High methane ebullition throughout one year in a regulated central European stream

Abstract

Ebullition transports large amounts of the potent greenhouse gas methane (CH ${}_4$ ) from aquatic sediments to the atmosphere. River beds are a main source of biogenic CH ${}_4$ , but emission estimates and the relative contribution of ebullition as a transport pathway are poorly constrained. This study meets a need for more direct measurements with a whole-year data set on CH ${}_4$ ebullition from a small stream in southern Germany. Four gas traps were installed in a cross section in a river bend, representing different bed substrates between undercut and slip-off slope. For a comparison, diffusive fluxes were estimated from concentration gradients in the sediment and from measurements of dissolved CH ${}_4$ in the surface water. The data revealed highest activity with gas fluxes above 1000 ml m ${}^{-2}$ d ${}^{-1}$ in the center of the stream, sustained ebullition during winter, and a larger contribution of ebullitive compared to diffusive CH ${}_4$ fluxes. Increased gas fluxes from the center of the river may be connected to greater exchange with the surface water, thus increased carbon and nutrient supply, and a higher sediment permeability for gas bubbles. By using stable isotope fractionation, we estimated that 12-44% of the CH ${}_4$ transported diffusively was oxidized. Predictors like temperature, air pressure drop, discharge, or precipitation could not or only poorly explain temporal variations of ebullitive CH ${}_4$ fluxes.

Figure 1

Experimental set-up. A map of the study area is provided in left and center of the top row. The map was compiled with ArcGIS Pro (version 3.0.3). In the middle row, average flow velocity v$_m^{}$ across the stream width is displayed as measured on June 28th, 2022. The cross section with four gas traps is displayed at the bottom. Bathymetry was measured at installation (2022) and removal (2023) of the gas traps. Heavy sedimentation covered parts of sampler B after a period of high flow in spring 2023. A detail in the top right schematically shows the sampling procedure.

Figure 2

Grain size distribution at the four sampling sites. Sediment cores were taken downstream of each gas trap. Site A represents the slip-off slope, sites B and C the center of the stream, and site D the undercut slope.

Figure 3

Sediment CH$_4^{}$ concentrations and stable carbon isotopic composition of CH$_4^{}$ in pore water. Site A represents the slip-off slope, site B the center, and site D the undercut slope. In panel (a), markers represent measured data and lines modeled CH$_4^{}$ concentrations. Lines in panel (b) connect measured values.

Figure 4

Summary of ebullition measurements. Panel (a) shows precipitation, discharge, surface water temperature, and the temperature in 20 cm and 45 cm depth. Gas volume fluxes are displayed in panel (b). Panel (c) shows CO$_2^{}$ (GC Micro Box) and panel (d) CH$_4^{}$ (GC-FID) contents in the gas samples. Two outliers in the CO$_2^{}$ data were removed from the data set of site B (see text). Panel (e) depicts CH$_4^{}$ fluxes, and panel (f) stable carbon isotopes of CH$_4^{}$. Error bars indicate the range of minimum and maximum measured values for gas contents and the standard deviation of repeated measurements for ${\delta }^{13}$C–CH$_4^{}$. For an easier description, the data was grouped into seasons which were chosen based on visual inspection of the data to represent specific seasonal patterns (summer: 15-06-2022 to 01-10-2022, autumn: 01-10-2022 to 15-12-2022, winter: 15-12-2022 to 01-04-2023, spring: 01-04-2023 to 15-06-2023).

Figure 5

Boxplots showing site-specific differences in CH$_4^{}$ fluxes and a comparison with diffusive fluxes across the water–air interface. Non-filled markers indicate that gas losses occurred during sampling. At site B, gray filling indicates that the gas trap was influenced by sedimentation (spring 2023).

Figure 6

Regression analysis of methane fluxes and relevant parameters. Precipitation represents the sum, all other parameters the average during the gas collection time. For surface water and sediment temperature, the modified Arrhenius model was fitted. For discharge, precipitation, and air pressure drop, Pearson correlation coefficients were calculated to test for linear regression. Correlations were statistically significant ($p<$ 0.05) only for discharge. There was also a statistically significant correlation between temperature (both T$_{\mathit{SW}}^{}$ and T$_{20cm}^{}$) and CH$_4^{}$ flux at sites A and C (Pearson correlation). At site B, high spring fluxes were not considered due to the reduced comparability after a sedimentation event, which partly buried the gas trap.