B11C-0448
Assimilating Paleoecological Data into a Forest Gap Model

Monday, 14 December 2015
Poster Hall (Moscone South)
Ann Raiho and Jason S McLachlan, University of Notre Dame, Notre Dame, IN, United States
Abstract:
Climate change is causing unprecedented shifts in forest composition, distribution, and function. Models are capable of capturing ecosystem consequences of these changes on an observable time scale, a century. However, forest processes operate at longer time scales. To better predict how they will respond to climate change, scientists need to understand whether they are simple extrapolations of short-term processes or function in a differently at longer time scales. The PalEON (Paleo-Ecological Observatory Network) project has developed multiple independent paleo-ecological climate and forest composition reconstructions, and millennium length ecosystem model runs to investigate long-term forest processes. If models capture the realities of long-term processes, the model's relationship between climate drivers and vegetation response should match the relationships we see in the data. We used generalized additive models to explore the relationship of independently reconstructed water table depth and gridded forest composition data estimated from networks of fossil pollen data. We calculated that compositional changes are highly correlated with water table depth over centennial time scales. We then tested the relationship between moisture and composition in a 1000 year run of LINKAGES, a model built specifically for understanding long-term forest processes. We found that composition changes have a stronger relationship to centennial scale climate than annual scale climate. The data shows that composition change varies regionally, at a coarse temporal resolution. These changes are not apparent in the model at annual resolution. The modeled composition change is not well constrained in the model at coarser (centennial) resolution, so the data are able to constrain the relationship more than the model by itself.