Modelling the GHG emission from hydroelectric reservoirs

Abstract

A mechanistic model has been constructed to compute the fluxes of CO₂ and CH₄ emitted from the surface of hydroelectric reservoirs. The structure of the model has been designed to be adaptable to hydroelectric reservoirs of different sizes and configurations and the reservoir can be partitioned into one, two or three vertical volumetric zones. Each zone may accommodate a number of influents and effluents including turbined flow and discharged flow. Each zone consists of a surface water layer (0–10 m) and a bottom water layer (>10 m). The model considers advective and diffusive mass transfers of dissolved CO₂ and CH₄ between zones and water layers, the rates of CO₂ and CH₄ produced from the decomposition of flooded vegetation and soil in the reservoir, and, mass transfer of CO₂ and CH₄ at the water-air interface. Global mass balance equations are solved to compute the magnitude of the advective flows between zones and water layers. Component mass balance equations are solved to compute the concentrations of CO₂ and CH₄ as a function of time in the surface and bottom water layers of each of the zones of the reservoir. The rates of CO₂ and CH₄ emitted from the surface water layer are computed using the two-film theory. Data from the Robert-Bourassa reservoir, a large operational hydroelectric reservoir, has been used as input data to the model. Results from the model were first compared with experimental data available for the calculation of dissolved CO₂ concentration in the surface water layer. Secondly, results from the model were compared with fluxes of CO₂ and CH₄ emitted from that reservoir as calculated from the experimental determination of dissolved CO₂ in water. Also, they were compared with direct measurements of the fluxes at the water-air interface. It has been observed that concentrations of CO₂ computed by the model are in the range of values reported for the surface water layer. No data was available for comparison with concentration of CH₄. Emissions of CO₂ computed by the model were in the range of fluxes calculated from the experimental determination of dissolved CO₂ in water. The computed flux as a function of reservoir age was also coherent with the CO₂ flux measurements data. The transitional emissions of CO₂ resulting from the decomposition of flooded vegetation and soil were found to be significant during not more than 6 to 8 years depending of the volumetric zones of the reservoir considered. Simulations were done under two distinct scenarios for the CO₂ content of the influents to the reservoir. The first scenario used data which reflected the contribution of carbon originating from the drainage basin. The second scenario assumed the CO₂ concentration in the influent water to be at equilibrium with the atmospheric CO₂. From the simulation results and the data available an important finding is that the main source of carbon contributing to the GHG emission from the hydroelectric reservoir after the transitional emissions of CO₂ due to the decomposition of the flooded vegetation and soil have faded away appears to be essentially the carbon originating from the drainage basin.

The results have also indicated that fluxes of CH₄ computed from the model are grossly underestimating the values reported from the direct measurements of CH₄ emissions. Analysis of the results have indicated the source of the discrepancies which lies with the very low production of CH₄ as indicated from the vegetation and soil decomposition data used by the model. Suggestions to improve the model forecasting of CH₄ emissions are indicated.