As for the median, we would need to know the depth, width and slope to determine whether a wingtip would contact before the gear. One wingtip contacting before the other would trigger high yaw and roll rates. I would be most pleasantly surprised if a similar aircraft remained in the median. The worst case would be being deflected into oncoming traffic.
The median barrier could be high enough to breach the wing tanks and add to forces directing the accident aircraft into oncoming traffic. Perhaps the median could work out for a high wing single.
As for the 700m pond, it was about a mile closer to the point where power was lost. People with access to appropriate sims could see if either large pond could be made from the power loss point. The NTSB might decide to research this on a sim. Often several crews are run through accident scenarios on a sim. The "success" rate would be interesting data.