unfortunately i only managed to find one thing (tangentially related) that i favorited at the time - this is more about "how, if at all, we can associate ancient descriptions of groups of peoples with modern ethnic groups", but i think you would probably find that interesting too, and it has a LOT of keyterms in it that would help the search for more information. for once in my life it seems like i didn't save enough links, instead of far too many
i originally got into the environmental-change thing through a "hey why did north africa become so much less important over 1000 years or so" search so the things i learned are related to that
basically to sum up what i can remember, the steady march towards desertification was due to a combination of the following:
roman agricultural and forestry practices sucking over the longterm (which was a thing elsewhere too)
groundwater being used up by said practices - when it ran out, the garamantes went bye-bye for example among other groups
actually i lied in the process of writing this post i found this, which again isn't as related as i would like but it goes over similar territory
also when a shitton of pastoralists moved in during the arab expansion plantlife got fucked over by grazing animals, as well as old infrastructure getting rekt
as for proto-AA, i think the more accepted linguistic theory is an urheimat around ethiopia? but frankly it's modeled to be spoken a long-ass time ago and it's not as well studied as proto-indo-european, which still has really big theoretical disputes (in who spoke it, where it was spoken, and what exactly it looked like). i don't know how much you know about historical linguistics but feel free to ask if you wanna get a crash course
it's quite late at night so forgive me if i'm less coherent than usual (which i admit isn't a comfortably high ceiling in the first place)