Ballistic reconstruction of HST observations of ejecta motion following Shoemaker-Levy 9 impacts into Jupiter

We present a detailed analysis of Hubble Space Telescope (HST) Wide Field and Planetary Camera 2 (WFPC 2) images of the first 25 min of the impacts of Comet Shoemaker-Levy 9 fragments A and G with Jupiter and the post impact ejecta patterns observed 1-2 h after each impact event, with a brief analysis of the available observations of the first 25 min of the E and W events. Via a ballistic model we (i) constrain the initial time of impact and the initial velocity and elevation angle of the apex of each observed impactor; (ii) reproduce the shape and the motion of the sunlit A and G ejecta clouds utilizing the initial velocities constrained for the apex of each ejecta cloud and initial elevation angles ranging from 80 degrees to 60 degrees +/- 5 degrees; and (iii) compare the ejecta pattern generated by particle trajectories consistent with the observed ejecta clouds to the observed postimpact ejecta pattern. We conclude that (i) the crescent patterns can be reconstructed from ballistic ejecta traveling on trajectories of elevation angles ranging from 70 degrees to 45 degrees and initial velocity equivalent to that constrained for the apex of each cloud; (ii) the streaks can be reconstructed from ballistic particles propelled up from the channel at velocities less than 8.5 +/- 1 km s(-1); and (iii) the altitude of the self-luminous radiation detected at the onset of the A, G, and W observations is inconsistent with the ballistic trajectories fit to the apex of the sunlit ejecta clouds. We believe that the discrepancy between the trajectories of the initial HST/WFPC 2 detections and the later images is an indication of distinctly different phenomena associated with the impacts of the comet fragments. The timing of these events indicates that the first event recorded in association with the G impact is consistent with the entry of the bolide and that the self-luminous radiation recorded similar to 120 s after the ballistically derived impact time for the A and G impacts may be consistent with the excitation of the jovian atmosphere and/or cometary debris in the path of shock fronts accelerating away from the respective impact fireballs. Given the uncertainty in the ballistically derived impact time, the initial detection of the W event may have corresponded to either the incoming bolide or atmospheric and/or cometary debris excitation initiated by the passing of a shock front associated with the impact fireball. (C) 2000 Academic Press.