If this can happen to an organisation ,on the surface well run and well managed and considered as just plain bad luck, then the only way to avoid such things in the future is to stop all scenic flights out of Sea World.
Lookleft, an honest thorough review of all relevant factors is indeed essential, but surely there's a possibility that no significant systemic problems will be found? Sometimes things just do happen, and not every occurrence warrants changes being made to systems.
I'm not saying that's the case here - there may well be important changes needed, and of course it's loss of human lives we're talking about, with all the terrible ramifications of that. My point is that safety systems should certainly have robust examination and ongoing improvement built in, but I believe there's a tendency for organisations, and outsiders, to conclude that 'something' must be done for the safety system to have done its job. The conclusion that no changes are necessary, even if it's supported by the investigation, often doesn't fly for those seeking answers.
That can be a good thing, if it roots out bad practices, but can also lead to institution of counter-productive (in the safety sense) measures, and in some cases, a smoke-screen warm and fuzzy feeling that 'we did something, so now we can forget about it.'
Edited to add: 'we did something' is also too often translated into 'we were seen to do something.'