There isn't a single reason for RC, but the biggest reason is that our local soil sucks and good soil to make 'well-engineered' earth structures is a limited resource.
Additionally, there's differences in engineering philosophy between public-works and the oil field (I do both), and it's probably the opposite of what most think. Especially wrt capitalization/ finance. Almost nothing is done cheapest, let alone cheap in the coastal oil field.
Oil wants it done for a known cost, even if higher up front, to have higher certainty in long-term risk/ or lower maintenance costs. (Also soil takes months longer to settle, and time might be big money, depending on the project.)
If that higher cost tanks the project numbers, fine; they have 100 other projects lined up to make that IRR without resorting cost/corner-cutting.
Other thoughts:
Stone that is suitable for armoring earthen structures isn't sourced locally. No mountains. Has to all be barged in.
And still wouldn't survive the constant battering of wind/waves (in both a long-term seasonal sense, as well as during an acute storm event) as well as concrete.
Subsidence of a pile supported concrete structure will match the adjacent pile supported facilities, and be minimal compared to an earthen structure, which will also have expected settling on top of subsidence.
Floodwalls can double as containment berms, and that adds many design considerations not typical for stand-alone flood structures, including chemical resistance and fire suppression.
But even something as simple as grass cutting might tip the scales. Lots of heli- or boat-only access facilities, and grass needs to be cut at least once per week during summer here. Maintaining grass/earth isn't necessarily cheaper when your yard guy needs a stack of govt twic/osha cards, a private ferry ride out, and probably transport to a private dock a few hours from town.
All of which just barely gets into the logistics involved in the industry.
>more likely to fail catastrophically
Maybe catastrophically, depending on how you want to define it, but less likely to fail overall. Concrete is more likely to fail in a predictable manor.
Even the Katrina concrete/steel 'floodwall' failures were actually soil failures, not RC or steel ones.