Simulation of the Seasonal and Interannual Variability of Carbon and Water Cycles Within Three Mid-latitude Forests Using a Dynamic Global Vegetation Model

Currently, a considerable challenge confronts the scientific community to acquire an improved understanding of the bi-directional exchange of carbon, water, and energy within the coupled biosphere-atmosphere system at the local scale using measurement strategies, with the goal of scaling these results to larger regions using modeling tools. The exchange of heat, moisture and trace gases that occur within the planetary boundary layer are dynamic and responsive to each other, occurring at a continuum of timescales ranging from seconds (e.g., plant physiology) to thousands of years (e.g., soil biogeochemistry). To address the need for more robust models of the biosphere that can capture processes which function across a continuum of timescales, a new class of dynamic biosphere models, termed Dynamic Global Vegetation Models (DGVMs), have been in continuous development and testing during the last decade. This group of models integrates biogeography, vegetation dynamics, soil biogeochemistry, and SVAT components into the same integrated framework, allowing for the continual update of vegetation and soil characteristics in response to plant competition, soil nutrients, and atmospheric forcing. The key objective of this study was to evaluate the ability of the Integrated BIosphere Simulator (or IBIS, a DGVM) to simulate seasonal and interannual surface-atmosphere exchanges of carbon and water within three mid-latitude AmeriFlux forest study sites (Walker Branch, Harvard Forest, and Niwot Ridge). We allowed for simulated vegetation structure to interact dynamically with changing atmospheric conditions in one set of simulations, while in another simulation, vegetation characteristics were fixed within the model. We used multiyear, half-hourly measurements of surface-atmosphere exchange of coupled carbon-water fluxes and observations of vegetation phenology (leaf area index), biomass, canopy height, soil carbon density, and surface runoff at these study sites to assess model realism. Specific questions we chose to address included: (1) Does the model satisfactorily simulate forest stand characteristics and associated seasonal and interannual carbon and water fluxes?; (2) How applicable are global scale model parameterizations at a local, field-site scale when simulations of forest stand characteristics and quantification of seasonal and interannual carbon and water fluxes are desired? Initial results indicated that under fixed vegetation conditions, the model satisfactorily captured the measured exchange of carbon and water between the surface and the atmosphere, suggesting that the SVAT component of the model was robust. Secondly, when simulated vegetation structure (leaf area, biomass, and canopy height) was allowed to interact dynamically with changing atmospheric conditions, the model reproduced vegetation height and total plant biomass within 20% for all three forests studied, while some improvement was noted in the simulated seasonality of carbon and water fluxes for broadleaved deciduous forests. We concluded that the use of a phenological modeling scheme that incorporates local climate characteristics is essential to capturing the seasonal and interannual variability in coupled carbon-water fluxes.