Summary: | Carbon (C) pools and fluxes in northern hemisphere forest ecosystems are attracting increasing attention concerning predicted climate change. This thesis studied C fluxes, particularly soil C dynamics, in spruce forest ecosystems in relation to interactions between physical/biological processes using a process-based ecosystem model (CoupModel) with data for Swedish conditions. The model successfully described general patterns of C and N dynamics in managed spruce forest ecosystems with both tree and field layers. Using regional soil and plant data, the change in current soil C pools was -3 g C m-2 yr-1 in northern Sweden and +24 g C m-2 yr-1 in southern Sweden. Simulated climate change scenarios resulted in increased inflows of 16-38 g C m-2 yr-1 to forest ecosystems throughout Sweden, with the highest increase in the south and the lowest in the north. Along a north-south transect, this increased C sequestration mainly related to increased tree growth, as there were only minor decreases in soil C pools. Measurements at one northern site during 2001-2002 indicated large soil C losses (-96 g C m-2 yr-1), which the model successfully described. However, the discrepancy between these large losses and substantially smaller losses obtained in regional simulations was not explained. A simulation based on Bayesian calibration successfully reproduced measured C, water and energy fluxes, with estimated uncertainties for major components of the simulated C budget. Site-specific measurements indicated a large contribution from field layer fine roots to total litter production, particularly in northern Sweden. Mean annual tree litter production was 66% higher at the most southerly site (240 g C m-2 yr-1 compared with 145 g C m-2 yr-1 in the north), but when field and bottom layers were included the difference decreased to 16% (total litter production 276 g C m-2 yr-1 and 239 g C m-2 yr-1 respectively). Regional simulations showed that decomposition rate for the stable soil C fraction was three times higher in northern ...
|