Spirit Mars Rover Mission: Overview and selected results
from the northern Home Plate Winter Haven to the side
of Scamander crater
R. E. Arvidson,1 J. F. Bell III,2 P. Bellutta,3 N. A. Cabrol,4,5 J. G. Catalano,1 J. Cohen,6
L. S. Crumpler,7 D. J. Des Marais,4 T. A. Estlin,3 W. H. Farrand,8 R. Gellert,9
J. A. Grant,10 R. N. Greenberger,1 E. A. Guinness,1 K. E. Herkenhoff,11 J. A. Herman,3
K. D. Iagnemma,12 J. R. Johnson,11 G. Klingelh?fer,13 R. Li,14 K. A. Lichtenberg,1
S. A. Maxwell,3 D. W. Ming,15 R. V. Morris,15 M. S. Rice,2 S. W. Ruff,16 A. Shaw,1
K. L. Siebach,1 P. A. de Souza,17 A. W. Stroupe,3 S. W. Squyres,2 R. J. Sullivan,2
K. P. Talley,3 J. A. Townsend,3 A. Wang,1 J. R. Wright,3 and A. S. Yen3
Received 23 April 2010; revised 6 June 2010; accepted 15 June 2010; published 30 September 2010.
[1] This paper summarizes Spirit Rover operations in the Columbia Hills, Gusev crater,
from sol 1410 (start of the third winter campaign) to sol 2169 (when extrication attempts
from Troy stopped to winterize the vehicle) and provides an overview of key scientific
results. The third winter campaign took advantage of parking on the northern slope of
Home Plate to tilt the vehicle to track the sun and thus survive the winter season. With the
onset of the spring season, Spirit began circumnavigating Home Plate on the way to
volcanic constructs located to the south. Silica?rich nodular rocks were discovered in the
valley to the north of Home Plate. The inoperative right front wheel drive actuator
made climbing soil?covered slopes problematical and led to high slip conditions and
extensive excavation of subsurface soils. This situation led to embedding of Spirit on the
side of a shallow, 8 m wide crater in Troy, located in the valley to the west of Home Plate.
Examination of the materials exposed during embedding showed that Spirit broke
through a thin sulfate?rich soil crust and became embedded in an underlying mix of sulfate
and basaltic sands. The nature of the crust is consistent with dissolution and precipitation
in the presence of soil water within a few centimeters of the surface. The observation
that sulfate?rich deposits in Troy and elsewhere in the Columbia Hills are just beneath the
surface implies that these processes have operated on a continuing basis on Mars as
landforms have been shaped by erosion and deposition.
Citation: Arvidson, R. E., et al. (2010), Spirit Mars Rover Mission: Overview and selected results from the northern Home Plate
Winter Haven to the side of Scamander crater, J. Geophys. Res., 115, E00F03, doi:10.1029/2010JE003633.
1. Introduction
[2] The Mars Exploration Rover, Spirit, touched down on
the volcanic plains of Gusev Crater on 4 January 2004.
Since landing, Spirit has conducted numerous traverses and
made extensive measurements on the plains that dominate
the floor of the crater and the soil?covered and rocky sur-
1Department of Earth and Planetary Sciences, Washington University in
St. Louis, St. Louis, Missouri, USA.
2Department of Astronomy, Cornell University, Ithaca, New York,
USA.
3Jet Propulsion Laboratory, California Institute of Technology,
Pasadena, California, USA.
4NASA Ames Research Center, Moffett Field, California, USA.
5SETI Institute, Mountain View, California, USA.
6Honeybee Robotics Spacecraft Mechanisms Corporation, New York,
New York, USA.
7New Mexico Museum of Natural History and Science, Albuquerque,
New Mexico, USA.
8Space Science Institute, Boulder, Colorado, USA.
9Department of Physics, University of Guelph, Guelph, Ontario,
Canada.
10Center for Earth and Planetary Studies, Smithsonian Institution,
Washington, DC, USA.
11U.S. Geological Survey, Flagstaff, Arizona, USA.
12Department of Mechanical Engineering, Massachusetts Institute of
Technology, Cambridge, Massachusetts, USA.
13Institut f?r Anorganische und Analytische Chemie, Johannes
Gutenberg?Universit?t, Mainz, Germany.
14Department of Civil and Environmental Engineering and Geodetic
Science, Ohio State University, Columbus, Ohio, USA.
15NASA Johnson Space Center, Houston, Texas, USA.
16School of Earth and Space Exploration, Arizona State University,
Tempe, Arizona, USA.
17Information and Communication Technologies Centre, CSIRO,
Hobart, Tasmania, Australia.
Copyright 2010 by the American Geophysical Union.
0148?0227/10/2010JE003633
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 115, E00F03, doi:10.1029/2010JE003633, 2010
E00F03 1 of 19
Figure 1. HiRISE subframe showing Home Plate with sites and traverses overlain since the third winter
campaign site on the northern slopes of Home Plate to embedding of Spirit at Troy. Times in sols for
selected locations are shown, along with place names for areas highlighted in this paper. Spirit is the white
feature located in Troy. The area called the sidewalk is located to the southeast of Spirit. Low Ridge is
where Spirit spent its second winter; Tyrone is the location (off to lower right beyond image coverage) of
sulfate?bearing soils, and Gertrude Weise is the location of opaline silica deposits investigated by Spirit.
HiRISE image ESP_0013499_1650_red.
Table 1. Spirit?s Payload Elementsa
Instrument Key Parameters
Mast?Mounted Science Instruments
Pancam: Panoramic Camera Multispectral imager (0.4 to 1.0 mm) with stereoscopic capability;
0.28 mrad IFOV; 16.8? by 16.8?FOV. Stereo baseline separation of
30 cm. External calibration target on rover deck.
MiniTES: Thermal Emission Spectrometer Emission spectra (5 to 29 mm, 10 cm?1 resolution) with 8 or 20 mrad FOV.
Internal and external blackbody calibration targets.
IDD?Mounted In Situ Package
APXS: Alpha Particle X?Ray Spectrometer 244Cm alpha particle sources, and X?ray detectors, 3.8 cm FOV.
MB: M?ssbauer Spectrometer 57Fe spectrometer in backscatter mode; 57Co/Rh source and Si?PIN diode
detectors; field of view approximately 1.5 cm2.
MI: Microscopic Imager 31 mm/pixel monochromatic imager (1024x1024) with 2 mm depth of field.
RAT: Rock Abrasion Tool Tool originally capable of preparing 5 mm deep by 4.5 cm wide surface on
rocks. For period covered in this paper: Grinding bit pads were
worn away by sol 416 and RAT primarily used for rock brushing.
Implemented soil ?grind? at Troy using steel pieces that held bit pads
in place. Encoder for grind motor stopped working on sol 1341 thus
replaced seek/scan with ?grind scan? that started from the Z home
position and implemented large steps with grind motor run at low voltage
to stall when RAT hit surfaces.
Engineering Cameras
Navigation Cameras (Navcam) Mast?mounted panchromatic stereoscopic imaging system with 0.77 mrad
IFOV;
45?FOV, and 20 cm stereo baseline separation.
Hazard Avoidance Cameras (Hazcam) Front and rear?looking panchromatic stereoscopic imaging systems with 2
mrad IFOV;
123?FOV, 10 cm stereo baseline separation.
aMagnets were also included on the spacecraft but not described.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
2 of 19
faces within the older Columbia Hills (see Arvidson et al.
[2006] and [2008] for mission summaries and science high-
lights and Table 1 for payload description). The purpose of
this paper is to summarize operations and present scientific
highlights from the time Spirit began its third winter cam-
paign on the north side of Home Plate on sol 1410 to the
start of its fourth winter stay on sol 2169 on the side of a
crater informally named Scamander. This crater is located in
the valley to the west of Home Plate (Figures 1?2 and
Table 2).
[3] Spirit?s right front wheel drive actuator failed on sol
779 and was declared inoperative on sol 787 [Arvidson et al.,
2008]. Driving up soil?covered slopes with tilts more than
several degrees has proven to be problematic and led to sig-
nificant wheel slippage, sinkage, and exposure of subsurface
soils, culminating in embedding in soils in Troy (Figure 1).
On the other hand, measurements of the soils excavated
during drives have provided new insights into the aqueous
history of the Inner Basin around Home Plate and are the
key science highlight presented in this paper. This overview
of operations and scientific results is meant to complement
papers that provide detailed findings about the geology,
chemistry, and mineralogy of Husband Hill and the Inner
Basin that are included as the fourth set of Mars Explo-
ration Rover papers published in the Journal of Geophysical
Research.
2. Five Wheel Drive Dynamics
[4] As noted in section 1, Spirit?s right front wheel drive
actuator failed relatively early during the mission. The wheels
were designed with permanent magnetic detents to keep
them from rolling (e.g., downhill) when not actuated. This
detent?induced resisting stress for the inoperative right front
wheel typically exceeded the soil shear strength during
drives, causing soil failure and exposing subsurface soils
(e.g., Figure 3). The drag stress also caused the vehicle to
yaw counterclockwise (viewed from above) when driving
backward and clockwise when driving forward. To better
understand the dynamics associated with these motions, a
two hundred element rigid body software model of the rover
developed by Lindemann and Voorhees [2005] was utilized
to calculate forces and torques associated with five? and six?
wheel driving across flat surfaces and slopes covered by
soils. The simulations assume a simplified model of wheel?
soil interactions that is dominated by frictional stresses using
Coulomb parameters for static and dynamic coefficients of
friction.
[5] Specifically, drives were simulated in which the max-
imum shear stress, Fs,max, that could be exerted by a wheel
on the soil before failure was set to the soil Coulomb shear
strength:
Fs;max ? C? Fn tanF
where C is the soil cohesion, Fn is the normal stress due to
the weight of the rover on the wheel, F is the soil angle of
internal friction, and tan F is static coefficient of friction.
Values for cohesion and angle of internal friction (i.e., to
determine the effective static coefficient of friction) of 1 to
5 kPa and 30 degrees, respectively, were used in the com-
putations. These values are similar to those retrieved from
trenching experiments conducted on the Gusev plains by R. J.
Sullivan et al. (Cohesions, friction angles, and other physical
properties of Martian regolith from MER wheel trenches and
wheel scuffs, submitted to Journal of Geophysical Research,
Figure 2. Timeline of events, including atmospheric optical depth, solar array energy, dust loading on
the array, seasons, and key activities for Spirit covering the period from when Spirit wintered on the
northern slopes of Home Plate to when Troy extrication drives stopped (sol 2165) because of the onset
of the fourth winter season. The first drive downhill from the third winter site occurred on sol 1772, fol-
lowed closely by acquisition of in situ data for the target Stapledon (Figure 3). A series of solar array dust
cleaning events coincided with the onset of the summer season while Spirit was embedded in Troy sandy
materials.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
3 of 19
Table 2. Major Activities for Spirit Organized by Sola
Earth Date Sol Activity Site
12/17/07?12/15/08 1410?1760 Winter campaign: IDD measurements of Chanute, Freeman,
Wendell Pruit rock targets, Arthur C. Harmon soil target and magnets;
Tuskegee Pancam panorama; atmospheric RS
133?135
2/27/08?10/19/09 1476?1705 Pancam 13F ?Bonestell? Panorama 134
3/26/2008 1503 Pancam superres ?Roger_Zellazny? target 134
4/2/08?4/25/08 1510?1532 MI, APXS, MB ?Arthur_C_Harmon? soil target 134
4/4/2008 1512 Pancam superres ?Arthur C Clarke? 134
5/14/08?10/6/08 1551?1692 IDD in winter hibernation position 134
10/23/08?11/8/08 1709?1724 Drive uphill to follow the sun 134?135
11/30/08?12/15/08 1746?1760 Solar Conjunction 135
12/16/08?12/17/08 1761?1762 RS and Health Check 135
12/18/08?12/23/08 1763?1768 Attempt to drive uphill out of winter position 135
12/14/08?12/26/08 1769?1771 Atmospheric RS 135
12/27/2008 1772 Attempt to drive downhill out of winter position 135
12/28/08?12/29/08 1773?1774 Atmospheric RS 135
12/30/08?1/1/09 1775?1777 Runout 135
1/2/09?1/5/09 1778?1781 Automode 135
1/6/2009 1782 Drive downhill out of winter position 135
1/8/09?1/14/09 1783?1789 RS surroundings; Pancam supperres ?Jack Williamson? 135
1/15/09?1/17/09 1790?1792 Runout 135
1/18/2009 1793 Drive to ?Stapledon? 135
1/19/09?1/20/09 1794?1795 RS; Pancam 13F ?Thunderbolt? 135
1/21/2009 1796 MI, APXS ?Stapledon? target 135
1/22/2009 1797 RS surroundings 135
1/23/2009 1798 Drive toward ?Home Plate? on ramp 135
1/24/09?1/30/09 1799?1805 Atmospheric RS 135
1/31/09?2/12/09 1806?1818 Attempt to drive onto ?Home Plate?; Atmospheric RS; excavated disturbed
soil ?C. L. Moore?
135
2/14/2009 1819 Atmospheric RS 135
2/15/2009 1820 Extrication from soft material 135
2/16/09?2/23/09 1821?1828 Attempt to drive uphill onto ?Home Plate?; Atmospheric RS 135
2/24/09?3/10/09 1829?1843 Attempt to drive to NE corner of ?Home Plate?; Excavated disturbed soil
?Edmund Hamilton?
135?136
3/11/09?3/21/09 1844?1854 Attempt to drive toward NW corner of ?Home Plate?; Targeted and
Atmospheric RS
136
3/22/09?3/29/09 1855?1861 Drive west around ?Home Plate?; RS 136
3/30/2009 1862 Atmospheric RS 136
3/31/09?4/2/09 1863?1865 MI, APXS, ?John Wesley Powell?; Pancam 8F ?Pioneer? 136
4/3/2009 1866 Pancam 13F ?John Wesley Powell?; Drive South 136
4/4/2009 1867 Pancam 13F ?York? 136
4/5/09?4/8/09 1868?1871 Drive into ?West Valley?; RS 136?137
4/9/09?4/20/09 1872?1885 Automode and Earth control 137
4/21/09?5/7/09 1886?1899 Attempts to drive South in West Valley and out of embedded position 137
5/8/09?5/24/09 1900?1916 Mobility diagnostics; RS 137
5/18/09?6/12/09 1910?1934 Calypso Panorama 137
5/25/09?5/29/09 1917?1921 RS surroundings 137
5/30/09?6/1/09 1922?1924 MI, MB ?Sackrider? 137
6/2/09?6/3/09 1925?1926 MI, APXS ?Sackrider? 137
6/4/2009 1927 MI ?Olive tree?; MI, APXS ?Olive Branch1? 137
6/5/09?6/10/09 1928?1932 APXS, MI, MB ?Olive Branch2? 137
6/11/2009 1933 Left middle wheel diagnostics 137
6/12/2009 1934 MI, APXS ?Olive Branch3? 137
6/13/2009 1935 MI, APXS ?Olive Branch4? 137
6/12/09?7/8/09 1934?1960 Twilight observations 137
6/14/09?6/17/09 1936?1939 MI, MB, APXS, Pancam ?Penina1? 137
6/18/2009 1940 MI, APXS ?Penina2? 137
6/19/2009 1941 MI, APXS ?Penina3? 137
6/20/09?6/22/09 1942?1944 MB ?Penina1? 137
6/23/2009 1945 MI, APXS and RS ?Cyclops Eye_a? undisturbed soil target 137
6/24/2009 1946 MI, APXS and RS ?Cyclops Eye2? undisturbed soil target 137
6/25/09?6/29/09 1947?1951 Mi, APXS, MB ?Cyclops Eye3? undisturbed soil target 137
6/30/2009 1952 RAT Grind Scan; RS ?Cyclops Eye3? 137
7/1/09?7/13/09 1953?1964 Atmospheric RS; APXS Argon 137
7/7/09?7/13/09 1959?1964 Targeted RS and superresolution imaging 137
7/12/2009 1963 RAT Grind Scan; MI, Pancam ?Cyclops Eye3? 137
7/14/2009 1965 RAT Brush ?Cyclops Eye3? 137
7/15/09?7/20/09 1966?1971 MI, APXS, MB, Pancam ?Cyclops Eye4? 137
7/16/2009 1967 MI, APXS ?Olive Branch5? 137
7/17/09?7/24/09 1968?1975 Navcam illumination experiment 137
7/21/2009 1972 RAT Grind Scan ?Cyclops Eye4?; APXS Argon 137
7/22/2009 1973 Attempt to RAT grind ?Cyclops Eye4? failed; APXS, Pancam ?Cyclops
Eye5?
137
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
4 of 19
Table 2. (continued)
Earth Date Sol Activity Site
7/23/2009 1974 Runout 137
7/24/2009 1975 MI ?Olive Branch6?; MI ?Cyclops Eye5? 137
7/25/09?7/27/09 1976?1978 Targeted and atmospheric RS; Pancam 13F tracks, ?Cyclops Eye5? 137
7/28/2009 1979 MI, APXS ?Olive Pit? 137
7/29/2009 1980 MI, APXS ?Polyphemus Eye1?; RAT grind scan 137
7/30/2009 1981 RAT Grind at ?Polyphemus Eye1?; MI, APXS ?Polyphemus Eye2? 137
7/31/2009 1982 MI ?Sackrider?; Pancam 13F ?Polyphemus Eye2? 137
7/31/09?8/3/09 1982?1985 MB ?Cyclops Eye5? disturbed soil target; RS 137
8/4/2009 1986 RAT grind scan; MI ?Penina2?; RS 137
8/5/2009 1987 Pancam 13F right track 137
8/6/09?8/7/09 1988?1989 MRO safe mode; Runout 137
8/8/2009 1990 MI rover undercarriage; RAT grind (unsuccessful), MI, Pancam ?Cyclops
Eye6?
137
8/9/09?8/10/09 1991?1992 Atmospheric RS 137
8/11/2009 1993 RAT grind scan ?Polyphemus Eye2? disturbed soil target; APXS Argon 137
8/12/2009 1994 APXS Argon; RS 137
8/13/09?8/14/09 1995?1996 RAT Grind, MI, APXS, MB ?Polyphemus Eye3? disturbed soil target; RS 137
8/15/2009 1997 MI, APXS ?Olive Leaf? target; RS 137
8/16/09?8/24/09 1998?2005 MB, Pancam ?Polyphemus Eye3? 137
8/25/2009 2006 RAT penetrometer experiment (?Ulysses Spear?) 137
8/26/2009 2007 MB ?Olive Leaf?; Pancam left middle wheel 137
8/27/2009 2008 Atmospheric RS; Pancam left middle wheel 137
8/28/2009 2009 Runout 137
8/29/09?8/30/09 2010?2011 MB ?Olive Leaf? target; Pancam 13F ?Scamander Plains? 137
8/31/2009 2012 Atmospheric RS; Pancam right middle wheel; Pancam 13F ?Scamander
Plains?
137
9/1/2009 2013 Runout 137
9/2/2009?9/3/09 2014?2015 MB ?Olive Leaf? target; Pancam 13F ?Scamander Plains? 137
9/4/2009 2016 APXS ?Olive Leaf?; Pancam 13F ?Scamander Plains? 137
9/5/09?9/11/09 2017?2023 APXS Integration on capture magnet; Pancam 13F work volume and
?Scamander Plains?
137
9/12/2009 2024 MI rover undercarriage; MI and APXS ?Penina4? 137
9/13/09?9/16/09 2025?2028 MB ?Penina4? 137
9/17/09?9/21/09 2029?2033 Atmospheric RS 137
9/22/09?9/24/09 2034?2035 Automode 137
9/25/09?9/26/09 2036?2037 Atmospheric RS; APXS Argon 137
9/27/09?9/28/09 2038?2039 Runout 137
9/29/09?9/30/09 2040?2041 Atmospheric RS 137
10/1/2009 2042 Runout 137
10/2/2009 2043 Atmospheric RS 137
10/3/2009 2044 MI Rover undercarriage; MI, MB ?Stratius? 137
10/4/2009?10/5/09 2045?2046 MB ?Stratius? 137
10/6/2009 2047 Runout 137
10/7/2009 2048 MB ?Stratius? 137
10/8/2009?10/10/09 2049?2051 APXS ?Stratius?; Pancam 13F ?Scamander Plains? 137
10/11/2009 2052 MI, MB ?Thoosa?; Pancam 13F ?Scamander Plains? 137
10/12/09?10/15/09 2053?2056 Runout 137
10/16/09?10/21/09 2057?2062 MB ?Thoosa?; Pancam 13F ?Scamander Plains? 137
10/20/2009 2061 MI rover undercarriage 137
10/22/2009 2063 MB ?Thoosa? 137
10/23/2009 2064 MB ?Thoosa?; Pancam 13F ?Scamander Plains? 137
10/24/09?10/28/09 2065?2069 Safe mode and recovery 137
10/29/09?10/30?09 2070?2071 APXS ?Thoosa? 137
11/1/2009 2072 MI ?BellyRock? 138
11/2/09?11/3/09 2073?2074 Engineering only, prep for flash reformat 138
11/4/2009 2075 Reformat flash; MI ?BellyRock? 138
11/5/2009 2076 MI Shoulder Joint 138
11/6/2009 2077 Atmospheric RS 138
11/7/09?11/9/09 2078?2080 Straighten wheels for drive; Pancam 13F ?Scamander Plains? 138
11/10/09 2081 MI ?BellyRock? 138
11/11/2009?11/16/09 2082?2087 Atmospheric RS; Pancam 13F ?Scamander Plains? 138
11/17/09 2088 Troy Extraction Attempt; MI ?BellyRock? 138
11/18/09 2089 Pancam 13F ?Scamander Plains? 138
11/19/09 2090 Troy Extraction Attempt; MI ?BellyRock? 138
11/20/09 2091 Atmospheric RS 138
11/21/09 2092 Troy Extraction Attempt; MI ?BellyRock? 138
11/22/09?11/23/09 2093?2094 Atmospheric RS 138
11/24/09 2095 Troy Extraction Attempt; MI ?BellyRock? 138
11/25/09?11/27/09 2096?2098 Atmospheric RS 138
11/28/09 2099 Troy Extraction Attempt; MI ?BellyRock? 138
11/29/09?11/30/09 2100?2101 Atmospheric RS 138
12/1/09 2102 Pancam Magnets 138
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
5 of 19
2010). The values for the effective dynamic coefficient of
friction for the driven wheels (0.3) and the inoperative
right front wheel (0.5) were chosen to replicate ?sandbox?
tests with the engineering test rover at the Jet Propulsion
Laboratory.
[6] Five?wheel forward drive simulations on flat surfaces
using the parameters described in the last paragraph show
that the pushed right front wheel underwent a periodic stick?
slip motion and encountered a resisting force of up to 500 N.
On the other hand, during backward drives, the value was
reduced to 75 N. The difference in values for frontward as
opposed to rearward driving is associated with increased
normal forces and thus soil strength on the right front wheel
due to compression of the forward rocker arm during the
forward motion of the rover. Considerable yaw was observed
during forward drives and only a slight yaw during backward
drives. Similar results were obtained for drive simulations
on slopes, with the addition of downhill slip magnitude
dependent on tilt angle and drive azimuth. Although more
complicated models with deformable soils could be em-
ployed, the simple frictional model replicates the dynamics
of the vehicle well enough on flat and tilted surfaces to
demonstrate that the preferred drive direction for Spirit is
backward.
[7] Even though yaw changes were minimized during
backward drives, visual odometry was often utilized during
five?wheel drive sessions to help keep Spirit on a predefined
course, both for fine maneuvering to targets and for long
drives in complex terrains [Maimone et al., 2007]. Visual
odometry consisted of repeated stereo imaging of nearby
targets and using those fixed target positions to update rover
motions to stay on course. This worked well, except when
climbing soil?covered slopes with tilts more than a few
degrees. As discussed in section 5 of this paper, significant
wheel slippage, sinkage, and sliding were encountered in this
latter case.
3. Winter Campaign on North Slope of Home
Plate
[8] Spirit is located at 14.5 degrees south areocentric lati-
tude [Arvidson et al., 2004]. Spirit was exploring the top of
Home Plate during its third southern hemisphere fall season
on Mars. When the approaching winter made it necessary to
Table 2. (continued)
Earth Date Sol Activity Site
12/2/09 2103 Atmospheric RS 138
12/3/09 2104 Troy Extraction Attempt; MI ?BellyRock? 138
12/4/09?12/7/09 2105?2106 Atmospheric RS; APXS Argon 138
12/9/09 2109 Troy Extraction Attempt; MI ?BellyRock? 138
12/10/09?12/12/09 2110?2112 Atmospheric RS 138
12/13/09 2113 Troy Extraction Attempt; MI ?BellyRock? 138
12/14/09 2114 Pancam 13F ?Von Braun,? ?Pioneer Mound? 138
12/15/09 2115 Pancam 13F tailings and ?Von Braun?; Pancam middle wheels 138
12/16/09 2116 MI ?BellyRock?; Pancam 13F ?Von Braun? 138
12/17/09 2117 Troy Extraction Attempt; MI ?BellyRock?; Pancam 13F ?Von Braun?;
Pancam middle wheels
138
12/18/09 2118 Troy Extraction Attempt; MI ?BellyRock?; Pancam 13F tailings 138
12/19/09 2119 Pancam 13F tailings 138
12/20/09 2120 Troy Extraction Attempt; MI ?BellyRock? 138
12/21/09 2121 Atmospheric RS 138
12/22/09 2122 Troy Extraction Attempt; MI ?BellyRock?; Pancam 13F tailings 138
12/23/09 2123 Pancam 13F tailings 138
12/24/09 2124 Pancam superres ?South Valley? 138
12/25/09 2125 Pancam superres ?Pioneer Mound? 138
12/26/09 2126 Troy Extraction Attempt; MI ?BellyRock?; Pancam 13F tailings 138
12/27/09 2127 Pancam 13F tailings 138
12/28/09?12/29/09 2128?2129 Runout 138
12/30/09 2130 Troy Extraction Attempt; MI ?BellyRock? 138
12/31/09 2131 Pancam 13F Deck 138
1/1/10 2132 Troy Extraction Attempt; MI ?BellyRock? 138
1/2/10?1/4/10 2133?2135 Atmospheric RS 138
1/5/10 2136 Troy Extraction Attempt; MI ?BellyRock? 138
1/6/10 2137 Atmospheric RS 138
1/7/10 2138 Troy Extraction Attempt; MI ?BellyRock? 138
1/8/10 2139 Atmospheric RS 138
1/9/10 2140 Troy Extraction Attempt; MI ?BellyRock? 138
1/10/10?1/11/10 2141?2142 Atmospheric RS 138
1/12/10 2143 Troy Extraction Attempt; MI ?BellyRock? 138
1/13/10 2144 Atmospheric RS 138
1/14/10 2145 Troy Extraction Attempt; MI ?BellyRock? 138
1/15/10 2146 Atmospheric RS 138
1/16/10 2147 Troy Extraction Attempt; MI ?BellyRock? 138
1/17/10?1/18/10 2148?2149 Atmospheric RS 138
1/19/10 2150 Troy Extraction Attempt 138
1/20/10 2151 MI ?BellyRock? 138
1/21/10 2152?2169 Troy Extraction Attempts, followed by preparations for winter 138
aRS, remote sensing; IDD, Instrument Deployment Device. Other acronyms defined in Table 1.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
6 of 19
tilt the rover toward the north as much as possible the
lowest?risk plan was to drive across the relatively flat top of
Home Plate and descend onto its northern slope, given the
rover?s limited ability to climb up slopes (Figure 1). The
successful descent on sol 1406 achieved a 13? northerly tilt.
Subsequent short downhill drives increased insolation. For
example, tilts of 22, 27, and 29 degrees were achieved on
sols 1440, 1463, and 1464, respectively. On sol 1763 the
first drive away from the winter site was undertaken by
driving uphill to get back onto Home Plate (Figure 1). The
intent was to drive across the relatively flat top of Home
Plate, with extensive bedrock exposures and relatively easy
driving conditions, to head south toward von Braun and
Goddard (Figure 1). Unfortunately, because of extensive
slippage and sliding, Spirit was unable to get back up onto
the top of Home Plate. The vehicle was then commanded to
drive downhill as an initial step to circumnavigate around
Home Plate by traversing the valleys that border it. On sol
1782, the rover drove into the valley to the north of
Home Plate.
[9] During its winter stay on the northern slope of Home
Plate, Spirit acquired the ?Tuskegee? and ?Bonestell? Pan-
cam panoramas, monitored the surface and atmosphere, and
acquired in situ observations for the rock targets Chanute,
Freeman, and Wendell Pruit. The RAT was used to brush
dust coatings from the rock targets. Grinding was precluded
because of bit wear (Table 1). The rock measurements
resemble observations made at other localities within the
upper stratigraphic section on Home Plate and show a dom-
inantly basaltic composition with a contribution from P2O5
and TiO2?rich rocks of the Wishstone and Watchtower
classes [e.g., Gellert et al., 2006]. In situ observations
were also acquired for the soil target Arthur C. Harmon and
closely resemble those for typical basaltic soil encountered
elsewhere within the Inner Basin.
4. Trip to Troy
[10] After extensive analyses by the science and mobility
teams, the route toward von Braun and Goddard was
selected to be clockwise in the northern valley to a low?tilt
?on ramp? that might allow an ascent onto Home Plate for
the drives to the south. MiniTES observations acquired from
the winter campaign site of low nodular outcrops in the
northern valley indicated the presence of silica?rich materials
(S. Ruff et al., Characteristics, distribution, and significance
of opaline silica in Gusev crater, manuscript in preparation,
2010). Thus, one of the first in situ targets after leaving the
winter site was Stapledon, a group of nodular outcrops that
the rover first drove over to expose relatively fresh materials
(Figure 3). APXS measurements showed a SiO2 content of
72%, confirming the presence of silica?rich materials in the
northern valley, in addition to the eastern valley surrounding
Home Plate (Figure 1) [Squyres et al., 2008; S. Ruff et al.,
manuscript in preparation, 2010].
[11] Mobility was found to be very difficult while
attempting to climb over soil?covered, northeast?trending
ridges in the valley on the northeastern side of Home Plate
(Figure 1). Spirit experienced significant wheel sinkage and
high slip as well as lateral and longitudinal sliding. One
Figure 3. Front Hazcam image acquired while Spirit was
in the valley to the north of Home Plate, after a backward
drive from the winter campaign site and maneuvering to
place the IDD within reach of the nodular, broken rock, Sta-
pledon. APXS data indicate that Stapledon consists of 72%
by weight of SiO2. Note similar easily broken material in
front of and to left side of the right front wheel. Hazcam
frame 2F285542810RSLAZAQP1214L0MZ acquired on
sol 1793.
Figure 4. Pancam false color image of disturbed area C. L.
Moore where whitish material was exposed by dragging the
right front wheel during backward drive maneuvers on sol
1813 to try and climb onto Home Plate. Note red nature of the
undisturbed surface in between patches of dark sand. Pancam
frames 2P287401133RSDAZJAP2544L2MZ (0.753mmcenter
wavelength as red), 2P287401156RSLAZJAP2544L5MZ
(0.535mmas green), and 2P287401169RSLAZJAP2544L7MZ
(0.432 mm as blue), acquired on sol 1814.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
7 of 19
illustrative example is C. L. Moore, an area where subsur-
face soil was excavated largely by the right front wheel on
sols 1811 and 1813 (Figure 4). While driving backward on
sol 1811, Spirit autonomously executed several forward and
backward arc turns to correct its heading. These motions
exhumed the left side of C. L. Moore and exposed bright
soil excavated from just beneath the surface. On sol 1813
Spirit executed another set of arc turns to rotate to the cor-
rect course heading before driving backward. These motions
further excavated bright soils and produced the linear seg-
ment shown near the lower right corner of Figure 4. The
right front wheel was dragged across the surface during the
arc turns with shear stress sufficient to cause soil failure and
excavation of underlying materials. The undisturbed soil
adjacent to C. L. Moore exposed between patches of rela-
tively dark windblown sand displays a red hue that was not
observed previously in soils located in the plains sur-
rounding the Columbia Hills, or even within most areas
within the Hills proper (Figure 4). Unfortunately, because of
time constraints, in situ measurements were not made for the
C. L. Moore target.
[12] A second example of significant soil excavation on the
northeastern corner of Home Plate is the Edmund Hamilton
area (Figures 5?6). When Spirit autonomously executed an
extensive set of turns in place on sols 1836 and 1837, the left
wheels produced a set of regularly spaced soil mounds as
slippage disturbed the soil and the wheels sank. In this case
the wheels created regularly spaced soil mounds when they
slipped extensively on loose or weakly consolidated soils.
The wheel cleats transported soil to the rear of the slipping
wheels and formed mounds with flanks at the angle of
repose of the soil. As the size of a mound increased, the
wheel gained additional thrust through shear along the
wheel?mound interface, allowing centimeter?scale forward
motion. As the rover moved in the drive direction, the soil
mound experienced slope failure, forming a peak and also
decreasing the available wheel thrust due to reduction in the
length of the wheel?soil interface. Wheel forward motion
slowed and the mound formation process began again,
forming periodic ?waves? of soil.
[13] In the two examples cited above, Spirit experienced
extensive slippage and increased resistance due to partial
wheel burials. This resistance can be modeled using classic
Bekker?Wong terramechanics theory [Bekker, 1969; Wong,
2003]. Two characteristic resistances are important. First the
contact between the wheel surface and soil provides a
compaction resistance to motion in the drive direction for a
driven wheel. This resistance increases as the extent of
wheel burial increases. Second, wheels that are driven,
pushed, or towed bulldoze soil in the drive direction, gen-
erating additional resistance. The summed resistances were
not sufficient to immobilize the vehicle at either C. L. Moore
or Edmund Hamilton. Unfortunately, in situ measurements
were not made for either the C. L. Moore or Edmund
Hamilton areas due to time constraints associated with the
need to get to von Braun and Goddard. The colors of the
soils excavated at both sites resemble those of the exca-
vated sulfate?rich soils at Tyrone and Arad [e.g., Wang et al.,
2008; Rice et al., 2010], indicating that similar materials
might have been exposed by wheel motions.
[14] Multiple attempts were made to drive Spirit clock-
wise around the northeastern side of Home Plate. Extensive
Figure 5. Navcam image of Edmund Hamilton disturbed
soil area generated when Spirit was driving and attempted
turns in place on sols 1836?1837. The left wheels excavated
bright materials as a set of ?sand waves.? Navcam frame
2N2899977045RSDBOEP0675L0MZ acquired on sol 1843.
Figure 6. Pancam false color image of Edmund Hamilton
disturbed area. The ?sand waves? are associated with left
wheel motions as described in the text. Note weak layering
in far wall of trench dug by the left wheels. Pancam frames
2P290424417RSLBOFP2555L2MZ, 2P290424450RSLBO-
F4P2555L5MZ, and 2P290424478RSLBOF4P2555L7MZ
acquired on sol 1848.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
8 of 19
sinkage, slippage, and yaw about the right front wheel,
combined with downhill slip, precluded climbing over
the northeasterly trending ridges extending from this side of
Home Plate. After extensive discussion between the science
and engineering teams, Spirit began to circumnavigate
counterclockwise around Home Plate through the northern
valley and then south through the western valley (Figure 1).
The western valley had been imaged in detail when Spirit
was on the western side of Home Plate, but there was still
concern that the soil and substrate characteristics for this
valley were uncertain.
[15] Drives from the northeastern side of Home Plate to
the west and then south into western valley were uneventful,
with the exception of exposing a long section of relatively
bright soil (exposed by dragging the right front wheel during
a backward drive) just after entering the western valley
(Figure 7). In situ observations on sol 1861 show that this
material (target John Wesley Powell, Figures 1 and 7) is
composed of sulfate?rich (28% SO3), fine?grained sands
mixed with basaltic materials. Mobility was nominal during
the drives through sol 1870 (Figure 1).
5. Embedding at Troy
[16] Drives beginning on sol 1871 through 1899 caused
Spirit to become embedded in the sands of Troy. The 1870
drive placed Spirit ?9 m to the northwest of an area sub-
sequently named Troy (Figure 8). The backward drive to
Troy on sol 1871 was intended to drive up a modest slope
and onto a plateau (informally named the sidewalk, Figure 1)
that extended from the southwestern side of Home Plate
(Figure 8). The drive indeed positioned Spirit on the slope,
but with an unexpected yaw. The Front Hazcam images
from sol 1871 showed that the left wheels uncovered bright
soil and sank more deeply than they had previously in the
western valley (Figure 9). Subsequent attempts were made
to drive onto the sidewalk and continue south. After Spirit?s
wheels became embedded further and the left middle wheel
drive actuator stalled on sol 1899 the MER project declared
a moratorium on further drives. Drives began again on sol
2078 after an extensive set of sand box experiments and
analyses focused on the optimum drive procedures for
extrication. For reference, on sol 2072 the vehicle?s yaw was
estimated from its accelerometers to be 340? clockwise from
Figure 8. Sol 1870 Navcam image looking to the south with
some of the same features shown as labeled on the Hazcam
images shown in Figures 9?10 and the topographic map in
Figure 11. Frame 2N292380548RSDB159P0691L0MZ.
Figure 7. Front Hazcam view showing relatively bright
disturbed soils excavated by dragging the right front wheel
backward. John Wesley Powell is the target for which the
MI data are shown in the lower left. APXS data show
28% SO3 by weight for this target. Hazcam frame
2F291581504RSLB0OAP1214L0MZ acquired on sol 1861.
Figure 9. Sol 1871 Front Hazcam image showing relatively
bright soils disturbed by left front wheel, along with features
also observed in the sol 1870 approachmosaic. Spirit was driv-
ing backward. Frame 2F292473825RSLB188P1212L0MZ.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
9 of 19
north, the pitch was 1? toward the northwest, and the roll
was 12? toward the southwest. Spirit at that point in time
was tilted 4? to the south.
[17] During the final drives prior to the drive stand down,
the active wheels experienced almost 100% slippage and
became substantially buried. In addition, the wheel cleats
became coated with fine?grained materials, lowering the
effective coefficient of friction between the wheels and soils.
The drives exposed bright soils around the left front wheel
in an area named Ulysses (Figure 10). During the final
backward drives, the left front wheel generated a set of soil
?waves.? As noted earlier, wheels that experience high
slippage and that sink into loose or weakly cohesive soils
typically form these ?waves.? Navcam images and locali-
zation data from sol 1870 indicate that Spirit had become
embedded with its left wheels within and right wheels out-
side of the wall and rim of a highly degraded crater, sub-
sequently named Scamander (Figure 11). Scamander crater
is approximately 8 m wide and 25 cm deep. The ?Rock
Garden? located under Spirit is one of several groups of
rocks on the crater rim.
[18] Two numerical simulations of Spirit?s embedding at
Troy provide insight into the available thrust from the drive
wheels and the resistance due to soils and any buried rocks
surrounding the wheels. For the first simulation it was
assumed that Coulomb frictional parameters limit the
thrust available from the five drive wheels, as discussed in
section 2 of this paper. The static coefficient of friction was
taken from the steepest slope (32 degrees) observed in the
Ulysses disturbed soils, with cohesion assumed to be min-
imal. Static wheel sinkage was modeled by placing the two
hundred element software model of the rover [Lindemann
and Voorhees, 2005] onto a digital elevation map gener-
ated from sol 1870 Navcam approach images and letting its
wheels settle into the subsurface according to a power law
relationship between normal force and sinkage. The resis-
tance due to soil around the wheels could not be explicitly
modeled using this approach. Rather, a simulated spring
representing this resisting force was attached to the rover
body and fixed to a static position. Drives were run in both
forward and backward directions. To simulate the nearly
100% wheel slip, the resisting force needed to be 30%
higher than the thrust available for the five driving wheels.
[19] The second method for calculating the resisting for-
ces used the Bekker?Wong theory, given estimates of the
degree of burial for each wheel that were derived from
Navcam and Hazcam image data, together with rover yaw,
pitch, roll from accelerometer data, and rocker and bogie
suspension angle values. Values for the Bekker?Wong
parameters were derived from a combination of information
from soil simulants used for MER single wheel tests
[Richter et al., 2006] and internal friction angle parameters
from Ulysses data. Again, the soil resistance derived from
this modeling exceeded the available drive thrust by a sig-
nificant amount. Within error the two analyses agree and
Figure 10. Front Hazcam view taken on sol 2072, before
extrication activities began, but after Spirit stopped driving
on sol 1899. Key targets are labeled. Ulysses is the exca-
vated area in front of the left front wheel. The press test tar-
get represents use of the RAT as a bevameter to push into
the soil to determine static sinkage as a function of applied
normal stress. The Ulysses trench in front of the left front
wheel is approximately 42 cm wide and 8 cm deep at its
deepest point, relative to the undisturbed surface to the east.
Frame 2F310307243RALB200P1214L0M1.
Figure 11. Topographic map for Scamander crater, approx-
imately 8 m wide and 0.25 m deep, generated from the sol
1870 drive direction Navcam image mosaic. Regional planar
tilt to the northwest was removed to emphasize local topo-
graphic variations. Relief is approximately 3 m with red col-
ors assigned to highest regions and black to lowest regions.
Sandals are a pair of rocks shown in Figures 8?10. The fea-
tures beneath the rover represent rocks within the Rock gar-
den, again as shown in Figure 8. The ?sidewalk? is the
relatively flat area to the southeast of Spirit and the destina-
tion for the post sol 1870 drives. During its last 10 extrication
drives Spirit moved 34 cm backward and to the southeast,
placing the left wheels further into Scamander crater.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
10 of 19
help to explain why Spirit became immobilized as its wheels
dug into the soil at high slip rates.
6. Troy Soil Measurement Campaign
[20] During the period when driving was prohibited, an
extensive measurement campaign was implemented to
characterize the materials exposed within the Instrument
Deployment Device (IDD) work volume and the surround-
ing areas (Figures 10 and 12). Multiple targets in Ulysses
were selected to investigate the range of materials evident in
the Pancam color images (e.g., Figure 12). The targets group
into three categories. The first was within the soil ?waves?
in Ulysses that the left front wheel excavated during its
backward drives before the sol 1899 drive stand down.
Sackrider, an area that appeared brown to gray in Pancam
false color images (e.g., Figure 12), was the first of these
sites for acquisition of Microscopic Imager (MI), Alpha
Particle X?Ray Spectrometer (APXS), and M?ssbauer
Spectrometer (MB) data (Table 1), with MB faceplate
imprints used to help confirm the surface location. Three
sites were examined within Penina, an exhumed area of soil
with a yellow?brown color. Each of three Penina targets was
Figure 12. Sol 1933 Pancam false color image showing
various IDD targets in Ulysses. The data were acquired dur-
ing the middle of the extensive IDD campaign in this area.
MB faceplate imprints are evident for target Sackrider, and
several imprints are evident with lateral displacements for
Olive Branch targets. Penina data were not yet acquired.
Olive leaf is a set of targets on bright crust exposed on the
eastern wall of Ulysses and just beneath the undisturbed sur-
face. The red layer occurs in a number of places and exists at
the surface where not covered by dark basaltic sands or rocks.
Pancam frames 2P297975975SRSLB1ESP2382L2M2,
2P297976112SRSLB1ESP2382L5M2, 2P297976177SRSL-
B1ESP2382L7M2.
Figure 13. Portion of an MI image of the Penina 3 target
acquired on sol 1986 when the target was shadowed. Image
covers ?1.5 cm across and shows angular, sand?sized
grains. Label points to one of several grains that are inter-
preted to be coated with a bright material. MI frame
2M302672578FFLB1E5P2956M2M1.
Figure 14. Grain size analyses presented as cumulative
percent of grains larger than some size as a function of grain
mean diameter in F units, which are ?log2(diameter in mm).
Graphic means, standard deviations, and skewness (skew
toward coarse or fine sizes relative to log normal distribu-
tion) computed using formulae of Lewis and McConchie
[1994] and presented in Table 3. The curves were derived
from counts using MI data, computing mean grain diameters
as the average of long and short dimensions. Penina targets
are skewed toward larger grains relative to the Cyclops data.
The later Penina 4 target is skewed even more toward larger
grains than the earlier target, Penina 1. Further, the Cyclops
targets are well sorted as compared to the other targets.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
11 of 19
displaced laterally from the others in order to provide
compositional mixing trends for APXS analyses. A fourth
Penina target was also investigated 83 sols after the other
Penina targets in order to search for textural and composi-
tional changes that might be caused by, for example, wind
erosion of the particles. Finally, the RAT was used as a
bevameter to press into the soil in Ulysses using a pre-
defined load to evaluate sinkage as a function of applied
normal stress.
[21] The second category of targets included MIs on Olive
tree and multiple Olive branch and Olive leaf targets for
which MI, APXS, and MB data were acquired. These targets
focused on the relatively bright material coming from the
top of the eastern ?cliff? bordering Ulysses and deposited on
the ?valley? wall (Figure 12). The third target category
included undisturbed soil surfaces in order to test the
hypothesis that the brown, yellow, and white material in
Ulysses exist beneath the surface to the east of Ulysses. The
first, Cyclops eye, was located ?10 cm to the east of the
Ulysses eastern wall where three laterally displaced sets of
MI and APXS measurements (Cyclops eye 1?3) were
acquired, along with one MB measurement (Figure 10). The
Cyclops eye 3 target was also brushed after a grind scan, and
MI, APXS, and MB data were acquired (Cyclops eye 4).
Finally, another grind scan was implemented followed by
another suite of measurements (Cyclops eye 5). The grind
scan was to be followed by an actual grind using the metal
grind pads as substitutes for the grind bits (Table 1). This
action failed, although the grind scan did expose relatively
bright material for measurements. An additional target was
selected 10 cm to the east of Cyclops eye. This target,
Polyphemus eye (Figure 10), was ground twice with mea-
surements after each RAT activity (Polyphemus eye 1?2). A
final undisturbed soil MI, APXS, and MB measurement set
was conducted on the target, Thoosa, located near the
Polyphemus eye targets.
[22] MI observations show that Sackrider and Penina
consist of angular, poorly sorted, sand?sized particles
(Figures 13?14 and Table 3) [Siebach, 2010]. Penina 4
exhibits the coarsest grains, interpreted to be due to win-
nowing of fine grains by winds during the period between
acquisition of Penina 3 and 4 MI observations. Further, the
particles are bimodal in brightness, with several examples in
which relatively bright material can be seen coating dark
grains. Also seen are dark grains without coatings and grains
composed of bright material. The Olive tree target shows
that the top of the Ulysses eastern cliff consists of a series of
?1 cm crust?like layers of cemented grains that provide
relatively bright material as scree down the sides of the
Ulysses trench (Figures 10 and 15). The Olive leaf targets,
again a crusty layer at the top of the Ulysses eastern cliff,
show similar textures as the Olive tree target. The crusts for
both targets are dominated by poorly sorted sands with sizes
similar to those found for Ulysses sands. Below the more
competent layer the surface slopes less steeply, consistent
with exposures of less indurated materials.
[23] As noted, the Cyclops, Polyphemus eye, and Thoosa
targets were located to the east of the Ulysses cliff and wall
(Figures 10 and 16). The MI data for the undisturbed sur-
faces show a particle size distribution that is skewed toward
finer sizes and indicates better sorting than the material
within Ulysses (Figure 14 and Table 3). A subpopulation of
dark, well?rounded grains of likely aeolian origin is also
evident for these undisturbed targets. The RAT brushing and
grinding in Cyclops eye, based on Pancam stereo data,
excavated ?0.5 cm beneath the surface by brushing and
exhumed an additional ?0.5 cm by using the RAT grind
scan capability. The final excavation exposed white, sand?
sized grains (Figures 16?17). These materials look similar to
the bright material on the cliff and slopes within Ulysses,
including brighter material coating darker grains. Pancam
and MI coverage of the Thoosa and both Polyphemus eye
targets (as deep as ?1 cm), on the other hand, look rather
similar to the undisturbed Cyclops eye targets.
[24] Mineralogical information derived from Pancam 13
filter spectral observations and MB observations of iron
mineralogy and oxidation state, together with elemental
abundances inferred from APXS data, provide important
additional information about materials encountered during
the in situ measurement campaign. The approach taken in
this paper is to first use the MB data to constrain iron
Table 3. Troy Target Grain Size Statisticsa
Parameter Olive Branch 1 Sackrider Penina 1 Penina 4 Cyclops Eye 1 Cyclops Eye 5
Mean 1.99 2.18 2.08 1.89 2.27 2.33
Graphic standard deviation, s 0.60 0.67 0.70 0.75 0.45 0.48
Skewness ?0.11 ?0.13 ?0.16 ?0.27 ?0.04 ?0.09
aUnits are F. Note that these are based on number of grains smaller than some fraction and not weight percentages, which is the typical method used for
reporting grain size analyses for terrestrial soils and sediments.
Figure 15. MI mosaic of Olive tree, a soil crust underlying
the cliff on the eastern side of the Ulysses disturbed soil
zone. See Figure 10 for location. The letters A, B, and C
are located on plateaus associated with discrete crust layers
exposed during Spirit?s embedding. Data acquired on sol
1927 and processed as a merged focal section.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
12 of 19
mineralogy and oxidation state and then to use these results to
help interpret Pancam spectral data. This is a valid approach
because the dominant features in Pancam data (0.4 to 1.0 mm)
result from the presence of iron bearing minerals [Bell et al.,
2003]. A ferric?bearing sulfate phase, with a similar MB
signature as found for the Arad sulfate?rich soils excavated
by Spirit in the northern part of the Inner Basin (R. V. Morris
et al., manuscript in preparation, 2010), dominates the iron
mineralogy for the Ulysses targets, particularly for the Olive
leaf target (Table 4). The values of the quadruple splitting
are not consistent with ferricopiapite as suggest by Lane et
al. [2008] for Paso Robles and by Johnson et al. [2007]
for Home Plate sulfate?rich soils. The reason is that the
two ferric doublets, as reported for ferricopiapite [Lane et al.,
2008], are not detected for the Mars data (R. V. Morris et al.,
manuscript in preparation, 2010). By analogy with Arad,
these sulfate?rich soils are hydrated on the basis of excess
light elements, perhaps with nine waters of hydration as
compared to the twenty needed for ferricopiate [Campbell
et al., 2008]. Detailed analyses of the likely phases are
reported by Morris et al. (manuscript in preparation, 2010).
[25] Correspondence analysis was applied to MB data set
shown in Table 4 to better understand dominant trends [e.g.,
Larsen et al., 2000; Arvidson et al., 2006, 2008]. Ninety?six
percent of the variance can be explained by the first two
factors, with 90% loaded onto the first factor (Figure 18).
Factor loadings define two groups of targets, with the
Cyclops eye 3, 4 and Polyphemus eye targets dominated by
olivine and pyroxene, with the addition of nanophase iron
oxide for the surface or near?surface targets. As noted,
Ulysses targets are dominated by a ferric sulfate phase. They
also show an affinity for hematite. The Cyclops eye 5 target
is located at the intersections of linear projections from the
Ulysses materials and the surface materials. Increasing iron
oxidation state trends follow the nanophase iron oxide
abundance trends for surface targets and increasing amounts
of hematite and ferric sulfate for the subsurface targets.
These results are consistent with the dominance of basaltic
materials and nanophase iron oxide dust or coatings at the
surface, with increased hematite and ferric sulfate abun-
dances just beneath the surface to the west of the Polyphe-
mus eye targets.
[26] Pancam 13 filter data acquired on sol 1933 (Figure 12)
were coregistered and calibrated to reflectance using the
Pancam calibration target following procedures defined by
Bell et al. [2003, 2006]. The seven left eye images (0.432
to 0.753 mm) were then run through a noise?weighted
principal components analysis and the first two components,
accounting for 98% of the variance of the data, were used to
Figure 16. Pancam false color views of the sequential boring into Cyclops eye using the RAT. The loca-
tion is shown in Figure 10, and the experiment was meant to search for the lateral extent of bright soils to the
east of Ulysses. (left) Frame was acquired on sol 1906, before any RAT activities, and shows the
approximate location of IDD measurements that define Cyclops eye 1 to 3, with offsets to provide com-
positional trends. (middle) The sol 1967 data were acquired after a RAT brush exposed a red layer (Cyclops
eye 4) similar to the layer observed on the undisturbed surface at the edge of Ulysses (see Figure 12). (right)
Image was acquired after a RAT grind scan removed the red layer and exposed white granular material
(Cyclops eye 5). RAT holes are approximately 4 cm in width.
Figure 17. Portion of an MI image of the bottom left sur-
face of Cyclops eye 5 acquired on sol 1990 when the target
was in shadow. Frame covers ?1.5 cm in width and mean
grain size is 2.03F (see Table 4). Label points to two
grains that are interpreted to be coated with a bright mate-
rial. MI frame 2M303036777EFFB1E5P2936M2F1.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
13 of 19
retrieve spectral end?members (Figure 19). Olive pit, a dark
pebble on the floor of Ulysses, plots as an end?member
along a mixing line, with red exposures on the undisturbed
surface to the east of Ulysses as the other end?member.
Dark sandy patches on the undisturbed surface, interpreted
to be basaltic sands, fall on the mixing line. Spectra for
Sackrider, Penina, and Olive leaf targets deviate from the
mixing line and form a diffuse mixing zone with Olive leaf
as an end?member.
[27] Pancam spectra for the targets labeled in Figures 12
and 19 are shown in Figure 20. All of the Pancam data
exhibit a ferric edge shortward of 0.7 mm that appears in
almost all spectra of Mars and indicates the presence of at
least minor amounts of nanophase iron oxide for all targets.
The dark patches on the undisturbed surface are consistent
with the dominance of olivine and pyroxene based on
shallow Fe+2 electronic absorptions at the longest wave-
lengths. Spectra for Olive pit (not shown) are consistent with
a largely unweathered basaltic mineralogy. The red layer on
the undisturbed surface is characterized by a very strong ferric
edge, consistent with a spectral dominance of nanophase
iron oxide.
[28] Olive leaf and Penina spectra show a subtle absorption
feature between 0.75 and 0.9 mm. The Sackrider spectrum
shows the subtle downturn at 0.75 mm evident in Olive leaf
and Penina spectra, but not the upturn at longer wavelengths.
Table 4. MB Iron Mineralogya
Sol, Type, and Name Ol Px Fe+3Sulfate Hm NanoFeOxide Fe+3/FeT
A1923SD Sackrider 25 10 58 7 0 0.65
A1930SD Olive branch 2 25 16 50 9 0 0.58
A1937SD Penina 1 (combined A1937 and A1943) 16 14 66 4 0 0.70
A1949SU Cyclops eye 3 43 37 0 3 17 0.19
A1970SR Cyclops eye 4 29 46 5 2 18 0.23
A1983SR Cyclops eye 5 28 36 21 6 9 0.35
A1997SR Polyphemus eye 2 23 44 0 8 26 0.30
A2010RD Olive leaf 1 6 8 79 7 0 0.85
A2026SD Penina 4 23 17 60 8 0 0.60
A2053SU Thoosa 34 35 0 8 23 0.31
aA refers to MER A or Spirit. The sol is next, followed by SD for disturbed soil, SU for undisturbed soil, SR for ratted soil, RD for disturbed rock
(actually soil crust). From Morris et al. (manuscript in preparation, 2010).
Figure 18. MB based results for the proportions of iron
bearing phases for Troy soil measurements (see Table 4)
were used to solve for and plot factor loadings using cor-
respondence analysis. The first two factors control 96% of
the variance of the data set. Numbers in parentheses repre-
sent Fe+3/Fetotal in percent. Cyclops eye 3 is an undisturbed
surface, Cyclops eye 4 is the brushed surface exposing red
material, and Cyclops eye 5 is the surface exposing white
material (see Figure 16). Polyphemus eye 2 is the deepest
hole for that RAT grinding experiment. Olive leaf and Olive
branch targets were on relatively bright material on the cliff
and scree slope of Ulysses whereas Penina and Sackrider
were in the interior of the Ulysses disturbed soil area. These
ferric sulfate?rich materials also show an affinity for
hematite.
Figure 19. Plot of the first two principal components
(expressed as minimum noise fractions or MNFs) for the
sol 1933 Pancam 6 filter left eye wavelengths (0.432 to
0.753 mm). The left eye data were chosen because they have
more variance than the right eye data and because geometric
warping of the right eye to the left eye produced data gaps
that lowered the number of pixels available for principal
components analysis. The first two factors control 85% of
the six?dimensional spectral data set variance. Mixing pat-
terns are discussed in the text.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
14 of 19
Continuum removed versions of the Olive leaf, Penina, and
Sackrider spectra indicate the presence of the subtle feature
between 0.75 and 0.9 mm in spectra for all three targets. As
noted, the Penina and Sackrider spectra shown in Figure 20
are not end?members, but lie between the Olive leaf position
and a relatively diffuse mixing zone defined by Olive pit and
red layer end?members. The positions of Sackrider and
Penina within the mixing space are probably governed by
local tilts, in addition to the amount of ferric sulfate material.
In particular, the Sackrider target was tilted away from the
sun relative to the Penina target and thus reflecting less
radiance toward Pancam when the data were acquired. The
subtle absorption feature between 0.75 and 0.9 mm is con-
sistent with spectra obtained experimentally by Ling et al.
[2009] for moderately hydrated ferric sulfates, e.g., ferric
hexahydrate. Thus, the combination of MB, APXS, and Pan-
cam data, using Arad results to help bootstrap inferences,
provide a self?consistent conclusion that Ulysses exposes
moderately hydrated ferric sulfates.
[29] APXS compositional information for relevant samples
from Troy is provided in Table 5. This extensive data set
(21 samples and 11 variables) was subject to correspondence
analysis to search for trends between samples and oxide
compositions. Results are shown in Figure 21 and exhibit
trends that are similar to those observed for the MI, MB, and
Pancam results. That is, the first factor controls most of the
data set variance and is represented by ferric sulfate?rich
samples on the right side and basaltic materials on the left
side. For example, the shallow Cyclops eye target, along
with Thoosa and Polyphemus eye targets, plot on the
basaltic side, whereas the Penina and Sackrider targets plot
on the ferric sulfate side. In fact, the Penina 3 target, located
furthest along the ferric sulfate vector direction, has the
highest concentration of SO3 (36%) of any of the APXS
targets for either Spirit or Opportunity. Cyclops eye 5 plots
between the two sets of end?members. This mixing line
controls 97% of the variance of the data set. The second
factor shows that the subsurface targets within the Ulysses
area in Scamander crater are rich in ferric sulfate. In addi-
tion, the shallow targets, Olive leaf and Olive branch, show
affinities for CaO and SiO2. This suggests the presence of
modest amounts of calcium sulfate and silica, although
Figure 20. Pancam spectra retrieved for targets in the sol
1933 Pancam data shown in Figure 12. The spectral region
shown is dominated by electronic transition absorption fea-
tures associated with iron, including the ubiquitous ferric
edge at short wavelengths and ferrous bands associated with
iron?bearing olivine and pyroxene at long wavelengths. Note
that the red layer (location shown in Figure 12) has a par-
ticularly strong ferric edge. Olive leaf shows a bright spectral
reflectance and a subtle absorption from 0.75 to 0.9 mm that is
consistent with the presence of ferric sulfates with modest
water content, as discussed in the text. This signature is also
evident in the spectra for Penina and Sackrider, particularly
for continuum removed spectra. Not corrected for cosine of
solar incidence angle.
Table 5. APXS Compositions as Oxide Weight Percentsa
Sol, Type, and Name Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 Cr2O3 MnO FeO
A1925SD Sackrider 1.38 4.46 4.92 28.4 0.79 33.03 0.30 0.14 5.75 0.43 0.43 0.20 19.7
A1927SD Olive branch 1 1.23 3.92 4.74 33.7 0.66 30.91 0.29 0.15 6.60 0.44 0.47 0.16 16.7
A1928SD Olive branch 2 1.38 4.35 5.27 35.0 0.68 28.64 0.31 0.17 5.85 0.48 0.47 0.19 17.1
A1934SD Olive branch 3 1.37 4.39 5.31 36.2 0.70 27.88 0.32 0.19 5.66 0.50 0.45 0.17 16.7
A1935SD Olive branch 4 1.45 4.52 5.71 36.7 0.70 26.66 0.33 0.21 5.48 0.51 0.44 0.19 17.1
A1939SD Penina1 1.08 4.07 4.79 28.7 0.73 34.74 0.29 0.15 5.64 0.37 0.50 0.15 18.7
A1940SD Penina2 0.96 3.91 4.61 28.6 0.76 35.56 0.28 0.12 5.28 0.36 0.51 0.15 18.9
A1941SD Penina3 1.02 3.80 4.50 28.2 0.76 35.82 0.27 0.12 5.42 0.41 0.49 0.15 18.9
A1945SU Cyclops eye 1 2.88 8.82 10.24 45.4 0.78 7.26 0.48 0.37 6.23 0.73 0.34 0.32 16.0
A1946SU Cyclops eye 2 2.83 8.89 10.37 46.0 0.78 6.76 0.46 0.35 6.26 0.72 0.33 0.31 15.9
A1947SU Cyclops eye 3 2.94 9.16 10.38 45.9 0.77 6.20 0.45 0.37 6.25 0.72 0.33 0.30 16.1
A1966SB Cyclops eye 4 2.80 8.16 9.16 45.7 0.87 8.78 0.59 0.36 5.75 0.89 0.37 0.31 16.2
A1967SD Olive branch 5 0.88 5.05 5.37 37.3 0.71 26.76 0.29 0.18 6.40 0.44 0.66 0.18 15.7
A1973SR Cyclops eye 5 2.31 6.85 8.15 40.4 0.80 15.93 0.42 0.32 5.70 0.83 0.46 0.29 17.4
A1979SD Olive pit 1.33 7.51 4.92 36.5 0.63 23.30 0.33 0.12 5.36 0.37 0.51 0.25 18.7
A1981SRPolyphemuseye2 2.65 8.45 9.01 42.9 0.87 9.36 0.65 0.36 6.03 1.02 0.44 0.36 17.7
A1995SRPolyphemuseye3 2.46 8.04 8.88 42.8 0.90 9.80 0.68 0.40 5.97 0.96 0.46 0.36 18.2
A2006RD Olive leaf 1 0.58 3.60 4.03 29.2 0.65 34.20 0.32 0.11 8.35 0.35 0.53 0.15 17.8
A2016RD Olive leaf 2 1.10 4.69 5.25 32.4 0.72 29.30 0.37 0.19 7.66 0.45 0.51 0.16 17.2
A2024SD Penina 4 1.46 5.37 6.16 33.9 0.78 27.94 0.33 0.18 5.63 0.50 0.54 0.18 17.0
A2071SUThoosa 2.87 8.77 9.61 46.1 0.85 6.21 0.62 0.39 6.11 0.80 0.37 0.31 16.9
aCr2O3 and TiO2 were not used in correspondence analysis because estimated measurement errors are comparable to retrieved abundances. Same
nomenclature as used for MB table. From Morris et al. (manuscript in preparation, 2010).
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
15 of 19
direct phase detection is not possible with the rover science
payload.
7. Troy Extrication Attempts
[30] It is not surprising that Spirit made little progress
during the initial extrication attempts that began on sol 2078
(see Table 2 for temporal summary). Motion resistance far
exceeded the available thrust and the wheels basically spun
in place. In addition, on sol 2104 the right rear wheel drive
actuator failed, forcing four wheel drives with only the
middle wheel operative on the right side. Additional pro-
blems were encountered when the left middle wheel on the
passive bogie suspension was raised off the ground because
of unbalanced torques generated by the other wheels. From
sols 2078 to 2142 twenty?one drives in the forward direction
produced only a few centimeters of lateral motion.
[31] The breakthrough for extrication was the decision to
perform wheel ?wiggles? for the outer wheels before each
drive segment. Wiggles consisted of using the steering actua-
tors to rotate the outer wheels periodically during drives. This
action emulated blades plowing though the soil around the
wheels and lowered the local compaction and bulldozing
resistance. Accordingly, Spirit made 34 cm of progress during
its last set of ten backward drives from sols 2145 to 2165 that
included wheel wiggles. In addition, the left front wheel
produced a set of sand ?waves? and excavated additional
bright soils during these last drives (Figures 22?23). After
the sol 2169 drive a stand down was called in order to prepare
Spirit for the winter, with the expectation that drives would
continue when solar energy became high enough during the
ensuing spring to resume activities. Vehicle yaw, pitch, and
roll just after the last drive on sol 2165 was 310, ?4, and 16?
respectively, with a 9? southerly tilt.
8. Pedogenic History
[32] The left side wheels excavated more bright soil
within Scamander crater during Spirit?s final 34 cm of
backward drives (see Figure 11). The dragging of the right
front wheel, on the other hand, did not excavate bright
material. Spirit was sitting on the side of Scamander crater
before extrication and the rearward drives kept the left
wheels within the crater. This indicates that Scamander has
been in?filled by a mix of ferric sulfate and basaltic mate-
rials. The Pancam false color mosaic generated after Spirit
stopped for the winter (Figure 23) shows a red layer to either
side of the sulfate?rich disturbed soils, extending onto the
floor of Scamander crater. The appearance of this material is
similar to that of the red layer found near C. L. Moore
(Figure 4) and a search of Pancam color data for Arad and
Tyrone sulfate soils exposed by wheel motions also suggests
the presence of a red layer.
[33] The shallow subsurface crusty soils exposed on the
eastern Ulysses cliff are enriched in hematite and ferric
sulfate, and perhaps calcium sulfate and silica, relative to the
bulk of the ferric sulfate and basaltic materials. The origin of
the sulfate and silica?rich materials in the Inner Basin in
association with fumarole and/or hydrothermal acid?sulfate
aqueous environments has been discussed in detail by Yen
Figure 21. APXS?based correspondence analysis as
applied to soils measured while Spirit was embedded at Troy.
Oxide compositions are given in Table 5. Factor one dom-
inates the data set variance and represents a mixture between
basaltic and sulfate?rich end?members. Factor two allows
subtle variations in composition to be inferred, including the
CaO and SiO2?rich Olive leaf and Olive branch materials.
Note also the displacement of Cyclops eye 5 away from
surface soils toward sulfate?bearing materials here and in
Figure 18. Olive branch 5 and Olive pit (shown in italics)
represent measurements with significant contamination by
dark pebbles of likely basaltic composition. Target names are
abbreviated on plot as follows: Ob, Olive branch; P, Penina;
Sr, Sackrider; Op, Olive pit; Ol, Olive leaf; Cy, Cyclops eye;
Pe, Polyphemus eye; Th, Thoosa.
Figure 22. Front Hazcam image taken on sol 2169 after
Spirit?s backward extrication maneuvers ceased because
of decreasing energy availability due to the onset of the
winter season. The last set of ten drives produced 34 cm
of motion, with additional excavation of material in Sca-
mander crater by the left front wheel, and exposure of part
of the Rock garden. Key targets are labeled. Frame
2F318929944RALB27MP1214L0M1.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
16 of 19
et al. [2008], Wang et al. [2008], Squyres et al. [2008],
and S. Ruff et al. (manuscript in preparation, 2010), and
will not be addressed in this paper. Rather, the focus is on
the processing of these materials after mixing with basaltic
materials and accumulation within Scamander crater. The
most likely explanation for the occurrence of ferric sulfates
and basaltic sands within the crater is that these minerals
were transported from other locations within the Inner Basin
and accumulated as aeolian deposits. Alternately they might
be primary fumarole or hydrothermal deposits. In any case the
data indicate that ongoing pedogenic processes have modi-
fied these materials since their emplacement.
[34] The addition of neutral to slightly acidic water to
material containing ferric sulfate minerals would result in
preferential ferric sulfate dissolution, iron oxide formation
(nominally as hematite), and generation of a sulfuric acid?
rich fluid (Figure 24). Coexisting basaltic materials would
also be unstable in such fluids, as exemplified by their
increasing solubility with decreasing pH (Figure 24). On the
other hand dissolution rates of basaltic materials would be
slow as compared to dissolution rates for ferric sulfates
[Brantley, 2008]. Competing with these dissolution reac-
tions would be the downward migration of the acidic fluid
driven by gravity and/or diffusion, which would be subse-
quently followed by sublimation or freezing. Provided that
the acidic fluid was not fully neutralized by reaction with
basaltic materials, the freezing or sublimation of this fluid
would result in the redeposition of ferric sulfates, forming
layers enriched in this material. Given the expected tem-
perature and pressure conditions, the latter process likely
occurred substantially faster than neutralization by reaction
with basaltic components.
[35] Thus it is postulated that neutral to slightly acidic
water has been added episodically to the soils in the valleys
surrounding Home Plate. Aqueous solution then migrated
into the subsurface, ultimately leaving behind relatively
insoluble species close to the surface and ferric sulfate?
enriched layers perhaps just centimeters beneath the surface.
The observation that the sulfate?rich deposits and the surface
crusts follow the topography is an important clue for under-
standing the time scale associated with the evolution of these
deposits. It implies that the dissolution and reprecipitation
must have been ongoing processes because the landforms in
the Inner Basin have continually evolved due to aeolian
Figure 24. The pH?dependent solubility of site?relevant
mineral phases. See Appendix A for details regarding the
solubility calculations and text for explanation.
Figure 23. Pancam false color mosaic from sols 2163 and
2177 showing Ulysses disturbed soils excavated during the
last ten backward extrication drives. Note ?sand waves?
excavated by the left front wheel. The brightness and color
indicate that these newly exposed deposits are also sulfate?
rich, providing further evidence that the floor of Scamander
crater is underlain by these materials. Key targets are shown,
including the location of a press test where the RAT was
used to determine soil deformation as a function of applied
normal stress. Note the white soil exposed on the cliffs to
the right of the newly excavated soils. The original Ulysses
area acquired a reddish color (e.g., compare to sol 1933
Pancam data, Figure 12), presumably due to dust deposi-
tion. The IDD instruments block part of the newly excavated
soil exposures.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
17 of 19
erosion and deposition [e.g., Grant et al., 2006]. We pos-
tulate that production of thin films of water associated with
frost formation and snowpack development, e.g., during
periods of high obliquity, have provided the necessary small
amounts of water. Reactions have not gone to completion
because basaltic minerals persist in the crusts and sands of
Troy and have not reacted completely with the ferric sulfates.
9. Conclusions and Implications
[36] Spirit spent its third winter on the northern slopes of
Home Plate. During the ensuing spring season the rover
drove down onto the valley to the north of Home Plate on its
way south to investigate von Braun and Goddard. An initial
target consisted of nodular rock that was found to be com-
posed primarily of silica. With its limited five?wheel drive
capability Spirit was unable to get clockwise around the
northeastern side of Home Plate or counterclockwise using
the valley to the west of Home Plate. Drives on soil?covered
slopes proved to be particularly problematical and led to
significant slippage, sinkage, and excavation of subsurface
soils. This ultimately led to embedding in the sands of Troy
in the valley to the west of Home Plate. On the other hand,
the excavated soils were found to be enriched in ferric
sulfate?bearing materials located only centimeters beneath
crusty surface layers enriched in these salts together with
iron oxides, calcium sulfate, and silica. The valleys around
Home Plate are thus special places in which soils contain
substantial deposits of highly soluble ferric sulfate materials
of likely fumarole or hydrothermal origin that have been
processed by ongoing aqueous processes to form thin crusts
overlying mixtures of sulfate?rich and basaltic sands. These
areas are intrinsically interesting but also hazardous for
driving a mobility?impaired vehicle.
[37] On Earth fumaroles and hydrothermal systems provide
the environmental conditions, water, nutrients and energy
sources needed to sustain robust microbial communities
[Walter and Des Marais, 1993]. For example, despite their
acidity, iron?rich sulfate hot springs in Yellowstone National
Park support thriving microbial communities [Innskeep and
McDermott, 2005]. Spirit has shown that deposits rich in
silica and/or ferric sulfates of likely fumarole or hydro-
thermal origins occur along most, if not all, of the perimeter
of Home Plate. It seems likely that the region in and around
Home Plate may have likewise supported a habitable envi-
ronment. Whether habitable environments accompanied the
more recent aqueous activity that redistributed the near?
surface ferric sulfate salts at Troy is less certain. The water
activities of solutions saturated in ferric sulfate may have
been too low to maintain microbial metabolism [e.g., Tosca
et al., 2008]. Thus the near?surface transient brines that
redistributed ferric sulfate in the sands of Troy probably did
not sustain habitable conditions.
Appendix A
[38] Calculations were performed using The Geochemist?s
Workbench? [Bethke, 2009]. The Lawrence Livermore
National Laboratory thermochemical database [Delany and
Lundeen, 1990] was employed with additional mineral sol-
ubility constants at 0?C obtained from the compilation by
Marion et al. [2008]. An extended Debye?Huckel activity
correction model was employed that is parameterized to be
accurate in up to 3 m NaCl solution and approximately
0.5?1 m ionic strengths of other electrolytes [Helgeson,
1969; Helgeson and Kirkham, 1974a, 1974b]. The solu-
bilities of the most soluble phases (magnesium sulfate,
halite, ferric sulfate, gypsum, sodium sulfate) were calcu-
lated using the extension of the thermochemical database of
Harvie et al. [1984] by Tosca et al. [2008], which employs a
Pitzer?type activity model to account for the effect of high
ionic strength solutions. Hematite solubility was also cal-
culated in this model to establish a relatively accurate
intersection with the ferric sulfate solubility curve. The main
hydrolysis species of iron were added to the Tosca et al.
[2008] model for this purpose.
[39] Mineral solubilities as a function of pH were calcu-
lated by first equilibrating the mineral phase with pure water
and then titrating the system using Ca(OH)2 or H2SO4. These
species were chosen to reflect a Ca?SO4 electrolyte that
should dominate at the site in question. Except at extreme pH
values the solubility curves are generally independent of
electrolyte concentration and composition as the activity
corrections and complexation by Ca2+ and SO4
2? were mini-
mal for the species considered. The more soluble phases were
titrated to high pH using the mineral cation to prevent
supersaturation with respect to gypsum or anhydrite.
[40] It is acknowledged that quantitative solubility calcu-
lations are system dependent because of activity corrections,
common ion effects, and complexation. However, the pur-
pose of the diagram shown in Figure 24 was not to produce
solubility curves to be employed directly in quantitative
chemical models. Rather, these calculations are intended to
illustrate how the solubilities of key minerals at this site can
differ by more than 13 orders of magnitude and also vary
substantially as a function of pH. Because system?specific
effects typically modify individual values by only an order of
magnitude or less, the larger differences shown in relative
solubilities displayed in Figure 24 illustrate the potential
consequences of aqueous processes in these deposits.
[41] Acknowledgments. We thank the capable team of engineers and
scientists at the Jet Propulsion Laboratory and elsewhere who made the
Spirit mission possible. We also thank support from NASA for the science
team.
References
Arvidson, R. E., et al. (2004), Localization and physical properties experi-
ments conducted by Spirit at Gusev crater, Science, 305, 821?824,
doi:10.1126/science.1099922.
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.
Arvidson, R. E., et al. (2008), Spirit Mars Rover Mission to the Columbia
Hills, Gusev Crater: Mission overview and selected results from the
Cumberland Ridge to Home Plate, J. Geophys. Res., 113, E12S33,
doi:10.1029/2008JE003183.
Bekker, M. G. (1969), Introduction to Terrain?Vehicle Systems, Univ. of
Mich. Press, Ann Arbor.
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.
Bell, J. F., III, et al. (2006), In?flight calibration and performance of the
Mars Exploration Rover Panoramic Camera (Pancam) Instruments,
J. Geophys. Res., 111, E02S03, doi:10.1029/2005JE002444.
Bethke, C. M. (2009), The Geochemist?s Workbench Release 8.0, Hydro-
geol. Program, Univ. of Ill., Urbana.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
18 of 19
Brantley, S. L. (2008), Kinetics of mineral dissolution, in Kinetics of
Water?Rock Interaction, edited by S. L. Brantley, J. D. Kubicki, and
A. F. White, pp. 151?210, Springer, New York, doi:10.1007/978-0-
387-73563-4_5.
Campbell, J. L., R. Gellert, M. Lee, C. L. Mallett, J. A. Maxwell, and J. M.
O?Meara (2008), Quantitative in situ determination of hydration of bright
high?sulfate Martian soils, J. Geophys. Res., 113, E06S11, doi:10.1029/
2007JE002959.
Delany, J. M., and S. R. Lundeen (1990), The LLNL thermochemical data-
base, Rep. UCRL?21658, Lawrence Livermore Natl. Lab., Livermore,
Calif.
Gellert, R., et al. (2006), Alpha Particle X?Ray Spectrometer (APXS): Re-
sults from Gusev crater and calibration report, J. Geophys. Res., 111,
E02S05, doi:10.1029/2005JE002555.
Grant, J. A., et al. (2006), Crater gradation in Gusev crater and Meridiani
Planum, Mars, J. Geophys. Res. , 111 , E02S08, doi:10.1029/
2005JE002465.
Harvie, C. F., N. M?ller, and J. H. Weare (1984), The prediction of mineral
solubilities in natural waters: The Na?K?Mg?Ca?H?Cl?SO4?OH?HCO3?
CO3?CO2?H2O system to high ionic strengths at 25?C, Geochim. Cos-
mochim. Acta, 48, 723?751, doi:10.1016/0016-7037(84)90098-X.
Helgeson, H. C. (1969), Thermodynamics of hydrothermal systems at ele-
vated temperatures and pressures, Am. J. Sci., 267, 729?804.
Helgeson, H. C., and D. H. Kirkham (1974a), Theoretical prediction of
thermodynamic behavior of aqueous electrolytes at high pressures and
temperatures. 1. Summary of thermodynamic?electrostatic properties of
solvent, Am. J. Sci., 274, 1089?1198.
Helgeson, H. C., and D. H. Kirkham (1974b), Theoretical prediction of
thermodynamic behavior of aqueous electrolytes at high pressures and
temperatures. 2. Debye?Huckel parameters for activity?coefficients and
relative partial molal properties, Am. J. Sci., 274, 1199?1261.
Innskeep, W. P., and T. R. McDermott (2005), Geomicrobiology of acid?
sulfate?chloride springs in Yellowstone National Park, in Geothermal
Biology and Geochemistry in Yellowstone National Park, edited by
W. P. Inskeep and T. R. McDermott, pp. 142?162, Montana State Univ.,
Bozeman.
Johnson, J. R., J. F. Bell III, E. Cloutis, M. Staid, W. H. Farrand, T. McCoy,
M. Rice, A. Wang, and A. Yen (2007), Mineralogic constraints on sulfur?
rich soils from Pancam spectra at Gusev crater, Mars, Geophys. Res. Lett.,
34, L13202, doi:10.1029/2007GL029894.
Lane, M. D., J. L. Bishop, M. D. Dyar, P. L. King, M. Parente, and B. C.
Hyde (2008), Mineralogy of the Paso Robles soils on Mars, Am. Mineral.,
93, 728?739, doi:10.2138/am.2008.2757.
Larsen, K. W., et al. (2000), Correspondence and least squares analyses of
soil and rock compositions for the Viking Lander 1 and Pathfinder land-
ing sites, J. Geophys. Res.105(E12)29,207?29,221, doi:10.1029/
2000JE001245.
Lewis, D.W., and D.McConchie (1994),Analytical Sedimentology, 197 pp.,
Chapman and Hall, New York.
Lindemann, R. A., and C. J. Voorhees (2005), Mars Exploration Rover
mobility assembly design, test and performance, IEEE Int. Conf. Syst.,
Man and Cybern., 1, 450?455, doi:10.1109/ICSMC.2005.1571187.
Ling, Z. C., A. Wang, and C. Li (2009), Comparative spectroscopic study
of three ferric sulfates: Kornelite, lausenite and pentahydrate, Lunar
Planet. Sci., XL, Abstract 1867.
Maimone, M., Y. Cheng, and L. Matthies (2007), Two years of visual odo-
metry on the Mars Exploration Rovers, J. Field Robot., 24(3), 169?186,
doi:10.1002/rob.20184.
Marion, G. M., J. S. Kargel, and D. C. Catling (2008), Modeling ferrous?
ferric iron chemistry with application to Martian surface geochemistry,
Geoch im. Cosmoch im. Ac ta , 72 , 242?266 , do i :10 .1016/ j .
gca.2007.10.012.
Rice, M. S., J. F. Bell III, E. A. Cloutis, A. Wang, S. Ruff, M. A. Craig, D. T.
Bailey, J. R. Johnson, P. A. de Souza Jr., andW. H. Farrand (2010), Silica?
rich deposits and hydrated minerals at Gusev Crater, Mars: Vis?NIR spec-
tral characterization and regional mapping, Icarus, 205, 375?395,
doi:10.1016/j.icarus.2009.03.035.
Richter, L., et al. (2006), A predictive wheel?soil interaction model for
planetary rovers validated in testbeds and against MER Mars rover per-
formance data, paper presented at 10th European Conference of the Inter-
national Society for Terrain?Vehicle Systems, Budapest, Hungary, 3?6 Oct.
Siebach, K. (2010), Recent Spirit results: Microscopic Imager analysis of
particle properties in Scamander crater, west of Home Plate, Lunar
Planet. Sci., XLI, Abstract 2548.
Squyres, S. W., R. E. Arvidson, S. Ruff, R. Gellert, R. V. Morris, D. W.
Ming, L. S. Crumpler, J. D. Farmer, D. J. Des Marais, and P. A. de Souza
(2008), Detection of silica?rich deposits on Mars, Science, 320, 1063?
1067, doi:10.1126/science.1155429.
Tosca, N. J., A. H. Knoll, and S. M. McLennan (2008), Water activity and
the challenge for life on early Mars, Science, 320, 1204?1207,
doi:10.1126/science.1155432.
Walter, M. R., and D. J. Des Marais (1993), Preservation of biological
information in thermal spring deposits: Developing a strategy for the
search for fossil life on Mars, Icarus, 101, 129?143, doi:10.1006/
icar.1993.1011.
Wang, A., et al. (2008), Light?toned salty soils and coexisting Si?rich species
discovered by the Mars Exploration Rover Spirit in Columbia Hills,
J. Geophys. Res., 113, E12S40, doi:10.1029/2008JE003126.
Wong, J. (2003), Theory of Ground Vehicles, 2nd ed., John Wiley, New
York.
Yen, A. S., et al. (2008), Hydrothermal processes at Gusev Crater: An eval-
uation of Paso Robles class soils, J. Geophys. Res., 113, E06S10,
doi:10.1029/2007JE002978.
R. E. Arvidson, J. G. Catalano, R. N. Greenberger, E. A. Guinness, K. A.
Lichtenberg, A. Shaw, K. L. Siebach, and A. Wang, Department of Earth
and Planetary Sciences, Washington University in St. Louis, St. Louis,
MO 63130, USA. (arvidson@rsmail.wustl.edu)
J. F. Bell III, M. S. Rice, S. W. Squyres, and R. J. Sullivan, Department
of Astronomy, Cornell University, Ithaca, NY 14853, USA.
P. Bellutta, T. A. Estlin, J. A. Herman, S. A. Maxwell, A. W. Stroupe,
K. P. Talley, J. A. Townsend, J. R. Wright, and A. S. Yen, Jet Propulsion
Laboratory, California Institute of Technology, Pasadena, CA 91125,
USA.
N. A. Cabrol and D. J. Des Marais, NASA Ames Research Center,
Moffett Field, CA 94035, USA.
J. Cohen, Honeybee Robotics Spacecraft Mechanisms Corporation,
460 West 34th St., New York, NY 10001, USA.
L. S. Crumpler, New Mexico Museum of Natural History and Science,
1801 Mountain Rd. NW, Albuquerque, NM 87104, USA.
P. A. de Souza, Information and Communication Technologies Centre,
CSIRO, Hobart, Tas 7001, Australia.
W. H. Farrand, Space Science Institute, 4750 Walnut St., Boulder, CO
80301, USA.
R. Gellert, Department of Physics, University of Guelph, Guelph, ONN1G
2W1, Canada.
J. A. Grant, Center for Earth and Planetary Studies, Smithsonian
Institution, PO Box 37012, Washington, DC 20013, USA.
K. E. Herkenhoff and J. R. Johnson, U.S. Geological Survey, 2255 North
Gemini Dr., Flagstaff, AZ 86001, USA.
K. D. Iagnemma, Department of Mechanical Engineering, Massachusetts
Institute of Technology, Cambridge, MA 02139, USA.
G. Klingelh?fer, Institut f?r Anorganische und Analytische Chemie,
Johannes Gutenberg?Universit?t, Mainz D?55099, Germany.
R. Li, Department of Civil and Environmental Engineering and Geodetic
Science, Ohio State University, Columbus, OH 43210, USA.
D. W. Ming and R. V. Morris, NASA Johnson Space Center, Houston,
TX 77058, USA.
S. W. Ruff, School of Earth and Space Exploration, Arizona State
University, Tempe, AZ 85287, USA.
ARVIDSON ET AL.: SPIRIT?OVERVIEW AND SELECTED RESULTS E00F03E00F03
19 of 19