Process-based ecosystem modeling to predict carbon dioxide fluxes in the newly flooded black spruce forest and peatland

We developed a process-based reservoir model (“flooded” version of Forest-DNDC) to project carbon fluxes from inundated black spruce forests and peatlands over the life-time of a hydroelectric reservoir located in the Boreal biome. The reservoir model was used to examine the changes of carbon dioxide (CO2) fluxes during the first four years after inundation and to evaluate the effects of the impoundments on CO2 fluxes from the Eastmain-1 reservoir in northern Quebec, Canada. The framework for the reservoir model was Forest-DNDC, a process-based terrestrial biogeochemistry model, which supports detailed soil carbon processes from considering redox chemistry and oxygen diffusion in flooded ecosystems. We modified this terrestrial model to represent the alteration of soil and vegetation carbon processes when they are located under a water column: soil decomposition parameters were adjusted for difference rates and temperatures due to submergence and the addition of new carbon via sedimentation. Using the measured environmental conditions from 2006 to 2009, modeled daily CO2 emissions from the flooded forest averaged 0.43 g C m-2 d-1 (ranging from 0.60 to 1.07 g C m-2 d-1), and those from the flooded peatland averaged 0.49 g C m-2 d-1 (ranging from 0.63 to 0.86 g C m-2 d-1). The simulated CO2 emissions decrease with the duration of flooded condition. Our simulations resulted in smaller values than those in CO2 flux measurements by the eddy-covariance system at the surface of the reservoir, but the changing pattern over time were similar. The disagreements would stem from the model structure and measurement method: the developed model certainly lacks some processes occurring in the open water portion of the reservoir, and the measured fluxes are a function of the actual turbulent transfer and are therefore somewhat removed in time and space from the actual fluxes of CO2.