It's mostly that the inbound energy doesn't even reach the reflector. An optical analog (yes, yes, EM true) would be trying to image something well behind a mir

It's mostly that the inbound energy doesn't even reach the reflector. An optical analog (yes, yes, EM true) would be trying to image something well behind a mirror. "Upward continuation": both gravity and (after certain manipulations) magnetostatic surveys obey Poisson's equation. The characteristic (Green's) function has a 1/R decay. If you will allow me another optical analog, upward continuation is closely analogous to a detailed image being convolved with a 1/R blurring function. That's the heart of why I (roughly) say you can only resolve stuff of characteristic length R by observing/measuring from below a height of R. Higher than R, and the short wavelengths are completely attenuated, and you lose ability to resolve them.