Oona (the author) knew about the existence of such things; it wasn't a mystery. The only reason it was strange is that because her radio had been knocked about during a move, it was strangely down-shifting the data channel ... she was not monitoring the FM (from what I could make out), but the actual audio from the radio.
This was just ('just' for her, impressive for me) an exercise of going down the rabbit hole, and then curating that tour for us.
Rather, I don't think the data channel was actually down-shifted, but that it wasn't filtered out (probably due to a broken analog filtering circuit after the fall), so it was still outside of the audible spectrum but available to the sound card and hence visible in her software spectrum analyzer. It also sounded like she was seeing some aliasing effect, i.e. that the data signal was probably out of band for the sound card ADC, but "folded" down into the sampler frequency range due to Nyquist folding "magic" (which is maybe what you meant with "down-shifting", now that I think about it?).
This was just ('just' for her, impressive for me) an exercise of going down the rabbit hole, and then curating that tour for us.