Gravity-dominated Collisions: A Model for the Largest Remnant Masses with Treatment for "Hit and Run" and Density Stratification.

We develop empirical relationships for the accretion and erosion of colliding gravity-dominated bodies of various compositions under conditions expected in late-stage solar system formation. These are fast, easily coded relationships based on a large database of smoothed particle hydrodynamics (SPH) simulations of collisions between bodies of different compositions, including those that are water rich. The accuracy of these relations is also comparable to the deviations of results between different SPH codes and initial thermal/rotational conditions. We illustrate the paucity of disruptive collisions between major bodies, as compared to collisions between less massive planetesimals in late-stage planet formation, and thus focus on more probable, low-velocity collisions, though our relations remain relevant to disruptive collisions as well. We also pay particular attention to the transition zone between merging collisions and those where the impactor does not merge with the target, but continues downrange, a "hit-and-run" collision. We find that hit-and-run collisions likely occur more often in density-stratified bodies and across a wider range of impact angles than suggested by the most commonly used analytic approximation. We also identify a possible transitional zone in gravity-dominated collisions where larger bodies may undergo more disruptive collisions when the impact velocity exceeds the sound speed, though understanding this transition warrants further study. Our results are contrary to the commonly assumed invariance of total mass (scale), density structure, and material composition on the largest remnants of giant impacts. We provide an algorithm for adopting our model into N-body planet formation simulations, so that the mass of growing planets and debris can be tracked.

Distribution of mid-latitude ground ice on Mars from new impact craters.

New impact craters at five sites in the martian mid-latitudes excavated material from depths of decimeters that has a brightness and color indicative of water ice. Near-infrared spectra of the largest example confirm this composition, and repeated imaging showed fading over several months, as expected for sublimating ice. Thermal models of one site show that millimeters of sublimation occurred during this fading period, indicating clean ice rather than ice in soil pores. Our derived ice-table depths are consistent with models using higher long-term average atmospheric water vapor content than present values. Craters at most of these sites may have excavated completely through this clean ice, probing the ice table to previously unsampled depths of meters and revealing substantial heterogeneity in the vertical distribution of the ice itself.