Modeling and experimental analysis of carbon exchange from artificially flooded forest and peatland ecosystems

Development of hydroelectricity in recent years has stirred an international debate in relation to greenhouse gas (GHG) emissions caused by flooding, which results from the creation of hydroelectric reservoirs. The debate focuses on whether hydroelectric reservoirs are negligible global GHG sources, particularly with regards to carbon dioxide (CO2) and methane (CH4). Most carbon (C) exchange studies applied to hydroelectric reservoirs have been based on irregular or sporadic field measurements and, therefore, hardly address the transient nature of reservoir C flux and the heterogeneity in flux that occurs across different types of ecosystems inundated with water. In this context, ecosystem modeling and laboratory experiments can improve our understanding of C exchange that takes place in flooded terrestrial ecosystems as a consequence of hydroelectric development. The aim of this research was to examine C exchange variation in boreal forest and peatland ecosystems prior to and after flooding as well as to project boreal ecosystem C exchange for the duration of the inundation period. The primary study area was the Eastmain-1 reservoir located in northern Quebec where impoundment was completed in 2006. For this research, a reservoir C model (FF-DNDC) was developed by modifying Forest-DNDC, a process-based biogeochemical model utilized for forest and wetland ecosystems. FF-DNDC was designed to replicate C processes that take place in submerged soil and the water column. It is used to simulate CO2 flux in flooded boreal forest and peatland ecosystems. The reliability of the Forest-DNDC simulation in relation to CO2 flux in black spruce forest and peatland ecosystems was tested before modifications to the software took place. This test showed that Forest-DNDC reasonably simulated CO2 flux and, as a result, supported the application of the model to simulate C dynamic changes after flooding occurs. Short-term incubation experiments using boreal soil and vegetation samples revealed that flooding decreased rates of CO2 production but increased rates of dissolved C production. The experiments quantified changes that occurred in C mineralization rates prior to and after flooding, which determined soil decomposition parameters under flooded conditions that were then applied to FF-DNDC. The Eastmain-1 reservoir flooded ecosystem simulations detected CO2 emissions from the water surface, and, hence, the direction in CO2 flux changed (from uptake to release) in comparison to flux that occurred in natural forest and peatland ecosystems. Simulated CO2 flux for both the flooded forest and peatland ecosystems decreased with the duration of inundation, and the forest ecosystem showed larger CO2 flux than the peatland ecosystem in the first decade after flooding was initiated. The trend of larger flux in the forest ecosystem was reversed after the first decade. Modeling and experimental results from this study emphasize the importance of spatial and temporal variation of C exchange in newly flooded boreal landscapes.