Data Set Overview : This data set presents the detailed plate shape model of comet 103P/Hartley2, as derived from the images of the comet that were obtained by the DeepImpact spacecraft around the time of closest approach. Additionalinformation about the model can be found in Thomas et al., 'Shape, density,and geology of the nucleus of Comet 103P/Hartley 2', Icarus 222, 550-558,2013 [THOMASETAL2013]. The Deep Impact Flyby (DIF) spacecraft encountered comet 103P/Hartley 2 onNovember 4, 2010. It approached to within 694 km of the nucleus, imagingthroughout the approach, encounter and departure time frame. Two cameras,the High Resolution Instrument (HRIV) and the Medium Resolution Instrument(MRI) obtained hundreds of images of the nucleus, providing stereoinformation that was used to derive the shape of the nucleus. The images used for producing the shape model are contained in the PDSarchive of EPOXI Data for Hartley 2.The data set IDs are: DIF-C-HRIV-5-EPOXI-HARTLEY2-DECONV-V1.0 DIF-C-MRI-2-EPOXI-HARTLEY2-V1.0  PLATE MODEL The shape model was derived in planetocentic coordinates with vertices atintervals of 2 degrees in both latitude and longitude. The radial distanceis given in kilometers and ranges from a minimum of 0.31 km to a maximum1.26 km. The center of the model is about 5 m from the coordinate origin inthe direction of 19.6N, 240.4E. (This may be corrected in a futureversion of the model, but it is currently smaller than the uncertainty in thevertex positions). The nearly cylindrical shape is such thatit is not well represented by a triaxial ellipsoid. The nucleus is in a state of non-principal axis rotation (Belton etal. 'The complex spin state of 103P/Hartley 2. Kinematics and orientationin space', Icarus 222, 595-609, 2013 [BELTONETAL2013]) that limits any mappingconvention that would be tied to rotation. Thus, the coordinate system hasbeen defined such that the 'poles' extend along the long axis of the body(with positive latitudes toward the smaller lobe). The prime meridian wasdefined by a dark feature, part of which is a block shadow, near the centerof the illuminated face of the nucleus. The orientation of the nucleus atthe time of closest approach is defined in the table below. Most of the useful MRI data were obtained at <1600 km range (16 m/pixel),and a few deconvolved HRI images were useful out to ranges of ~2500 km.These data were obtained during an interval when a simplified model ofrotation predicts that the object orientation changed by ~0.6 deg. Theroutine mapping was done with a simple rotation model that introducesrelative position errors of ~0.3 deg between ends of the object. Determination of the shape and accurate relative positioning of the imagesrelies upon a set of 244 stereo control points distributed around the 50%of the nucleus that was illuminated and observed during the encounter.Limb outlines provide additional constraints on the shape, but have anambiguity on where the surface is intersected by the line of sight. Two versions of the shape model are included. The first is the originalmodel (HARTLEY2_2012_PLAN.TAB) derived as noted above, given inplanetocentric coordinates. The table includes a list of 16022 vertices,with connectivity for forming 32040 plates. The format follows thestandard PDS shape model definition (see PLATE_SHAPE_DEFINITION.ASC in thedocuments directory). In addition, each vertex includes a code thatdenotes the primary constraint on its derivation (control points, limbsilhouette, or not well constrained) providing a measure of the accuracy ofthe derived radius at that point. A second version of the model (HARTLEY2_2012_CART.WRL), was derived fromthe planetocentric version and presents the vertices in cartesiancoordinates. In this coordinate system, the X axis is defined in thedirection of the intersection of the equator and the prime meridian, the Zaxis lies along the positive pole and the Y axis completes the right-handcoordinate system. The cartesian coordinate version of the model is presented with a VRMLwrapper that allows it to be displayed with existing VRML viewers that arefreely available (e.g., INSTANT PLAYER, OCTAGA, CORTONA, etc.) Differentcamera angles and illumination conditions have been embedded in the WRLfile to display the nuclues as it would appear during the approach, closestapproach point, and departure, with the associated solar illumination(though the twist angle around the line of sight may not match what isrecorded in the corresponding images). There are also embedded viewpointsfrom all of the principal axis directions, with the 'headlight'illumination. Note that the capabilities of different viewers may limitthe ability to display some or all of these viewpoints. Look for 'cameras'or 'navigation' items in the menus.  TABLE: Shape Model Characteristics  Area: 5.24 km^2 Volume: 0.809 +/- 0.077 km^3 Radius of eq. volume sphere: 0.58 +/- 0.018 km Radius Range: 0.31-1.26 km Gravity: 0.0019-0.0044 cm s^2 (for mean density 300 kg m^3) Model moments: 5.302E4, 31.26E4, 31.92E4 m^2 Model moment ratios (A,B,C are the principal moment of inertia axes): A/C: 0.166 +/- 0.004 B/C: 0.979 +/- 0.002 Body model moment orientations: A: 89.73 deg, 207.56 deg E B: 0.05 deg, 106.14 deg E C: 0.25 deg, 16.14 deg E Time of Closest Approach: 2010-11-04T13:59:47.7 Directions of Axes at Closest Approach: RA (deg) Dec (deg) (J2000) +Z (Long) axis: 226.12 39.37 +X axis: 237.10 -50.10 +Y Axis: 320.59 5.42   Author------The descriptions in this file were written by Tony Farnham, based oninformation from [THOMASETAL2013] and communications with PeterThomas, the shape model creator.

Confidence Level Overview : In the regions that are well-defined by control points, the radialuncertainty of the plate model is approximately 10 m. In regions notdefined by control points, silhouettes provide the constraints, withuncertainties of ~30 m dominated by the ambiguity of where the limb liesalong the line of sight. The mean uncertainty is calculated to be ~18 m(1-sigma).  Review :The plate model was reviewed March 5, 2013.