Topband: elevated radials dilemma

[Grant Saviers]

My top loaded "T" top band vertical is up.  85' vertical, 43' each side 
"T" and 3 elevated 123' radials at 10 to 12'.  Resonance Z is currently 
36 ohms and 90KHz 2:1 swr BW.    EZNEC+ shows a resonance Z of 25 ohms 
with 4 elevated radials, so I presume the delta is ground loss.  The 
antenna is tree supported in a 40% grass/60% forest (hemlock. cedar, 
fir) area with a very wet, some standing water ground surface.  Buried 
radials are not feasible.

I can/will add 4 more elevated 123' radials for 7 total, but am debating 
what to do about a 4000 sq ft all steel 13' tall nearly flat roof 
building within 1 quadrant of the radial field, about 60' from the 
antenna base.  "Don't ground elevated radials" makes sense to keep 
antenna return currents out of the earth.  However, when a large area of 
"perfect" grounded conductor is available, why not connect to it? (the 
building has steel side walls and a Ufer perimeter footing ground).

Then a secondary question is "how many elevated wires to connect to the 
building?"  Three makes sense to me, given the angular spacing of the wires.

Can any modeling software handle such a configuration?

Whatever is up gets a try out on the CQ 160m SSB this weekend.

Grant KZ1W