On farming - it's worth keeping in mind that the trophic pyramid biomass distribution for land and sea look pretty much completely different. On land you get the 10-1 ratio, where you have 10 biomass of plants for every biomass of plant eater, 10 biomass of plant eaters for every biomass of primary predator, etc. (keeping in mind that the pyramid is messy and there's lots of things that occupy multiple levels.

The ocean doesn't work like that, instead being top heavy. The 10-1 ratio only applies to flows, where 10 times as much biomass grows and is eaten in level N as in level N+1, mostly starting with zooplankton and algae at the bottom. For stocks you get a more inverted pyramid, where there is way more biomass in fish than algae, and in big predators than in smaller prey fish. This also tends to apply to rivers, though there's more variation there.

So, how does this relate to farming? It suggests that the farm shouldn't be the basis of food - instead, that should be the managed fishery. Set up large net based fisheries on currents, where food is swept through (tiny zooplankton), but fish aren't. Slightly inland you also have lake fisheries, wetland fisheries (though this might be an exception to the rule, where farming actually works), etc.

Note that I'm assuming that they're not just sea faring, but effectively amphibious. If they're land based but just marginally better swimmers this analysis applies a lot less.