PARENT SESSION
Organized Oral Session 34: Incorporating ecological processes at many scales into biogeochemical and global climate change models
Organizer(s): WM Post, JS Olson, and C Peng
Wednesday, August 10, 1:30 PM - 5:00 PM, Meeting Room 510b, Level 5, Palais des congrès de Montréal

Representing vegetation dynamics in a regional model of ecosystem biogeochemistry.

Smith, Benjamin^*,1, Hickler, Thomas¹, Koca, Deniz¹, Morales, Pablo¹, Sykes, Martin¹, ¹ Geobiosphere Science Centre, Lund, Sweden

ABSTRACT- LPJ-GUESS is a process-based ecosystem model optimised for regional studies. The model incorporates physiology-based formulations of plant growth and ecosystem biogeochemistry similar to dynamic global vegetation models (DGVMs). Population and community dynamics are modelled as an outcome of neighbourhood-scale interactions among individual trees and a herbaceous ground layer. Disturbances, silvicultural management and migration can further influence vegetation dynamics in the model. Compared with a similar model that represents vegetation dynamic processes in a more generalised way, LPJ-GUESS tends to be more successful at reproducing the observed vegetation structure, especially for vegetation in earlier stages of stand or successional development. The models likewise differ in their ability to reproduce seasonal patterns of carbon cycling by European forests, with the more detailed model generally yielding a better overall agreement between modelled and measured monthly net ecosystem carbon exchange (NEE). LPJ-GUESS was applied to assess potential changes in forest species composition and carbon balance in Sweden under future climate and atmospheric CO2 concentrations. The predicted vegetation changes include northward and upslope vegetation zone shifts, primarily driven by increased winter temperatures, but also an increased broadleaved component in the boreal forest reflecting changed competitive balance among tree species differing in phenology and life history. Taken together, the results presented generally support the view that predictions of how ecosystems will respond to environmental changes will be sensitive to the representation of vegetation and vegetation dynamic processes in the model applied. Many current global models, which incorporate highly simplified vegetation dynamics - or none at all - may be too simplified to adequately describe responses at the regional scale.

Key words: Vegetation dynamics, Succession, Ecosystem model, Biogeochemical cycling