"Roy Lewallen" wrote
2. Making elevated radials too long, even if the elevation consists of
being just above the ground, can seriously reduce the antenna efficiency.
Buried radials, on the other hand, are insensitive to length provided
they're sufficiently long. This latter fact is well known. I've found in
other modeling I've done that making elevated radials shorter than a
resonant quarter wavelength doesn't negatively impact the efficiency. So
if you have to guess, guess on the short side.
_____________

NEC studies of a 1/4-wave vertical radiator working against three 1/4-wave
horizontal radials at 120 degrees, when all elements are elevated 12 feet
above a perfect ground plane show virtually identical peak gain as when the
same radiator minus the radials is mounted with its base at the perfect
ground plane, and connects to it though two ohms (about the same ground loss
as produced by 120 buried radials, each 1/4-wave long).

As few as four elevated radials have been used at AM broadcast sites where a
typical system of 120 buried radials was impossible due to rocky terrain.
The FCC "efficiency" of these radiator systems meets/exceed FCC
requirements for radiation at 1 km.

Conclusion: a few elevated radials can be the electrical equivalent of a
classic "Brown, Lewis & Epstein" system of 113 (or 120)buried radials.

RF

"Richard Fry"
NEC studies of a 1/4-wave vertical radiator working against three 1/4-wave
horizontal radials at 120 degrees, when all elements are elevated 12 feet
above a perfect ground plane... etc
____________

I neglected to include that the frequency in these studies was 1 MHz, so
elevation of the system was about 0.012 wavelengths (Reg).

RF