The shape, topography, and geology of Tempel 1 from ... - CiteSeerX
Icarus 187 (2007) 4–15 www.elsevier.com/locate/icarus
The shape, topography, and geology of Tempel 1 from Deep Impact observations Peter C. Thomas a,∗ , J. Veverka a , Michael J.S. Belton b , Alan Hidy c , Michael F. A’Hearn d , T.L. Farnham d , Olivier Groussin d , Jian-Yang Li d , Lucy A. McFadden d , Jessica Sunshine d , Dennis Wellnitz d , Carey Lisse e , Peter Schultz f , Karen J. Meech g , W. Alan Delamere h a Center for Radiophysics and Space Research, Cornell University, Ithaca, NY 14853, USA b Belton Space Exploration Initiatives, 430 S. Randolf Way, LLC, Tucson, AZ 85716, USA c Department of Geology, Utah State University, Logan, UT 84322, USA d Department of Astronomy, University of Maryland, College Park, MD 20742, USA e Applied Physics Laboratory, 11000 Johns Hopkins Rd., Laurel, MD 20723, USA f Department of Geological Sciences, Brown University, Providence, RI 02912, USA g Institute for Astronomy, University of Hawaii, Honolulu, HI 96822, USA h Delamere Support Services, 525 Mapleton Avenue, Boulder, CO 80306, USA
Received 7 August 2006; revised 18 December 2006 Available online 8 January 2007
Abstract Deep Impact images of the nucleus of Comet Tempel 1 reveal pervasive layering, possible impact craters, flows with smooth upper surfaces, and erosional stripping of material. There are at least 3 layers 50–200 m thick that appear to extend deep into the nucleus, and several layers 1–20 m thick that parallel the surface and are being eroded laterally. Circular depressions show geographical variation in their forms and suggest differences in erosion rates or style over scales >1 km. The stratigraphic arrangement of these features suggests that the comet experienced substantial periods of little erosion. Smooth surfaces trending downslope suggest some form of eruption of materials from this highly porous object. The Deep Impact images show that the nucleus of Tempel 1 cannot be modeled simply as either an onion-layer or rubble pile structure. © 2007 Elsevier Inc. All rights reserved. Keywords: Comets, nucleus; Comet Tempel-1
1. Introduction Deep Impact returned the highest resolution images yet obtained of a cometary nucleus. About one-quarter the surface was imaged at well under 10 m/pixel, and small areas were imaged at less than 1 m/pixel. A’Hearn et al. (2005a) identified some of the most important aspects of the surface characteristics of Tempel 1, including the diversity and complexity of the surface, the occurrence of crater-like depressions, layering, evidence of extensive surface erosion, and the unexpected oc* Corresponding author. Address for correspondence: 422 Space Sciences, Cornell University, Ithaca, NY 14850, USA. E-mail address: [email protected] (P.C. Thomas).
0019-1035/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.icarus.2006.12.013
currence of very smooth terrains. Another major result from Deep Impact which affects the interpretation of surface features is the very low average density of the nucleus, ∼0.3 gm/cm3 (Richardson and Melosh, 2006a, 2006b). In this paper we present a more comprehensive survey of surface features on Tempel 1 and use an improved shape model to derive local slopes and gravitational gradients over the wellimaged portion of the nucleus. Such information makes it possible in many cases to refine the geological interpretation. Conventions established by the International Astronomical Union (IAU) (Seidelmann et al., 2005) for coordinate systems of comets and asteroids are based on positive and negative latitudes with respect to right-hand rule for spin, and positive longitudes. We also apply the terms north, south, east, and west here to facilitate local directional descriptions. North is the positive
Geology of Tempel 1
5
Fig. 1. Control points and shape model on MRI image MV173728441. Negative (south) pole is to lower left. Sub-spacecraft point −44◦ , 303◦ ; range 1238 km.
pole, positive longitudes run east on Tempel 1. Surface coordinates are body-centric. 2. Shape of the nucleus of Tempel 1 Before the arrival of Deep Impact, ground-based studies (summarized in Fernandez et al., 2003 and Belton et al., 2005) indicated a nucleus with a mean radius of about 3 km, and an irregular shape. These studies suggested axial ratios of at least 1.4 and possibly as high as 3. The spin period was determined to be 41.85 h (Belton et al., 2005). The orientation of the spin vector was poorly constrained before the flyby, but subsequent analyses of spacecraft and ground-based data have shown convergence of solutions (A’Hearn et al., 2005a). Given the flyby speed, the comet’s slow rotation, and the approach phase angle, less than half the object was well-resolved, and shape and spin vector determination required lightcurve comparisons and some assumptions about the homogeneity of the object. An approximate shape model was first developed with limb and terminator positions in the inbound and outbound images (Figs. 1 and 2) obtained with the Medium Resolution Instrument (MRI). The instruments on Deep Impact are described in Hampton et al. (2005). The sense of rotation was visible in the low-resolution inbound images, but images that revealed more than a few surface features spanned only a few degrees of rotation of the nucleus. With this preliminary model, control points were obtained for the region seen at pixel scales of 50 m or better (Fig. 1). Solution of a control point requires measurement in at least three images, and convergence angles greater than 10◦ .
Fig. 2. Outbound image 173730609 showing ejecta plume slightly over the horizon and full disk silhouette. View from +22◦ , 104◦ ; range 20,930 km.
Control points used are circular forms, bright spots, and other distinctive small features. Our solution has 189 points, using 1520 measurements in 70 images, with an average residual of 0.4 pixels. The pixel residuals are slightly larger than for objects with well-known spin vectors and sharp crater rims that are easily centroided, but this solution provides useful topography with an average radial uncertainty of approximately 20 m at the control point locations. Some of the control points are in an area near the south pole of the nucleus that is illuminated only by light scattered from the impact ejecta.
6
P.C. Thomas et al. / Icarus 187 (2007) 4–15
Fig. 3. Image map of Tempel 1, simple cylindrical projection. ITS and MRI images used.
With 30% of the object’s shape well-constrained and the rest very approximately estimated, a comparison of the shape model’s maximum moment orientation (assuming a homogeneous object) and the assumed spin vector showed a difference of more than 15◦ . The lightcurve observed on approach (Belton et al., 2006) also suggested errors in either the shape or spin vector. We then used a new spin vector along the direction of the calculated maximum moment vector, and set the coordinate origin at a model center of mass, again assuming a homogeneous interior. A small change to one part of the model not seen at high resolution, but facing the spacecraft in Fig. 2, resulted in some improvement in the lightcurve match, but further modification may be warranted. Some additional adjustments of the limb and terminator matches were made for the current model, which involved another iteration of checking moment orientation and center of mass. The current spin model is a positive pole at RA = 294◦ , Dec ◦ 73 , with an uncertainty of approximately 5◦ (this vector differs from the preliminary value in A’Hearn et al., 2005a). The prime meridian is set at the center of a 350-m crater west of the impact site (Fig. 3), such that W = 252.63◦ + 212.064◦ d, where d is the number of days since the standard epoch (JD = 2451545.0). MRI and Impactor Targeting Sensor (ITS) images have been projected to produce an image map of part of the surface of Tempel 1 (Fig. 3). In Fig. 4 we display the image map projected on the shape model as viewed from four different directions, a presentation that shows which portions of the nucleus were imaged well by Deep Impact. The shape model (Table 1) has a mean radius of 3.0 ± 0.1 km. This is the radius of a sphere of equivalent volume. The uncertainty in the volume is estimated by the amount of additional volume that can be added or subtracted with-
out violating either the limb views (look-back images such as Fig. 2 are crucial here), or generating physically implausible shapes, such as holes to the center of the object in the “unseen” longitudes. The mean radius compares well with ground-based measurements, but the elongation is less than most pre-encounter estimates. For comparison, the nucleus of Wild 2 has a mean radius of ∼2 km (Duxbury et al., 2004; Kirk et al., 2005), and that of Borrelly is about 3 km but is more elongate than Tempel 1 (Britt et al., 2004). Richardson and Melosh (2006b) have used observations of the evolution of the solid ejecta plume to model local gravitational acceleration, and thereby derive the mean density of a homogeneous nucleus. Their result, ρ = 0.35 ± 0.25 g cm−3 (2σ uncertainty), suggests an extremely underdense object. The data make it possible to derive gravitational topography and slopes (Thomas, 1993) on the well-imaged portion of the nucleus (Fig. 5). The slopes on this object, in the wellcontrolled regions, are slightly greater than those measured on well-imaged asteroids and small satellites (Fig. 6). The higher slopes may result from the lack of a smoothing effect from the redistribution of regolith by impacts, seismic effects, and other downslope transport. Slopes of individual features and local areas are discussed below. Modeled gravitational heights for the nucleus as a whole are shown in Fig. 7. These assume a homogeneous interior. Uncertainties in the shape in poorly imaged areas create uncertainties in the local gravity vectors, but these are small compared to the overall pattern of topography and slopes. Uncertainties in the mean density have little effect on the slope calculations because the very slow rotation adds only very small accelerations. Much of the relief in the well-imaged part of this object derives from the somewhat faceted appearance of this region, with ridges bounding flatter areas being the gravitationally high areas.
Geology of Tempel 1
7
Fig. 4. Four views of the shape model with image mosaic (Fig. 3) projected on surface. Table 1 Shape model of Tempel 1 Mean radius Diameter range Gravity Area Range of gravitational heights
3.0 ± 0.1 km 5.0–7.5 km 0.024–0.030 cm s−2 119 km2 0.73 km
3. Surface features and geology We begin this discussion by classifying the various morphologic forms and albedo features visible on the nucleus of Tempel 1. Keys to surface features are shown in Fig. 8; panel B shows regions of occurrence of those features that can be classed as morphologic units, distinct from those that are specific forms that may occur on wider units. The other panels of Fig. 8 give selected examples of specific morphologies. The classification below is arranged in pseudo-stratigraphic order, oldest first, with the caveat that because interpretation of the forms is uncertain, the sequence is also. a: Linear outcrops. Darker and brighter banding runs approximately NNE–SSW for >3.5 km, exposed over widths as small as a few 10’s of m to over 200 m. The scarp bounding part of unit h parallels some of these bands. b: Intermediate roughness pitted surface. Much of the western facet appears less rough and perhaps less eroded than unit c.
c: Pitted and rough surface. This unit is perhaps the roughest part of Tempel 1, and has what may be many merged depressions 10’s to 100 m across. It appears somewhat darker than much of unit h and the smooth areas. Its boundary with parts of unit h are poorly defined. d: Isolated, rimless depressions. These occur in the western facet, in unit b. They range from 100 to 400 m in diameter, have no raised rims, usually have flat floors, and display a variety of roughly concentric albedo markings. Depths are probably well under 50 m. e: Rim remnants. These occur in unit h and reach diameters of 350 m. They appear to be nearly circular raised rims of varying width, darker than their surroundings, with interior fill similar to material exposed outside. f: Close-packed depressions. Two zones, that may be within unit c, have round depressions and arcuate scarps indicating closely packed depressions 100–200 m in diameter, and
The shape, topography, and geology of Tempel 1 from Deep Impact observations Peter C. Thomas a,∗ , J. Veverka a , Michael J.S. Belton b , Alan Hidy c , Michael F. A’Hearn d , T.L. Farnham d , Olivier Groussin d , Jian-Yang Li d , Lucy A. McFadden d , Jessica Sunshine d , Dennis Wellnitz d , Carey Lisse e , Peter Schultz f , Karen J. Meech g , W. Alan Delamere h a Center for Radiophysics and Space Research, Cornell University, Ithaca, NY 14853, USA b Belton Space Exploration Initiatives, 430 S. Randolf Way, LLC, Tucson, AZ 85716, USA c Department of Geology, Utah State University, Logan, UT 84322, USA d Department of Astronomy, University of Maryland, College Park, MD 20742, USA e Applied Physics Laboratory, 11000 Johns Hopkins Rd., Laurel, MD 20723, USA f Department of Geological Sciences, Brown University, Providence, RI 02912, USA g Institute for Astronomy, University of Hawaii, Honolulu, HI 96822, USA h Delamere Support Services, 525 Mapleton Avenue, Boulder, CO 80306, USA
Received 7 August 2006; revised 18 December 2006 Available online 8 January 2007
Abstract Deep Impact images of the nucleus of Comet Tempel 1 reveal pervasive layering, possible impact craters, flows with smooth upper surfaces, and erosional stripping of material. There are at least 3 layers 50–200 m thick that appear to extend deep into the nucleus, and several layers 1–20 m thick that parallel the surface and are being eroded laterally. Circular depressions show geographical variation in their forms and suggest differences in erosion rates or style over scales >1 km. The stratigraphic arrangement of these features suggests that the comet experienced substantial periods of little erosion. Smooth surfaces trending downslope suggest some form of eruption of materials from this highly porous object. The Deep Impact images show that the nucleus of Tempel 1 cannot be modeled simply as either an onion-layer or rubble pile structure. © 2007 Elsevier Inc. All rights reserved. Keywords: Comets, nucleus; Comet Tempel-1
1. Introduction Deep Impact returned the highest resolution images yet obtained of a cometary nucleus. About one-quarter the surface was imaged at well under 10 m/pixel, and small areas were imaged at less than 1 m/pixel. A’Hearn et al. (2005a) identified some of the most important aspects of the surface characteristics of Tempel 1, including the diversity and complexity of the surface, the occurrence of crater-like depressions, layering, evidence of extensive surface erosion, and the unexpected oc* Corresponding author. Address for correspondence: 422 Space Sciences, Cornell University, Ithaca, NY 14850, USA. E-mail address: [email protected] (P.C. Thomas).
0019-1035/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.icarus.2006.12.013
currence of very smooth terrains. Another major result from Deep Impact which affects the interpretation of surface features is the very low average density of the nucleus, ∼0.3 gm/cm3 (Richardson and Melosh, 2006a, 2006b). In this paper we present a more comprehensive survey of surface features on Tempel 1 and use an improved shape model to derive local slopes and gravitational gradients over the wellimaged portion of the nucleus. Such information makes it possible in many cases to refine the geological interpretation. Conventions established by the International Astronomical Union (IAU) (Seidelmann et al., 2005) for coordinate systems of comets and asteroids are based on positive and negative latitudes with respect to right-hand rule for spin, and positive longitudes. We also apply the terms north, south, east, and west here to facilitate local directional descriptions. North is the positive
Geology of Tempel 1
5
Fig. 1. Control points and shape model on MRI image MV173728441. Negative (south) pole is to lower left. Sub-spacecraft point −44◦ , 303◦ ; range 1238 km.
pole, positive longitudes run east on Tempel 1. Surface coordinates are body-centric. 2. Shape of the nucleus of Tempel 1 Before the arrival of Deep Impact, ground-based studies (summarized in Fernandez et al., 2003 and Belton et al., 2005) indicated a nucleus with a mean radius of about 3 km, and an irregular shape. These studies suggested axial ratios of at least 1.4 and possibly as high as 3. The spin period was determined to be 41.85 h (Belton et al., 2005). The orientation of the spin vector was poorly constrained before the flyby, but subsequent analyses of spacecraft and ground-based data have shown convergence of solutions (A’Hearn et al., 2005a). Given the flyby speed, the comet’s slow rotation, and the approach phase angle, less than half the object was well-resolved, and shape and spin vector determination required lightcurve comparisons and some assumptions about the homogeneity of the object. An approximate shape model was first developed with limb and terminator positions in the inbound and outbound images (Figs. 1 and 2) obtained with the Medium Resolution Instrument (MRI). The instruments on Deep Impact are described in Hampton et al. (2005). The sense of rotation was visible in the low-resolution inbound images, but images that revealed more than a few surface features spanned only a few degrees of rotation of the nucleus. With this preliminary model, control points were obtained for the region seen at pixel scales of 50 m or better (Fig. 1). Solution of a control point requires measurement in at least three images, and convergence angles greater than 10◦ .
Fig. 2. Outbound image 173730609 showing ejecta plume slightly over the horizon and full disk silhouette. View from +22◦ , 104◦ ; range 20,930 km.
Control points used are circular forms, bright spots, and other distinctive small features. Our solution has 189 points, using 1520 measurements in 70 images, with an average residual of 0.4 pixels. The pixel residuals are slightly larger than for objects with well-known spin vectors and sharp crater rims that are easily centroided, but this solution provides useful topography with an average radial uncertainty of approximately 20 m at the control point locations. Some of the control points are in an area near the south pole of the nucleus that is illuminated only by light scattered from the impact ejecta.
6
P.C. Thomas et al. / Icarus 187 (2007) 4–15
Fig. 3. Image map of Tempel 1, simple cylindrical projection. ITS and MRI images used.
With 30% of the object’s shape well-constrained and the rest very approximately estimated, a comparison of the shape model’s maximum moment orientation (assuming a homogeneous object) and the assumed spin vector showed a difference of more than 15◦ . The lightcurve observed on approach (Belton et al., 2006) also suggested errors in either the shape or spin vector. We then used a new spin vector along the direction of the calculated maximum moment vector, and set the coordinate origin at a model center of mass, again assuming a homogeneous interior. A small change to one part of the model not seen at high resolution, but facing the spacecraft in Fig. 2, resulted in some improvement in the lightcurve match, but further modification may be warranted. Some additional adjustments of the limb and terminator matches were made for the current model, which involved another iteration of checking moment orientation and center of mass. The current spin model is a positive pole at RA = 294◦ , Dec ◦ 73 , with an uncertainty of approximately 5◦ (this vector differs from the preliminary value in A’Hearn et al., 2005a). The prime meridian is set at the center of a 350-m crater west of the impact site (Fig. 3), such that W = 252.63◦ + 212.064◦ d, where d is the number of days since the standard epoch (JD = 2451545.0). MRI and Impactor Targeting Sensor (ITS) images have been projected to produce an image map of part of the surface of Tempel 1 (Fig. 3). In Fig. 4 we display the image map projected on the shape model as viewed from four different directions, a presentation that shows which portions of the nucleus were imaged well by Deep Impact. The shape model (Table 1) has a mean radius of 3.0 ± 0.1 km. This is the radius of a sphere of equivalent volume. The uncertainty in the volume is estimated by the amount of additional volume that can be added or subtracted with-
out violating either the limb views (look-back images such as Fig. 2 are crucial here), or generating physically implausible shapes, such as holes to the center of the object in the “unseen” longitudes. The mean radius compares well with ground-based measurements, but the elongation is less than most pre-encounter estimates. For comparison, the nucleus of Wild 2 has a mean radius of ∼2 km (Duxbury et al., 2004; Kirk et al., 2005), and that of Borrelly is about 3 km but is more elongate than Tempel 1 (Britt et al., 2004). Richardson and Melosh (2006b) have used observations of the evolution of the solid ejecta plume to model local gravitational acceleration, and thereby derive the mean density of a homogeneous nucleus. Their result, ρ = 0.35 ± 0.25 g cm−3 (2σ uncertainty), suggests an extremely underdense object. The data make it possible to derive gravitational topography and slopes (Thomas, 1993) on the well-imaged portion of the nucleus (Fig. 5). The slopes on this object, in the wellcontrolled regions, are slightly greater than those measured on well-imaged asteroids and small satellites (Fig. 6). The higher slopes may result from the lack of a smoothing effect from the redistribution of regolith by impacts, seismic effects, and other downslope transport. Slopes of individual features and local areas are discussed below. Modeled gravitational heights for the nucleus as a whole are shown in Fig. 7. These assume a homogeneous interior. Uncertainties in the shape in poorly imaged areas create uncertainties in the local gravity vectors, but these are small compared to the overall pattern of topography and slopes. Uncertainties in the mean density have little effect on the slope calculations because the very slow rotation adds only very small accelerations. Much of the relief in the well-imaged part of this object derives from the somewhat faceted appearance of this region, with ridges bounding flatter areas being the gravitationally high areas.
Geology of Tempel 1
7
Fig. 4. Four views of the shape model with image mosaic (Fig. 3) projected on surface. Table 1 Shape model of Tempel 1 Mean radius Diameter range Gravity Area Range of gravitational heights
3.0 ± 0.1 km 5.0–7.5 km 0.024–0.030 cm s−2 119 km2 0.73 km
3. Surface features and geology We begin this discussion by classifying the various morphologic forms and albedo features visible on the nucleus of Tempel 1. Keys to surface features are shown in Fig. 8; panel B shows regions of occurrence of those features that can be classed as morphologic units, distinct from those that are specific forms that may occur on wider units. The other panels of Fig. 8 give selected examples of specific morphologies. The classification below is arranged in pseudo-stratigraphic order, oldest first, with the caveat that because interpretation of the forms is uncertain, the sequence is also. a: Linear outcrops. Darker and brighter banding runs approximately NNE–SSW for >3.5 km, exposed over widths as small as a few 10’s of m to over 200 m. The scarp bounding part of unit h parallels some of these bands. b: Intermediate roughness pitted surface. Much of the western facet appears less rough and perhaps less eroded than unit c.
c: Pitted and rough surface. This unit is perhaps the roughest part of Tempel 1, and has what may be many merged depressions 10’s to 100 m across. It appears somewhat darker than much of unit h and the smooth areas. Its boundary with parts of unit h are poorly defined. d: Isolated, rimless depressions. These occur in the western facet, in unit b. They range from 100 to 400 m in diameter, have no raised rims, usually have flat floors, and display a variety of roughly concentric albedo markings. Depths are probably well under 50 m. e: Rim remnants. These occur in unit h and reach diameters of 350 m. They appear to be nearly circular raised rims of varying width, darker than their surroundings, with interior fill similar to material exposed outside. f: Close-packed depressions. Two zones, that may be within unit c, have round depressions and arcuate scarps indicating closely packed depressions 100–200 m in diameter, and
