Earth and Planetary Science Letters 400 (2014) 130–144
Contents lists available at ScienceDirect
Earth and Planetary Science Letters
www.elsevier.com/locate/epsl
The effects of weathering on the strength and chemistry of Columbia
River Basalts and their implications for Mars Exploration Rover Rock
Abrasion Tool (RAT) results
B.J. Thomson a,∗ , J.A. Hurowitz b , L.L. Baker c , N.T. Bridges d , A.M. Lennon d , G. Paulsen e ,
K. Zacny e
a
Center for Remote Sensing, Boston University, 725 Commonwealth Ave. Rm. 433, Boston, MA 02215, USA
Department of Geosciences, Stony Brook University, Stony Brook, NY 11794, USA
c
Department of Plant, Soil and Entomological Sciences, University of Idaho, Moscow, ID 83844, USA
d
Johns Hopkins University, Applied Physics Laboratory, Laurel, MD 20723, USA
e
Honeybee Robotics, Pasadena, CA 91103, USA
b
a r t i c l e
i n f o
Article history:
Received 21 August 2013
Received in revised form 16 April 2014
Accepted 10 May 2014
Available online xxxx
Editor: C. Sotin
Keywords:
Mars
Earth
Mars Exploration Rovers
basalt weathering
rock strength
physical properties
a b s t r a c t
Basalt physical properties such as compressive strength and density are directly linked to their chemistry
and constitution; as weathering progresses, basalts gradually become weaker and transition from intact
rock to saprolite and ultimately, to soil. Here we quantify the degree of weathering experienced by the
Adirondack-class basalts at the Mars Exploration Rover Spirit site by performing comparative analyses on
the strength and chemistry of a series of progressively weathered Columbia River Basalt (CRB) from
western Idaho and eastern Washington. CRB samples were subjected to compressive strength tests,
Rock Abrasion Tool grinds, neutron activation analysis, and inductively coupled plasma optical emission
spectroscopy. Analyses of terrestrial basalts indicate linked strength-chemical changes, as expected.
Weathering sufficient to induce the loss of more than 50% of some cations (including >50% of MgO
and MnO as well as ∼38% of Fe2 O3 and 34% of CaO) was observed to weaken these samples by as much
as 50% of their original strength. In comparison with the terrestrial samples, Adirondack-class basalts
are most similar to the weakest basalt samples measured in terms of compressive strength, yet they do
not exhibit a commensurate amount of chemical alteration. Since fluvial and lacustrine activity in Gusev
crater appears to have been limited after the emplacement of flood basalt lavas, the observed weakness
is likely attributable to thin-film weathering on exposed, displaced rocks in the Gusev plains (in addition
to some likely shock effects). The results indicate that Adirondack-class basalts may possess a several
mm-thick weak outer rind encasing an interior that is more pristine than otherwise indicated, and also
suggest that long rock residence times may be the norm.
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
Understanding the material strength of rock, outcrop, and soil
targets encountered by landed spacecraft on Mars is an essential
part of our knowledge of present and past surface and subsurface
environmental conditions. When considered together with a rock’s
chemistry and mineralogy, information about the physical properties of materials (including parameters such as texture, strength,
hardness, and particle size, shape, and sorting) provides an additional means to help reconstruct its geologic history. Of particular
*
Corresponding author. Tel.: +1 617 353 5148.
E-mail address: [email protected] (B.J. Thomson).
http://dx.doi.org/10.1016/j.epsl.2014.05.012
0012-821X/© 2014 Elsevier B.V. All rights reserved.
relevance is the role of rock alteration. As chemical weathering
progresses, primary igneous minerals are altered and replaced by
secondary phases to a degree dictated by limiting factors such as
temperature, time, kinetics, and/or the availability of water. The
complex assemblage of primary and secondary materials created
by these processes yields a unique fingerprint that can be used
to constrain the modification processes. The strength and physical
properties of geologic materials vary in tandem with these mineralogical alterations, but, until now, such strength changes have
received little attention.
The purpose of this study is to investigate the linked chemical and strength changes undergone by basalt during weathering. Basalt is an important crustal component that represents a
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
131
significant fraction of the surface area of terrestrial bodies (Basaltic
Volcanism Study Project, 1981). Martian basalts have been identified from orbital remote sensing measurements (e.g., Mustard et
al., 1997; Christensen et al., 2000; Bandfield, 2002), chemical, spectral, and textural analyses from landed spacecraft (e.g., Clark et al.,
1982; Bell et al., 2004; Christensen et al., 2004; Gellert et al., 2004;
Herkenhoff et al., 2004; McSween et al., 2004), and within the
martian meteorite suite (e.g., McSween, 1994). Indeed, basalt is the
most abundant rock type on Mars (McSween et al., 2009). Given
its wide abundance on Earth, Mars, and the other terrestrial planets, basalt makes an ideal target material for comparative physical property studies. Establishing an empirical linkage between
strength and chemistry of basalt undergoing weathering will permit greater constraints on the nature and degree of weathering
processes on Mars as inferred from remote measurements.
1.1. Previous work
The chemical and mineralogical effects of basalt weathering
have been the subjects of numerous investigations. These include
studies of water-limited basalt weathering (e.g., Elwood Madden et
al., 2004; Hurowitz and McLennan, 2007), acid weathering (e.g.,
Tosca et al., 2004; Niles and Michalski, 2009; Hurowitz et al.,
2010), high water–rock ratio weathering in surficial environments
(e.g., Eggleton et al., 1987; Nesbitt and Wilson, 1992), hydrothermal alteration (e.g., Kristmannsdottir, 1979; Jakobsson and Moore,
1986; Baker et al., 2000), and impact-induced alteration (e.g.,
Newsom, 1980). Yet the specific effects of these processes on basalt
strength that occur with mineralogical changes have been little addressed.
In a prior study, we used data from Rock Abrasion Tool (RAT)
grinds to infer the compressive strength of rocks at the Mars Exploration Rover Spirit site in Gusev crater (Thomson et al., 2013).
The RAT is a small rotary grinder carried at the end of the each
rover’s robotic arm (additional details are given in Section 2.3.2).
Designed to abrade and remove the outer layers of rocks in order to expose fresher interior surfaces (Gorevan et al., 2003),
the power consumed during grinding combined with the grinding time can be linked to rock strength (Myrick et al., 2004;
Thomson et al., 2013). The results from our initial study constrain
the range of strengths over which the RAT effectively operates, and
provides a means to link the specific grind energy to compressive strength. At the MER Spirit site, the results confirm and expand previous initial indications that many intact martian materials are significantly weaker than their fresh terrestrial counterparts
(Arvidson et al., 2004; Haskin et al., 2005; McSween et al., 2006;
Squyres et al., 2006; Wang et al., 2006), implicating alteration effects. We focus here on Adirondack-class rocks, which are basaltic
in composition and are interpreted as having a volcanic origin (e.g.,
Herkenhoff et al., 2004; McSween et al., 2004). Specifically, they
have mm-sized phenocrysts (possibly olivine) in a matrix whose
constituent particles are not discernible at the best Microscopic
Imager (MI) resolution of 31 μm, they contain irregular vesicles
and vugs, they commonly exhibit conchoidal fracture (which is
consistent with a lack of cleavage planes), and they have no evident sedimentary characteristics such as laminae, banding, or gradation. All of these characteristics are suggestive of a glassy or
aphanitic matrix such as would be expected in a volcanic basalt.
Whereas our prior results demonstrate a robust method for inferring bulk strength, the specific nature of the weathering was not
addressed. Herein, we derive a relationship between the type and
extent of weathering and the measured strength, a correlation that
holds promise for inferring alteration processes from drilling and
abrasion tools on current and future planetary rovers.
Fig. 1. Map of the northwestern United States with the extent of Columbia River
Basalt (CRB) given in orange (from Burns et al., 2010) and sample localities marked
with black circles. The flows themselves are more extensive (e.g., Reidel et al., 1989),
but the outline covers the main central area of thickest CRB. (For interpretation
of the references to color in this figure legend, the reader is referred to the web
version of this article.)
1.2. Goals of this study
The goal of this effort is to understand and quantify the geochemical and geomechanical changes in typical basalt weathering
profiles as a proxy for Mars. The sampling locale is the Columbia
River Basalt Province, a flood basalt province with numerous interflow weathering horizons. This study extends our previous work in
which the RAT specific grind energy was used infer the strength of
rock targets (Thomson et al., 2013). Here, we quantify the systematic changes in basalt physical properties with degree of weathering. The results from this study will provide independent constraints on the nature and vigor of martian weathering processes.
2. Approach
2.1. Sample localities
All of the samples described in this work were obtained from
flows of the Columbia River Basalt Group (CRBG) located in eastern Washington, Oregon, and western Idaho (Fig. 1). Erupted during the Miocene, the Columbia River Basalt Group is divisible into
four main stratigraphic formations. From oldest to youngest, these
are the Imnaha Basalt Formation (erupted ∼17.5 Ma), the Grand
Ronde Basalt Formation (∼16.5 and 15.6 Ma), the Wanapum Basalt
Formation (∼15.6 and 14.5 Ma), and the Saddle Mountain Basalt
Formation (∼14 and 6 Ma) (Swanson et al., 1979; Camp and
Hooper, 1981; Tolan et al., 1989). Grande Ronde Basalts constitute
the overwhelming majority (87 vol%) of the total erupted volume
(1.74 × 105 km3 ) (Tolan et al., 1989), and have been divided into
four magnetostratigraphic units: R 1 , N 1 , R 2 , and N 2 , from oldest
to youngest. Here, the designators R and N refer to reverse and
normal magnetic polarity, respectively.
Within this stratigraphy, our focus is in the entombed paleosol
horizons in the Grande Ronde and Wanapum Basalts. These paleosols were developed on top of basalt flows during interflow intervals, and their subsequent burial in later flows provides a unique
snapshot of weathering processes from that interflow interval (e.g.,
Sheldon, 2003; Hobbs, 2010). Compared to other paleosol surfaces,
interflow paleosols confer several advantages. First, the protolith
material can be directly determined, rather than inferred from the
132
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
general surroundings. Second, the element of time is also tightly
constrained: paleosol development is bracketed by the ages of the
host basalt and the successive, capping flow. Finally, as these deposits are presently exposed only in recent road cuts, there has
been effectively no modern re-activation of these paleoweathering
horizons. Limited post-burial alteration effects include compaction
due to the overburden and oxidation of organic matter in the uppermost soil layers (e.g., Retallack, 1991).
Idaho. Two parallel rock saw cuts were sliced from each block to
extract hand samples bounded by flat planes. The resulting hand
samples were approximately 2 cm in thickness and 10 cm in minimum extent to accommodate the outer butterfly contact points of
the RAT instrument (Gorevan et al., 2003). Sample bulk densities
were recorded using the glass bead displacement method (Macke
et al., 2010, 2011).
2.3. Experimental testing procedures
2.1.1. Lawyer Canyon
The Lawyer Canyon site is located south of the town of Craigmont, ID along US-95. Hosted by the N 1 magnetostratigraphic unit
of the Grande Ronde Basalt, the paleosol layer consists of a welldefined, 2 m thick paleosol underlain by a saprolite with a gradational lower boundary into progressively less weathered host rock.
The Basalt of Icicle Flat, equivalent to the Basalt of Dodge (Eckler Mountain Member of the Wanapum Basalt (Kauffman, 2004,
2009)), caps the paleosol layer (Bush et al., 2004). This broad, gradual transition presents an ideal locale for numerous macroscopic
samples (in other localities, the saprolite-basalt transition is commonly compact, i.e., less than several cm across). Basalt corestones
are observed both in the paleosol layer and the saprolite layer;
in cross-section these are rounded, isolated masses of less weathered protolith material. The total exposed stratigraphic thickness is
about 10 m.
2.1.2. Kendrick grade
At this locality, a ∼2 m thick paleosol layer (montmorillonite–
kaolinite) developed on Grande Ronde N1 Basalt and is capped
by the Priest Rapids Member of the Wanapum Basalt (Lewis et
al., 2005). Beneath the soil layer, a thick saprolite zone transitions
down over about 2 m through weathered corestones into unaltered
basalt parent material. Vesicularity increases upward through the
saprolite. The paleosol is capped with a blackened layer that may
be composed of vegetation carbonized during emplacement of the
overlying basalt flow.
2.1.3. Upper Shumaker
At Upper Shumaker, a thin, 3–10 cm paleosol layer is present
within the Eckler Mountain Member of the Wanapum Basalt Formation. The weathered sequence is capped by the Shumaker Creek
Member of the Wanapum (the intervening Frenchman Spring
Member is not present at this locality; it pinches out well to the
west). Beneath the thin paleosol, a broad saprolite zone is present
that is more highly vesicular (∼20%) than the basalt or saprolite present at Lawyer Canyon. Protolith material is accessible only
as basalt corestones, not as an intact layer of massive, unaltered
basalt. The total exposed stratigraphic thickness at this exposure is
about 4 m.
2.2. Sample acquisition strategy and preparation procedure
At each road cut site, relatively intact blocks of basalt were selected at multiple heights. These were photodocumented in place,
and the surrounding context was noted. Blocks were then dislocated and transported down to road level where drill cores were
extracted using a portable gasoline-powered, wet core drill (the
core drill used water to cool the diamond bit segments and remove cuttings). Two people were required to extract the cores:
one to operate the drill and another to supply water using a manual pump. At least two drill cores were extracted per sampled
block. This approach was deemed safer than moving the drill and
water-cooling assembly up and down the road cut face due to the
unstable nature of the surface slopes.
To avoid complications in the grind tests induced by irregular rock textures, the samples were prepared at the University of
2.3.1. Compressive strength tests
Cylindrical cores were cut from the basalt samples using a diamond coring bit. The resulting cores had a mean diameter of
24.6 mm (±0.11 mm). Following the procedure established in our
initial study (Thomson et al., 2013), the ends of each core were
trimmed and ground such that they were flat, parallel, and roughly
perpendicular to the cylindrical axis of the core. No additional
grinding was performed on the cylindrical wall of the core. The
target length-to-diameter (L / D) ratio for these specimens was 2.0,
but fracturing of the cores resulted in a range of final L / D ratios
of 0.87 to 1.86. The difference in L / D ratio introduces an error of
approximately 4% (Thuro et al., 2001). Grinding was accomplished
using a diamond masonry blade mounted in a wet saw with a linear translation table, and a custom jig for fixing the cylindrical axis
of the specimen perpendicular to the direction of the table’s travel
through the blade.
All of the specimens were individually measured for length and
diameter with a digital caliper. Each nominal dimension was taken
as the mean of either three (for diameter) or four (for length)
measurements at different locations and orientations. The nominal specimen diameter was used to calculate the nominal crosssectional area for converting compressive force into compressive
stress. The nominal length was used to check for correlation between the deviation of specimen failure strength from the mean
and L / D ratio. No significant correlation was found, indicating that
it is unlikely that end effects skewed the compressive strength
results more for short specimens than for longer specimens. The
nominal length was also used as the gauge length for converting
compressive displacements into compressive strain.
Compressive strength of the basalt samples was measured using an Instron 8502 servohydraulic load frame equipped with a
250 kN load cell and 64 mm (2.5 ) diameter spherical compression
platens. The compression tests were conducted without radial confinement, so the failure stress of the specimens is reported as the
unconfined compressive strength (UCS). Tests were run in displacement control at a constant strain rate of 10−4 s−1 , which required
modifying the displacement rate for each test based on the specimen length, and a data sampling rate of 10 Hz.
2.3.2. RAT instrument overview
The Rock Abrasion Tool is an engineering instrument and vital component of the in situ sensing portion of the Athena science
payload carried onboard the Mars Exploration Rovers (Squyres et
al., 2003). Attached to the end of the robotic arm (IDD or Instrument Deployment Device), the RAT was designed to abrade
and remove the outer weathered layers of rocks, exposing fresher
material beneath (Gorevan et al., 2003). The RAT fits within a cylindrical volume 8.5 cm in diameter and 12.8 cm in length, and has
a mass of 0.687 kg. It consists of three actuators: one to rotate a
diamond impregnated “paddle wheel” at high speed (about 3000
rpm) against the target; the second to revolve the paddle wheel
about the RAT’s central axis at lower speeds (<2 rpm); and the
third to raise or lower the grinding assembly against the target
(termed the Z -axis motor). Typical step sizes of the Z -axis motor
are 0.05 mm per revolution for dense targets. Prior to operation,
the IDD must be placed firmly against a target with a preload force
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
of about 10–40 N (typically 30 N). When the rover is on a slope,
lower preload is used to reduce the risk of rover slip downhill. A
successful grind operation produces an abraded region 4.5 cm in
diameter and nominally 2–5 mm deep, though the actual depth
attained depends on the rock hardness.
Numerous engineering data are recorded during grind operations, including the current draw from the grinding motor. The
integrated power consumption over the last 0.25 mm of a grind
combined with the grinding time has been termed Specific Grind
Energy (SGE, units J/mm3 ) and is an estimation of the target’s
strength and resistance to abrasion.
2.3.3. Rock Abrasion Tool grind tests
To quantify the grinding operation, two terms are used: Specific
Grind Energy (SGE) and the G-Ratio. The SGE-term is the energy
used by a grinder to remove a unit volume of material; its units
are J/mm3 (Teale, 1965). This energy parameter is a function of a
number of system-specific variables such as grinding parameters,
the state of the abrasive pad, rock physical properties, cleaning efficiency, cuttings removal, and friction. Because SGE is a composite
quantity that is a function of many variables, it does not necessarily capture the full spectrum of rock mechanical properties by
itself, and different rock types may yield non-unique, overlapping
SGE values.
To partially address this issue, we also consider the G-ratio,
which is a ratio of the volumetric wear of rock divided by a volumetric wear of abrasive material (i.e., the pads of the grinding
bit itself; G-ratio units are mm3 /mm3 ). For example, when grinding an abrasive rock (e.g., sandstone) and a non-abrasive rock (e.g.,
limestone), the SGE values may be the same. However, because
the grinding bit wears faster in an abrasive rock, the G-ratio derived for sandstone would be lower than for limestone. Thus, the
G-ratio helps differentiate between different types of rocks with
similar SGE values.
Measuring wear of a grinding bit and volume of rock removed
on Mars is complicated by the lack of direct measurement of these
parameters (Myrick et al., 2004); they must be inferred from visual
images or more preferably, from stereo image-derived topography
(e.g., Herkenhoff et al., 2006). In addition, the values of SGE are
RAT-specific, and as such, any data interpretation has to be calibrated just for the RAT.
For this project, we used the RAT brassboard model to conduct
grind tests. From a design and performance standpoint, the RAT
brassboard, Engineering Model (EM), and the Flight Models (FM)
are all the same. The main difference is that the EM and FM have
flight-qualified actuators and vacuum-grade lubricants. In terms of
the critical factor of power consumption, however, the models are
identical in performance.
Grinding test protocols followed the same procedure as RAT
grinds conducted on Mars, including the same software and command sequences. Initially, the RAT executed a ‘seek-scan’ routine
whereby the grinding wheel rotates at a low speed while the
Z -axis moves the bit towards the surface. Once an elevated current is detected on the grinding axis (meaning the grinding wheel
has contacted and stalled on the rock surface), the Z -axis stops
and backs off the surface a small distance until the grinding wheel
starts spinning again. At this point the scan part of the routine is
executed. During the scan routine, the grinding wheel continues
to spin at a slow rate. Meanwhile, the revolve axis starts to rotate at 1 rpm to sweep or “scan” the surface for the highest point
on the rock. If a stall is detected on the grinding axis during this
scan, the revolve axis stops and the Z -axis backs off the surface
until the grinding axis is free to rotate. Once the grinding axis begins to rotate, the revolve motion begins again. The stall detection
and retract process repeats until one full revolution is made. This
Z -axis position is then set to zero depth and the grinding opera-
133
tion commences. The grinding operation consists of a bit spinning
at 3000 rpm and revolving at up to 2 rpm (the actual revolving
rate is dependent on torque applied by the grinding wheel). After each revolution, the Z -axis advances the bit 50 μm. A typical
grinding operation lasts 1–3 h. Once the grinding is complete, the
grinding telemetry is analyzed to determine SGE. In a few cases the
G-ratio was also acquired by estimating the volume of the ground
hole as well as measuring the wear of the grind pad. The latter
was achieved by taking a photo of the profile of the grind bit before and after the test. However, the steps taken to acquire the
G-ratio presented additional operation complexity, took extra mission time, and were not very accurate due to insufficient spatial
resolution for bit imaging. For that reason this is no longer being
done on the MER mission though bit profile images are still being
captured to monitor bit life over several grinds.
2.3.4. Major and trace element chemistry, mineralogy
Sample slabs were prepared using a rock saw at the University
of Idaho Department of Geological Sciences from the large blocks
of sample that were cored in the field (Section 2.3). When possible,
slabs were collected from immediately adjacent to locations on the
rock where slabs were collected for the purposes of Rock Abrasion
Tool grind tests (Section 2.3.3). When this was not possible (generally due to difficulty orienting the sample block for adjacent cuts),
slabs were collected from a different location on the same block
that was used for preparation of a sample for RAT grind testing.
In this manner, we have attempted to maximize the similarity between sample geochemistry and physical properties (as deduced
from rock crushing and RAT grinding experiments). In addition to
competent slab samples, two samples of loose, disaggregated paleosol material that were too friable for mechanical strength or RAT
testing were collected from Lawyer Canyon and Shumaker Grade
for geochemical analysis.
Whole sample slabs, weighing between 10 and 20 g, were sent
to Activation Laboratories of Lancaster, Ontario, CA for preparation
and analysis of major, minor, and trace elements. Sample preparation entailed crushing the entire sample to minus 10 mesh (1.7
mm) followed by pulverization to at least 95% minus 150 mesh
(106 μm). All sample crushing and grinding equipment is made of
mild steel, which can add up to 0.2 wt% Fe; pure quartz sand was
run through the sample preparation equipment between each sample in order to ensure cleanliness and minimize sample-to-sample
cross contamination.
After powder preparation, samples were analyzed by a combination of instrumental neutron activation analysis (INAA, for As,
Au, Br, Ce, Co, Cr, Cs, Eu, Hf, Hg, Ir, La, Lu, Nd, Rb, Sb, Sc, Se, Sm,
Ta, Tb, Th, U, W, Yb, Zr) and inductively coupled plasma optical
emission spectroscopy (ICP-OES). Sample data are tabulated in Tables 4–6; detection limits for each element are also reported.
For INAA, a 1 g aliquot is encapsulated in a polyethylene vial
and irradiated with flux wires and an internal standard at a thermal neutron flux of 7 × 1012 n cm−2 s−1 . After a 7-day period
to allow for the decay of Na-24 the samples are counted on a
high purity Ge detector with resolution of better than 1.7 keV for
the 1332 keV Co-60 photopeak. Using the flux wires, the decaycorrected activities are compared to a calibration developed from
multiple certified international reference materials. The standards
are only a check on accuracy and are not used for calibration purposes. The method is described in detail in Hoffman (1992).
For analysis by ICP-OES, a 0.2 g sample is mixed with a mixture
of lithium metaborate/lithium tetraborate and fused in a graphite
crucible. The molten mixture is poured into a 5% nitric acid solution and shaken until dissolved (∼30 min). The samples are run for
major oxides (Si, Al, Fe, Mg, Mn, Ca, Ti, Na, K, P) and selected trace
elements (Ba, Be, Sr, V, Y, Zr) on a combination simultaneous/sequential Thermo Jarrell–Ash Enviro II ICP-OES. For the remaining
134
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
Table 1
Compressive strength test results.
Sample
ID
Length
(mm)
Diameter
(mm)
Peak load
(kN)
Stress area
(mm2 )
Peak stress
(MPa)
Failure mode
UCS
(mean ± std dev)
LC-01a
LC-01b
LC-02a
LC-02b
LC-03a
LC-03b
LC-03c
LC-04a
LC-04b
LC-05a
LC-05b
LC-06a
LC-06b
LC-06c
LC-07a
LC-07b
SH-01a
SH-01b
SH-02a
SH-02b
SH-03
KG-01b
KG-01a
KG-02a
KG-02b
38.20
39.19
42.61
37.91
23.15
36.13
26.33
24.36
32.28
38.32
28.51
44.30
39.69
37.25
36.11
24.78
41.82
46.04
32.32
22.15
n/a
28.40
40.54
21.95
21.62
24.52
24.77
24.67
24.75
24.65
24.64
24.64
24.71
24.53
24.77
24.53
24.71
24.68
24.50
24.64
24.70
24.64
24.72
24.34
24.66
n/a
24.72
24.74
24.57
24.71
99.4
95.0
108.0
108.0
109.0
81.2
98.5
60.1
71.0
88.1
51.7
62.7
53.3
33.8
55.4
73.0
107.0
102.0
31.4
35.1
n/a
94.3
64.7
86.8
119.0
472.2
482.0
477.9
481.0
477.1
476.9
476.8
479.6
472.5
482.0
472.8
479.6
478.4
471.4
476.9
479.1
476.9
480.0
465.4
477.7
n/a
480.0
480.7
474.1
479.6
210.5
197.1
226.0
224.5
228.5
170.3
206.6
125.3
150.3
182.8
109.4
130.7
111.4
71.7
116.2
152.4
224.4
212.5
67.5
73.5
n/a
196.4
134.6
183.1
248.1
axial splitting
axial splitting
shear failure
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
axial splitting
total destructiona
axial splitting
axial splitting
axial splitting
n/a
axial splitting
axial splitting
axial splitting
total destructiona
204 ± 9
a
225 ± 1
202 ± 29
138 ± 18
146 ± 52
105 ± 30
134 ± 26
218 ± 8
70 ± 4
n/a
166 ± 44
216 ± 46
Failure mode of ‘total destruction’ refers to complete sample failure along multiple shear planes.
trace elements (Ag, Bi, Cd, Cu, Mo, Ni, Pb, S, Zn) a 0.25 g sample
is digested with four acids beginning with hydrofluoric, followed
by a mixture of nitric and perchloric acids, heated using precise
programmer controlled heating in several ramping and holding cycles which takes the samples to dryness. After dryness is attained,
samples are brought back into solution using hydrochloric acid and
analyzed using a Varian Vista ICP-OES. With this digestion, certain phases may be only partially solubilized. These phases include
zircon, monazite, sphene, gahnite, chromite, cassiterite, rutile and
barite, and only sulfide sulfur is solubilized. Loss on Ignition (LOI)
was determined by measuring the mass lost from samples after
heating them for 2 h at 1050 ◦ C in an open crucible, providing
information on the amount of H2 O, CO2 , S, and other volatile compounds that are present in the samples.
Mineral phase identification and quantification was carried out
on a powdered sample of LC-01 that was mixed with a powdered
alumina standard in known proportions. The mixture was then analyzed on a Bruker HiStar X-ray diffractometer using Cu k-alpha
radiation (40 kV, 20 mA), over a 2-theta range of 10–70◦ . Powder diffraction data were then reduced using the Crystallographica
search-match software program to determine what mineral phases
were present and what their abundances were.
3. Results
3.1. Individual strength test results
A total of 24 compressive strength tests were conducted (Table 1): 7 sample sets from Lawyer Canyon road cut with >2
sample cores per set (N = 16 crushing tests); 2 samples from
Kendrick Grade (N = 4 crushing tests); and 2 samples from Shumaker Grade (N = 4 crushing tests). A third sample from Shumaker
Grade (SH-03) was too friable for compressive strength testing and
failed prior to loading. During the course of Rock Abrasion Tool
testing, 42 grinding tests were performed in 3 different rock types
(see summary Table 2). These included 7 sample sets from Lawyer
Canyon road cut (N = 22 grind tests); 2 samples from Kendrick
Grade (N = 7 grind tests); and 3 samples from Shumaker Grade
Fig. 2. Plot of unconfined compressive strength (UCS) test results ordered by decreasing mean value. KG = Kendrick Grade samples; LC = Lawyer Canyon samples;
and SH = Shumaker Grade samples. Mean values are given by horizontal bars.
(N = 13 grind tests). The number of tests per rock sample ranged
from 3 to 7. These data sets were used to determine the average
and standard deviation for each set of conditions.
Figs. 2 and 3 present univariate data plots of compressive
strength and grind energy, respectively, ordered by decreasing
mean value. In the compressive strength results (Fig. 2), there is
reasonably good separability between samples that have experienced different degrees of weathering (assigned here based on
qualitative visual and textural field evidence). Mean UCS values
vary between 225 and 70 MPa from the freshest to most weathered samples, spanning the rock mechanical strength categories of
“very strong” (100–250 MPa) to “strong” (50–100 MPa) (using categories of Hoek and Brown, 1997). The seven sample points from
the Lawyer Canyon site define an array or sequence going from
stronger, less weathered samples (LC-01, LC-02) to progressively
more weathered ones (LC-06, LC-07). Samples from Kendrick Grade
tend to fall in the intermediate to high range, while samples from
Shumaker Grade span the range of observed strengths. The most
weathered samples from each locality have compressive strengths
that range from ∼30 to 50% of the most pristine samples (i.e., 46%
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
135
Table 2
RAT grinding results over the last 0.25 mm of grind.
Target ID
Mean grind
voltage
[V]
Mean grind
current
[mA]
Duration
[min]
Rotate energy
[J]
Z grind
depth
[mm]
Material
removed
[mm3 ]
SGE
[J/mm3 ]
SGE
(mean ± std dev)
SH-01-a-1
SH-01-a-2
SH-01-a-3
SH-01-a-4
SH-01-a-5
SH-01-a-6
SH-01-a-7
SH-03-a-1
SH-03-a-2
SH-03-a-3
KG-01-a-1
KG-01-a-2
KG-01-a-3
KG-01-a-4
SH-02-a-1
SH-02-a-2
SH-02-a-3
KG-02-a-1
KG-02-a-2
KG-02-a-3
LC-01-a-1
LC-01-a-2
LC-01-a-3
LC-02-a-1
LC-02-a-2
LC-02-a-3
LC-02-a-4
LC-03-a-1
LC-03-a-2
LC-03-a-3
LC-04-a-1
LC-04-a-2
LC-04-a-3
LC-05-a-1
LC-05-a-2
LC-05-a-3
LC-06-a-1
LC-06-a-2
LC-06-a-3
LC-07-a-1
LC-07-a-2
LC-07-a-3
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
23
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
287
289
364
297
304
304
293
255
282
259
270
294
301
287
268
300
292
299
274
278
276
257
271
268
278
280
257
304
284
289
282
291
294
279
248
263
276
278
273
273
269
261
157
40
85
120
120
60
47
105
120
106
120
120
120
120
120
120
120
180
120
120
120
60
60
120
60
60
44
120
60
60
120
79
60
120
60
60
120
60
60
120
60
60
37759
10825
34792
44948
46417
22852
17209
20427
14095
10112
34382
45766
47587
43882
34104
44222
45383
71383
39039
39745
42227
14728
34556
41062
18857
18444
10811
41511
19071
24268
37599
27619
20798
39623
14987
17257
40109
18589
17739
35615
18282
15956
1.48
0.77
0.82
0.56
0.41
0.36
0.36
1.74
1.58
2.04
1.33
0.51
0.31
0.56
1.22
0.61
0.51
0.46
0.97
0.87
1.02
0.82
0.66
1.48
0.77
0.71
0.82
0.97
0.72
0.51
1.12
0.61
0.56
1.07
0.92
0.72
1.28
0.52
0.77
1.23
0.71
0.87
232
292
293
308
308
308
308
361
330
320
388
388
388
388
358
358
358
317
395
395
325
357
397
154
225
256
288
395
394
394
393
392
393
380
387
387
324
392
391
383
382
383
163
37
119
146
151
74
56
57
43
32
89
118
123
113
95
124
127
225
99
101
130
41
87
267
84
72
38
105
48
62
96
71
53
104
39
45
124
47
45
93
48
42
107 ± 51
Fig. 3. Plot of Rock Abrasion Tool (RAT) grinding test results given as specific grind
energy (SGE, units J/mm3 ) ordered by decreasing mean value. KG = Kendrick Grade
samples; LC = Lawyer Canyon samples; and SH = Shumaker Grade samples. Mean
values are given by horizontal bars.
for Lawyer Canyon, 77% for Kendrick Grade, and 32% for Shumaker
Grade).
In Fig. 3, there is much less separation evident between grind
values for different rock types compared to Fig. 2, and also less
separation between samples from a given sample locality. Kendrick
Grade samples are the most grind resistant (1st and 4th strongest
in Fig. 3), whereas in terms of compressive strength they are 3rd
44 ± 13
111 ± 15
115 ± 18
142 ± 72
86 ± 45
115 ± 103
72 ± 30
73 ± 22
63 ± 36
72 ± 45
61 ± 28
and 6th strongest, respectively (Fig. 2). Repeat grinds of individual
targets have a high degree of scatter. Overall, the grind results have
an average standard deviation of 44.5% (expressed as a percentage
of the mean value), compared to an average standard deviation of
15.8% for the compressive strength tests.
The weakest SGE values of different targets have a high degree of overlap; in contrast, the strongest recorded values for each
sample show greater consistency. Despite the relatively high uncertainties in Fig. 3, a general trend in the Lawyer Canyon samples
is evident, i.e., LC-02, LC-01 > LC-05, LC-07. As with the compressive strength data, samples from Shumaker Grade span the range
of grind values. However, the weakest UCS samples, SH-02, is the
second strongest in terms of SGE. (Note sample SH-03 is not represented in the crushing tests as it was too weak for this type of
test). Weathered samples for each locality have grind values from
∼40 to 80% of the most pristine sample (53% for Lawyer Canyon,
78% for Kendrick Grade, and 38% for Shumaker Grade).
It is noteworthy that the degree of scatter exhibited by the SGE
values of this suite of basalts is higher than for the rock assemblage measured previously (Thomson et al., 2011, 2013). A likely
controlling factor of this variability is the high overall strength of
these samples. Our prior work indicates that the advancement of
the RAT grinding bit into the target rock operates with reduced effectiveness for rocks with UCS values greater than ∼150 MPa. For
136
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
samples at or exceeding this strength threshold, the RAT’s rate of
penetration into the target is minimal. As a check on the grind
depth values recorded by the Z -axis motor, independent depth
measurements were made using a depth micrometer. A comparison of these two sets of depth values is given in Table 3. It is
apparent that the independently measured depths are significantly
shallower than those recorded by the Z -axis stage (they are, on
average, 73% shallower). This difference is likely attributable to
compliance in the RAT system, i.e., strain accommodated in the
RAT motor, housing, and mounting stage assembly. In two thirds
of the grind tests, the micrometer measurements indicate that a
grind depth of 0.25 mm was not achieved. This means that the actual SGE values are higher than those reported in Table 2. Yet we
report the Z -axis depths and associated SGE values because these
are most comparable to the results from Mars. System compliance
is an issue with data from both the brassboard and flight models,
and therefore cannot be summarily excluded from consideration.
We also note that the specific grind energies we derived here
and in Thomson et al. (2013) are significantly lower (∼20–30%
lower) than previously reported SGE values (e.g., Arvidson et al.,
2004; McSween et al., 2006; Squyres et al., 2006; Wang et al.,
2006; Herkenhoff et al., 2009). A combination of factors accounts
for the differences, and they relate to how data reduction is performed. Previous analyses used a standard 26.5 V as the operating
voltage for the grind axis. Here, we compute the voltage based
on the average velocity of the grind axis, resulting in a value that
is typically around 10% lower. Also, the average no-load current
for the previously computed data were calculated based on a fit
produced over the minimum currents recorded in the grind data
product (as opposed to the calibration data product) until this minimum value settled. The difference for some cases is as high as
40% over the last 0.25 mm. In the data presented here, the noload current is strictly from the calibration data collected prior to
the physical grinding of the rock. SGE values for the lab tests presented here are thus consistent with both the lab and martian SGE
values reported in our initial work (Thomson et al., 2013).
3.2. Comparison of grind and compressive strength tests
When comparing the grind test results to the compressive
strength test results, the samples from Lawyer Canyon are rea-
Table 3
Comparison of grind depth values recorded by Z -axis motor versus independently.
Target ID
Z grind depth
(mm)
Independent depth
(mm)a
SH-01-a-1
SH-03-a-2
SH-03-a-3
KG-01-a-1
KG-01-a-2
KG-01-a-3
KG-01-a-4
SH-02-a-2
SH-02-a-3
KG-02-a-1
KG-02-a-2
KG-02-a-3
LC-01-a-1
LC-01-a-2
LC-01-a-3
LC-02-a-1
LC-02-a-2
LC-02-a-4
LC-03-a-1
LC-03-a-2
LC-03-a-3
LC-04-a-1
LC-04-a-2
LC-04-a-3
LC-05-a-1
LC-05-a-2
LC-05-a-3
LC-06-a-1
LC-06-a-2
LC-06-a-3
LC-07-a-1
LC-07-a-2
LC-07-a-3
1.479
1.581
2.040
1.327
0.510
0.306
0.561
0.612
0.510
0.458
0.969
0.868
1.021
0.816
0.663
1.478
0.765
0.817
0.970
0.715
0.510
1.123
0.612
0.561
1.073
0.917
0.715
1.276
0.515
0.765
1.225
0.714
0.867
0.762
0.762
1.549
0.432
0.076
0.075
0.001
0.508
0.152
0.127
0.127
0.178
0.203
0.178
0.025
0.762
0.203
0.152
0.254
0.127
0.025
0.356
0.050
0.051
0.457
0.025
0.051
0.508
0.203
0.127
0.533
0.127
0.127
a
Gray shaded fields indicate grind tests where a depth of 0.25 mm was not attained.
sonably consistent with the previously determined correlation
(Thomson et al., 2013) (Fig. 4). As expected, stronger samples
(LC-01, LC-02) plot toward the upper right portion of the plotted
field, while weaker samples (LC-05, LC-07) tend to fall toward the
lower left. Samples from Kendrick Grade (denoted by filled red circles in Fig. 4) conform to this general trend but are offset to higher
SGE values. The most anomalous data point is SH-02, a slightly
Fig. 4. Mean compressive strength values (in MPa) plotted against mean specific grind energy (SGE) values (in J/mm3 ). Left-hand figure is a duplicate of right-hand figure
with error bars omitted for clarity. Error bars on right-hand figure are ± one standard deviation of the mean value. Dashed and dot-dashed lines give power law and linear
fits to the calibration data of Thomson et al. (2013). Shaded field on right-hand plot is measured SGE values of Adirondack-class basalts in Gusev crater measured by the
MER Spirit; concomitant compressive strength values calculated from fits to calibration data.
741
2
324
396
33
174
Ba
Be
Sr
V
Y
Zr
Notes: ∗ Sample designator ‘PS’ = paleosol samples. ND = Not Detected. Oxide values are reported in %; elemental values are in ppm. Detection limits are 0.01% for SiO2 , Al2 O3 , Fe2 O3(T ) , MgO, CaO, Na2 O, K2 O, and P2 O5 ; 0.001% for
TiO2 and MnO; 1 ppm for Ba, Be, and Y; 2 ppm for Sr and Zr; and 5 ppm for V.
1071
1
96
306
15
150
515
ND
139
358
38
119
442
ND
156
292
26
102
377
ND
328
333
23
86
285
ND
357
339
25
97
726
1
339
408
33
176
748
1
367
414
36
185
439
2
72
254
21
465
853
2
375
417
38
223
870
2
362
413
46
214
780
2
344
409
33
184
820
2
367
400
38
204
784
2
332
402
33
177
54.38
2.313
13.23
14.11
0.207
3.26
7.12
3.19
1.73
0.45
ND
99.7
SiO2
TiO2
Al2 O3
Fe2 O3(T )
MnO
MgO
CaO
Na2 O
K2 O
P2 O5
LOI
Total
757
2
329
387
36
196
41.55
1.492
17.16
13.2
0.184
2.54
4.14
0.83
0.2
0.7
16.1
98.09
41.88
1.491
17.52
14.83
0.172
3.32
4.85
0.99
0.2
0.27
13.3
98.83
47.23
1.252
14.29
12.63
0.263
4.62
10.66
2.19
0.21
0.27
6.02
99.63
49.78
1.251
14.49
11.42
0.161
6.44
10.62
2.58
0 .3
0 .3
3.25
100.6
53.26
2.126
13.14
12.66
0.197
3.75
7 .7
2.77
1.51
0.42
1.18
98.71
52.8
2.268
14.34
12.9
0.191
3.85
7.86
2.88
1.14
0.43
2.11
100.8
40.28
2.278
26.28
12.69
0.059
0.33
0.66
0.08
0.05
0.13
17
99.85
55.32
2.516
14.34
10.73
0.089
1.85
5.82
3.04
1.86
0.52
3.04
99.13
55.92
2.399
14.01
11.26
0.124
1.97
5.92
2.99
1.98
0.52
2.97
100.1
52.73
2.401
13.52
13.77
0.118
1.83
5.82
2.77
1.65
0.45
4.33
99.39
54.18
2.458
14.41
12.79
0.132
2.41
6.18
3.04
1.77
0.39
2.56
100.3
54.29
2.347
13
12.96
0.195
3.15
7.05
2.9
1.87
0.46
0.57
98.8
KG-01
LC-PS∗
LC-07
LC-06
LC-05
LC-04
LC-03
LC-02
LC-01
Systematic trends in bulk density were also noted with increasing degree of weathering (Fig. 6), particularly for the Lawyer
Canyon (LC) samples. The difference between the most pristine
and most weathered samples is about 0.4 g/cm3 , equivalent to a
loss of ∼15% of the initial bulk density of 2.7 g/cm3 . This density
change is comparable to the difference in density between fresh
137
Oxide/
element
3.4. Bulk density
Table 4
Elements analyzed by ICP-OES following fusion digestion.
To help further quantify the degree of alteration, we calculated the chemical index of alteration (CIA = molar Al2 O3 /(Al2 O3 +
Na2 O + CaO + K2 O)) from the elemental abundances determined
by ICP/OES (Table 4). Figs. 5a–b give the CIA as a function of
compressive strength and grind energy, respectively. The highest
recorded CIA values were found in paleosol layers; these are excluded from Fig. 5 because they were not sufficiently competent to
conduct crushing tests or grind tests. However, based on their observed field characteristics (i.e., they crumbled under blows by the
point of a geologic hammer), their UCS value could be generally
inferred to be less than about 5 MPa, and their expected SGE values would be accordingly very low, ∼1–2 J/mm3 , consistent with
very weakly consolidated material. Hence, these paleosols would
plot above the upper left corner of both Figs. 5a and 5b.
In Fig. 5a, there is a distinct separation between less weathered samples (LC-01, LC-02, LC-03) and more weathered samples
(LC-04 to LC-07). This general trend of weathering is consistent
with the position of most weathered samples (paleosols) plotting
above the upper left corner of the plot. In comparison, although in
Fig. 5b some separation is still evident between the most weathered samples (LC-05, LC-07) and least weathered (LC-01, LC-02),
there is much more overlap for intermediate cases (LC-03, 04, 06)
compared to Fig. 5a. CIA values, defined as the difference between the most pristine and most weathered sample from a given
locality, are 5.2 for Lawyer Canyon, 2.0 for Kendrick Grade, and 0.3
for Shumaker Grade (SH-02 minus SH-01). Samples SH-03 has a
CIA value of 62.2, which is intermediate between the rock samples and the paleosols at Lawyer Canyon and Shumaker Grade that
have CIA values in the range from 87 to 95.
Sample SH-03 was too weak to conduct compressive tests on
but sufficiently competent for grind tests; its value lies in between
paleosols and more competent basalts. But apart from this sample,
there is a large gap in terms of strength parameters and CIA values
between paleosols and intact but weathered basalt samples. It is
likely that the collection and measurement of additional samples
at a finer sampling interval (e.g., every 2 cm below paleosol) might
reveal more transitional states.
KG-02
3.3. Chemical index of alteration (CIA)
54.95
2.299
13.06
13.34
0.192
3.17
6.92
3.15
1.74
0.41
0.07
99.32
SH-01
SH-02
SH-03
SH-05
SH-PS∗
more weathered sample from the Upper Shumaker site. This sample has a low compressive strength yet unusually high grind resistivity. It may be that macroscopic weakness (such as porosity or
pre-existing fractures) lowered the compressive strength but did
not comparably affect its grindability. Opaline silica was also observed to cement the paleosol soils at the Upper Shumaker site,
and it is possible that opal was deposited in pore spaces of this
rock sample and affected its grind resistivity.
Also given in the right-hand panel of Fig. 4 is a shaded rectangular field indicating the measured SGE values of Adirondack-class
basalts and interpreted UCS values from Thomson et al. (2013).
The Adirondack-class basalts Adirondack, Humphrey, and Mazatzal
have recorded SGE values that range from ∼30 to 45 J/mm3 ; their
inferred compressive strengths range from about 70–130 MPa. In
terms of strength, Adirondack-class basalts are similar to samples
LC-05 and LC-07, which are the most weathered but intact basalt
samples measured in this study (excluding paleosols or saprolite).
35.59
1.529
19.26
15.32
0.049
1.13
1.45
0.06
0.11
0.21
24.68
99.38
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
138
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
Table 5
Elements analyzed by INAA.
Analyte
symbol
LC-01
LC-02
LC-03
LC-04
LC-05
LC-06
LC-07
LC-PS∗
KG-01
KG-02
SH-01
SH-02
SH-03
SH-05
SH-PS∗
Au
As
Br
Co
Cr
Cs
Hf
Hg
Ir
Rb
Sb
Sc
Se
Ta
Th
U
W
La
Ce
Nd
Sm
Eu
Tb
Yb
Lu
Au
ND
ND
ND
36.1
8 .3
1 .2
5 .4
ND
ND
60
ND
34
ND
1 .2
5 .5
1
ND
30.8
56
25
7.12
2.33
1 .2
3.47
0.55
ND
ND
1
ND
36.7
ND
1.6
5.2
ND
ND
40
ND
33.9
ND
1
5.8
2
ND
30.4
56
23
7.15
2.4
1.4
3.53
0.57
ND
ND
ND
ND
39.2
5.7
2.9
5.2
ND
ND
50
0.3
33.4
ND
1
5.3
1.9
ND
30.1
55
25
6.97
2.34
1.4
3.3
0.53
ND
ND
ND
ND
33.8
4.5
2
5.5
ND
ND
50
ND
33.9
ND
ND
6.6
1.8
ND
32
56
29
7.28
2.5
1.2
3.61
0.58
ND
ND
ND
ND
30
ND
2.4
5.8
ND
ND
50
0 .3
33
ND
0 .9
6.5
1.3
ND
33.1
63
31
7.78
2.68
1.3
3.5
0.52
ND
10
3
ND
43.6
ND
2.1
5.8
ND
ND
70
3.5
35.9
ND
ND
6.7
1.8
ND
35.1
65
32
8.26
2.88
1.6
4.7
0.78
10
ND
ND
ND
41.9
ND
2.8
5.7
ND
ND
50
ND
35.5
ND
1
6.5
2
ND
34.5
61
31
8.12
2.77
1.8
3.75
0.57
ND
2
4
ND
20.7
10.4
ND
11.3
ND
ND
ND
0 .9
33.9
ND
1.5
11.4
2
ND
35.3
87
35
9.72
2.79
ND
2.89
0.44
2
ND
ND
1.9
41.7
11.8
ND
4.8
ND
ND
30
0.2
38.9
ND
0.9
4.6
1
ND
25
46
19
6.72
2.43
1.3
3.5
0.56
ND
ND
2
ND
38.3
14.1
2
4 .6
ND
ND
40
ND
36.7
ND
ND
4 .3
ND
ND
24.2
44
21
6.38
2.19
1 .2
3.44
0.53
ND
ND
ND
ND
41.3
218
ND
2.4
ND
ND
ND
ND
44.7
ND
ND
0.7
ND
ND
13.2
26
15
4.29
1.58
ND
2.68
0.44
ND
ND
ND
ND
41.3
230
1 .2
2 .3
ND
ND
ND
ND
43.7
ND
ND
0.6
ND
ND
12.8
27
14
4.14
1.53
0 .8
2.82
0.41
ND
ND
ND
ND
49.7
436
ND
3
ND
ND
ND
ND
53.6
ND
0 .4
0.9
ND
ND
13.1
30
18
5 .2
1.85
1
3 .4
0.51
ND
3
ND
ND
41.9
382
ND
3.1
ND
ND
10
ND
52.3
ND
0.5
0.7
ND
ND
16
33
22
5.57
2.05
1.1
3.98
0.65
3
ND
2
ND
29.8
141
1.1
3.9
ND
ND
ND
ND
48.4
ND
0.5
2
0 .9
ND
12.2
29
15
3.71
1.22
0 .8
3.44
0.56
ND
Notes: ∗ Sample designator ‘PS’ = paleosol samples. ND = Not Detected. Elements Au and Ir are reported in ppb; remainder are in ppm. Detection limits are 10 ppm for Rb;
1 ppm for As, Ce, Hg, Nd, and W; 0.5 ppm for Br, Cr, and Se; 0.3 ppm for Ta; 0.2 ppm for Cs and Hf; and 0.1 ppm for Co, Sb, Tb, Th, and U; 0.05 ppm for Eu, La, and Yb;
0.01 ppm for Lu, Sc, and Sm; and 1 ppb for Au and Ir.
Table 6
Elements analyzed by ICP-OES following acid digestion.
Analyte
symbol
LC-01
LC-02
LC-03
LC-04
LC-05
LC-06
LC-07
LC-PS∗
KG-01
KG-02
SH-01
SH-02
SH-03
SH-05
SH-PS∗
Ag
Bi
Cd
Cu
Mo
Ni
Pb
S
Zn
ND
ND
ND
62
ND
12
9
0.058
120
ND
ND
ND
26
ND
12
10
0.035
114
ND
ND
ND
27
ND
14
9
0.034
121
0.5
7
ND
30
ND
11
6
0.027
131
ND
ND
ND
24
ND
11
11
0.025
146
ND
ND
ND
40
ND
10
10
0.036
179
ND
ND
ND
24
ND
10
11
0.028
147
0.6
ND
ND
50
ND
31
29
0.004
92
ND
ND
0.6
23
ND
12
10
0.036
122
ND
ND
ND
20
ND
12
ND
0.041
111
ND
ND
ND
87
ND
50
ND
0.044
76
ND
ND
ND
76
ND
50
ND
0.04
81
ND
ND
ND
93
ND
50
ND
0.02
94
ND
4
ND
77
ND
38
ND
0.016
80
ND
ND
ND
79
ND
83
5
0.032
101
Notes: ∗ Sample designator ‘PS’ = paleosol samples. ND = Not Detected. Elements S is reported in %; remainder are in ppm. Detection limits are 0.001% for S; 5 ppm for Pb;
2 ppm for Bi and Mo; 1 ppm for Cu, Ni, and Zn; and 0.5 ppm for Ag and Cd.
Fig. 5. (a) Chemical index of alteration versus compressive strength. (b) Chemical index of alteration versus grind energy.
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
139
Fig. 6. (a) Compressive strength versus bulk density. (b) Grind energy versus bulk density. (c) Chemical index of alteration versus bulk density.
and slightly weathered samples of 2 Ma old basalts (with densities
of 2.8 and 2.1 g/cm3 , respectively) measured by Moon and Jayawardane (2004).
3.5. Depth profile and mobility ratios
In this section, we focus our attention on the samples from
Lawyer Canyon (LC). Samples LC-01 to LC-07 were obtained in a
vertical profile of progressively weathered basalt, starting at the
bottom from LC-01, the least weathered sample of the protolith,
extending 8 m upwards to sample LC-07, the most weathered.
However, because weathering into the intact basalt proceeded in a
dendritic pattern along fractures and defect planes (i.e., the weathering front is not strictly planar), the sequence of weathering states
may not exactly match the stratigraphic height-ordered sequence.
Above these rock samples, the paleosol layer was also sampled (LCpaleosol).
Fig. 7 gives major oxides as a function of depth for the Lawyer
Canyon series. The profile exhibits a steady loss of MgO, CaO, and
K2 O as one approaches the paleosol layer, consistent with downward leaching and loss of mobile cations. Total iron exhibits a
more complicated pattern. It generally declines, but it is slightly
elevated in samples LC-05 and in the paleosol relative to the materials directly beneath them (though still below the initial value
in the protolith). Al2 O3 steadily increases in proportion toward the
paleosol layer, presumably as more mobile cations are preferentially removed. Silica is strongly depleted in the paleosol layer but
slightly enhanced in the two samples below it, suggesting perhaps
both mobilization and deposition of silica at different rates in different portions of the column.
An alternative method of examining elemental changes or
fractionation is to compute the constitutive mass balance (e.g.,
Brimhall et al., 1992; Sheldon, 2003). By indexing the concentration of a given element with respect to a relatively immobile one,
the gains and losses of mobile elements compared to the immobile element can be determined. The mass transport function for
the jth element in the weathered sample (w) is defined as follows:
τ j, w =
ρ w C j, w
(εi , w + 1) − 1
ρ p C j, p
(1)
where ρ p and ρ w are the densities of the parent and weathered
material, respectively, and C j , p and C j , w are the weight percentages of the jth element in the parent and weathered material,
respectively. Strain (ε )i , w is defined by the volumetric change during weathering:
εi , w =
V
V
=
ρ p C i, p
− 1.
ρ w C i, w
(2)
Substituting and rearranging equations 1–2 yields an expression
for the mobility ratio (MR) of the jth element without density or
volume-dependent terms:
MR j , w =
C j , w /C i , w
C j , p /C i , p
− 1.
(3)
In Eq. (3), we use the highly immobile element Zr as the index element C i . The resulting mobility ratios plotted as a function of sample depth (Fig. 8) indicate that all major cations have
been depleted relative to the least weathered parent material. Total loss relative to Zr due to dissolution in the most weathered
140
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
Fig. 8. Mobility ratio calculated according to Eq. (3) for Lawyer Canyon (LC) samples indicating cation gain or loss compared to least weathered material (LC-01)
arranged in order of sample depth below top of paleosol surface. Index element
(assumed to be immobile) is Zr.
Fig. 7. Major oxides versus depth below top of paleosol surface for Lawyer Canyon
samples.
non-paleosol sample (LC-07) includes 65% MnO; 54% MgO; 38%
Fe2 O3(T ) ; 34% CaO; 23% Na2 O; 17% SiO2 ; and 12% Al2 O3 . Minor
oxides K2 O, TiO2 , and P2 O5 also exhibit losses of 13%, 12%, and 6%,
respectively.
Supporting this interpretation is a ternary diagram designed to
evaluate the major element changes that occur with weathering.
Fig. 9 gives the molar proportions of Al2 O3 , (CaO + Na2 O + K2 O),
and (FeO T + MgO). The weathering trend observed in the Lawyer
Canyon series follows two distinct stages or segments. In the first
segment, the trend is radially away from the FeO T + MgO apex,
a trend consistent with early removal of ferromagnesian minerals
(clinopyroxene or groundmass olivine) or glass during incipient alteration. In the second segment, the trend is nearly orthogonal to
the first segment, consistent with bulk removal of soluble elements
and production of clays.
3.6. Mineralogy from X-ray diffraction
Fig. 9. Ternary Al2 O3 , (CaO + Na2 O + K2 O), and (FeOT + MgO) diagram, data plotted in mole percent. Solid black arrow labeled ‘A1’ indicates a weathering trend in
Lawyer Canyon samples from LC-01 to LC-07 consistent with dissolution of clinopyroxene phenocrysts or groundmass olivine. Dashed black arrow labeled ‘A2’ indicates weathering trend required to reach Lawyer Canyon paleosol composition that
is consistent with feldspar removal and clay deposition. Dashed black arrow labeled
‘B1’ is weathering trend displayed by Shumaker grade samples; no indication of an
early weathering phase dominated by loss of ferromagnesian minerals is evident.
4. Discussion
4.1. Styles of alteration
Analysis of three separate aliquots of sample LC-01 (the presumed protolith) reveals that the sample is composed predominantly of glass (58–70%), clinopyroxene (19–25%), plagioclase
(10–17%), and ilmenite (0–2%). The structural data for clinopyroxene, plagioclase, and ilmenite that best fit the LC-01 XRD pattern
are from Takeda (1972), Stewart et al. (1966), and Wechsler and
Prewitt (1984), respectively.
Several contrasting views of martian weathering processes have
been advanced. Based on the fact that the secondary mineral budget at the MER landing sites is dominated by Mg, Fe, and Ca
sulfates plus Fe-oxides, one view is that limited chemical weathering proceeded under a low-pH, sulfuric acid-rich environment
(e.g., Tosca et al., 2004; Ming et al., 2006; Hurowitz and McLennan,
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
2007; Hausrath et al., 2008). Alteration is inferred to have occurred
in water-limited, rock-dominated conditions, akin to a closed system model. An alternative perspective is a pedogenic model where
aqueous alteration of the soil driven by atmospherically-sourced
water, resulting in top-down weathering and vertical partitioning of atmospheric solutes (Amundson et al., 2008). Such an open
system of chemical alteration is inferred to have been more active earlier in Gusev history, but with continued accumulation of
soluble SO3 and Cl-bearing salts. A third perspective stresses the
importance of physical fractionation processes in accounting for
certain observed gains or losses of iron and magnesium in local
soils (McGlynn et al., 2011, 2012).
In terms of style of alteration, some aspects of the weathering
trend at Lawyer Canyon are consistent with open system alteration. All cations are not fully accounted for, meaning that cation
loss from the system via migrating fluids seems inevitable. Yet the
high water:rock ratios necessary to mobilize Al in circum-neutral
pH fluids are not necessarily required to account for the degree
of weathering-induced weakening observed in the vertical profile
below the saprolite layers. True open system weathering is more
like that required to form the (heavily weathered) soils and paleosol layers, and comparable, deeply weathered soils (e.g., fines
dominated by clay minerals) do not appear widespread on Mars.
Kaolinite-bearing areas have been identified in regions like Mawrth
Vallis, but outside a limited number of regional exposures, they appear restricted to isolated outcrops associated with impact craters.
As calculated by McGlynn et al. (2012), martian soils cannot be
completely explained by open system or acid–sulfur alteration;
some compositional modification by hydrodynamic sorting and admixture of secondary components is necessary (i.e., a non-local
origin).
What can we infer about the nature of weathering on Mars
from the strength and chemistry of Adirondack basalts? In terms
of mechanical strength, Adirondack basalts are most similar to the
weakest, most weathered non-paleosol samples measured in our
study (LC-05, LC-07). Strength measurements indicate that comparably weathered terrestrial basalt samples from Lawyer Canyon
only retain 40–50% of their intact strength, 15% of their initial bulk
density, and have also lost upwards of half of some mobile cations,
such as 54% of MgO, 38% of Fe2 O3(T ) , and 34% of the initial CaO.
To assess the degree of chemical alteration that Adirondack
basalts have experienced, we recast the mobility ratios using Ti
as an index element in Fig. 10 to facilitate direct comparison with
the CRB samples. While Ti is not completely immobile, it is among
the least mobile minor oxides (e.g., Amundson et al., 2008). The
gray shaded field in Fig. 10 gives the range of mobility ratio values of the brushed versus RAT-abraded surfaces of Adirondack and
Humphrey; most cations are within plus or minus 0.2 (excluding
enrichments in K2 O, which are interesting in their own right). Considering Mg as our most sensitive tracer of alteration, Adirondackclass basalts are therefore intermediate in Mg loss (mobility ratio
of −0.14) between LC-03 and LC-04 (mobility ratios of −0.05 and
−0.3, respectively). It is clear from these data that the brushed vs.
abraded surfaces of Adirondack-class basalts do not encompass the
same degree of alteration evidenced by the transition from LC-01
to LC-07.
4.2. Mismatch between physical and chemical weathering
We can envision three possible explanations for the observed
weakness of the Adirondack class basalts. First, the weakening may
be entirely attributable to physical weathering such as shock effects (e.g., French, 1998) during the emplacement process or subsequent salt weathering (Malin, 1974). Second, Adirondack basalts
may be samples from a soil-like weathering profile developed in a
basalt layer reflective of deep (i.e., meter length scale) downward
141
Fig. 10. Mobility ratio calculated according to Eq. (3) for Lawyer Canyon (LC) samples indicating cation gain or loss compared to least weathered material (LC-01).
Index element (assumed to be immobile) is Ti. Gray shaded field represents maximal cation gain or loss (excluding K2 O) of Adirondack-class basalts comparing
brushed surfaces to those abraded by the RAT. Specifically, the samples considered
are Adirondack brush and RAT, Humphrey brush and RAT1, and Humphrey brush at
RAT2. All data are normalized to 100% to a LOI-free basis for Lawyer Canyon, and
SO3 - and Cl-free basis for the Adirondack-class basalt. Vertical dashed lines on right
mark positive mobility ratio or Tau values for K2 O.
percolation of fluids (e.g., Amundson et al., 2008), followed by later
disaggregation into isolated blocks. A third alternative is that the
observed weakening is limited to a several mm to cm-thick rind
on exposed basaltic clasts sampled by Spirit. In this case, we are
looking at a moderately weathered, weak layer that encases more
pristine (and unsampled) corestones.
Although we consider it unlikely that shock-induced weakening
is solely responsible due to geometric considerations (especially
given the self-similarity of the four Adirondack basalt samples),
some degree of shock-related weakening is to be expected. Some
inferences on the strength of shocked samples can be gleaned
from terrestrial meteorite strength data. Although such data is
sparse (due to an understandable reluctance to perform destructive
strength tests on rare meteorite samples), the limited data available suggests average compressive strength values of ∼200 MPa for
stony meteorites (e.g., Buddhue, 1942; Petrovic, 2001; Popova et
al., 2011). Indeed, the compressive strength of the strongest stony
meteorite samples overlaps with the weakest iron meteorites, despite the fact that iron meteorites are typically more than twice as
strong as stony meteorites. Strength data for meteorites is clearly
somewhat biased toward higher values since only fragments that
survive ejection, atmospheric entry, and landing are available for
study. Nonetheless, the results suggest that the low to moderate
levels of shock do not uniformly weaken the target material, and
thus support our contention that not all of the observed weakening is necessarily shock-induced.
The effects of salt weathering on individual rock clasts are more
difficult to assess. Dissolved salts can percolate into even nominally nonporous media, and subsequent pressure from salt crystal growth, absorption of water by anhydrous salts, or thermal
142
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
cycling can exert pressure and induce cracking on infiltrated surfaces (e.g., Rodriguez-Navarro and Doehne, 1999). Small veins filled
with light-toned material were observed on brushed and abraded
surfaces of Adirondack basalts (e.g., Herkenhoff et al., 2004), and
polygonal cracks filled with secondary minerals observed in Meridiani sandstones have been suggested to be the result of cyclical
salt-atmosphere water exchange, i.e., repeated episodes of condensations, dissolution, and reprecipitation (Chavdarian and Sumner,
2006). Salt weathering may therefore play a role in the weakening
of the outer surfaces of rocks on Mars (e.g., Malin, 1974), though
little data exists to help quantify the potential magnitude of this
process.
Assuming that all of the weakening is not attributable to physical weathering, the reason that a distinction between the second
and third scenarios matters is one of timing. If Adirondack basalts
were weathered and weakened in a soil-like context, then these
effects are likely related to an early phase of weathering on Mars
(but still in the Hesperian since it had to have occurred after the
lava flows that filled Gusev crater were emplaced (Milam et al.,
2003)). However, if the third scenario is correct, this provides little to no age control for the weathering. It could be recent, or
it could be significantly older. Exactly how old it could be is unclear, as this has to do with how long one would expect a rock to
last on the surface of Mars. On the Moon, this age is only a few
Ma due to micrometeorites (e.g., Horz, 1977), but the atmosphere
of Mars shields surface rocks to some degree and therefore much
older rock ages are likely permissible. Using a reaction-transport
model, a minimum cumulative exposure age of 2.2 × 105 years
has been estimated for the formation of the weathering rind on
Humphrey (an Adirondack-class basalt) (Hausrath et al., 2008). If
the material assumed to be the unaltered protolith is itself weathered to the degree such that it is mechanically similar to samples
LC-05 and LC-07 from this work, the required weathering time
would be longer, perhaps at least by a factor of two (and potentially >106 yr).
A key discriminator between the second and third scenarios is
the mismatch between strength and chemical data. If the physical
and chemical weathering were due to the development of a soillike weathering profile in a basalt layer, we would not expect a
mismatch between the degree of weathering implied by the mechanical strength versus that implied by the degree of chemical alteration. Though such a process may have been active, one or both
of the other two explanations (i.e., additional physical weathering
or thick rinds) are necessary to explain the totality of available evidence.
In summary, this third hypothesis (i.e., Adirondack basalts are
encased in a several mm-thick weak weathering rind) appears
most consistent with the available evidence and known history
of activity in Gusev crater. Considering this latter possibility, we
can infer some general characteristics of this hypothetical parent
composition, but acknowledge that our ability to make firm inferences about its precise nature is hampered by several factors.
First, the parental and weathered bulk compositions of the CRB
and Adirondack basalt are different, and second, their mineralogies
are likely dissimilar as well. These two factors mean that alteration
would proceed along different geochemical pathways in the two
rocks even if the style of weathering were identical, particularly
if weathering proceeded under water-limited, low pH conditions
(e.g., Nesbitt and Wilson, 1992; Hurowitz and Fischer, 2014).
If we set aside these caveats and assume all of the observed
weakening is due to substantially similar chemical weathering, the
inferred Adirondack protolith would have higher Fe, Mg, Mn, Si and
proportionally lower amounts of the remaining cations. Total iron
would exceed 20 wt%, and MgO and MnO would roughly double.
In a total alkali versus silica diagram, the inferred compositions
would fall within the picrobasalt field (40–45 wt% SiO2 ). McSween
et al. (2006) inferred similar but not-quite-as-silica-poor ‘endmember’ compositions for Adirondack, Humphrey, and Mazatzal with
SiO2 values ranging from 45.3 to 45.9 wt%. These endmember compositions were attained by extrapolating linear trends from the
brushed to RAT-abraded surfaces such that residual sulfur was reduced to 0.3 wt%, akin to subtracting a portion of the brushed
composition values from the abraded values. On Earth, comparable
volcanic rocks with magnesium-rich, low-silica compositions are
largely confined to the Archean period (e.g., komatiites), and are
associated with large degrees of partial melting of an undepleted
mantle source (e.g., Nisbet, 1982). Without independent corroborating evidence, however, it would be an overstatement to conclude that these results constituted evidence for martian komatiites. Nevertheless, the results hint at the fact that the unweathered
cores of basalts on Mars may trend towards more ultramafic compositions than otherwise indicated.
4.3. Suggestions for grinders on future Mars missions
Given the successful operation of the RAT instrument on both
Spirit and Opportunity, we recommend including a grinder on future missions (e.g., Mars2020) in order to better interrogate rock
coatings and weathering rinds. The Mars Science Laboratory rover
Curiosity posses a brush but no grinder (Grotzinger et al., 2012),
and the rover’s drill system does not retain sample cuttings from
the first ∼1.5 cm of target penetration (Anderson et al., 2012).
Based on MER RAT experience, future grinders should be designed
to penetrate to a minimum of ∼2–3 mm depth in all rock types,
which will facilitate removal of cuttings and allow for more accurate estimates of specific grind energy (a proxy for rock hardness). In addition, rock hardness can be further quantified using
by computing G-ratio values based on abrasive pad loss. In general, high-resolution images of both the RAT hole (to assess the
volume abraded using stereo photogrammetry-derived DEMs) and
the grinding pads (to calculate G-ratio based on pad wear) would
provide more information about the target and permit more careful monitoring of grinding pad status. For the latter calculation,
images with a spatial resolution <30 μm per pixel (but preferably
higher) of the grinding pads are recommended, especially in targets with SGE values >20 J/mm3 where more bit wear is expected.
5. Conclusions
Investigations of the linked chemical and strength changes experienced by Columbia River Basalts help inform our understanding of martian weathering processes. Our results provide a basis
for comparison with the MER Spirit results, and there is a notable
mismatch in the degree of alteration implied by rock strength inferred from RAT grinds versus that indicated by APXS data recast
into mobility ratios (with the former indicative of greater alteration
than the latter). These results suggest that Adirondack-class basalts
on the plains of Gusev crater have experienced a significant degree
of weathering-induced weakening (in addition to some likely shock
effects and potentially salt weathering). Underneath a thin rind of
sulfur-rich dust, the basalts sampled by Spirit appear encased in a
weaker, reduced-density rind that is at least several mm thick.
Acknowledgements
We thank Bill Phillips and the Idaho Geologic Survey for loaning us a portable drill corer for sample collection, Bill Rember and
Kevin Hobbs (U. Idaho) for assistance in field site identification
and sample collection, and Robert Macke (Boston College/Vatican)
for conducting density measurements. Constructive comments by
reviewer Paul Niles and an anonymous reviewer helped sharpen
several aspects of this study. This work was supported by a NASA
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
Fundamental Research Program grant to BJT (award NNX11AP46G).
Special thanks to David Nava (NASA/Goddard) for having twice
helped facilitate the transfer of funds from one institution to another.
References
Amundson, R., Ewing, S., Dietrich, W., Sutter, B., Owen, J., Chadwick, O., Nishiizumi, K., Walvoord, M., McKay, C., 2008. On the in situ aqueous alteration of
soils on Mars. Geochim. Cosmochim. Acta 72 (15), 3845–3864. http://dx.doi.org/
10.1016/j.gca.2008.04.038.
Anderson, R.C., Jandura, L., Okon, A.B., Sunshine, D., Roumeliotis, C., Beegle, L.W.,
Hurowitz, J., Kennedy, B., Limonadi, D., McCloskey, S., Robinson, M., Seybold,
C., Brown, K., 2012. Collecting samples in Gale crater, Mars an overview of the
Mars; Science Laboratory Sample Acquisition, Sample Processing and Handling
System. Space Sci. Rev., 57–75.
Arvidson, R.E., Anderson, R.C., Bartlett, P., Bell, J.F., Blaney, D., Christensen, P.R., Chu,
P., Crumpler, L., Davis, K., Ehlmann, B.L., Fergason, R., Golombek, M.P., Gorevan,
S., Grant, J.A., Greeley, R., Guinness, E.A., Haldemann, A.F.C., Herkenhoff, K., Johnson, J., Landis, G., Li, R., Lindemann, R., McSween, H., Ming, D.W., Myrick, T.,
Richter, L., Seelos, F.P., Squyres, S.W., Sullivan, R.J., Wang, A., Wilson, J., 2004.
Localization and physical properties experiments conducted by Spirit at Gusev
Crater. Science 305 (5685), 821–824.
Baker, L.L., Agenbroad, D.J., Wood, S.A., 2000. Experimental hydrothermal alteration
of a martian analog basalt: Implications for martian meteorites. Meteorit. Planet.
Sci. 35 (1), 31–38.
Bandfield, J.L., 2002. Global mineral distributions on Mars. J. Geophys. Res. 107 (E6),
5042. http://dx.doi.org/10.1029/2001JE001510.
Basaltic Volcanism Study Project, 1981. Basaltic, Volcanism on the Terrestrial Planets. Pergamon Press, Inc., New York. 1286 pp.
Bell, J.F., Squyres, S.W., Arvidson, R.E., Arneson, H.M., Bass, D., Blaney, D., Cabrol, N.,
Calvin, W., Farmer, J., Farrand, W.H., Goetz, W., Golombek, M., Grant, J.A., Greeley, R., Guinness, E., Hayes, A.G., Hubbard, M.Y.H., Herkenhoff, K.E., Johnson, M.J.,
Johnson, J.R., Joseph, J., Kinch, K.M., Lemmon, M.T., Li, R., Madsen, M.B., Maki,
J.N., Malin, M., McCartney, E., McLennan, S., McSween, H.Y., Ming, D.W., Moersch, J.E., Morris, R.V., Noe Dobrea, E.Z., Parker, T.J., Proton, J., Rice, J.W., Seelos,
F., Soderblom, J., Soderblom, L.A., Sohl-Dickstein, J.N., Sullivan, R.J., Wolff, M.J.,
Wang, A., 2004. Pancam multispectral imaging results from the Spirit Rover at
Gusev Crater. Science 305 (5685), 800–806.
Brimhall, G.H., Chadwick, O.A., Lewis, C.J., Compston, W., Williams, I.S., Danti, K.J.,
Dietrich, W.E., Power, M.E., Hendricks, D., Bratt, J., 1992. Deformational masstransport and invasive processes in soil evolution. Science 255 (5045), 695–702.
Buddhue, J.D., 1942. The compressive strength of meteorites. Contrib. Soc. Res. Meteor. 3 (8), 390–391.
Burns, E.R., Morgan, D.S., Peavler, R.S., Kahle, S.C., 2010. Three-dimensional digital geomodel of the Columbia Plateau Regional Aquifer system, Idaho, Oregon,
and Washington, U.S. U.S. Geological Survey Scientific Investigations Report,
2010-5246. USGS, Reston, VA.
Bush, J.H., Kauffman, J.D., Schmidt, K.L., 2004. Geologic Map of the Craigmont
Quadrangle, Lewis and Idaho Counties, Idaho, 1:24,000. Idaho Geologic Survey,
Moscow, Idaho.
Camp, V.E., Hooper, P.R., 1981. Geologic studies of the Columbia Plateau: Part 1. Late
Cenozoic evolution of the southeast part of the Columbia River Basalt Province.
Geol. Soc. Am. Bull. 92 (9), 659–668.
Chavdarian, G.V., Sumner, D.Y., 2006. Cracks and fins in sulfate sand: evidence for
recent mineral-atmospheric water cycling in Meridiani Planum outcrops? Geology 34 (4), 229–232. http://dx.doi.org/10.1130/G22101.1.
Christensen, P.R., Bandfield, J.L., Smith, M.D., Hamilton, V.E., Clark, R.N., 2000. Identification of a basaltic component on the Martian surface from Thermal Emission
Spectrometer data. J. Geophys. Res. 105 (E4), 9609–9621.
Christensen, P.R., Ruff, S.W., Fergason, R.L., Knudson, A.T., Anwar, S., Arvidson, R.E.,
Bandfield, J.L., Blaney, D.L., Budney, C., Calvin, W.M., Glotch, T.D., Golombek, M.P.,
Gorelick, N., Graff, T.G., Hamilton, V.E., Hayes, A., Johnson, J.R., McSween, H.Y.,
Mehall, G.L., Mehall, L.K., Moersch, J.E., Morris, R.V., Rogers, A.D., Smith, M.D.,
Squyres, S.W., Wolff, M.J., Wyatt, M.B., 2004. Initial results from the Mini-TES
Experiment in Gusev Crater from the Spirit Rover. Science 305 (5685), 837–842.
Clark, B.C., Baird, A.K., Weldon, R.J., Tsusaki, D.M., Schnabel, L., Candelaria, M.P.,
1982. Chemical composition of Martian fines. J. Geophys. Res. 87, 10059–10067.
Eggleton, R.A., Foudoulis, C., Varkevisser, D., 1987. Weathering of basalt: changes in
rock chemistry and mineralogy. Clays Clay Miner. 35 (3), 161–169.
Elwood Madden, M.E., Bodnar, R.J., Rimstidt, J.D., 2004. Jarosite as an indicator of water-limited chemical weathering on Mars. Nature 431, 821–823.
http://dx.doi.org/10.1038/nature02971.
French, B.M., 1998. Traces of Catastrophe: A Handbook of Shock-Metamorphic Effects in Terrestrial Meteorite Impact Structures. Lunar and Planetary Institute,
Houston, 120 pp.
Gellert, R., Rieder, R., Anderson, R.C., Bruckner, J., Clark, B.C., Dreibus, G., Economou,
T., Klingelhoefer, G., Lugmair, G.W., Ming, D.W., Squyres, S.W., d’Uston, C.,
Wanke, H., Yen, A., Zipfel, J., 2004. Chemistry of rocks and soils in Gusev Crater
from the alpha particle X-ray spectrometer. Science 305, 829–833.
143
Gorevan, S.P., Myrick, T., Davis, K., Chau, J.J., Bartlett, P., Mukherjee, S., Anderson,
R., Squyres, S.W., Arvidson, R.E., Madsen, M.B., Bertelsen, P., Goetz, W., Binau,
C.S., Richter, L., 2003. Rock Abrasion Tool: Mars Exploration Rover mission. J.
Geophys. Res. 108 (E12), 8068. http://dx.doi.org/10.1029/2003JE002061.
Grotzinger, J.P., Crisp, J., Vasavada, A.R., Anderson, R.C., Baker, C.J., Barry, R., Blake,
D.F., Conrad, P., Edgett, K.S., Ferdowski, B., Gellert, R., Gilbert, J.B., Golombek, M.,
Gómez-Elvira, J., Hassler, D.M., Jandura, L., Litvak, M., Mahaffy, P., Maki, J., Meyer,
M., Malin, M.C., Mitrofanov, I., Simmonds, J.J., Vaniman, D., Welch, R.V., Wiens,
R.C., 2012. Mars Science Laboratory Mission and Science Investigation. Space Sci.
Rev. 170, 5–56.
Haskin, L.A., Wang, A., Jolliff, B.L., McSween, H.Y., Clark, B.C., Des Marais, D.J., McLennan, S.M., Tosca, N.J., Hurowitz, J.A., Farmer, J.D., Yen, A., Squyres, S.W., Arvidson,
R.E., Klingelhöfer, G., Schröder, C., de Souza, P.A., Ming, D.W., Gellert, R., Zipfel, J.,
Brückner, J., Bell, J.F., Herkenhoff, K., Christensen, P.R., Ruff, S., Blaney, D., Gorevan, S., Cabrol, N.A., Crumpler, L., Grant, J., Soderblom, L., 2005. Water alteration
of rocks and soils on Mars at the Spirit rover site in Gusev crater. Nature 436
(7047), 66–69.
Hausrath, E.M., Navarre-Sitchler, A.K., Sak, P.B., Steefel, C.I., Brantley, S.L., 2008. Basalt
weathering rates on Earth and the duration of liquid water on the plains of
Gusev Crater, Mars. Geology 36 (1), 67–70.
Herkenhoff, K.E., Squyres, S.W., Arvidson, R., Bass, D.S., Bell, J.F., Bertelsen, P., Cabrol,
N.A., Gaddis, L., Hayes, A.G., Hviid, S.F., Johnson, J.R., Kinch, K.M., Madsen, M.B.,
Maki, J.N., McLennan, S.M., McSween, H.Y., Rice, J.W., Sims, M., Smith, P.H.,
Soderblom, L.A., Spanovich, N., Sullivan, R., Wang, A., 2004. Textures of the soils
and rocks at Gusev Crater from Spirit’s Microscopic Imager. Science 305 (5685),
824–827. http://dx.doi.org/10.1126/science.3050824.
Herkenhoff, K.E., Squyres, S.W., Anderson, R., Archinal, B.A., Arvidson, R.E., Barrett,
J.M., Becker, K.J., Bell, J.F., Budney, C., Cabrol, N.A., Chapman, M.G., Cook, D.,
Ehlmann, B.L., Farmer, J., Franklin, B., Gaddis, L.R., Galuszka, D.M., Garcia, P.A.,
Hare, T.M., Howington-Kraus, E., Johnson, J.R., Johnson, S., Kinch, K., Kirk, R.L.,
Lee, E.M., Leff, C., Lemmon, M., Madsen, M.B., Maki, J.N., Mullins, K.F., Redding, B.L., Richter, L., Rosiek, M.R., Sims, M.H., Soderblom, L.A., Spanovich, N.,
Springer, R., Sucharski, R.M., Sucharski, T., Sullivan, R., Torson, J.M., Yen, A., 2006.
Overview of the Microscopic Imager Investigation during Spirit’s first 450 sols
in Gusev crater. J. Geophys. Res., Planets 111 (E02), E02S04. http://dx.doi.org/
10.1029/2005JE002574.
Herkenhoff, K.E., Golombek, M.P., Guinness, E.A., Johnson, J.B., Kusack, A., Richter, L.,
Sullivan, R.J., Gorevan, S., 2009. In situ observations of the physical properties of
the Martian surface. In: Bell, J. (Ed.), The Martian Surface. Cambridge Planetary
Science, pp. 451–467.
Hobbs, K.M., 2010. Global climate change in the Miocene recorded in Columbia River
Basalt-hosted paleosols. M.S. thesis, Univ. of Idaho, 74 pp.
Hoek, E., Brown, E.T., 1997. Practical estimates of rock mass strength. Int. J. Rock
Mech. Min. Sci. 34 (8), 1165–1186.
Hoffman, E.L., 1992. Instrumental neutron activation in geoanalysis. J. Geochem. Explor. 44 (1–3), 297–319.
Horz, F., 1977. Impact cratering and regolith dynamics. Phys. Chem. Earth 10, 3–15.
Hurowitz, J.A., McLennan, S.M., 2007. A ∼3.5 Ga record of water-limited, acidic
weathering conditions on Mars. Earth Planet. Sci. Lett. 260 (3–4), 432–443.
Hurowitz, J.A., Fischer, W., Tosca, N.J., Milliken, R.E., 2010. Origin of acidic surface
waters and the evolution of atmospheric chemistry on early Mars. Nat. Geosci. 3
(5), 323–326. http://dx.doi.org/10.1038/ngeo831.
Hurowitz, J.A., Fischer, W.W., 2014. Contrasting styles of water–rock interaction
at the Mars Exploration Rover landing sites. Geochim. Cosmochim. Acta 127,
25–38.
Jakobsson, S.P., Moore, J.G., 1986. Hydrothermal minerals and alteration rates at
Surtsey Volcano, Iceland. Geol. Soc. Am. Bull. 97 (5), 648–659.
Kauffman, J.D., 2004. Geologic Map of the Gifford Quadrangle, Nez Perce County,
Idaho, 1:24,000. Idaho Geological Survey, Moscow, Idaho.
Kauffman, J.D., 2009. Revised stratigraphy for several Saddle Mountains and Wanapum Basalt units, Columbia River Basalt Group. Paper Presented at Geological
Society of America Annual Meeting. GSA, Portland, OR, 18–21 October 2009.
Kristmannsdottir, H., 1979. Alteration of basaltic rocks by hydrothermal activity at
100–300 ◦ C. In: Mortland, M.M., Farmer, V.C. (Eds.), Developments in Sedimentology. Elsevier B.V., pp. 359–367.
Lewis, R.S., Bush, J.H., Burmester, R.F., Kauffman, J.D., Garwood, D.L., Myers, P.E., Othberg, K.L., 2005. Geologic Map of the Potlatch 30 × 60 Minute Quadrangle, Idaho,
1:100,000. Idaho Geological Survey, Moscow, ID.
Macke, R.J., Britt, D.T., Consolmagno, G.J., 2010. Analysis of systematic error in “bead
method” measurements of meteorite bulk volume and density. Planet. Space
Sci. 58 (3), 421–426.
Macke, R.J., Britt, D.T., Consolmagno, G.J., 2011. Density, porosity, and magnetic susceptibility of achondritic meteorites. Meteorit. Planet. Sci. 46 (2), 311–326.
Malin, M.C., 1974. Salt weathering on Mars. J. Geophys. Res. 79 (26), 3888–3894.
McGlynn, I.O., Fedo, C.M., McSween Jr., H.Y., 2011. Origin of basaltic soils at Gusev
crater, Mars, by aeolian modification of impact-generated sediment. J. Geophys.
Res. 116, E00F22. http://dx.doi.org/10.1029/2010JE003712.
McGlynn, I.O., Fedo, C.M., McSween Jr., H.Y., 2012. Soil mineralogy at the Mars
Exploration Rover landing sites: an assessment of the competing roles of physical sorting and chemical weathering. J. Geophys. Res., Planets 117, E01006.
http://dx.doi.org/10.1029/2011JE003861.
144
B.J. Thomson et al. / Earth and Planetary Science Letters 400 (2014) 130–144
McSween, H.Y., Arvidson, R.E., Bell, J.F., Blaney, D., Cabrol, N.A., Christensen, P.R.,
Clark, B.C., Crisp, J.A., Crumpler, L.S., Des Marais, D.J., Farmer, J.D., Gellert, R.,
Ghosh, A., Gorevan, S., Graff, T., Grant, J., Haskin, L.A., Herkenhoff, K.E., Johnson,
J.R., Jolliff, B.L., Klingelhoefer, G., Knudson, A.T., McLennan, S., Milam, K.A., Moersch, J.E., Morris, R.V., Rieder, R., Ruff, S.W., de Souza, P.A., Squyres, S.W., Wanke,
H., Wang, A., Wyatt, M.B., Yen, A., Zipfel, J., 2004. Basaltic rocks analyzed by the
Spirit Rover in Gusev Crater. Science 305, 842–845.
McSween, H.Y., Wyatt, M.B., Gellert, R., Bell, J.F., Morris, R.V., Herkenhoff, K.E., Crumpler, L.S., Milam, K.A., Stockstill, K.R., Tornabene, L.L., Arvidson, R.E., Bartlett,
P., Blaney, D., Cabrol, N.A., Christensen, P.R., Clark, B.C., Crisp, J.A., Des Marais,
D.J., Economou, T., Farmer, J.D., Farrand, W., Ghosh, A., Golombek, M., Gorevan, S., Greeley, R., Hamilton, V.E., Johnson, J.R., Joliff, B.L., Klingelhöfer, G.,
Knudson, A.T., McLennan, S., Ming, D., Moersch, J.E., Rieder, R., Ruff, S.W.,
Schröder, C., de Souza, P.A., Squyres, S.W., Wänke, H., Wang, A., Yen, A., Zipfel,
J., 2006. Characterization and petrologic interpretation of olivine-rich basalts
at Gusev Crater, Mars. J. Geophys. Res. 111 (E02), E02S10. http://dx.doi.org/
10.1029/2005JE002477.
McSween, H.Y., Taylor, G.J., Wyatt, M.B., 2009. Elemental composition of the martian
crust. Science 324 (5928), 736–739. http://dx.doi.org/10.1126/science.1165871.
McSween Jr., H.Y., 1994. What we have learned about Mars from SNC meteorites.
Meteoritics 29, 757–779.
Milam, K.A., Stockstill, K.R., Moersch, J.E., McSween, H.Y., Tornabene, L.L., Ghosh, A.,
Wyatt, M.B., Christensen, P.R., 2003. THEMIS characterization of the MER Gusev
crater landing site. J. Geophys. Res., Planets 108 (E12), 8078. http://dx.doi.org/
10.1029/2002JE002023.
Ming, D.W., Mittlefehldt, D.W., Morris, R.V., Golden, D.C., Gellert, R., Yen, A., Clark,
B.C., Squyres, S.W., Farrand, W.H., Ruff, S.W., Arvidson, R.E., Klingelhöfer, G.,
McSween, H.Y., Rodionov, D.S., Schröder, C., de Souza, P.A., Wang, A., 2006. Geochemical and mineralogical indicators for aqueous processes in the Columbia
Hills of Gusev crater, Mars. J. Geophys. Res. 111 (E02), E02S12. http://dx.doi.org/
10.1029/2005JE002560.
Moon, V., Jayawardane, J., 2004. Geomechanical and geochemical changes during
early stages of weathering of Karamu Basalt, New Zealand. Eng. Geol. 74, 57–72.
Mustard, J.F., Murchie, S., Erard, S., Sunshine, J., 1997. In situ compositions of Martian
volcanics: implications for the mantle. J. Geophys. Res. 102, 25605–25616.
Myrick, T.M. Carlson, L. Chau, C.J. Powderly, J. Bartlett, P. Davis, K., Gorevan, S., 2004.
The RAT as a Mars rock physical properties tool. Paper Presented at AIAA Space
2004 Conference & Exhibit, AIAA-2004-6096, AIAA, San Diego, CA.
Nesbitt, H.W., Wilson, R.E., 1992. Recent chemical weathering of basalts. Am. J.
Sci. 292 (10), 740–777.
Newsom, H.E., 1980. Hydrothermal alteration of impact melt sheets with implications for Mars. Icarus 44 (1), 207–216.
Niles, P.B., Michalski, J., 2009. Meridiani Planum sediments on Mars formed through
weathering in massive ice deposits. Nat. Geosci. 2, 215–220.
Nisbet, E.G., 1982. The tectonic setting and petrogenesis of komatiites. In: Arndt,
N.T., Nisbet, E.G. (Eds.), Komatiites. George Allen & Unwin, London, pp. 501–520.
Petrovic, J.J., 2001. Review mechanical properties of meteorites and their constituents. J. Mater. Sci. 36 (7), 1579–1583.
Popova, O., Borovicka, J., Hartmann, W.K., Spurny, P., Gnos, E., Nemtchinov, I., TrigoRodiguez, J.M., 2011. Very low strengths of interplanetary meteoroids and small
asteroids. Meteorit. Planet. Sci. 46 (10), 1525–1550. http://dx.doi.org/10.1111/
j.1945-5100.2011.01247.x.
Reidel, S.P., Tolan, T.L., Hooper, P.R., Beeson, M.H., Fecht, K.R., Bentley, R.D., Anderson,
J.L., 1989. The Grande Ronde Basalt, Columbia River Basalt Group; Stratigraphic
descriptions and correlations in Washington, Oregon, and Idaho. In: Reidel, S.P.,
Hooper, P.A. (Eds.), Volcanism and Tectonism in the Columbia River Flood-Basalt
Province. Geological Society of America, Boulder, CO, pp. 21–54.
Retallack, G.J., 1991. Untangling the effects of burial alteration and ancient soil formation. Annu. Rev. Earth Planet. Sci. 19, 183–206.
Rodriguez-Navarro, C., Doehne, E., 1999. Salt weathering: influence of evaporation
rate, supersaturation and crystallization pattern. Earth Surf. Process. Landf. 24,
191–209.
Sheldon, N.D., 2003. Pedogenesis and geochemical alteration of the Picture Gorge
subgroup, Columbia River basalt, Oregon. Geol. Soc. Am. Bull. 115 (11),
1377–1387. http://dx.doi.org/10.1130/B25223.1.
Squyres, S.W., Arvidson, R.E., Baumgartner, E.T., Bell, J.F., Christensen, P.R., Gorevan, S., Herkenhoff, K.E., Klingelhöfer, G., Madsen, M.B., Morris, R.V., Rieder,
R., Romero, R.A., 2003. Athena Mars rover science investigation. J. Geophys.
Res. 108, 8062. http://dx.doi.org/10.1029/2003JE002121.
Squyres, S.W., Arvidson, R.E., Blaney, D.L., Clark, B.C., Crumpler, L., Farrand, W.H.,
Gorevan, S., Herkenhoff, K.E., Hurowitz, J., Kusack, A., McSween, H.Y., Ming,
D.W., Morris, R.V., Ruff, S.W., Wang, A., Yen, A., 2006. Rocks of the Columbia
Hills. J. Geophys. Res., Planets 111 (E02), E02S11. http://dx.doi.org/10.1029/
2005JE002562.
Stewart, D.B., Walker, G.W., Wright, T.L., Fahey, J.J., 1966. Physical properties of calcic
labradorite from Lake County, Oregon. Am. Mineral. 51, 177–197.
Swanson, D.A., Wright, T.L., Hooper, P.R., Bentley, R.D., 1979. Revisions in stratigraphic nomenclature of the Columbia River Basalt Group. U.S. Geol. Surv.
Bull. 1457-G, 1–59.
Takeda, H., 1972. Structural studies of rim augite and core pigeonite from lunar rock
12 052. Earth Planet. Sci. Lett. 15 (1), 65–71.
Teale, R., 1965. The concept of specific energy in rock drilling. Int. J. Rock Mech.
Min. Sci. Geomech. Abstr. 2 (1), 57–73.
Thomson, B.J., Bridges, N.T., Cohen, J., Hurowitz, J., Lennon, A., 2011. Estimating rock
strength parameters from Rock Abrasion Tool (RAT) grinds. In: Lunar and Planetary Science Conference, vol. 42, Abstract #2567.
Thomson, B.J., Bridges, N.T., Cohen, J., Hurowitz, J.A., Lennon, A., Paulsen, G.,
Zacny, K., 2013. Estimating rock compressive strength from Rock Abrasion
Tool (RAT) grinds. J. Geophys. Res. 118, 1233–1244. http://dx.doi.org/10.1029/
2012JE004240.
Thuro, K., Plinninger, R., Zäh, S., Schütz, S., 2001. Scale effects in rock strength properties. Part 1: Unconfined compressive test and Brazilian test. Paper Presented
at ISRM Regional Symposium, EUROCK, 169–174.
Tolan, T.L., Reidel, S.P., Beeson, M.H., Anderson, J.L., Fecht, K.R., Swanson, D.A., 1989.
Revisions to the estimates of the areal extent and volume of the Columbia River
Basalt Group. In: Reidel, S.P., Hooper, P.A. (Eds.), Volcanism and Tectonism in the
Columbia River Flood-Basalt Province. Geological Society of America, Boulder,
CO, pp. 1–20.
Tosca, N.J., McLennan, S.M., Lindsley, D.H., Schoonen, M.A.A., 2004. Acid-sulfate
weathering of synthetic Martian basalt: the acid fog model revisited. J. Geophys.
Res., Planets 109 (E5), 5003. http://dx.doi.org/10.1029/2003JE002218.
Wang, A., Korotev, R.L., Jolliff, B.L., Haskin, L.A., Crumpler, L., Farrand, W.H., Herkenhoff, K.E., de Souza, P., Kusack, A.G., Hurowitz, J.A., Tosca, N.J., 2006. Evidence
of phyllosilicates in Wooly Patch, an altered rock encountered at West Spur,
Columbia Hills, by the Spirit rover in Gusev crater, Mars. J. Geophys. Res. 111
(E2), E02S16. http://dx.doi.org/10.1029/2005JE002516.
Wechsler, B.A., Prewitt, C.T., 1984. Crystal structure of ilmenite (FeTiO3) at high temperature and at high pressure. Am. Mineral. 69 (1–2), 176–185.