Fluvial ecosystems play an outsized role in the aquatic carbon cycle, and are especially important loci of methane (CH₄) production, transport and processing. Methane concentrations and fluxes span a large range in streams and rivers, reflecting a complex set of environmental drivers, which operate at multiple spatial scales. Experimental and within-site studies have shown strong temperature dependence of CH₄ emissions, yet global-scale syntheses have emphasized the need to consider the effect of temperature in the context of landscape drivers such as slope, soil carbon storage and local hydrology. In this study, we leveraged a new fluvial CH₄ database (GRIMeDB) to evaluate the relative strength of temperature and stream discharge in predicting CH₄ concentrations from 172 sites with more than 20 observations of gas concentration coupled to one or both of the predictor variables (n_obs = 4887), using hierarchical linear models. We then used coupled spatial datasets to investigate the role of in situ and landscape-level predictors on the strength and direction of these relationships. Our results suggest that the global relationships with either discharge or temperature were weak (adjusted r² = 0.02 and 0.001, respectively), and when site was included as a random effect, approximately 75% sites did not have a directional relationship between either temperature or discharge and CH₄ concentration. These results emphasize the complex controls on river CH₄ cycling and value of multi-year datasets in teasing apart interactive drivers of ecosystem-level patterns.

Primary Presenter: Nora Casson, University of Winnipeg (n.casson@uwinnipeg.ca)

Authors:

Luke Loken, U.S. Geological Survey  (lloken@usgs.gov)

Samantha Oliver, U.S. Geological Survey  (soliver@usgs.gov)

Gerard Rocher-Ros, Swedish University of Agricultural Sciences  (gerard.rocher.ros@slu.se)

Ryan Sponseller, Umeå University  (ryan.sponseller@umu.se)

Emily Stanley, University of Wisconsin-Madison  (ehstanley@wisc.edu)