Environmental drivers of seasonal and hourly fluxes of methane and carbon dioxide across a lowland stream network with mixed catchment

Streams serve as open windows for carbon emissions to the atmosphere due to the frequent supersaturation of carbon dioxide (CO 2 ) and methane (CH 4 ) that originates from large carbon input during runoff and associated in-stream processes. Due to the high spatial and temporal variability of the underlying environmental drivers (e.g., concentrations of dissolved CO 2 and CH 4 , turbulence, and temperature), it has remained difficult to address the importance and upscale the emissions to annual whole-system and regional values. In this study, we measured concentrations and calculated emissions of CO 2 and CH 4 at diel and seasonal scales at 15 stations in a 1.4 km 2 stream network that drains a mixed lowland catchment consisting of agriculture (210 km 2 ), forest (56 km 2 ), and lakes, ponds, and wetlands (22 km 2 ) in the upper River Odense, Denmark to evaluate environmental drivers behind the spatiotemporal variability. We used automatically venting floating chambers to calculate hourly diffusive fluxes of CO 2 and CH 4 and CH 4 ebullition. We found: 1) highly supersaturated CO 2 and CH 4 concentrations (median: 175 and 0.33 µmol L −1 , respectively) and high diffusive fluxes of CO 2 and CH 4 (median: 3,608 and 19 µmol m −2 h −1 , respectively); 2) lower daytime than nighttime diffusive emissions of CO 2 in spring and summer, but no diel variability of CH 4 ; 3) higher concentrations and emissions of CH 4 at higher temperatures; and 4) higher emissions of CH 4 at stations located in sub-catchments with higher agricultural coverage. Ebullition of CH 4 peaked at two stations with soft organic sediment and low summer flow, and their ebullition alone constituted 30% of total annual CH 4 emissions from the stream network. Mean annual CO 2 emissions from the hydrological network (37.15 mol CO 2 m −2 y −1 ) exceeded CH 4 emissions 100-fold (0.43 mol CH 4 m −2 y −1 ), and their combined warming potential was 1.83 kg CO 2 e m −2 y −1 . Overall, agricultural sub-catchments had higher CH 4 emissions from streams, while lakes and ponds likely reduced downstream CH 4 and CO 2 emissions. Our findings demonstrate that CO 2 and CH 4 emissions data at high spatial and temporal resolution are essential to frame the heterogeneous stream conditions, understand gas emissions regulation, and upscale to annual values for hydrological networks and larger regions.