 |
Center for Nonlinear Studies |
The magnitude of the feedback between the terrestrial biosphere and the atmosphere represents a large uncertainty in climate predictions. Recent model inter-comparisons indicate that choice of Dynamic Global Vegetation Model (DGVM) alone can change the size of the climate-carbon cycle feedback such that the 2100 atmospheric CO2 concentration varies by as much as 227ppm. Existing DGVMs are skilled at simulating the contemporary carbon cycle, but differ significantly in their responses future climate and CO2 concentrations. To move on from this situation, we must intensify our efforts to interface existing ecological data and experiments with model construction and parameterisation. However this is not straightforward using existing DGVMs, as these typically require many parameters which are not easily linked to observations. I will discuss the development of a new DGVM based on the 'Ecosystem Demography' (ED) approach. Using the individual tree based ED allows for the first time, in a global context, the simulation of vertical competition for light between plant types, heterogeneity of the land surface resulting from disturbance and ultimately the mechanistic representation of timescales determining the velocity of biome shifts. This new 'second generation' DGVM, which is coupled to the land surface exchange scheme of the Hadley Centre family of climate models, is parameterised using observational data and as such represents a major shift in capacity for integration of ecological understanding into climate-relevant vegetation modeling.
Host: Nate McDowell (EES) and Todd Ringler (T)