NEAR Shoemaker images reveal widespread occurrence of tectonic landforms on asteroid 433 Eros. Hinks Dorsum is a ridge that extends for about 18 km around the asteroid and strongly resembles thrust fault structures on the terrestrial planets. Tectonic landforms can provide information on the mechanical properties of asteroids, a subject of much controversy. Modeling constrained by topographic data shows that Hinks Dorsum can be accounted for by a shallow rooted thrust fault no greater than 250 m in depth with ∼90 m of cumulative slip. Strength envelopes based on frictional and rock mass strength criteria suggest the near-surface shear strength of Eros is from ∼1 to 6 MPa. A spatial correlation is found between Shoemaker crater, a transition from low to high crater density, and Hinks Dorsum. This spatial relation along with the estimated strength of the asteroid suggests the thrust fault was formed by impact induced compression.
 Eros is a near-Earth, S-type asteroid that comes within 0.13 astronomical units of Earth's orbit [Veverka et al., 2000]. It is an elongated body measuring roughly 34 km by 11 km by 11 km [Veverka et al., 2000]. The NEAR Shoemaker spacecraft orbited Eros for a year before it became the first spacecraft to land on an asteroid [Veverka et al., 2001a]. The multispectral imager (MSI) obtained 5 m/pixel or better resolution images of the entire surface and acquired hundreds of images with sub-meter pixel scales [Veverka et al., 2001b]. The shape of the asteroid was determined from image-based techniques [Thomas et al., 2002a] and laser ranging from the NEAR Laser Rangefinder (NLR) [Zuber et al., 2000; Miller et al., 2002]. NLR profiles resolve small-scale landforms with only a few meters of relief [Cheng et al., 2001, 2002].
 Hinks Dorsum is similar in morphology to lobate scarps found on Mercury and Mars [cf. Watters, 2003; Watters et al., 2009], and it is similar in scale to lunar lobate scarps [cf. Watters et al., 2010]. Lobate scarps on the Moon, Mercury and Mars often cut and offset the walls and floors of impact craters and are interpreted to be the surface expressions of thrust faults [Watters, 2003; Watters et al., 2002, 2009, 2010]. Topographic profiles derived from NLR data confirm that the cross-sectional geometry of Hinks Dorsum is weakly to strongly asymmetric [also see Cheng et al., 2002] with the vergent side to the north (Figure 1). The segments of Hinks Dorsum near Tutanekai crater have a maximum relief of ∼60 m (Figures 1 and 2a). Image and topographic data reveal two smaller subsidiary scarps flanking Hinks Dorsum (Figure 1). The smaller of the subsidiary scarps has a maximum relief of about 25 m (Figure 2b), while the larger scarp has about 50 m of relief. Measured slopes of the northern, vergent-side of Hinks Dorsum and the secondary scarp are ∼20°, comparable to maximum slopes of scarp faces of large-scale lobate scarps on Mercury and Mars [Watters, 2003; Watters et al., 1998]. An analysis of topographic profiles across the segment of Hinks Dorsum in Himeros indicates the northern side is also the vergent, steeper-side of the scarp. The consistent northward vergence suggests the underlying thrust fault extends below the southern flanks of Hinks Dorsum along its entire length.
 We simulated deformation over a thrust fault that propagates upward toward the surface using elastic dislocation modeling [Lin and Stein, 2004; Toda et al., 2005]. The magnitude and sense of slip is specified, and the stresses and material displacements are determined using the stress functions for an elastic half-space [Okada, 1992]. The fault surface is defined as a rectangular plane having a dip (θ), and vertical depth of faulting (T1, T2) (Figure 2c). We assumed that deformation occurred above a blind thrust fault (non-surface breaking) because there is no photogeologic evidence of vertical offset at the base of the vergent side (scarp face) of Hinks Dorsum. We estimated the magnitude of the slip D from the height of the scarp and then adjusted D, along with the other free parameters, until the best fit with the topography was achieved. The relative-slip distribution along the model fault is tapered with the minimum slip at the fault tips. Although the geometry and depth of faulting cannot be uniquely determined, the modeling gives a good fit to the topography within a relatively narrow range of parameters [see Watters et al., 2002].
 Modeling of Hinks Dorsum suggests the underlying fault has a dip of ∼40°, a maximum depth of ∼240 m, and a slip of ∼90 m (Figure 2a). Modeling of one of the secondary ridges indicates an underlying fault with a dip of ∼35°, a maximum depth of ∼200 m, and a slip of ∼40 m (Figure 2b). These estimates suggest that Hinks Dorsum and the secondary ridges are the surface expression of shallow-rooted thrust faults with thrust slip directions to the north.
 Because the modeling is constrained by topography, factors such as the orientation of the orbital track relative to the strike of Hinks Dorsum or post-deformation modification of the structure will influence the model results. NLR data tracks rarely cross segments of Hinks Dorsum at the optimal angle (perpendicular to its orientation) (Figure 1b). The orientation of the profiles used to constrain the modeling, however, are nearly orthogonal to the scarp segments (Figures 1b and 2) and thus, minimize the error. Evidence of downslope movement or slumping on the vergent-side of the scarp indicates that the post-deformation slopes of Hinks Dorsum have been modified (Figure 1). Relatively high albedo material appears to have moved down the slope of the scarp face forming debris aprons on the foot wall-side of some segments (Figure 1). Debris aprons from downslope regolith movement are common on Eros [see Robinson et al., 2002]. The slope of the scarp face strongly influences the best-fit depth of the upper fault tip T1. Thus, it is possible that the depth of the upper tip of the blind thrust fault is nearer to the surface than the best-fit value would suggest. However, there is no evidence of debris aprons on the segments modeled (Figure 1b).
4. Stresses and Strength
 The origin of the compressional stresses that formed Hinks Dorsum is not well understood. Surveys of tectonic features on Eros show only a few sets of structures radial to impact or other features [Prockter et al., 2002; Thomas et al., 2002b; Buczkowski et al., 2008]. Impact craters with squared outlines, however, are cited as evidence of structural control [Cheng et al., 2002; Prockter et al., 2002]. The global distribution of grooves, troughs, and ridges on Eros suggests a pervasive fabric of fractures [Prockter et al., 2002; Thomas et al., 2002b], supporting the interpretation that the asteroid is a coherent body held together by material strength rather than by gravitational attraction alone [Zuber et al., 2000; Prockter et al., 2002; Thomas et al., 2002b]. Proposed expressions of this fabric are a Calisto Fossae-Hinks Dorsum plane and a prominent facet of Eros [see Thomas et al., 2002b]. Our analysis indicates that the dip of the Hinks Dorsum fault is not along these planes and thus the fault was likely not influenced by the putative fabric. This does not exclude the possibility that the Hinks Dorsum fault exploited another preexisting zone of weakness, conjugate to the Calisto and surface facet planes.
 The stress necessary to initiate thrust faulting on Eros is determined by its near-surface strength. Two approaches are used to evaluate the near-surface shear strength of the asteroid, frictional strength and rock mass strength. The frictional strength is controlled by the resistance to brittle failure by sliding on randomly oriented, through going fractures (Byerlee's law). In terms of the maximum σ1 and minimum σ3 principal effective stresses Byerlee's law is expressed by σ1 ≃ 5σ3 for σ3 < 110 MPa where σ3 = ρgz and ρ is the density, g acceleration due to gravity, and z is the depth [see Brace and Kohlstedt, 1980]. Although g on Eros varies from 0.24 to 0.56 cm s−2 [Veverka et al., 2001a], 0.5 cm s−2 is a good average in the area of Hinks Dorsum [see Robinson et al., 2001, Figure 2]. Using the mean density of Eros ρ = 2.67 g/cm2 [Veverka et al., 2000], the shear strength at a depth of 0.25 km is ∼1.3 MPa (Figure 3). The frictional strength criteria is a lower limit on the shear strength for near-surface rocks because it assumes the cohesion Co = 0 [see Schultz, 1993; Scholz, 2002]. The strength of near-surface rocks with non-zero cohesion is better represented by the Hoek-Brown failure criterion given by
where σ1 and σ3 are the effective principal stresses, σc is the uniaxial compressive strength of intact rock, and m, s, and a are material constants related to the Geologic Strength Index (GSI) and a rock mass disturbance factor d [Hoek et al., 2002; Hoek and Diederichs, 2006]. Using σc = 200 MPa, the mean compressive strength of intact stony meteorites [Petrovic, 2001], a GSI = 45 (consistent with highly jointed basalt rock mass), and d = 1.0 (where 0 = undisturbed, 1 = very disturbed) [Hoek et al., 2002], the shear strength at a depth of 0.25 km is ∼6 MPa (Figure 3). Based on these two approaches, we estimate the near-surface shear strength of Eros to be from ∼1 to 6 MPa. This shear strength is well above the maximum expected from tidally induced stresses [e.g., Dobrovolskis, 1982] and far exceeds recent estimates of the strength of Phobos, Pandora, and Epimetheus (0.01 to 0.1 MPa) [Morrison et al., 2009]. Thermally induced stresses could exceed 5 MPa [Dombard and Freed, 2002], however, the presence of a relatively thick, uncompacted regolith [Veverka et al., 2001b; Robinson et al., 2002; Thomas and Robinson, 2005] will likely insolate the interior from their effects. The most likely source of stress capable of forming thrust faults on Eros is impact-induced compression [Melosh, 1989; Richardson et al., 2004; Thomas et al., 2001].
 The relatively young impact crater Shoemaker (there is no IAU approved name for this crater), which contains Charlois Regio, is ∼7.6 km in diameter and is superposed on the rim of the larger Himeros crater [Thomas et al., 2002a; Thomas and Robinson, 2005]. The relatively young age of both Shoemaker crater and Hinks Dorsum, combined with the likely need for impact-induced stress in the formation of Hinks, raises the question of a possible genetic relation. Shoemaker has been associated with most of the large ejecta blocks on Eros [Thomas et al., 2001], and has removed much of the population of craters <0.5 km in diameter within ∼9 km straight line distance of the center of Shoemaker [Thomas and Robinson, 2005]. A global structural feature map of Eros shows that Hinks Dorsum is roughly circumferential to the center of Shoemaker [see Thomas et al., 2002b, Figure 1], and the northward thrust slip direction of the fault is away from the crater center. Following the approach of Thomas and Robinson , the straight-line distance from the center of Shoemaker to points along Hinks Dorsum are plotted (Figure 4a). The plot indicates that Hinks Dorsum has a simple geometric relation to Shoemaker crater. Over half its length is within a very narrow range of distance from the center of Shoemaker (∼8.3 to 8.6 km), with a maximum distance of ∼10.3 km at the eastern end of the structure (Figure 4a). While the loss of craters within a certain distance of Shoemaker can be related to a level of impact induced acceleration or energy deposition, the formation of a thrust fault at a near-constant distance may have involved compressional stresses that exceeded 6 MPa. We note that segments of Hinks Dorsum are very close to the transition from low to high crater density, particularly in Himeros (Figure 4b). This spatial correlation suggests that the Hinks Dorsum thrust faults acted as a mechanical discontinuity where seismic shaking was attenuated. The crater density effects reach ∼9 km from Shoemaker around the entire asteroid, so any discontinuity at Hinks is not the only control on the crater density transition. Impact-induced compression from the Shoemaker event may have formed the Hinks Dorsum thrust fault, and subsequent seismic shaking likely resulted in the additional slip on the fault and the growth of the structure.
 We thank the two anonymous reviewers for helpful comments that improved the manuscript. This work is supported by the National Aeronautics and Space Administration under grants issued through the Office of the Planetary Geology and Geophysics Program.