Hi Bruce. If you have the means to suspend a wire well above ground level, unless it's unavoidably surrounded by a thick stand of tall trees, I'd recommend avoiding loop antennas altogether. It's generally more effective, foot for foot, to use a vertical monopole.Even where a loop is the only alternative because of terrain, I would utterly avoid a horizontal loop in any case, regardless of elevation. At LF, groundwave propagation is effectively shorted out by horizontal polarization. Also, since soil is a better conductor at LF than at HF, the out-of-phase reflection from the earth beneath a horizontal antenna (dipole or loop) tends to cancel most of the skywave, too. Same basic reason HF loops-on-ground work better at higher frequencies than they do at, say, 160 meters. At 1750 m, performance would be a factor of 10 to 11× worse.
Your surmise about how Warren's dipole works is substantially correct, according to my understanding of the matter. The current down into and back up from the earth forms a broad approximation of a vertical loop. This accounts for why it is a better radiator off the ends of the dipole rather than broadside, among other things. The principle dates back to the work of Ferdinand Braun at the end of the 19th century.
But Warren is not using this type of antenna for radiation efficiency. That's definitely not one of its strong points. Braun was seeking ways to achieve short to medium range tactical communication over land for the German army, without the need to either extend (and maintain) fragile telegraph lines across no-man's-land, or to waste resources erecting substantial antennas that would immediately become bombing targets. I believe the greatest range Warren achieved before the SWR problem was 40 miles with 45 watts...many times the output of a Part 15 beacon.