The FCC's propagation models for the FM broadcast band seem grotesquely inaccurate in some places, where stations can be received that the models say are absolutely out of range. I'm not even talking about DX conditions like ducting or skip, I'm talking about baseline. When I lived in Fayetteville, AR for instance, I could often listen to 99.5 FM from Tulsa, OK over the local station on the same frequency, even though the FCC data said this should be impossible by 3-4 orders of magnitude in signal strength! Clarly there is a lack of field observations to test the limits of these models/approximations. They are presumably only tested under a narrow range of conditions, not on non-ideal, "real world" scenarios with buildings, hills, city, suburbs, persistent atmospheric temperature inversions, etc. Why has there not been a bigger push for better observations?