During the 21st century, global climate change is expected to become a significant force redefining global biospheric boundaries and vegetation dynamics. In the northern hardwood–boreal forest transition forests, it should, at the least, control reproductive success and failure among unmanaged mixed forest stands. One means by which to predict future responses by the mixed forests is to examine the way in which they have responded to climate changes in the past. We used proxy climate data derived from Holocene (past 10 000 years) pollen records in the western Upper Peninsula of Michigan to drive forest gap models, in an effort to define regional prehistoric vegetation dynamics on differing soils. The gap models mimic forest reproduction and growth as a successional process and, hence, are appropriate for defining long-term tree and stand dynamics. The modeled period included a mid-postglacial period that was warmer than today's climate. Model failures, made apparent from the exercise, were corrected and the simulations were repeated until the model behaved credibly. Then, the same gap model was used to simulate potential future vegetation dynamics, driven by projections of a future climate that was controlled by greenhouse gases. This provided us with the same "measure" of vegetation in the past, present, and future, generating a continuously comparable record of change and stability in forest composition and density. The resulting projections of vegetation response to climate change appear to be affected more by the rate than by the magnitude of climate change.
Supplemental Material: Central eastern China hydrological changes and ENSO-like variability over the past 1800 yr