I don't know, cost of disaster recovery isn't in the analyses I've read. For nuclear power the failure modes for different fuels and generations is different, so any analysis would need to take that into account. If you find something would you please post it?