Lunar-forming impacts: High-resolution SPH and AMR-CTH simulations

Abstract We present results of the highest-resolution simulations to date of potential Moon-forming impacts using a Lagrangian, particle-based method (smooth particle hydrodynamics, or SPH) and an Eulerian, grid-based method with adaptive mesh refinement (AMR-CTH). We consider a few candidate impacts advocated by recent works, directly comparing simulations performed at varying resolutions and with both numerical methods and their predictions for the properties of resulting protolunar disks. For a fixed set of impact conditions, simulations with either method and with different resolutions yield very similar results for the initial impact and the first few hours of the post-impact period. The subsequent disk properties in the ∼5–20 h time period can vary substantially from case-to-case, depending on the orbits of and mutual interactions between large bound clumps of ejecta that often form after the initial impact. After such clumps have completed at least one orbit (which typically requires ∼25–50 h), the predicted protolunar disk mass and its angular momentum converge to within about 10% for simulations of very similar impact conditions using different resolutions or methods. The disks produced by the CTH simulations are consistently about 10% less massive than those produced by SPH simulations, due presumably to inherent differences between the codes. The two methods predict broadly similar values for the fraction of the protolunar disk that originates from the target vs. the impactor, and for the initial disk radial surface density and temperature profiles.