The NASA Mars Exploration Rover (MER) Project has been considering a landing-site ellipse designated EP78B2 in southeastern Utopia Planitia, southwest of Elysium Mons. The site appears to be relatively safe for a MER landing site because of its predicted low wind velocities in mesoscale atmospheric circulation models and its low surface roughness at various scales as indicated by topographic and imaging data sets. Previously, the site's surface rocks have been interpreted to be marine sediments or lava flows. In addition, we suggest that Late Noachian to Early Hesperian collapse and mass wasting of Noachian highland rocks contributed to the deposition of detritus in the area of the ellipse. Furthermore, we document partial Late Hesperian to Early Amazonian resurfacing of the ellipse by flows and vents that may be of mud or silicate volcanic origin. A rover investigation of the Utopia landing site using the MER Athena instrument package might address some fundamental aspects of Martian geologic evolution, such as climate change, hydrologic evolution, and magmatic and tectonic history.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The Mars Exploration Rover (MER) mission presents exciting new opportunities to explore the Martian landscape. However, engineering and safety considerations, data availability, and the character of Mars itself limit the number of serious potential landing sites. At the time of this writing, four sites were under consideration for exploration by the two separate MER A and B rovers (Figure 1): Meridiani Planum (formerly known as Terra Meridiani; 2.06°S, 353.77°E), Gusev (14.64°S, 175.06°E), Isidis (4.22°N, 87.91°E) and Elysium (11.73°N, 123.96°E) [Golombek et al., 2003] (all latitude values in this paper are planetocentric). The Elysium ellipse covers ∼1640 km2 and has a major axis of 155 km oriented N81–86°W (depending on launch date) and a minor axis of 16 km. The site ranges from −3224 to −2669 m elevation. Because of recent redefinition of Martian regional geographic names based on Mars Orbiter Laser Altimetry (MOLA) topographic mapping, the Elysium site no longer falls within Elysium Planitia but now occurs within southeastern Utopia Planitia [U.S. Geological Survey, 2002]. For consistency with current usage by the MER mission, we will continue to refer to the location as the “Elysium site”.
 The Elysium site was selected as a potential landing site due to its apparently benign attributes given surface characteristics and predicted low wind velocities that would minimize risks in the landing and operation of a MER rover. Here, we explore the geologic setting of this site with Viking, Mars Orbiter Camera (MOC) narrow-angle, and Mars Odyssey Thermal Emission Imaging System (THEMIS) daytime and nighttime infrared (DIR and NIR) images and MOLA digital elevation mosaics (DEM's) and derivative products. We address the following questions: What is the geologic setting of the site? What science issues could a MER rover investigate here? Has water played a role in surface processes?
2. Topographic Setting
 The MER Elysium site, as expressed in a digital elevation model (DEM) derived from MOLA data, lies on the gentle upper slopes of the highland/lowland boundary in southeastern Utopia Planitia (Figure 1). To the south, the cratered highlands of northwestern Terra Cimmeria are broken up into a wide swath of irregular mesas and knobs known as Nepenthes Mensae (Figures 2a, 2b, and 3). The southeastern edge of Utopia Planitia is defined by Hyblaeus Dorsa, a series of NE-trending scarps and ridges that form a topographic divide between Utopia and Elysium Planitiae. The dorsa extend several hundred kilometers to the broad Elysium rise, capped by Elysium Mons. The central basin floor of Utopia occurs some 1400 to 2300 km north of the MER ellipse.
 The MER ellipse (Figure 4) is on the eastern part of a subtle bench that is 200–250 km wide (north to south) and ∼1500 km long (west to east) and that has a fairly narrow upper elevation range of −2200 to −2000 m where it adjoins Nepenthes Mensae and Terra Cimmeria. The bench's lower elevation ranges between −3000 and −2600 m in the region of the ellipse, where it is defined by the tops of a series of arcuate, moderately sloping scarps, demarcating a lower plain characterized by complex, irregular depressions tens of kilometers wide and tens to hundreds of meters deep. Many of the depressions appear fully enclosed, but some only partly. The bench disappears east of the ellipse, where the complex lower plain disappears and the topography rises across Hyblaeus Dorsa. Regional, basinward slopes along the bench range from 0.1 to 0.3°. Northward, the complex plain transitions into a smooth plain at −3300 to −3100 m elevation. The topography of the ellipse is summarized in Table 1.
Table 1. Topographic Characteristics of MER Elysium Ellipse Based on a MOLA 1/128° Digital Elevation Model and a Slope Map Derived From the Elevation Model
Elevation Range, m
Elevation Mean, m
Elevation median, m
Slope Range, deg
Slope Mean, deg
Slope Median, deg
−3224 to −2669
0 to 6.75
3. Ellipse Area Landforms
 Landforms within and near the MER ellipse are surprisingly varied given that the site represents a relatively smooth area of the planet on the scale of the ellipse. Different aspects of surface features are revealed by Viking, MOC, and THEMIS images and various representations of MOLA topography data, including shaded relief, slope, contour, and high-pass filtered topography maps, point measurements, and profiles. We use a MOLA DEM at 1/128°/pixel resolution (∼460 m/pixel at the ellipse center) in an Equidistant Cylindrical projection using the latest Mars spheroid [Seidelmann et al., 2002] to which we have approximately registered the available MOC narrow-angle (NA) and THEMIS images (Table 2 and Figures 4d–4f).
Table 2. MOC Images of the MER Elysium Ellipse
Sparse dunes, varying albedo
Subtle ridge forms, crater ejecta rays
Minor, N-trending wrinkle ridge; larger dunes in ∼650-m crater
South half of table mesa with central knob and slightly crenulated SW scarp that extends into plain; sparse dunes
Gently rolling, cratered topography; subdued craters and ridges
East-facing lobate scarp
Complexly crenulated western wrinkle ridge
Same as E18-00898
Nearly same as E18-00429
West part of 3-km crater, part of >1 km table mesa with 300-m knob; also a 500 × 800 m knob/dome NW of crater
E-trending scarps, parts of 2 table mesas
South-facing lobate scarp; sparse dunes
Edge of subdued hill in center of ellipse; subdued craters
E-trending scarps; table mesa
Lobate scarps; part of crater no. 3
Subdued ridges and craters
 Wrinkle ridges tens to hundreds of kilometers long occur throughout the region's highland and plains areas. Two dominantly NE-trending wrinkle ridges crossing near the center and in the western part of the ellipse include broad arches 5–11 km wide. The western ridge is continuous through the ellipse and rises 80 to 140 m above the adjacent plain. The central ridge has a sinuous form, rises a few tens of meters, and continues northeast into lower plains where its southeastern flank displays somewhat greater relief (100 to 160 m) than the northwestern flank (50 to 100 m). The western ridge also extends northeast of the ellipse, but four large craters and some irregular depressions modify its expression. Regionally, NE-trending wrinkle ridges and scarps of southern Utopia, including those of Hyblaeus Dorsa just east of the Elysium ellipse, have greater relief and lengths than those trending at right angles to them (Figure 4c). Locally, narrow crenulations form the crests of the ridges as well as secondary ridges on the ridge flanks (Figure 5). Most of the MOC images contain subtle ridge forms of various orientations that appear to consist of crenulations and perhaps low-relief arches of modest-size wrinkle ridges.
 The plain surrounding the large wrinkle ridges varies moderately in MOLA topography across the ellipse (Figures 4a–4c). West of the western wrinkle ridge, the plain appears smooth and level in MOLA and THEMIS data, with local slopes mostly <0.5°. Between the ridges, the plain gently undulates with local slopes steepening to between 1.0 and 2.0°. The lower part of this area forms an ∼12-km-wide trough snaking in and out of the ellipse. Part of this trough bounds the eastern flank of the western ridge. East of the central ridge, the gently rolling topography has MOLA slopes up to 3° and is broken by a shallow depression along part of the ridge margin. The eastern tip of the ellipse lies in the northern, deeper part of a roughly circular, 40-km-diameter, 200-m-deep depression. A particularly deep depression occurs 35 km northwest of the ellipse center and ranges from 400 to 600 m deep; its floor lies about 500 m below the ellipse center surface.
 Knobs resolved in MOLA data are scattered through the ellipse and surrounding plain (Figure 4). Diameter and height measurements made from basal elevations determined by lowest, relatively rounded, enclosed contours that seem to exclude other landforms show 21 knobs >5 m high and several hundred meters to several kilometers in diameter (Table 3, Figure 4a). Some of the knobs have slopes extending below much of their basal elevation contour. Because of the low spatial resolution of the MOLA DEM and its interpolation of local, broad gaps between data tracks, significant uncertainty and smoothing occurs in the sub-kilometer plan view details of the knob forms (Figures 4a–4c). Thus high-resolution images of the features are needed to characterize their detailed form (Figure 6). Eight knobs occur along the crests of the two wrinkle ridges, and one forms part of a rim of an impact crater. We also note an exceptionally large knob (no. 10) south of the ellipse about 820 m high lying just east of the eastern ridge. This knob forms a steep, triangular peak above a ∼200 meter high gently sloping bench surrounding the knob's base. The knob occurs on the southern end of a broader rise ∼300 to 400 m high centered on the central ridge. The northern margin of the rise forms a relatively steep slope within the ellipse, with local slope values ranging from <1 to 3.6°.
Table 3. Knobs of the MER Elysium Ellipse
Geography and Geomorphology
1.7 × 2.4
On western wrinkle ridge; crest south of ellipse
On western wrinkle ridge; extends down ridge
1.2 × 2.5
On western wrinkle ridge
2.9 × 4.7
On western wrinkle ridge
1.3 × 1.7
Part of minor wrinkle ridge?
3.7 × 5.8
NW part of 8-km diameter table mesa
1.5 × 2.2
May be crest of minor wrinkle ridge
1.3 × 2.3
East rim of impact crater
1.4 × 2.6
North-trending ridge line atop ∼300 m high prominence on central wrinkle ridge; 2 km south of ellipse
5.0 × 6.4
Large knob atop broader hilly region; basal bench ∼200 m high; peaks at −1931 m, 15 km south of ellipse
1.4 × 2.0
Part of central wrinkle ridge
1.2 × 1.9
Part of central wrinkle ridge
3.1 × 3.5
Table mesa in I01868002
0.9 × 1.3
NW part of ∼2.8 km diameter table mesa in I01868002 and E22-00479
3.2 × 4.6
On rolling plain
2.4 × 2.7
Table mesa ∼2.1 km across in E21-01405
1.3 × 1.7
Table mesa ∼1.6 km across in E21-01405
1.3 × 1.8
On floor of depression
1.5 × 2.1
On margin of depression; south flank extends downward another ∼100 m
1.3 × 1.6
On floor of depression
2.0 × 2.7
Table mesa on floor of depression; ∼3.5 km wide in I01843008
 Some knobs consist of relatively broad table mesas topped by narrow peaks as resolved in MOC (Figures 6 and 7) and Viking and THEMIS (Figure 4e) images. Four of the mesas observed in MOC NA and THEMIS DIR images range from 1.6 to 8 km in width and 8 to 84 m in height (knob nos. 6, 13, 14, and 21). A dozen or so additional table mesas similar in size are scattered within a few tens of kilometers of the ellipse as seen in THEMIS DIR images (Figure 4e). Additional table mesas may occur in areas not imaged by THEMIS DIR and perhaps make up knobs in the eastern part of the ellipse in the MOLA DEM (Figure 4a, Table 3). A narrow ridge extends north of the pronounced dome designated knob no. 13.
 Lobate scarps in and around the ellipse embay most of the knobs and appear to outline flow material of locally varying texture in THEMIS DIR images (Figure 4e). The east flank of the eastern wrinkle ridge appears embayed as well. In plan view, the scarp orientations broadly reflect northward flow, down the regional gradient. Scarps demarcating two apparently abutting flows west of the ellipse center occur in both THEMIS DIR and MOC NA images (Figure 7).
 A possible channel occurs near a sinuous scarp in the ellipse. The feature is partly resolved in a MOC stereo elevation model (Figure 8) and may originate near the base of the scarp. The scarp forms a ridge in places and appears to be a minor, asymmetric wrinkle ridge.
 The MER ellipse avoids larger impact craters having steep rims and deep bowls as seen in Viking images and the MOLA DEM. The bowls of five pronounced impact craters >1 km in diameter within and along the margin of the ellipse appear in the MOLA DEM (Figure 4a; Table 4). Crater rims generally are not well resolved in the MOLA DEM due to the pattern and density of MOLA shot points, and at least one of the knobs in the ellipse (no. 8 in Table 2) is probably a partly resolved rim. MOC images in and near the ellipse depict many craters >10 m in diameters, and craters as much as a few hundred meters in diameter are common. In THEMIS DIR images (Figure 4e), crater no. 3 in particular, and perhaps a few of the other larger craters as well, appear to have dark, or relatively cool, ejecta indicative of blockier material. In THEMIS NIR images (Figure 4f), craters 2 and 4 and other craters show bright ejecta that may also be made up of blocky material; however, crater no. 1 displays only subtle, bright rays. Smaller craters also vary in ejecta signature. Crater floors mostly appear dark in THEMIS NIR images, and the darker the floor, the less contrast between the crater rim and ejecta and the surrounding plain. Three craters 15 to 16 km in diameter occur within 50 km of the ellipse; the rim of the closest one is 39 km north of the ellipse center.
Table 4. Kilometer-Size Impact Craters in the MER Elysium Ellipse
Maximum Slope, deg
4. Regional Geologic Units
 We adapted a preliminary geologic map of the northern plains of Mars based largely on MOLA data [Tanaka et al., 2003] to the landing-site region (Figures 2a, 2b, and 3); previous geologic mapping includes that of Greeley and Guest  based on Viking images. Our map includes undivided highland material (unit HNu), knobby unit (unit HNk), and boundary plains units (units Hb1a,1b,2). A schematic cross-section of this geology is shown in Figure 9, according to the interpretations that follow. Densities of craters >5 km in diameter (Table 5) provide mean relative ages for the units (however, the data do not address modest resurfacing that would tend to obliterate only smaller craters).
Table 5. Crater Densities of Regional Geologic Units of the Southern Utopia Planitia and Northern Terra Cimmeria Regiona
Crater densities are cumulative number per 106 km2. Units are from the database of N. G. Barlow (personal communication, 2002). Crater locations were modified as needed to correct for registration errors in Viking photomosaics versus the MOLA DEM.
N(X), number of craters > X km in diameter per 106 km2.
N, Noachian; H, Hesperian; A, Amazonian; E, Early; L, Late.
Count is likely high, because some craters embay rather than superpose the unit as seen in some MOC NA images.
 Highland material, undivided (unit HNu) forms heavily cratered terrain of northern Terra Cimmeria of mainly Noachian age. The unit's rugged topography consists of (1) large crater forms, most of which are in states of degradation in which rim and ejecta features including ramparts and secondaries have been eroded or buried, (2) broad ridges that may have tectonic or erosional origins, and (3) valleys that debouch into local basins floored by sedimentary and perhaps volcanic materials. The composition of highland material can only be surmised broadly at present. MGS Thermal Emission Spectrometer (TES) results indicate a generally basaltic to andesitic or weathered basaltic mineralogy [Bandfield et al., 2000; Wyatt and McSween, 2002]. The meteorite ALH84001, likely derived from highland material somewhere on Mars, represents a mafic crystalline igneous rock [e.g., Treiman, 1995]. Early Mars had higher heat flow, thus volcanic and igneous rocks may make up much of the highland material [e.g., Scott et al., 1986; Tanaka, 2000]. MOC images reveal that much of the highland material is layered and may include crater ejecta and sedimentary deposits [e.g., Scott et al., 1986; Malin and Edgett, 2001]. The youngest highland material mainly occupies intermontane basins and crater floors and likely includes some Hesperian rocks [Scott et al., 1986]. Valleys debouch into many of the basins, indicating that they may include alluvial and lacustrine sediments.
 Along the highland/lowland boundary, the knobby unit (unit HNk) consists of highland material broken up into large knobs and mesas typically a few to tens of kilometers across and rising hundreds of meters to more than a kilometer above adjacent plains. The density of both superposed and degraded craters >5 km in diameter (Table 5) indicates that the unit began to form at the end of the Noachian, coincident with Viking-based relative-age dating of the HLB [Tanaka, 1986]. The interknob plains were emplaced from the Late Noachian into the Early Hesperian and are crossed by networks of wrinkle ridges and scarps likely resulting from folding and thrusting due to contraction [e.g., Watters, 1993]. Local depressions tens to hundreds of kilometers across marking the unit's surface resulted in part from tectonic deformation, but others have arcuate outlines unrelated to the ridges. The latter may result from collapse due to the expulsion of subsurface, H2O and perhaps CO2-charged slurries [Tanaka et al., 2003]. The mesa topography of the knobby unit resembles topography produced by extensive fluvial dissection on Earth (e.g., Monument Valley in Arizona, U.S.A.), but sparse valley forms cut into highland material (unit HNu) surfaces do not appear to continue across the interknob plains. The low slopes of the interknob plains preclude most dry or ice-assisted landslide and mass-wasting processes. However, debris may flow across flat surfaces if liquefied by water and perhaps CO2 [e.g., Nummedal and Prior, 1981; Hoffman, 2000; Tanaka et al., 2001], and some flow-like aprons surrounding knobs and mesas occur near the highland boundary several hundred kilometers southeast of the landing site (Figure 10). No obvious volcanic or intrusive landforms are evident that may suggest magmatic/ground-ice interactions [cf. Squyres et al., 1987; Wilhelms and Baldwin, 1989; Carruthers and McGill, 1998], although such features could be difficult to detect.
 The knobby unit grades into boundary plains unit 1a (unit Hb1a), as the knobs become sparser and the interknob plains dominate. This material has a gentle basin-ward slope of 0.1 to 0.3° and is marked by networks of wrinkle ridges and other features described in detail in the previous section. Prior to MGS, this unit was interpreted to be lava flows on the basis of the presence of wrinkle ridges [e.g., Scott et al., 1986] or marine sediments, assuming that the unit Hb1a surface had been submerged beneath a plains-filling ocean [Parker et al., 1989, 1993; Baker et al., 1991]. Following MGS, these proposals are still favored by some workers [e.g., Clifford and Parker, 2001; Head et al., 2002]. In addition, using MOC images and MOLA topography, Tanaka et al.  interpreted some landforms widely associated with plains-forming units as erosional, collapse, and ground-ice and/or groundwater-related features. In any case, unit Hb1a apparently embays unit HNk.
 The topographic form of the western wrinkle ridge crossing the ellipse has been modeled numerically by Okubo et al. . The steeper, southeastern slopes indicate a buried, primary thrust fault dipping 9–10° to the northwest. Above ∼2 km depth, the faults steepen to 10 to 22° through significantly weaker material thereby producing the crenulations prominent along the southeastern margin of the ridge (Figure 5). Thus, at the time of deformation, a pronounced, sharp interface was apparently present, possibly defined by resistant basement rock overlain by weaker detritus or volcanic flows.
 Here, we distinguish a lower member, boundary plains unit 1b (unit Hb1b) that unit Hb1a grades into where the plains material becomes marked by depressions and scarps, as described in the previous section. In the MOLA DEM (Figures 2a, 2b, and 3) and Viking and THEMIS images, this unit includes stepped scarps and depressions of tens of kilometer length scales and relief that attains a few hundred meters. These features look like they may be formed by collapse processes, likely involving liquefaction causing the plains material to lose strength and settle, perhaps due to loss of pore space and volatile removal. However, flows exiting the depressions are not evident.
 At the base of unit Hb1b, boundary plains unit 2 (unit Hb2) forms smooth plains marked by asymmetric ridges that connect with wrinkle ridges in unit Hb1. Tanaka et al.  proposed that this unit is the sediment produced by mass wasting and erosion of unit Hb1. In turn, below unit Hb2, the Vastitas Borealis Formation (unit AHv, not in study area) is characterized by a rugged topography in the MOLA DEM that corresponds to low, sub-kilometer-size cones, many chained into ridges. The Vastitas Borealis Formation, as remapped by Tanaka et al.  adapted from original mapping by Scott et al. , roughly coincides with the putative Contact 2 shoreline of Parker et al. [1989, 1993].
5. Geologic Scenarios, MER Elysium Ellipse
 The gently sloping plains of the MER Elysium landing-site ellipse are typical of much of the highland-lowland boundary of Mars, where rugged, highland terrain commonly grades into knobs and mesas and then plains (Figures 1–3). Here we discuss possible geologic scenarios for the materials and morphologic features in and near the ellipse that might be investigated by a rover. These scenarios are based on previous hypotheses for the origin of the highland-bounding plains, with additional details drawn from our observations as well as from studies of other locales on Mars with geologic similarities. Many aspects of these different scenarios are not mutually exclusive; multiple origins and overlapping events appear possible. Because of this potential complexity, interpretation of the geology of the ellipse region remains controversial and thus impacts the possible science results from a MER investigation.
5.1. Lava/Volcanic Flows
 The lava-flow origin is based on the presence of wrinkle ridges, which generally has been thought to be a good indicator of lavas or at least a stratigraphy of more competent layers overlying less competent material [e.g., Zuber and Aist, 1990; Watters, 1993]. On Mars, some wrinkle-ridge plains occur in association with volcanic vents and/or lobate flow fields, such as Hesperia, Syrtis Major, Malea, Lunae, and Thaumasia Plana and likely represent lavas [e.g., Scott et al., 1986; Schaber, 1982; Dohm et al., 2001]. Thus the discovery of extensive wrinkle ridges deep into the northern plains from MOLA data has been interpreted as indicative of extensive lavas at or near the surface [Head et al., 2002]. However, the mechanics of wrinkle-ridge formation appears to be better understood now in light of studies of terrestrial blind thrust faults, and wrinkle ridges may form in many diverse, geologic materials whose mechanical characteristics may influence deformation style and hence morphology [e.g., Schultz, 2000]. Specifically, crenulations may represent backthrusts and forethrusts that may form as long as bedding plane slip occurs. What appears to be required is a discontinuity in which material of lower frictional strength overlies higher strength material, such as poorly indurated, rounded, porous, sorted granular material or lava-flow rock masses overlying tightly packed, angular sediments or indurated rock masses [Okubo and Schultz, 2002; Okubo et al., 2003]. Lobate flows appear in THEMIS DIR images and MOC image E18-01455 (Figures 4e and 7). Potentially, some of the small knobs in and near the ellipse could be volcanic vents that could have contributed to broader resurfacing in the ellipse. Overall, however, it seems unlikely that lavas would make up a large part of the higher elevation materials within the northern plains, as associated vent structures and lobate flows generally appear to be lacking.
5.2. Marine Sediments
Parker et al. [1989, 1993] and Clifford and Parker  interpreted a deep, northern-plains ocean rising above the MER Elysium ellipse on the basis of a proposed shoreline level at the topographic base of the exposed highland material (unit HNu) and knobby unit (HNk), which might be viewed as an equipotential surface (Parker et al.'s Contact 1). The large knob (no. 10) south of the center of the ellipse is partly surrounded by a low bench; such benches common in the northern plains have been interpreted as wave-cut terraces produced within a northern plains ocean [Parker et al., 1993]. Also, higher-standing layered deposits in Meridiani Planum have also been suggested as marine sediments [Edgett and Parker, 1997]; however, a volcanic origin for this material has also been proposed [Chapman and Tanaka, 2002]. Testing for possible morphologic and mineralogic signatures supportive of the marine hypothesis seems to be difficult from orbiting spacecraft, as the features that might develop for marine deposition and coastal processes on Mars would be subtle and perhaps largely destroyed by impact gardening. For example, tides would generally have been feeble, and bodies of water might have frozen and sublimated within ∼106 years [Kreslavsky and Head, 2002], which would preclude formation of significant shoreline features. MOC imaging of Parker et al.'s  Contact 2, which would be less impact gardened than their Contact 1, does not show any preserved shorelines [Malin and Edgett, 1999]. Furthermore, other putative shoreline terraces and ridges occurring on highland slopes [Clifford and Parker, 2001] might be produced by slope processes instead [e.g., Tanaka et al., 2003]. Testing by MOLA topographic data, while suggestive of a possible ancient ocean for Parker et al.'s Contact 2 (which largely corresponds to the outline of the Vastitas Borealis Formation [Tanaka et al., 2003] below the Elysium site), appears to be much less likely for Contact 1 and thus over the MER Elysium site [Head et al., 1999].
 Another relevant matter that is as yet unsettled is the interpretation of TES spectra for mineralogic signatures. Bandfield et al.  showed that much of the northern lowlands have a distinct spectral signature consistent with unweathered andesite, in contrast to spectra indicative of unweathered basalt over much of the southern highlands. However, Wyatt and McSween  proposed that the northern plains spectra could also be produced by low-temperature, aqueously altered basalts. While most of this signature occurs within the confines of the Vastitas Borealis Formation, higher lowland areas that include the occurrence of the boundary plains units also appear dominated by the “andesite” component, where the surface is not overprinted by dust.
5.3. Colluvium and Alluvium
 South of the MER Elysium site, Noachian highland material has degraded into inlier knobs surrounded by plains-forming material (Figure 10). Tanaka et al. [2001, 2003] suggested that extensive mass wasting erosion of highland material occurred along the HLB, which may have been caused by effusive and perhaps violent discharge of volatiles. Unit Hb1a material making up most of Elysium site then would be the colluvium and alluvium derived from highland material resulting from this activity. In the Libya Montes/Isidis Planitia region, which is another parts of the HLB having similar boundary plains, possible mass flows originating from steep-sided, arcuate depressions can be detected in MOC NA and THEMIS images [Crumpler and Tanaka, 2003].
 Also, the table mesas, lobate material, and irregular depressions may be the vents, flow deposits, and subsidence features of related but younger sedimentation. The table mesas could be mud volcanoes produced by venting of pressurized slurries from depth, as a residual product from the emplacement of debris flows, as has been proposed for similar features elsewhere in the northern plains [Tanaka, 1997]. The irregular trough in the center of the ellipse and the depression at the eastern tip of the ellipse may relate to deformation caused by liquefaction and removal of buried, volatile-rich material associated with development of unit Hb1b. Finally, the bench surrounding the knob south of the ellipse might result from collapse of plains material surrounding the knob [Tanaka, 1997; Tanaka et al., 2003].
5.4. Other Activity
 Wrinkle ridges throughout the ellipse indicate that contractional deformation occurred following the emplacement of the older plains material and the younger flows. The ridges include broad arches, narrow crenulations (Figure 5), and scattered small, commonly elongate knobs (as resolved in MOLA DEM's) along the shoulders of the ridges (see Table 3). These morphologies are consistent with anticlines above buried thrust faults; the crenulations and knobs may express the deformation resulting from secondary forethrusts or backthrusts [Schultz, 2000; Okubo and Schultz, 2002]. Given that the southeast sides of the ridges are steeper and lower than the northwest sides, the anticlines verge to the southeast indicating southeastward thrusting along faults that dip to the northwest. The ridges extend hundreds of kilometers to the northeast, where they acquire a more asymmetric cross-sectional profile. These ridges seem to make up the westernmost part of the northeast-trending set of ridges making up Hyblaeus Dorsa. Across these ridges to the east, the topography rises about a kilometer, forming a topographic break between Utopia and Elysium Planitiae. The ridges extend into lower plains materials, including the upper part of the Vastitas Borealis Formation (VBF). However, within Utopia basin, contractional ridges appear subdued where polygonal troughs dissect the plains, and Amazonian lobate and channeled flows bury the ridges (J. A. Skinner and K. L. Tanaka, Geology and materials of Utopia Planitia, Mars based on Mars Global Surveyor data, submitted to Icarus, 2003). Shorter, lower ridges trend roughly perpendicular (NW-SE) to the ellipse ridges in the vicinity of the ellipse within units HNu, HNk, and Hb1a (Figure 1b). However, such ridges appear diminished to absent in the lower plains units. We suggest that contractional ridge deformation had begun by the time of Hb1a emplacement, if not beforehand [Thomson and Head, 2001; Head et al., 2002]. Minor deformation probably continued into the Late Hesperian or Early Amazonian following Late Hesperian resurfacing in the ellipse area.
 The MER ellipse was carefully placed to avoid the hazards associated with terrain of large impact craters, so few large, rugged ones occur within the ellipse (Table 4). However, given the Early Hesperian age of the surface, a fair number of craters hundreds of meters in diameter occur that preserve steep slopes that pose local hazards (Figure 8). In addition, much of the surface may be impact gardened to depths of tens of meters [Hartmann et al., 2001], but less so in parts of the ellipse that appear resurfaced by younger material. Densities of sub-kilometer craters in MOC NA images of the ellipse area (Figure 11) indicate that much of the ellipse has been resurfaced during the Late Hesperian and Amazonian. The oldest detected surface in the ellipse is that forming the arch of the western wrinkle ridge (MOC E19-00178). This surface thus apparently stood above the younger flows that later buried the lower, adjacent plains.
6. Potential MER Athena Science Investigations
 The MER investigation seeks to determine the history of climate and water at two sites on Mars where conditions may once have been favorable to life. Considering the diverse proposed geologic interpretations for the materials and landforms at the MER Elysium ellipse, some potentially exciting scientific investigations might be performed by an Athena rover, with its array of geologic tools designed to image and to analyze for mineralogy, chemical composition, and physical characteristics of materials and structures. We might expect the following science issues and their associated implications to be addressed at the MER Elysium ellipse:
 • What is the origin of the unit Hb1a material within the ellipse: volcanic, marine, mass wasting, a combination, or other? This unit encircles and perhaps underlies most of the northern plains of Mars, or perhaps as much as 30% of the planet. Thus an understanding of the make up and processes associated with resurfacing of the northern plains will be better understood, and to what degree they might be related to activity of water versus volcanism. The dominant lithology of lowland material, which may be either andesite or weathered basalt according to TES spectra [Bandfield et al., 2000; Wyatt and McSween, 2002], should be resolved by mineralogic analyses performed by the Athena science instruments. In addition, processes and materials evident here may characterize surfaces within other basins of Mars, including Hellas, Argyre, and other intracrater and intercrater basins within ancient highland regions of Mars.
 • What is the mineralogy and composition of ancient highland rocks on Mars? If the material investigated by the MER rover constitutes coarse debris weathered from the highlands, and/or the knob south of the ellipse is a highland remnant that might be imaged spectrally by the rover instruments, then aspects of the mineralogy and composition of about half of the planet's surface might be characterized. Potentially, a diversity of rock types having undergone various amounts of weathering and other modification under as much as 4 billion years of exposure to the Martian climate may be investigated. Significantly, was the climate warmer and wetter during the Noachian Period?
 • What is the record of post unit Hb1a resurfacing and modification within the ellipse? A MER rover would likely see up close a variety of craters excavating surface materials over a period of as much as 3.5 billion years since emplacement of unit Hb1a [Hartmann and Neukum, 2001]. Systematic changes in morphology and perhaps mineralogy among crater forms and ejecta investigated by the MER rover might document changing climatic conditions over about the last 3/4 of the planet's history. Some of the table mesas and thin flows in and near the ellipse may be landforms produced by silicate and/or mud volcanism that might be visited or imaged by the MER rover if it landed near one (Figure 12). Possible sites of spring discharge may occur along wrinkle ridges (Figure 8) [Okubo et al., 2003]. Additionally, landforms and materials visible at an actual landing site might record other processes unrecognized at present from orbital data sets.
 • What is the detailed form of wrinkle ridges? The MER Elysium ellipse offers an opportunity to record detailed topographic characteristics of two apparently typical Martian wrinkle ridges. The morphologies may help elucidate near-surface fault and fold geometries, which might be governed by the mechanical stratigraphy of the deformed rocks [Schultz, 2000; Okubo and Schultz, 2002; Okubo et al., 2003].
7. Comparison With Other MER Candidate Landing Sites
 After submission of this paper, the Meridiani and Gusev sites had been selected for the MER landings. Nevertheless, the selection of the Elysium site as one of the top four for the MER project out of a couple hundred initial candidates indicates that it someday may be considered as an optimal site for a future mission, depending on the dictates of particular engineering and science requirements. This site is distinctive geologically from the sites at Meridiani Planum and Gusev crater, but has similarities with the one at Isidis Planitia. The Meridiani Planum site consists of layered deposits superposed on cratered highland rocks having a hematite spectral signature in MGS Thermal Emission Spectrometer data. Diverse origins for the hematite, mostly involving water, have been suggested [Christensen et al., 2000, 2001]. Gusev crater forms a basin into which the Ma'adim Vallis channel system debouches, and the MER site there may sample sediments laid down within a temporary lake [Cabrol et al., 2003]. Like the Elysium site, the Isidis site occurs on a plain north of and below highstanding ancient materials, including massifs related to the Isidis basin-forming impact that may be among the oldest materials on Mars [Crumpler and Tanaka, 2003]. At both the Elysium and Isidis sites, water likely was involved in removing subsurface material and eroding and transporting ancient clastic debris. Environmental conditions may have been rather clement at times if fluvial, debris-flow, or marine activity occurred at or near the Elysium site. Also, water may have welled up from deep, ancient aquifers that may have constituted habitable environments for microbes.
 Landing-site selection will always remain a complex attempt to balance science issues with safety and operational considerations of the lander and availability of orbital data. Commonly, safer-appearing sites such as the Elysium site have less desirable science. However, safe sites may well be entertained in the final stages of site selection for future missions and perhaps occasionally be selected for actual landings. Thus it makes sense to collect orbital and reconnaissance data and perform scientific analyses in advance of possible landings for such sites, along with studies of the more scientifically interesting ones, so that the most thorough understandings possible are achieved.
 Much of this work resulted from discussions with many colleagues and participation in Mars Exploration Rovers landing-site selection workshops that were led by Matthew Golombek and John Grant. We thank the two reviewers whose comments helped us to sharpen the text and figures. Particular credit goes to Randy Kirk for supplying us with the MOC stereo model used in Figure 8 and to Nadine Barlow for use of her crater database. Our work was supported by grants from NASA's Mars Data Analysis Program.