Continuous measurement and modeling of CO2 losses from a peatland stream during stormflow events

While streams draining peatland and wetland systems are known to be supersaturated in CO2 with respect to the atmosphere, relatively little is known about short-term temporal variability in response to extreme hydrological events. Here, we use submerged, nondispersive infrared (NDIR) sensors to make continuous measurements of CO2 concentrations during 18 storm events in a Scottish peatland stream. Individual storms exhibited 3 distinct types of hysteresis loop. We suggest that differences in loop form may be due to differences in the relative contributions of soil water or differences in the contributing catchment source area. We found a negative concentration-discharge relationship over the full study period, suggesting that CO2-rich deep peat/groundwater was the major source of aquatic CO2 under low flow conditions. By removing the effect of dilution and estimating additions and losses of CO2, we also show the importance of both surface peat CO2 inputs into the stream and evasion loss during stormflow. The best model of temporal variability in CO2 was achieved by separating the data set into stormflow'' and dry periods.'' Downstream CO2 export during the study period was dominated by stormflow events (71%), highlighting the importance of accurately accounting for high-flow CO2 sources.