You don't seem to understand: there's no indication that the population that coalesced with the hypothetical BE component came from the north.
So Fu et al were wrong when they present a model with 32% Basal en 68% Malta? Malta being the prime example of a population called Ancient
North-Eurasians? I mean, you posted that graph and I don't see it modeled as admixed with anything else but Basal (split of before Ust Ishim) and something from which came from the same source as Malta did.
Now you may have the opinion that this is not the case and maybe that's right. But this graph is not your piece of evidence for it. May you should provide links which make the case for it.
Fair enough, that constitutes the Balkans and Anatolia
That's just ridiculous. A bottleneck tells us nothing about the gene flow that occurred before and after this particular event. If a very mixed population persists during a cataclysmic event and while other populations are culled, that's still a bottleneck. That's particularly true when most models that yield LGM & Late Glacial bottlenecks are primarily based on the study of mtDNA.
From Lararidis 2014: "
Through comparison of Loschbour’s two chromosomes we find that this low diversity is not due to recent inbreeding but instead due to a population bottleneck in this individual’s more distant ancestors (Extended Data Fig. 2)."
EDIT: I just noticed that we've veered off of your original hypothesis of Aurignacian genetic structure, so let's just cut it here.
But this is exactly why I came up with that idea. Migration from M.E. is virtually impossible because of the Basal issue. So we have two possibilities left:
1) S.E. Europe. However, a number of odd things make it less likely candidate to host the
original admixture which was partly ancestral to WHG. First, we have a number of adjacent samples from 30.000 + something years ago, from Romania. They show no affinity to WHG (see Fu et al). We know that during the Mesolithic Greek HG's were of a different mtDNA: K1b.
2) Italy and South-East France. Both El Miron and Paglicci71 are mtDNA U5b, which is said to be a tracer for WHG/Villabruna cluster (Krause) and are situated at the end of LGM. El Miron has part WHG admixture and has the affinities that WHG exposes at well.
Now what culture could that have been? Perhaps the only one not sampled in Fu et al, the Solutreans. But they can't have been Gravettians since that would mean WHG would nicely and sweetly model as Gravettians. Hence something pre-Gravettian must have survived from before. And that makes it Aurignacian.
Now, completely independent from that we read that Aurignacian toolkits from the same area that the Solutrean came look like an intrusive population in the Levant some 33.000 years ago.
Now the latter is just an idea I have. But the case against M.E admixture is simple: No Basal.