for their tireless effort
in operating these rovers and obtaining the data that made this paper
possible.
References
Arvidson, R. E., et al. (2006), Overview of the Spirit Mars Exploration
Rover Mission to Gusev Crater: Landing site to Backstay rock in the
Columbia Hills, J. Geophys. Res., 111, E02S01, doi:10.1029/
2005JE002499.
Bell, J. F., III, et al. (2003), Mars Exploration Rover Athena Panoramic
Camera (Pancam) investigation, J. Geophys. Res., 108(E12), 8063,
doi:10.1029/2003JE002070.
Blaney, D. L., J. F. Bell III, N. Cabrol, P. Christensen, W. H. Farrand,
D. Ming, J. Moersch, S. Ruff, and the Athena Science Team (2005), Spec-
tral diversity at Gusev Crater from coordinated mini-TES and Pancam
observation, Lunar Planet. Sci. Conf., XXXVI, Abstract 2064.
Bond, C. E., A. D. Gibbs, Z. K. Shipton, and S. Jones (2007), What do you
think this is? ??Conceptual uncertainty?? in geoscience interpretations,
GSA Today, 17, 4?10, doi:10.1130/GSAT01711A.1.
Cabrol, N. A., E. A. Grin, R. Landheim, R. O. Kuzmin, and R. Greeley
(1998), Duration of the Ma?adim Vallis/Gusev Crater hydrogeologic sys-
tem, Mars, Icarus, 133, 98?108, doi:10.1006/icar.1998.5914.
Cabrol, N. A., et al. (2003), Exploring Gusev Crater with Spirit: Review of
sciences objectives and testable hypotheses, J. Geophys. Res., 108(E12),
8076, doi:10.1029/2002JE002026.
Christensen, P. R., et al. (2003), The Miniature Thermal Emission Spectro-
meter for the Mars Exploration Rovers, J. Geophys. Res., 108(E12),
8064, doi:10.1029/2003JE002117.
Clark, B. C., III, et al. (2007), Evidence for montmorillonite or its composi-
tional equivalent in Columbia Hills, Mars, J. Geophys. Res., 112,
E06S01, doi:10.1029/2006JE002756.
Cohen, B. A. (2006), Quantifying the amount of impact ejecta at the MER
landing sites and potential paleolakes in the southern Martian highlands,
Geophys. Res. Lett., 33, L05203, doi:10.1029/2005GL024963.
Crisp, J. A., M. Adler, J. R. Matijevic, S. W. Squyres, R. E. Arvidson, and
D. M. Kass (2003), Mars Exploration Rover mission, J. Geophys. Res.,
108(E12), 8061, doi:10.1029/2002JE002038.
Crumpler, L. S., T. McCoy, and the Athena Science Team (2006), MER
surface geologic transect mapping in the Plains and Hills, Gusev Crater,
Lunar Planet. Sci. Conf., XXXVII, Abstract 1685.
Crumpler, L. S., T. McCoy, and M. Schmidt (2007), Spirit: Observations of
very vesicular basalts in the Columbia Hills, Mars and significance for
primary lava textures, volatiles and paleoenvironment, Lunar Planet. Sci.
Conf., XXXVIII, Abstract 2298.
Edwards, L., and M. Broxton (2006), Automated 3D surface reconstruction
from orbital imagery, paper presented at AIAA Space 2006, San Jose,
Calif., Sept.
Edwards, L., J. Bowman, C. Kunz, D. Lees, and M. Sims (2005), Photo-
realistic terrain modeling and visualization for Mars Exploration Rover
science operations, paper presented at IEEE SMC 2005, SPONSOR,
Waikoloa, Hawaii, Oct.
Golombek, M. P., et al. (2003), Selection of the Mars Exploration Rover
landing sites, J. Geophys. Res., 108(E12), 8072, doi:10.1029/
2003JE002074.
Golombek, M. P., et al. (2006a), Geology of the Gusev cratered plains from
the Spirit rover traverse, J. Geophys. Res., 111, E02S07, doi:10.1029/
2005JE002503.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
13 of 14
E06S03
Golombek, M. P., et al. (2006b), Erosion rates at the Mars Exploration
Rover landing sites and long-term climate change on Mars, J. Geophys.
Res., 111(E12), E12S10, doi:10.1029/2006JE002754.
Gorevan, S. P., et al. (2003), The rock abrasion tool, J. Geophys. Res.,
108(E12), 8068, doi:10.1029/2003JE002061.
Grant, J. A., S. A. Wilson, S. W. Ruff, M. P. Golombek, and D. L. Koestler
(2006), Distribution of rocks on the Gusev Plains and on Husband Hill,
Mars, Geophys. Res. Lett., 33, L16202, doi:10.1029/2006GL026964.
Greeley, R., B. H. Foing, H. Y. McSween Jr., G. Neukum, P. Pinet,
M. van Kan, S. C. Werner, D. A. Williams, and T. E. Zegers (2005), Fluid
lava flows in Gusev Crater, Mars, J. Geophys. Res., 110, E05008,
doi:10.1029/2005JE002401.
Greeley, R., et al. (2006), Gusev Crater: Wind-related features and pro-
cesses observed by the Mars Exploration Rover Spirit, J. Geophys.
Res., 111, E02S09, doi:10.1029/2005JE002491.
Haldemann, A. F. C., L. Crumpler, J. A. Grant, M. P. Golombek, B. A.
Cohen, and J. W. Rice Jr. (2006), Mapping and interpreting the cratering
record in the Columbia Hills with Spirit, Lunar Planet. Sci. Conf.,
XXXVII, Abstract 1231.
Hale, W., and J. W. Head (1980), Central peaks in Mercurian craters:
Comparisons to the moon, Proc. Lunar Planet. Sci. Conf., 11th, 2191?
2205.
Herkenhoff, K. E., et al. (2003), The Athena microscopic imager investiga-
tion, J. Geophys. Res., 108(E12), 8065, doi:10.1029/2003JE002076.
Hodge, P. (1994), Meteorite Craters and Impact Structures of the Earth,
132 pp., Cambridge Univ. Press, Cambridge, U.K.
Hurowitz, J. A., S. M. McLennan, H. Y. McSween, P. A. DeSouza, and
G. Klingelho?fer (2006), Mixing relationships and the effects of secondary
alteration in theWishstone andWatchtower classes of HusbandHill, Gusev
Crater, Mars, J. Geophys. Res., 111, E12S14, doi:10.1029/2006JE002795.
Klingelhofer, G., et al. (2003), The Athena MIMOS II Mossbauer spectro-
meter investigation, J. Geophys. Res., 108(E12), 8067, doi:10.1029/
2003JE002138.
Kuzmin, R. O., R. Greeley, R. Landheim, N. A. Cabrol, and J. D. Farmer
(2000), Geologic map of the MTM-15182 and MTM-15187 Quadran-
gles, Gusev Crater?Ma?adim Vallis region, Mars, in Geologic Atlas of
Mars, U.S. Geol. Surv. Geol. Invest. Ser. I-2666.
Maki, J. N., et al. (2003), Mars Exploration Rover engineering cameras, J.
Geophys. Res., 108(E12), 8071, doi:10.1029/2003JE002077.
Martinez-Alonso, S., B. M. Jakosky, M. T. Mellon, and N. E. Putzig (2005),
A volcanic interpretation of Gusev Crater surface materials from thermo-
physical, spectral and morphological evidence, J. Geophys. Res., 110,
E01003, doi:10.1029/2004JE002327.
McSween, H. Y., et al. (2004), Basaltic rocks analyzed by the Spirit rover in
Gusev Crater, Science, 305, 842?845, doi:10.1126/science.3050842.
McSween, H. Y., et al. (2006a), Alkaline volcanic rocks from the Columbia
Hills, Gusev Crater, Mars, J. Geophys. Res., 111, E09S91, doi:10.1029/
2006JE002698.
McSween, H. Y., et al. (2006b), Characterization and petrologic interpreta-
tion of olivine-rich basalts at Gusev Crater, Mars, J. Geophys. Res., 111,
E02S10, doi:10.1029/2005JE002477.
Melosh, H. J. (1989), Impact Cratering: A Geologic Process, 254 pp.,
Oxford Univ. Press, Oxford, UK.
Milam, K. A., K. R. Stockstill, J. E. Moersch, H. Y. McSween Jr., L. L.
Tornabene, A. Ghosh, M. B. Wyatt, and P. R. Christensen (2003), THE-
MIS characterization of the MER Gusev Crater landing site, J. Geophys.
Res., 108(E12), 8078, doi:10.1029/2002JE002023.
Ming, D. W., et al. (2006), Geochemical and mineralogical indicators for
aqueous processes in the Columbia Hills of Gusev Crater, Mars, J. Geo-
phys. Res., 111, E02S12, doi:10.1029/2005JE002560.
Mittlefehldt, D. W., R. Gellert, T. McCoy, H.Y. McSween Jr., R. Li, and the
Athena Science Team (2006), Possible Ni-rich mafic-ultramafic mag-
matic sequence in the Columbia Hills: Evidence from the Spirit Rover,
Lunar Planet. Sci. Conf., XXXVII, Abstract 1505.
Rice, J. W. (2005), The origin of the Columbia Hills, Eos Trans. AGU,
86(52), Fall Meet. Suppl., Abstract P21A-0133.
Rieder, R., R. Gellert, J. Bru?ckner, G. Klingelho?fer, G. Dreibus, A. Yen, and
S. W. Squyres (2003), The new Athena alpha particle X-ray spectrometer
for the Mars Exploration Rovers, J. Geophys. Res., 108(E12), 8066,
doi:10.1029/2003JE002150.
Ruff, S. W., P. R. Christensen, D. L. Blaney, W. H. Farrand, J. R. Johnson,
J. R. Michalski, J. E. Moersch, S. P. Wright, and S. W. Squyres (2006),
The rocks of Gusev Crater as viewed by the Mini-TES instrument,
J. Geophys. Res., 111, E12S18, doi:10.1029/2006JE002747.
Smith, D. E., et al. (2001), Mars Orbiter Laser Altimeter: Experiment
summary after the first year of global mapping of Mars, J. Geophys.
Res., 106, 23,689?23,722.
Squyres, S. W., and the Athena Science Team (2005), Recent results from
the Spirit Mars Exploration Rover mission, Lunar Planet. Sci. Conf.,
XXXVI, Abstract 1918.
Squyres, S. W., et al. (2003), Athena Mars rover science investigation,
J. Geophys. Res., 108(E12), 8062, doi:10.1029/2003JE002121.
Squyres, S. W., et al. (2004), The Spirit Rover?s Athena science investiga-
tion at Gusev Crater, Mars, Science, 305, 794 ? 799, doi:10.1126/
science.3050794.
Squyres, S. W., et al. (2006), Rocks of the Columbia Hills, J. Geophys.
Res., 111, E02S11, doi:10.1029/2005JE002562.
Squyres, S. W., et al. (2007), Pyroclastic activity at Home Plate in Gusev
Crater, Mars, Science, 316, 738?742, doi:10.1126/science.1139045.
Yen, A. S., et al. (2005), An integrated view of the chemistry and mineralogy
of Martian soils, Nature, 436, 49?54, doi:10.1038/nature03637.
Yen, A. S., et al. (2008), Hydrothermal processes at Gusev Crater: An
evaluation of Paso Robles class soils, J. Geophys. Res., 113, E06S10,
doi:10.1029/2007JE002978.

R. E. Arvidson, Department of Earth and Planetary Sciences, Washington
University, Campus Box 1169, One Brookings Drive, St. Louis, MO
63130, USA.
D. L. Blaney and M. P. Golombek, Jet Propulsion Laboratory, California
Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109,
USA.
B. A. Cohen, Institute of Meteoritics, University of New Mexico,
Albuquerque, NM 87131, USA.
L. S. Crumpler, New Mexico Museum of Natural History and Science,
1801 Mountain Road NW, Albuquerque, NM 87104, USA.
P. A. de Souza Jr., Tasmanian ICT Centre, CSIRO, Castray Eslanade,
Hobart, Tas 7000, Australia.
C. d?Uston and E. Tre?guier, CESR/OMP, 9 avenue C. Roche, F-31400
Toulouse, France.
L. Edwards and M. Sims, Intelligent Systems Division, NASA Ames,
Mail Stop 269-3, Moffett Field, CA 94035, USA.
J. A. Grant, Center for Earth and Planetary Studies, National Air and
Space Museum, Smithsonian Institution, Washington, DC 20560, USA.
A. F. C. Haldemann, ESA/ESTEC HME-ME, P.O. Box 299, NL-2200
AG Noordwijk ZH, Netherlands.
T. J. McCoy and M. E. Schmidt, Department of Mineral Sciences,
National Museum of Natural History, Smithsonian Institution, Washington,
DC 20560-0119, USA. (mccoyt@si.edu)
H. Y. McSween Jr., Department of Earth and Planetary Sciences,
University of Tennessee, Knoxville, TN 37996, USA.
J. W. Rice Jr., Department of Geological Sciences, Arizona State
University, Campus Box 871404, Tempe, AZ 85287-6305, USA.
L. A. Soderblom, Branch of Astrogeology, U.S. Geological Survey, 2255
North Gemini Drive, Flagstaff, AZ 86001, USA.
S. W. Squyres, Department of Astronomy, Cornell University, 428 Space
Science Building, Ithaca, NY 14853, USA.
L. L. Tornabene, Lunar and Planetary Laboratory, University of Arizona,
Tucson, AZ 85721, USA.
E06S03 MCCOY ET AL.: STRUCTURE AND ORIGIN OF COLUMBIA HILLS
14 of 14
E06S03