in
ho
fDepartment of Terrestrial Magnetism, Carnegie Institution of Washington, Washington, DC 20015, USA
a r t i c l e i n f o
its full extent (Murchie et al., 2008; Fassett et al., 2009; Watters
et al., 2009). Because of high Sun (low incidence) angles in the ba-
sin area, however, the images do not give much indication of the
topographic relief and surface morphology. Moreover, craters with
bright ejecta rays are abundant in the area and mask some of the
details of the basin morphology. MESSENGER?s onboard laser
altimeter collected two elevation pro?les, one during each of the
tectonic evolution and provides important constraints on the ther-
mal state and mechanical properties of Mercury?s crust and litho-
sphere over time.
2. Image data
2.1. Camera
The Mercury Dual Imaging System (MDIS) (Hawkins et al., 2007,
2009) consists of a pair of wide- and narrow-angle cameras,
* Corresponding author at: German Aerospace Center (DLR), Rutherfordstr. 2, D-
12489 Berlin, Germany.
Icarus 209 (2010) 230?238
Contents lists availab
r
.e lE-mail address: Juergen.Oberst@dlr.de (J. Oberst).the long-wavelength topography of the model cannot be excluded, con?rmation of these undulations
must await data from MESSENGER?s orbital mission phase.
 2010 Elsevier Inc. All rights reserved.
1. Introduction
The Caloris basin (centered at 30N, 165E) was discovered in
images obtained by Mariner 10 during its three ?ybys of Mercury
in 1974?1975. Owing to the orbital resonance of Mercury and
the spacecraft, the same hemisphere of the planet was imaged un-
der similar illumination conditions during each ?yby, and the Mar-
iner 10 cameras were able to view only the eastern portion of
Caloris. Images obtained during the ?rst ?yby of Mercury by the
MErcury Surface, Space ENvironment, GEochemistry, and Ranging
(MESSENGER) spacecraft (Solomon et al., 2008) show the basin at
?rst two Mercury ?ybys, but these were in near-equatorial areas
(Zuber et al., 2008) and did not transect the Caloris basin area.
In this paper, we use stereo images obtained during the ?rst
MESSENGER ?yby of Mercury to derive a topographic model of
the Caloris basin, from which we examine the basin morphology
in detail. At a basin ring diameter of 1550 km (Murchie et al.,
2008; Fassett et al., 2009), Caloris is one of the largest impact ba-
sins known in the Solar System and among the youngest of the
large impact basins on Mercury. Its structure is relevant to the pro-
cesses of basin formation and subsequent modi?cation. Addition-
ally, the morphology of the basin re?ects its volcanic andArticle history:
Received 30 October 2009
Revised 1 March 2010
Accepted 10 March 2010
Available online 16 March 2010
Keywords:
Mercury
Cratering
Geological processes0019-1035/$ - see front matter  2010 Elsevier Inc. A
doi:10.1016/j.icarus.2010.03.009a b s t r a c t
A digital terrain model (1000-m effective spatial resolution) of the Caloris basin, the largest well-charac-
terized impact basin on Mercury, was produced from 208 stereo images obtained by the MESSENGER nar-
row-angle camera. The basin rim is far from uniform and is characterized by rugged terrain or knobby
plains, often disrupted by craters and radial troughs. In some sectors, the rim is represented by a single
marked elevation step, where height levels drop from the surroundings toward the basin interior by
approximately 2 km. Two concentric rings, with radii of 690 km and 850 km, can be discerned in the
topography. Several pre-Caloris basins and craters can be identi?ed from the terrain model, suggesting
that rugged pre-impact topography may have contributed to the varying characteristics of the Caloris
rim. The basin interior is relatively smooth and shallow, comparable to typical lunar mascon mare basins,
supporting the idea that Caloris was partially ?lled with lava after formation. The model displays long-
wavelength undulations in topography across the basin interior, but these undulations cannot readily
be related to pre-impact topography, volcanic construction, or post-volcanic uplift. Because errors inCenter for Earth and Planetary Studies, National Air and Space Museum, Smithsonian Institution, Washington, DC 20560, USA
dDepartment of Geological Sciences, Brown University, Providence, RI 02912, USA
eDepartment of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139-4307, USAThe morphology of Mercury?s Caloris bas
topographic models
J?rgen Oberst a,*, Frank Preusker a, Roger J. Phillips b, T
Sean C. Solomon f
aGerman Aerospace Center, Institute of Planetary Research, D-12489 Berlin, Germany
b Planetary Science Directorate, Southwest Research Institute, Boulder, CO 80302, USA
c
Ica
journal homepage: wwwll rights reserved.as seen in MESSENGER stereo
mas R. Watters c, James W. Head d, Maria T. Zuber e,
le at ScienceDirect
us
sevier .com/ locate/ icarus
coaligned on a pivot platform. Both cameras are equipped with
identical 1024  1024 format charge-coupled devices (CCDs) fea-
turing 14-lm pixels and can deliver images with a pixel dynamic
intensity range of 12 bits. Unlike the wide-angle camera (WAC),
the narrow-angle camera (NAC) possesses a single ?lter only
(700?800 nm). The NAC, featuring a compact off-axis optical sys-
tem, suffers from geometrical distortions. Laboratory data from
the ground as well as in?ight stellar calibration data have been
combination with changes in the orientation of the spacecraft. This
substantially (8?14) among respective image pairs, and were low-
est (8) for areas near the equator.
block adjustment techniques. This type of least-squares analysis
J. Oberst et al. / Icarus 20produces solutions for spacecraft position and camera pointing
for each image on the basis of large numbers of tie-point measure-
ments. For the tie points, ground coordinates are computed in the
process (Zhang et al., 1996). The a priori errors for angles were
450 lrad; errors in the spacecraft positions were assumed to be
small (50 m).
Table 1
NAC sub-mosaics used to produce the DTM.
Mosaic Time No. of images Resolution (m/pixel)
H1 E + 15 min 4  17 120?170
H2 E + 27 min 9  11 230?310
D1 E + 50 min 11  9 470?5703. Stereo processing
The stereo processing follows algorithms and software realiza-
tions that have been used extensively on previous planetary image
data sets (Albertz et al., 2005; Scholten et al., 2005; Heipke et al.,
2007; Gwinner et al., 2009, 2010). However, adaptations of the
software were required to account for MESSENGER camera param-
eters and image data formats. Also, software elements had to be
combined in shell scripts. The processing had several stages, as de-
scribed in the following.
3.1. Pointing data correction
The navigation data correction was carried out using bundlepowerful capability was used to acquire several contiguous image
mosaics during MESSENGER?s ?rst Mercury ?yby (M1) on 14 Janu-
ary 2008 (Solomon et al., 2008). The digital terrain model (DTM)
mosaic was produced from 208 NAC stereo images, which formed
three sub-mosaics (H1, H2, and D1; Table 1).
The viewing geometry and the stereo conditions during the ?y-
bys were far from optimum, which aggravated DTM production.
The basin was located near the planetary limb, and viewing (emis-
sion) angles near the northeastern margin of the basin were typi-
cally larger than 60. Stereo angles were also small, differedused to determine a geometric distortion map (F.S. Turner, per-
sonal communication, 2009), which has been utilized for this anal-
ysis. Indeed, once calibrated, CCD cameras are considered
geometrically more stable than the Mariner 10 vidicon cameras
and better suited for the photogrammetric measurement tech-
niques central to this paper.
2.2. Images and mosaics
Image mosaics are obtained by scans of the pivot platform inE (time of closest approach): 19:04 UTC, 14 January 2008. Time refers to ?rst image
taken.First, the images were subjected to automatic image matching
with a coarse matching grid to collect tie points. Only those tie
points that were matched on at least three images were selected.
In total, 21,975 line/sample coordinates were collected from the
208 images, representing 6475 ground points. Next, a starting
model was derived. Three-dimensional Cartesian body-centered
surface coordinates (x, y, z) of the tie points were computed from
the nominal navigation data. While the horizontal coordinates
were computed from classical ray intersections, the radius was
kept ?xed.
Next, full three-dimensional coordinates were derived by bun-
dle block adjustment, beginning the iterations with the starting
model. The inversion required solving the equations of observation
for more than 24,500 unknowns (i.e., three spacecraft position
components and three pointing angles for each image as well as
three coordinates for each point). The adjustment converged after
four iterations. Root mean square intersection errors were reduced
from 10 km originally to 220 m. The intersection errors were visu-
ally inspected to verify that the model was free of stresses and
topography bias. The high quality of the starting model prevented
the iterations of the bundle block adjustment from converging
erroneously on a local minimum of the residual ?eld. Small errors
in the starting model were found not to affect the outcome of the
adjustment.
3.2. Image matching
Images were pre-recti?ed using the pointing data from the bun-
dle block adjustment and subjected to automated image matching.
A total of 241 individual matching runs were carried out on dou-
ble- or triple-overlapping images. The pre-recti?cation step al-
lowed searches for tie points to be limited to small areas. Hence,
point misidenti?cations and gaps were reduced to a minimum.
3.3. Assembly of the DTM
Ray intersections of all matched points were computed to de-
rive surface coordinates of object points. This process resulted in
a total of 150 million object points. The object points were then
interpolated to form a contiguous DTM grid at a size of 12.8 million
pixels and a spatial resolution of 1000 m. While this spatial resolu-
tion is typically chosen to match the effective resolution of the
model (a factor of three larger than the resolution of the images,
according to a rule of thumb), a comparably small spatial resolu-
tion was chosen to avoid the loss of detail. No adjustments for
absolute height or trend were applied. Though some erroneous
height trend cannot be ruled out, gross errors in long-wavelength
topography of the model are not likely, owing to the large number
of concatenated images that are involved.
The DTM covers 12% of Mercury, or 8.8  106 km2 (Fig. 1).
Heights of the DTM are given with respect to a reference sphere
of radius 2440 km. For further discussion, the portion of the object
points that contained Caloris were extracted and projected sepa-
rately onto a stereographic map (Fig. 2) centered on Caloris
(30N, 165E). A variety of DTM representations were prepared,
including the raw DTM (grey scale, Fig. 2), as well as shaded-relief
maps color-coded by height (see Fig. 1 for an example). Selected
height pro?les were also computed for Caloris (Fig. 3) and areas
of interest (Figs. 4?6).
The DTM has 8 km of total relief. The highest point within Cal-
oris (2.5 km above the Mercury reference radius) is represented by
an unnamed isolated peak (42.6N, 149.9E, Fig. 4). The lowest
point within Caloris, and indeed of the DTM (5 km below the refer-
9 (2010) 230?238 231ence radius), is within the crater Atget in the southern basin ?oor
(25.6N, 166.4E). At 100 km in diameter Atget is the largest crater
within the Caloris basin (Fig. 5).
s 20232 J. Oberst et al. / Icaru4. Basin morphology
The DTM is a rich source of information on diverse aspects of
Caloris geology, in particular the Caloris basin rim, its ring struc-
ture, basin in?lling, and possible pre-impact topography.
4.1. Basin rim-crest topography
The Caloris basin rim is well covered (except for parts of the
southeastern sector, located beyond the planet limb) and can be
clearly identi?ed in the DTM. Along the northeastern margin, the
rim is characterized by rugged massifs, which may represent what
was identi?ed as ??Caloris Montes? in Mariner 10-based terrain
classi?cation schemes (Guest and Greeley, 1983; Spudis and Guest,
1988). These massifs have elevations above their surroundings of
1.5?3 km. Along the northwestern basin margin, the rim is repre-
sented by knobs within comparably smooth plains (??knobby
plains?). Inspection of individual knobs indicates that these have
typical horizontal scales of several kilometers and elevations of
1?2 km above their surroundings.
In other parts of the basin margin there is a lack of any obvious
elevated ring structure at the resolution of the DTM. Rather, there
appears to be a sudden drop from the surrounding elevation level
toward the ?at basin ?oor. This geometry is best seen along the
meandering southern margin of the basin, where elevations fall
from the surrounding terrain to the basin ?oor by 2?3 km over a
linear distance of 20 km (Fig. 6). On the northern margin, the rim
appears similar, but it is disrupted by craters and troughs.
Ridges and troughs oriented radially to the basin center disrupt
the basin rim in several places. In the Mariner 10 classi?cation
schemes, these landforms have been grouped within the ??Van Eyck
Formation? (Guest and Greeley, 1983; Spudis and Guest, 1988; Fas-
Fig. 1. Caloris area digital terrain model (DTM). The model has a grid size of 1000 m and
of radius 2440 km.covers an area of approximately 8.8  106 km2. Heights are with respect to a sphere
9 (2010) 230?238sett et al., 2009). The troughs represent a combination of secondary
crater chains and graben that originated during basin formation.
The most prominent of these features is located on the northwest-
ern margin of Caloris [Fig. 4; see also Fig. 7a of Fassett et al. (2009)]
and is characterized by two parallel troughs ?anked by steep (5?6
slope) ridges on either side that extend for about 400 km and rise
above the trough ?oors by approximately 2000 m (Fig. 4). Along
the trough ?oor (Fig. 4, pro?le cc0), slopes cannot be distinguished
from zero. The mean depth of the trough is 1.4 km below the ref-
erence datum.
4.2. Basin ring structure
Whereas the Mariner 10 images of the eastern margin sug-
gested a diameter of the main Caloris basin ring of 1300 km (Guest
and Greeley, 1983; Spudis and Guest, 1988), the MESSENGER
images reveal a larger diameter of 1550 km (Murchie et al.,
2008). Moreover, Fassett et al. (2009) argued for an elliptical shape
for Caloris with long and short diameters of 1525 km and 1315 km,
respectively.
Although the heavily eroded ring structure is dif?cult to identify
in image data, owing to illumination conditions, the DTM gives us a
fresh look at the basin structure. The DTM hints at residual ring
structures in several areas (see arrows in Fig. 2 or rings in Fig. 7),
but these features are not easily matched to other parts of a rim
represented by a single circular or elliptical shape. An isolated
chain of mountain peaks at the northwestern margin, the highest
point within Caloris, is a prominent example (Fassett et al.,
2009). We suggest that elevated topography of the basin rim re-
gion can be represented by two concentric circles, having diame-
ters of 1380 km and 1700 km (Fig. 7). The mean of the two
values (1540 km) agrees with the 1550-km diameter reported by
us 20J. Oberst et al. / IcarMurchie et al. (2008). The two rings are centered at 31.7N,
161.5E.
There is further evidence for two or perhaps three circular con-
centric topographic highs within the basin (see Fig. 3, marked
points A?C) at radii of approximately 315, 530, and 640 km,
respectively. These highs represent modest variations in elevation
with a horizontal scale of 50?100 km and with amplitudes of
approximately 300 m. Although these highs can be dif?cult to
identify in single pro?les, the circular pattern becomes visible in
the two-dimensional representation of the DTM. This pattern pro-
vides tentative con?rmation of earlier suggestions of inner rings
from Mariner 10 images (Spudis and Guest, 1988; Murchie et al.,
2008). At the resolution and limited coverage of the DTM, we do
not ?nd clear evidence of rings outward of those identi?ed here,
as suggested by Spudis and Guest (1988) from Mariner 10
observations.
4.3. Pre-Caloris topography?
From inspection of the area around Caloris, it is dif?cult to
determine a mean pre-impact elevation level, consistent with the
possibility that the Caloris impactor hit rugged terrain with sub-
stantial topography in the form of pre-existing basins or large cra-
Fig. 2. Grey-scale representation of the Caloris basin DTM. Black and white arrows m9 (2010) 230?238 233ters. We suggest that some of the Caloris basin-rim topography and
its lack of continuity may be related to pre-Caloris topography. The
rim structure shows signi?cant asymmetry in level of preservation
(see discussion above) as a function of radial range and sector, sim-
ilar to what has been reported from studies of lunar impact basins,
such as Imbrium (Head, 1982). From lunar studies it is known that
where the pre-existing topography was large, the basin rim crest
can be poorly developed. Some pre-Imbrium-basin lows are known
to have in?uenced the subsequent Imbrium mare ?ll. We ?nd
traces of several pre-Caloris basins and smaller craters, tentatively
marked in Fig. 7.
4.4. Caloris interior plains
The DTM provides critical data for assessing the nature of post-
Caloris plains. It indicates that the interior topography of Caloris is
smoother on average than its surroundings. Smooth areas follow
the geologic boundaries of the basin rim. These characteristics sup-
port the idea that the interior of the basin was in?lled by extensive
volcanism (Murchie et al., 2008; Head et al., 2008).
It is instructive to compare the morphology of Caloris with
those of lunar impact basins. Single pro?les across prominent ba-
sins were selected from Clementine lunar topography (Smith
ark segments of the outer and inner rings, respectively, of the basin rim region.
wn
234 J. Oberst et al. / Icarus 209 (2010) 230?238Fig. 3. Selected topographic pro?les and locations of areas shoet al., 1997). Caloris is shallower than the lunar Orientale basin,
which is similar in diameter but has a depth from rim crest to basin
?oor of 6 km (Fig. 8). The Caloris pro?le is more similar to those of
Serenitatis and Imbrium, both of which are comparably shallow
(approximately 1?2 km from rim crest to ?oor) and have ?at basin
?oors (Fig. 8). Unlike Orientale, Serenitatis and Imbrium are ?lled
with mare basalt to depths of several kilometers. The comparison
suggests that Caloris was likewise in?lled by large volumes of lava.
While the resolution of Clementine data pro?les are certainly lim-
ited, we suggest that further comparisons should be carried out
Fig. 4. Selected topographic pro?les along and across a prominent trough located on th
pro?le dd0 , interpreted as part of the inner Caloris basin ring, marks the highest point oin greater detail in later ?gures (identi?ed by ?gure number).using the vast volume of new altimeter data from lunar orbital
missions now becoming available (Araki et al., 2009; Smith et al.,
2010).
4.5. Long-wavelength topography
Within the Caloris basin interior, the DTM displays prominent
long-wavelength variations in topographic height, particularly in
the north?south direction (Fig. 3). A broad high dominates large
parts of the northern basin ?oor and at its highest points exceeds
e northwestern rim of, and radial to, the Caloris basin. The topographic high along
f the DTM.
crat
J. Oberst et al. / Icarus 209 (2010) 230?238 235the elevation of the basin rim by 1 km. A second, more modest
high, seen in parts of the southern basin ?oor, is ?anked by lows
to the north and south (Fig. 3). The northern of those lows, which
trends more or less east?west across the basin center, covers an
area at least as large as the southern high. The magnitudes of these
long-wavelength undulations in ?oor height reach 1?1.5 km (see
pro?le bb0 in Fig. 3). It is possible that these topographic patterns
are artifacts, as discussed below, but ?rst we consider the basin
?oor topographic model at face value.
Physical mechanisms to account for the long-wavelength
undulations in the Caloris basin ?oor displayed in the DTM are
all problematic. One possibility is that the highs are relics of ini-
tially high topography in the pre-Caloris target area, perhaps
Fig. 5. Selected topographic pro?les across Apollodorus (aa0) and Atget (cc0 and dd0)
interior of Atget marks the lowest point of the DTM.smaller analogs of the extensive high region to the west of the
basin (Fig. 1). The prominent Apennine?Archimedes area of the
Imbrium basin provides an example of this scenario; the Archi-
Fig. 6. Selected topographic pro?les across the southern rim of the Caloris basin. Note the
arrow is identi?ed as part of the outer Caloris rim.medes plateau and the Apennine peaks remained high during la-
ter volcanic ?lling and constitute one of the least ?ooded areas
of the Imbrium interior (Head, 1982). The Caloris interior is per-
vasively ?lled by plains material characterized by high re?ec-
tance and a steeper than average (or ??redder?) slope to the
re?ectance from visible to near-infrared wavelengths (Murchie
et al., 2008; Robinson et al., 2008; Denevi et al., 2009). Suf?-
ciently large younger craters in the basin ?oor have excavated
low-re?ectance material with shallower (or ??bluer?) spectral
slope (Murchie et al., 2008; Denevi et al., 2009), indicating that
such low-re?ectance material underlies the high-re?ectance
plains at a depth of several kilometers (Ernst et al., 2010). There
is no indication, however, that the thickness of high-re?ectance
ers and a prominent trough (bb0) within the Pantheon Fossae graben complex. Theplains is markedly less in areas of high ?oor topography on
the DTM, as would be expected if such areas were also high
prior to the time of volcanic in?lling.
steep drop in elevation across the rim. The topographic feature marked by the black
s 20236 J. Oberst et al. / IcaruTwo other possible mechanisms for the formation of highs
within the Caloris basin ?oor are volcanic construction and uplift,
but observations do not provide support for either. The high areas
in the DTM lack evident volcanic source regions, and physical vol-
canological models for eruptions on Mercury favor high-effusion-
rate eruptions that tend to produce broad plains rather than dis-
tinct edi?ces (Wilson and Head, 2008). Extensional fault structures,
widespread within the interior of Caloris, provide ample evidence
for ?oor uplift (Murchie et al., 2008; Watters et al., 2009), but
the mapped distribution of extensional faults in the area does
not show any correlation with the topographic highs in the DTM
(Fig. 9). Notably, the large complex of radially oriented graben in
the Pantheon Fossae area (see pro?le bb0 in Fig. 5 or marked point
D in Fig. 3) are located within a central low in the DTM, opposite to
the expectation if these structures formed in response to uplift by
tectonic or magmatic processes (Murchie et al., 2008; Head et al.,
2008; Watters et al., 2009).
5. Discussion
The DTM presented in this paper was determined solely on the
basis of geometric (image disparity) effects and does not rely on
Fig. 7. Identi?cation of Caloris basin rings (white) and tentative locations of large impact
DTM. The two identi?ed Caloris rings have radii of 690 and 850 km.9 (2010) 230?238assumptions regarding surface photometric properties. Owing to
small stereo and oblique viewing angles for the available images,
small errors in pointing angles and in the image matching will
introduce comparatively large height errors, which we expect to
correlate to some extent with the viewing geometry.
With some remaining errors in spacecraft navigation and cam-
era calibration information, we cannot rule out offsets in absolute
height and errors in long-wavelength trends in the model. It may
be worthwhile to carry out a full study of the propagation of errors
from navigation and camera parameters to the point accuracies for
this particular imaging geometry with Monte Carlo techniques, but
such an analysis is beyond the scope of this paper to perform. Ow-
ing to the large number of interlinked images, we nonetheless ex-
pect that the relative orientations of the images within the block
are comparably stable. In addition, the analysis of residuals of con-
trol point coordinates does not show evidence of any ??stresses? in
the image block, i.e., displacements or vertical offsets of the indi-
vidual DTM pieces. Consequently, any systematic errors within
the model should be small.
It is important to assess whether the unusual long-wavelength
topographic undulations within the ?oor of Caloris may be associ-
ated with errors in the long-wavelength stereo topography. To
craters and basins (black) that predated the formation of Caloris, as inferred from the
obtain an additional check, we investigated MESSENGER WAC
images that were obtained during the same ?yby. Though the
acquisition of laser altimeter data from the Caloris area during
MESSENGER?s orbital mission phase.
The terrain model derived from MESSENGER stereo images
gives us a fresh look at the morphology of Caloris and sets the stage
for the observations of this largest impact basin on Mercury during
MESSENGER?s orbital mission phase. For the orbital phase of the
mission, stereo sequences with global coverage from closer range
and under appropriate lighting and viewing conditions are being
planned. These observations will result in topographic models that
improve the spatial resolution of the current Caloris model by a
factor of two. With a more favorable viewing geometry, the noise
in the terrain models will be substantially reduced and visibility
into model details will be improved. Moreover, laser topographic
pro?les with their superior height precision will provide ??ground
truth? and calibration to reduce remaining ambiguities regarding
absolute elevations and trends in long-wavelength topography.
Once data from the orbital mission phase are in hand, it will be
worthwhile to study in detail the morphologies of other impact ba-
sins on Mercury to place Caloris into a global context. Comparisons
of long-wavelength topography and gravitational anomaly obser-
vations obtained from MESSENGER radio tracking, together with
high-resolution imaging and spectroscopy of geological units and
tectonic features, will permit the large-scale structure and associ-
ated geological evolution of each of the larger basins to be dis-
cerned and compared.
Acknowledgments
Fig. 8. Topographic pro?les across selected lunar impact basins and Caloris. Lunar
basin topography was taken from Clementine lidar data (Smith et al., 1997). Pro?les
pass through the basin centers and are in the north?south direction for the lunar
basins; the pro?le for Caloris extends from northwest to southeast.
J. Oberst et al. / Icarus 209 (2010) 230?238 237wide-angle data are limited by low resolution, the data (with
images taken through a different optical system and differently ar-
ranged footprints) hint at similar topographic patterns and trends.
However, given the problems faced by physical mechanisms to ac-
count for the model?s long-wavelength undulations across the Cal-
oris basin ?oor, we suggest that this topic should await theFig. 9. Comparison of Caloris DTM with a map of prominent faults in the area, including
2009).We wish to thank two anonymous reviewers whose thoughtful
comments improved this manuscript. The MESSENGER project is
supported by the NASA Discovery Program under contracts
NASW-00002 to the Carnegie Institution of Washington and
NAS5-97271 to the Johns Hopkins University Applied Physics
Laboratory.graben (in black) and wrinkle ridges (in white) (Murchie et al., 2008; Watters et al.,
References
Albertz, J., 24 colleagues, and the HRSC Co-Investigator Team, 2005. HRSC on Mars
Express ? Photogrammetric and cartographic research. Photogramm. Eng.
Remote Sens. 71, 1153?1166.
Araki, H., and 10 colleagues, 2009. Lunar global shape and polar topography derived
from Kaguya-LALT laser altimetry. Science 323, 897?900.
Denevi, B.W., and 10 colleagues, 2009. The evolution of Mercury?s crust: A global
perspective from MESSENGER. Science 324, 613?618.
Ernst, C.M., Murchie, S.L., Barnouin-Jha, O.S., Robinson, M.S., Denevi, B.W., Blewett,
D.T., Head, J.W., Izenberg, N.R., Solomon, S.C., 2010. Exposure of spectrally
distinct material by impact craters on Mercury: Implications for global
stratigraphy. Icarus, in this issue.
Fassett, C.I., Head, J.W., Blewett, D.T., Chapman, C.R., Dickson, J.L., Murchie, S.L.,
Solomon, S.C., Watters, T.R., 2009. Caloris impact basin: Exterior
geomorphology, stratigraphy, morphometry, radial sculpture, and smooth
plains deposits. Earth Planet. Sci. Lett. 285, 297?308.
Guest, J.E., Greeley, R., 1983. Geologic map of the Shakespeare quadrangle of
Mercury. Map I-1408, Misc. Investigations Ser., US Geological Survey, Denver,
Colo.
Gwinner, K., Scholten, F., Spiegel, M., Schmidt, R., Giese, B., Oberst, J., Jaumann, R.,
Heipke, C., Neukum, G., 2009. Derivation and validation of high-resolution
digital terrain models from Mars Express HRSC-data. Photogramm. Eng. Remote
Sens. 75, 1127?1142.
Gwinner, K., Scholten, F., Preusker, F., Elgner, S., Roatsch, T., Spiegel, M., Schmidt, R.,
Oberst, J., Jaumann, R., Heipke, C., 2010. Topography of Mars from global
mapping by HRSC high-resolution digital terrain models and orthoimages:
Characteristics and performance. Earth Planet. Sci. Lett. 294, 506?519.
Hawkins III, S.E., and 24 colleagues, 2007. The Mercury Dual Imaging System on the
MESSENGER spacecraft. Space Sci. Rev. 131, 247?338.
Hawkins III, S.E., and 18 colleagues, 2009. In-?ight performance of MESSENGER?s
Mercury Dual Imaging System. In: X-ray, UV, Visible, and IR Instrumentation for
Planetary Missions, SPIE International Symposium on Optical Engineering and
Applications, Paper 7441A-3, 12pp. San Diego, CA, August 2?6, 2009.
Head, J.W., 1982. Lava ?ooding of ancient planetary crusts ? Geometry, thickness,
and volumes of ?ooded lunar impact basins. Moon Planets 26, 61?88.
Head, J.W., and 10 colleagues, 2008. Volcanism on Mercury: Evidence from the ?rst
MESSENGER ?yby. Science 321, 69?72.
Heipke, C., 21 colleagues, and the HRSC Co-Investigator Team, 2007. Evaluating
planetary digital terrain models ? The HRSC DTM test. Planet. Space Sci. 55,
2173?2191.
Murchie, S.L., Watters, T.R., Robinson, M.S., Head, J.W., Chapman, C.R., Solomon, S.C.,
McClintock, W.E., Prockter, L.M., Domingue, D.L., Blewett, D.T., 2008. Geology of
the Caloris basin, Mercury: A view from MESSENGER. Science 321, 73?76.
Robinson, M.S., and 12 colleagues, 2008. Re?ectance and color variations on
Mercury: Regolith processes and compositional heterogeneity. Science 321, 66?
69.
Scholten, F., Gwinner, K., Roatsch, T., Matz, K.-D., W?hlisch, M., Giese, B., Oberst, J.,
Jaumann, R., Neukum, G., and the HRSC Co-Investigator Team, 2005. Mars
Express HRSC data processing ? Methods and operational aspects. Photogramm.
Eng. Remote Sens. 71, 1143?1152.
Smith, D.E., Zuber, M.T., Neumann, G.A., Lemoine, F.G., 1997. The global topography
of the Moon from the Clementine lidar. 1997. J. Geophys. Res. 102, 1591?1611.
Smith, D.E., Zuber, M.T., Neumann, G.A., Lemoine, F.G., Mazarico, E., Torrence, M.H.,
Duxbury, T.H., Head III, J., and the LOLA Science and Instrument Team, 2010.
LOLA observations of the Moon. Lunar Planet. Sci. 41. Abstract 1993.
Solomon, S.C., and 10 colleagues, 2008. Return to Mercury: A global perspective on
MESSENGER?s ?rst Mercury ?yby. Science 321, 59?62.
Spudis, P.D., Guest, J.E., 1988. Stratigraphy and geologic history of Mercury. In: Vilas,
F., Chapman, C.R., Matthews, M.S. (Eds.), Mercury. University of Arizona Press,
Tucson, pp. 118?164.
Watters, T.R., Murchie, S.L., Robinson, M.S., Solomon, S.C., Denevi, B.W., Andr?, S.L.,
Head, J.W., 2009. Emplacement and tectonic deformation of smooth plains in
the Caloris basin, Mercury. Earth Planet. Sci. Lett. 285, 309?319.
Wilson, L., Head, J.W., 2008. Volcanism on Mercury: A new model for the history of
magma ascent and eruption. Geophys. Res. Lett. 35. L23205. doi:10.1029/
2008GL035860.
Zhang, W., Giese, B., Oberst, J., Jaumann, R., 1996. Methods of the bundle block
adjustment of planetary image data. Int. Arch. Photogramm. Remote Sens. 31,
1004?1009.
Zuber, M.T., and 13 colleagues, 2008. Laser altimeter observations from
MESSENGER?s ?rst Mercury ?yby. Science 321, 77?79.
238 J. Oberst et al. / Icarus 209 (2010) 230?238