Modeling the refugia of biodiversity under a changing climate is a big challenge, and needs to be accomplished in order to design effective conservation measures for threatened species.
Refugia have been under discussion mostly in the lowlands since the eighties, but I do not know somebody modelling refugia. This also because of the density of sample points (for pollen diagrams) is not suffiently dense enough.
I guess this also is true for tropical Africa and Asia.