[Previous] | [Session 4] | [Next]
J. Harrington (Cornell), D. Deming (NASA GSFC)
We model the plume flight and atmospheric response phases of the Shoemaker-Levy 9 impacts, producing synthetic impact-site images and lightcurves that are consistent with observations. We initialize a ballistic Monte-Carlo plume model with the final velocity distribution of Zahnle and Mac Low's (1995, J. Geophys. Res. 100, 16,885--16,894) entry-response model. Our model ejects the plume into a cone with adjustble tilt and opening angles, and calculates where and when the mass, energy, and momentum land. It also has parameterized material sliding, since the observed plumes slid after impact (Hammel et al. 1994, Science 267, 1288--1296). The model reproduces the main features of the large impact sites, including the previously-unexplained expanding ring seen at 3.2 microns by McGregor et al. (1996, Icarus 121, 361--388). The model's output feeds a radiative-hydrodynamic atmosphere model based on the ZEUS 3D code of Stone and Norman (1992, Astrophys. J. Suppl. 80, 753--790). This model produces synthetic lightcurves and explains the physics of the ``third precursors'', ``main event'', ``bounces'', ``flare'', and several spectroscopic observations, including hot CO. The maximum temperature is 2500 K. The principal cooling mechanism following the main event is adiabatic horizontal expansion, not radiation. We cannot reproduce many lightcurve features without a "vanguard" of high-density, fast material preceeding the main plume. This is a little-noted feature of Zahnle and Mac Low's (1995) published velocity distribution. Our models represent the closure of the scientific method on models of fragment impact and breakup, as they carry the final conditions of impact models forward in time and produce synthetic observables capable of distinguishing among those models. We offer to initialize our models with the final conditions of any published impact model to perform this test. Work funded by the NASA Planetary Atmospheres Program.