Inland waters are important global sources, and occasional sinks, of CO2, CH4, and N2O to the atmosphere, but relatively little is known about the contribution of GHGs of constructed waterbodies, particularly small sites in agricultural regions that receive large amounts of nutrients (carbon, nitrogen, phosphorus). Here, we quantify the magnitude and controls of diffusive CO2, CH4, and N2O fluxes from 20 agricultural reservoirs on seasonal and diel timescales. All gases exhibited consistent seasonal trends, with CO2 concentrations highest in spring and fall and lowest in mid-summer, CH4 highest in mid-summer, and N2O elevated in spring following ice-off. No discernible diel trends were observed for GHG content. Analyses of GHG covariance with potential regulatory factors were conducted using generalized additive models (GAMs) that revealed CO2 concentrations were affected primarily by factors related to benthic respiration, including dissolved oxygen (DO), dissolved inorganic nitrogen (DIN), dissolved organic carbon (DOC), stratification strength, and water source (as δ18Owater). In contrast, variation in CH4 content was correlated positively with factors that favoured methanogenesis, and so varied inversely with DO, soluble reactive phosphorus (SRP), and conductivity (a proxy for sulfate content), and positively with DIN, DOC, and temperature. Finally, N2O concentrations were driven mainly by variation in reservoir mixing (as buoyancy frequency), and were correlated positively with DO, SRP, and DIN levels and negatively with pH and stratification strength. Estimates of mean CO2-eq flux during the open-water period ranged from 5,520 mmol m−2 year1 (using GAM-predictions) to 10,445 mmol m−2 year−1 (using interpolations of seasonal data) reflecting how extreme values were extrapolated, with true annual flux rates likely falling between these two estimates.

Small inland waterbodies are well-known hotspots of greenhouse gas (GHG) emissions (Cole et al., 2007;Tranvik et al., 2009) owing to their cumulative abundance in many regions (Downing et al., 2006) and because they often release carbon dioxide (CO 2 ) and methane (CH 4 ) at higher rates than larger inland waters (Downing, 2010). However, GHG concentrations are extremely variable, both spatially and temporally in these small surface waters (

Inland waters are the largest natural source of methane (CH4), a potent greenhouse gas, but models and estimates of aquatic CH4 cycling and emissions were developed in soft-water ecosystems and may not apply to globally abundant salt-rich inland waters. Here we show that elevated salinity constrains microbial CH4 cycling restricting aquatic emissions at large scales. Our survey of the Canadian Prairie ecozone demonstrates that salinity interacts with organic matter availability to shape CH4 patterns across aquatic networks (rivers, lakes, wetlands, and agricultural ponds). Current empirical models, biased toward solute-poor surface waters, overestimated CH4 concentrations and emissions measured in hardwater systems by up to several orders of magnitude, with discrepancies strongly linked to salinity. Models were particularly inaccurate for bubble-mediated emissions from small lentic systems, one of the largest sources of aquatic CH4 globally. Elevated salinity reduced aquatic CH4 emissions by an estimated 81 % in the Canadian Prairies, and could result in a 7.8 % overestimation of global lentic emissions. Widespread salinization of inland waters under future land use and climate regimes could further restrict methane emissions.

Inland waters are hotspots of greenhouse gas (GHG) emissions, and small water bodies are now well known to be particularly active in the production and consumption of carbon dioxide (CO 2 ), methane (CH 4 ), and nitrous oxide (N 2 O). High variability in physical, chemical, and environmental parameters affect the production of these GHG, but currently the mechanistic underpinnings are unclear, leading to high uncertainty in scaling up these fluxes. Here, we compare the relative magnitudes and controls of emissions of all three major GHG in twenty pairs of natural wetland ponds and constructed reservoirs in Canada's largest agricultural region. While gaseous fluxes of CO 2 and CH 4 were comparable between the two waterbody types, CH 4 ebullition was greater in wetland ponds. Carbon dioxide levels were associated primarily with metabolic indicators in both water body types, with primary productivity paramount in agricultural reservoirs, and heterotrophic metabolism a stronger correlate in wetland ponds. Methane emissions were positively driven by eutrophication in the reservoirs, while competitive inhibition by sulfur-reducing bacteria may have limited CH 4 in both waterbody types. Contrary to expectations, N 2 O was undersaturated in both water body types, with wetlands a significantly stronger and more widespread N 2 O sink than were reservoirs. These results support the need for natural and constructed water bodies for regional GHG budgets and identification of GHG processing hotspots.