Efficacy of climate models for projecting future phenology changes

Background/Question/Methods Output from model simulations of future climate provide a valuable resource for prognostic ecological modeling, but to date they have been little used for phenology simulations. We test the ability of these models to simulate 20th century phenology, as well as future changes, using a long term phenological dataset of plant flowering at Mohonk Lake, NY. For each species, a model for onset of flowering is calibrated and validated on in situ daily weather data, using a growing degree day approach. We then use output from the IPCC archive for 20th century climate simulations, to test the ability of various climate models to simulate onset of flowering using an analogue climate approach. Finally, for those models that reliably reproduce the variability in flowering onset for the 20th century, we use the simulations of 21st century climate to project into the future. Results/Conclusions This study will provide insight into the potential for joining climate and species level phenological modeling, and discuss the uncertainties, flaws, and strengths inherent in this approach.