Non-locality in many-worlds when spacelike-separated observers decohere

[Derek P]

The reason the wavefunction must be ontic (which that I mean maps 1-to-1 onto reality) is given by the PBR and Colbeck and Renner theorems.

Indeed so but that's a weaker sense of ontic than I was using. I was just meaning that there is no reason why reality must actually be the wavefunction. A 1-1 map would be fine.