Is it permissible on FM to use a directional antenna so the signal exceeds 250 µV/m in some azimuths and elevations (and is attenuated in others), so long as the 3D/spherical RMS/average value does not exceed 250 µV/m? For example could you have 2,000 µV/m peak on a 30° beam, average 10 µV/m on the "back" 300° (leaving 30° for the transition from lobe to null), 40-60 dB attenuation at elevations greater than 15° (which pattern I'm guessing would still have an RMS below 250 µV/m) and be in compliance? And if so, what about using multiple synchronized co-channel transmitters with directional antennas, each aimed in a different direction?
Or, if I wanted to cover my entire ~ 90 by 250 foot (1/2 acre) yard (conductivity "8" according to the M3 map but is probably actually a lot less although I don't have a way of measuring it) with a signal that indicates at least 89 to 94 dBµ on a Tecsun DSP PL-xxx using only the built-in antenna (or at least 72-88 dBµ on the 2nd harmonic even though that'd severely overload the fundamental), would I be better off using LW (under 15.217), MW (under 15.219) or SW (under 15.225)?