The role of landscape-scale edaphic variation in forecasting regional-scale forest ecosystem dynamics

Landscape-scale variation in edaphic conditions has long been known to have a strong influence on forest composition and productivity. Furthermore, the direct responses of ecosystems to climate change, as well as their interactions with disturbances, are expected to vary strongly depending upon edaphic factors. However, models of global and regional vegetation have difficulty capturing this variation because they average over the heterogeneity of large regions (e.g. 1X1 degree cells). Capturing landscape-scale variability in these models is imperitive both because of the inherent nonlinearities in the aggregate response of heterogenous ecosystems, as opposed to the response of a single “average” ecosystem, as well as the need to understand compositional shifts among landscape positions (e.g. upslope migration). We report on changes to the Ecosystem Demography (ED) model aimed at capturing sub-grid scale variation in soils, topography, and hydrology in a spatially implicit manner. The refined model is then validated against long-term data on forest composition, ecosystems processes, and hydrology for the Hubbard Brook and Coweeta LTERs. The refined model captures much of the spatial heterogeneity in forest composition and productivity and the temporal patterns of watershed discharge and soil moisture. The model is then scaled up to eastern North America and compared to regional-scale patterns of composition and productivity in the USFS Forest Inventory and Analysis (FIA). Finally, preliminary results suggest that edaphic variability helps stabalize ecosystem responses to interannual variability.