I like the idea of experimentally building the elevated radial system described by MRAM, and I'm going to do a feasibility study for both medium wave and 13.568 shortwave.
From a slight distance away it would be hardly noticable.
MRAM'S COMMENT: My point was that here is a simple elevated radial system for the AM band using only six radials which apparently works as it satisfies FCC performance standards.
MY COMMENT: It ONLY satisfies FCC for licensed broadcast. Not Part 15.
Licensed AM broadcast stations have no limitations on the ground system configuration they use. It can be whatever configuration they need/want to use, as long as the radiation efficiency of their antenna system meets the minimum value for their class of license. In some cases, 4 or 6 elevated radials are used because the earth around the monopole is too rocky, marshy etc to use buried radials. The elevated radials can give the same system performance as if the station was using 120 x 1/4-wave buried radials.
But elevated radials might be considered by an FCC inspector as necessary to include in the 3-meter radiating length allowed for Part 15 AM under §15.219(b) -- whether or not those elevated radials produce any useful, far-field radiation.
An NOUO issued for such might be dismissed based on successful technical arguments and proof, but probably not without some effort and expense.
For indoor antenna placement, I have experimented with baseboard radials, and they do add boost to the resulting signal propagation, but would be visible to an inspector and perhaps also at risk of being counted as "length" of total wire.
But the next example is another question, namely dropping a very short ground lead down the baseboard into the basement and clamped to an I-beam, part of the house ground to both the electrical ground rod and water pipe. Is "below the floor" the same as "below the ground?"
Mram, the elevated radial broadcast antenna system you referenced at http://www.mwpersons.com/nott-elevated-ground.html is indeed simple in concept and very likely performs as stated.
The killer issue is that the 6 radial wires must be a specific length to tune them to a specific operating frequency. That length is close to 1/4 wavelength. They state the length at 1700 kHz as being 150 ft. I suspect you can't just plug the frequency into a standard formula to get the length due to the significant reactive near field effects of the radials being only 18 ft high. The actual tuned length could be found by trial and error, or the Nott company can use their own formulas developed for that specific antenna. Note that it says "FREE system design with purchase of an elevated radial system".
To duplicate this particular antenna, a Part 15er would need to install 6 specific-length radial wires each 18 ft. high. The radials would need to be stretched between 18 ft poles, not draped across a roof or over trees.
The key distinction between ground level radials and elevated radials is that ground-level radials have zero or near-zero reactance while elevated radials have a widely varying reactance depending on their length, height and operating frequency. The reactance would go to zero at some point about 1/4 wavelength long. Inductors in series with each radial could allow shorter radials, but it isn't clear if they would be as effective, and adjusting the inductances for resonance would be a bear. And then, after all that, consider what would need to be done to change to a different frequency.
I suppose a separate thread devoted to "elevated radials" would be worthwhile, but look out, it would be a lightning rod for controversy. There is already a lot of stuff on the web about elevated ground planes, with the typical mix of opinions and valid experimental results. Much of the good information applies to HF and VHF where dimensions are much smaller.
Carl, you are the only one that can answer your own question, unless somewhere out there you can find a Doctoral candidate that would be willing to develop a supercomputer model of every last detail of your house. By detail I mean every wire including its length, diameter and exact routing, every appliance, every metallic doodad or decoration, conductivity and dielectric constants of all the building materials, etc, etc. etc.
The effectiveness of any sort of indoor grounding scheme or radial system is totally unpredictable. Experimentation is you only recourse.