Reservoir water level drawdown as a novel, substantial, and manageable control on methane release to the atmosphere

Abstract:

Reservoirs constitute a globally important source of atmospheric methane (CH₄). Although it is reasonably well-established that hydrostatic and barometric pressure can influence rates of CH₄ release from lake and tidal sediments, the relationship between water-level manipulation and CH₄ release from man-made impoundments has not been quantified or characterized. Furthermore, cross-system controls on CH₄ production and release to the atmosphere have not been established. We collected CH₄ emission (diffusion and ebullition) data for 8 reservoirs in the U.S. Pacific Northwest that are subject to a range of trophic conditions and water level management regimes. Our aim was to: (1) characterize CH₄ emissions from these systems, and (2) quantify effects of water level management and eutrophication on CH₄ fluxes. Results indicate very high fluxes, in some cases the highest reported reservoir emission rates, and a strong correspondence between lake level reduction and CH₄ emissions, including quantitatively important bursts of CH₄ bubbling. In one reservoir, drawdown-associated CH₄ fluxes accounted for over 25% of annual CH₄ emissions in a period of just 16 days (4% of the year). Average CH₄ ebullition rates in a reservoir managed for hydropower peaking were nearly three-fold higher than in a paired upstream reservoir managed to maintain a constant water level (528 mg CH₄ m^-2 d^-1 and 187 mg CH₄ m^-2 d^-1 respectively). Highest gas fluxes were observed during the water level drawdown component of the hydropower peaking cycle (14.3 g CH₄ m^-2 d^-1). In addition we observe a strong, positive relationship between eutrophication (as indicated by surface Chl a concentrations) and CH₄ production (r² = 0.88; P<0.001) and between eutrophication and the sensitivity of CH₄ emissions to drawdown (r² = 0.84; P<0.001). This work suggests that manipulation of water levels can significantly affect CH₄ emissions from reservoirs to the atmosphere, and that sampling programs that miss drawdown periods may substantially underestimate CH₄ fluxes. It also suggests that controlling nutrient loading may reduce greenhouse gas fluxes from surface waters to the atmosphere.