Oh, I like your Fermi model! (And also my above comment was horribly incorrect -- see the subsequent discussion with Constant.)

What I was wondering however is whether it might be off too much even by Fermi problem standards, i.e. by multiple orders of magnitude. The trouble is that if the target creatures are vastly better or poorer conductors than the surrounding medium, this greatly influences how the flow of energy through and around them is distributed, possibly making the model based on uniform energy flow across all angles too inaccurate even for a Fermi calculation. (To give an extreme example, a metal wire connecting the poles of a battery draws nearly all energy flow in the circuit through itself, despite being a negligible part of the spatial cross-section.)

Or to put it more precisely, the way a human distorts the flow of electrical energy when surrounded by ground and air may well be extremely different, and possibly go in a totally different direction, from the way a fish distorts it when surrounded by seawater, so your generalization from humans to fish might be problematic.

My initial idea was to attempt another Fermi approach based on guesstimating V(r) and its derivative, but the poor conductivity of fish relative to seawater seems to complicate that one too.