Redistribution of methane emission hot spots under drawdown conditions

Sci Total Environ. 2019 Jan 1;646:958-971. doi: 10.1016/j.scitotenv.2018.07.338. Epub 2018 Jul 26.

Abstract

In the context of reservoirs, sediment trapping, and aquatic greenhouse gas (GHG) production, knowledge about the distribution of hot and low spots is essential for improved measurement strategies. It is also a key to a precise assessment of the GHG emissions of each reservoir. Large numbers of reservoirs are used mainly for hydroelectric power generation and, hence, affected by strong changes in water level. Drawdown events may lead to significant changes in spatial sediment and organic carbon distribution and, consequently, strongly alter the GHG emission patterns of the water body. We combined hydroacoustic sediment classification, sediment magnitude detection, and ebullition flux assessment with in-situ pore water investigations and sediment coring to detect ebullition distribution patterns after strong reservoir drawdown. The research was conducted in the Capivari Reservoir in the southeast of Brazil, which was affected by up to 15 m of drawdown within the last 10 years. Results show severe changes in sediment accumulation and composition. The focusing of sediment divides the reservoir in extreme hot and low spots. Methane pore water concentrations are highly correlated with acoustic backscatter values (r2 = 0.97) as well as with the organic carbon content (r2 = 0.55) and allow for a precise detection of the newly created emission patterns. Highly productive sediment could be acoustically distinguished from non-productive areas. Only 23.6% of the reservoir surface produced 64% of the detected bubbles. An organic carbon content in the sediment of 2.4% was found to be a prerequisite for the formation of GHG emission hot spots. These findings may help to complement the still insufficient knowledge of methane ebullition fluxes from reservoirs with changing water levels.

Keywords: Drawdown; Hot spot; Methane emission; Reservoir; Sediment.