Jessica Blois (former geography postdoc) and Prof. Jack Williams have published an article in the Proceedings of the National Academy of Sciences (PNAS) on their research. They use global climate models and fossil pollen data from the last deglaciation to test a fundamental assumption in biodiversity modeling: that the spatial relationships between biodiversity and climate relationships can be used to predict climate-driven changes in biodiversity over time. Good news: it can, at least for the recent past, with space-for-time models achieving 72% predictive skill.
Read more about their research in this UW news story.
Or see the PNAS article abstract and download a PDF of the full text here.