After squeezing my neurones for many days I have found a reasonable (albeit still hypothetical) timeline for Eurasian mtDNA based on SNP count. I have used a very simplified tree/sequence, that focuses in West Eurasian lineages, as these are the easiest ones to match with the known archaeo-cultural record, at least for me.
I used the mtDNA tree at PhyloTree.org as reference because it seems best kept and updated.
Obviously this sequence is not a genealogical tree but a mere lineal count of some representative nodes (epysodes) in the Eurasian genetic tree. The different lineages have been blended in a single "temporal" line for working purposes only. Each ">" symbol represents 1SNP in relation to the previous node and to the general L3 node (that includes all Eurasian plus some African mtDNA).
Of course, my assumption is that each SNP should represent an approximately similar period of time in the evolution of the Eurasian mtDNA.
"Present" is considered here by taking some well studied K subclade. I decided to use it after preliminary attempts made modern K sublineages to be well "in the future". At this stage it has not be contrasted with other lineages, what may produce some relevant fine tuning later on.
I tried hard (I swear) to make the H "explosion" fit with Gravettian but that only pushed R0 and U too late into the UP, and made the M,N,R "supernode" to be also incredibly too recent. Also contemporary U6 and U8 are way too likely to be of Aurignacian (Dabban in the case of U6) times. So the H/U6/U8 is here fixed to 41,000 BP, being the other reference point.
Likely contemporary nodes not directly considered are listed in [brackets] for reference. Not all other major lineages/nodes are coincident with the ones considered here; the ones listed are just for a quick reference of the wider Eurasian picture.
Notice that D and G are listed among the oldest East Asian nodes to branch and they are also the ones considered oldest in Central Asia (derived subclades, not considered), what suggests that the oldest colonization of this region had an Oriental origin.
Overall the impression is that:
1. The "fast coastal migration" scenario, via Bab el-Mandeb, seems ratified, as the dates could not fit with an older Eurasian coalescence. Would I try to force that, the H "explosion", for instance, would have to be much older than any European AMH presence, what makes absolutely no sense. This seems to contradict my previous estimates for Y-DNA, so I will have to review them.
2. When the R "explosion" happened, there were already N-lineage people in East Asia (N9) and Australia (S), and most M sublineages (16/26) had already began branching out in their respective destinations (exceptions include the only West Eurasian M sublineages: M1 and M48). This probably means that the expansion of R implies a second distinct pulse of the Eurasian expansion of AMH, I guess that associated with that of Y-DNA K.
3. AMHs arrived to West Asia and expanded there c. 48-45 kya. The later date may imply also the colonization of parts of the Balcans (Bakho-Kiro).
4. The main expansion of West Asians into Europe and North Africa happened c. 41 kya (this is presumed but should pose no problem as we know that well enough on archaeological grounds). H, U6 and U8 formed and began branching out ("exploding" in the case of H) at that time.
5. The timeline of U5 approaches that of Gravettian (original at c. 28 kya) but not well enough. It must be noted anyhow that the western expansion of Gravettian is much more recent, closer to the Solutrean origins, what gives some margin for speculation. Eastern Europe was basically colonized in this period, what explains well why U5 is more common towards the East, while H is towards the West of Europe.
6. K appears to have a Solutreo-Magdalenian time frame, what would mean that U8 as a whole, also U8b'K, probably coalesced in Europe and not in West Asia. At first I suspected that K would end up being Neolithic but, after some reality check (mutations to present), it seems not. West Asian U8b (also found in Italy) and K is therefore product of a back-migration therefore. I am at the moment uncertain but suspecting that T2 might have migrated to Europe also in this prehistorical context.
And for those who worship the TRMCA (i.e. Tandem Repeat Molecular Clock [Hypothesis] Age [Estimate]) model, I suggest them to re-read my favorite Grimm brothers' tale: The Emperor's New Clothes, which is very illustrative of what happens when people trust what they can't see, and when blind acceptance of authority on the fear of being called a fool goes mainstream, even if it's an obvious nonsense. In any case, as this estimate of mine clearly conradicts many of these elaborate guesses (even if it fits surprisingly well with known Prehistory), I'll be glad to discuss their concerns in the comments section, like always. .