| The global net uptake of carbon (C) by the terrestrial biosphere averaged 1.4 Gt C y-1 during the 1990s. Although temperate forests have been identified as significant contributors to this uptake, the underlying biological mechanisms through which this uptake has occurred have remained a matter of debate, largely because many of the spatiotemporal scales on which these mechanisms operate are not yet accessible to direct, continuous measurement. To fill this gap in our understanding, in this thesis I develop a new terrestrial biosphere model, version 2 of the Ecosystem Demography model (ED2).; I used ED2 to simulate carbon dioxide (CO2) fluxes, water fluxes, and tree growth and mortality rates at Harvard Forest. These simulation results were compared to eddy covariance data and biometry data, and model parameters were adjusted until the model was in accord with the data. The optimized model gave greatly improved predictions of net ecosystem productivity and of growth and mortality rates at both Harvard Forest (hardwood-dominated) and at Howland Forest (conifer-dominated).; I also used the model to predict regional vegetation dynamics and C fluxes. The simulated region encompassed the northeastern U.S. and southern Quebec (from 41°N to 52°N and from 80°W to 65°W). The modeled regional growth and mortality rates compared favorably to the observed rates.; I then performed a factor analysis to determine mechanisms driving interannual variability in the regional C fluxes. I found that meteorological variability had a strong, non-linear effect, suppressing both photosynthesis and respiration. The interannual variability in the total land-atmosphere C flux was primarily controlled by interannual variability in heterotrophic respiration. In particular, the years in which the land most strongly acted as a C sink were the years in which dry summers most strongly suppressed heterotrophic respiration.; Finally, I investigated the atmospheric signatures of CO2 fluxes using ED2 coupled to a new, mass-conserving version of the Regional Atmospheric Modeling System developed here. The coupled model successfully predicted altitudinal gradients of CO2 as observed by aircraft under a variety of meteorological conditions, but it underpredicted the gradient on several sunny, warm days. |
