Otter,
I do kind of question the impeller versus air myself. I'm not sure how pump loads are tested for this purpose, but one would think that when run in a circulatory capacity the load would be very low. You don't really have to "lift" anything you just have to move it. Pumps are typically thought of as moving water against a load or restriction and lifting in the atmosphere.
The key points would be plumbing size, complication, and having electricity in the water.
I wondered about taking (one for starters, possibly several) small 12v pumps (battery with solar collector) that move approx. 5GPM ea and run submersed to see what it would do. I think a guy would want the discharge into a fairly larger diameter pipe that extended to near the surface and possibly some sort of diffuser to spread the discharge.
I still don't think it would match the air pump, but it's intriguing.
