Methane emissions from an anthropogenically modified freshwater continuum
Inland waters are known to emit non-negligible greenhouse gas amounts to the atmosphere contributing to global warming. The potent methane (CH4) is produced naturally by degradation of organic matter in aquatic sediments and is released towards surface waters. Yet, the various anthropogenic activities such as dam building, river channelization and diversion disrupt and interrupt the course of organic matter along the freshwater continuum leading to a large spatial heterogeneity among CH4 sources. Additionally, the accuracy of CH4 emission estimates from freshwaters is currently questionnable due to the lack of high resolution data and process coverage. Methane emissions from the aquatic realm are thus regionally and locally variable. With one of the highest demographic density, Switzerland produces more than 50% of its electricity through hydropower reservoirs. Hydroelectric production is considered as a renewable source of energy although it is not carbon free due to the potential of reservoir to store large amounts of organic matter in front of dams enhancing subsequent CH4 production. Delta regions also constitute sites where organic matter accumulates that potentially release large amount of CH4 but are rarely taken into account in CH4 emission estimates.
The main hypothesis of the present study is that the Swiss aquatic landscape that is severly impacted by the human footprint may release non-negli-gible amounts of CH4 in the atmosphere with high local and regional variability. Variability in CH4 dynamics, focussing specifically on CH4 emissions, was investigated along the Swiss freshwater continuum. Methane distribution within the delta of the largest western European lake, Lake Geneva, was studied. The impact of the oldest Swiss Alpine reservoir, Lake Klöntal, on CH4 emissions was established. Finally, CH4 dynamics along the dam-impacted Aare River that drains the largest watershed of Switzerland have been investigated. Conclusions varied greatly depending on site locations and revealed possible future approaches in order to quantify CH4 emissions with more accuracy.
Several sampling strategies were used to fulfill the objectives of this work. We took sediment samples at diverse locations of a delta to measure CH4 concentrations and infer local CH4 production and variability. An eddy covariance (EC) system mounted on a floating platform was deployed to quantify CH4 flux at a high frequency year round. Samples at the surface water were also taken to measure CH4 concentrations and calculate fluxes based on diffusion law for further comparison with the EC. Simultaneously, we used hydroacoustic methods to detect CH4 bubbles known to contribute largely to the overall CH4 flux when present. Finally, by combining floating chamber surveys, CH4 flux calculations based on diffusion law to mass-balance approaches, we were able to consider all relevant fluxes within limited amount of time.
The link between sedimentation rate and terrigenous carbon with CH4 production was demonstrated in chapter 2. Methane production rate within the numerous underwater canyons of the Rhone delta (Lake Geneva) was greater at locations of high river inflow compared to regions that received less inflow. Studying the C:N ratio allowed to identify the region where the Rhone inputs were the highest and linked to potentially large CH4 source. However, CH4 production rate was low close to the river mouth where hydrodynamics are greater due to sediment remobilization that prevent high methanogenesis rates. This study showed the importance of research focuses orientated towards hotspots of CH4 production such as regions with high sedimentation rates.
Regional and local variability of CH4 dynamics and emissions was very large as shown by studies conducted in the Alpine reservoir, Lake Klöntal and along the damimpacted Aare River. Although we measured a high diel and seasonal variability in CH4 emissions from Lake Kl¨ontal, overall low CH4 emissions were detected from this Alpine reservoir compared to downstream lowland regions where water carried higher amounts of organic matter. In both regions, however, CH4 production was controlled by temperature with the highest emission rates observed in summer. Additionally, comparing CH4 flux detected by the EC system, CH4 accumulation rates measured using chambers and flux calculation using surface concentrations and parameterized turbulence showed large discrepancies. The studies conducted in Lake Klöntal and along the Aare River highlight the importance of a good quality in data collection taken at high spatio-temporal resolution and the need to obtain realistic estimates of CH4 emissions using several approches.
The first study along the Aare demonstrated the inability of gas exchange models based on a turbulent diffusion approach to correctly estimate CH4 emissions from very dynamic systems such as rivers. Using a mass-balance approach revealed CH4 local hotspots that were not identified when calculating fluxes using surface concentration and surface turbulence. Locations such as major tributary confluences with the Aare were recognized and confirmed by the second study along the Aare as sites of high turbulence where degassing is substantially underestimated of at least one order of magnitude using traditional approaches. We also suggested that trace gas experiments, using SF6 or He for instance, may allow future research to measure more accurately the degassing at such turbulent locations in rivers. Additionally, the second study showed that the more pristine reach of the river supported CH4 concentrations along the river, while oxidation and degassing counter-balanced CH4 sources in the channelized part leading to overall CH4 losses.
Finally, the various sampling strategies were shown to be determining in regard to more accurate estimates of CH4 emissions from freshwater. The current work also highlighted importance of CH4 emissions from lowland hydropower schemes compared to Alpine reservoirs. Ultimately, the identification of new hotspot locations such as intense degassing occuring at tributary confluences with the main river questionned the accuracy of current estimates suggesting that they may be highly underestimated. It is, therefore, necessary to revise CH4 emissions from freshwater using the aforementioned sampling strategies and recommendations.