Eddy Covariance Data Reveal That a Small Freshwater Reservoir Emits a Substantial Amount of Carbon Dioxide and Methane

Abstract

Small freshwater reservoirs are ubiquitous and likely play an important role in global greenhouse gas (GHG) budgets relative to their limited water surface area. However, constraining annual GHG fluxes in small freshwater reservoirs is challenging given their footprint area and spatially and temporally variable emissions. To quantify the GHG budget of a small (0.1 km2) reservoir, we deployed an Eddy covariance (EC) system in a small reservoir located in southwestern Virginia, USA over 2 years to measure carbon dioxide (CO2) and methane (CH4) fluxes near‐continuously. Fluxes were coupled with in situ sensors measuring multiple environmental parameters. Over both years, we found the reservoir to be a large source of CO2 (633–731 g CO2‐C m−2 yr−1) and CH4 (1.02–1.29 g CH4‐C m−2 yr−1) to the atmosphere, with substantial sub‐daily, daily, weekly, and seasonal timescales of variability. For example, fluxes were substantially greater during the summer thermally stratified season as compared to the winter. In addition, we observed significantly greater GHG fluxes during winter intermittent ice‐on conditions as compared to continuous ice‐on conditions, suggesting GHG emissions from lakes and reservoirs may increase with predicted decreases in winter ice‐cover. Finally, we identified several key environmental variables that may be driving reservoir GHG fluxes at multiple timescales, including, surface water temperature and thermocline depth followed by fluorescent dissolved organic matter. Overall, our novel year‐round EC data from a small reservoir indicate that these freshwater ecosystems likely contribute a substantial amount of CO2 and CH4 to global GHG budgets, relative to their surface area.

Toward more accurate estimates of carbon emissions from small reservoirs

“…A similar amplitude (0.56 g CO 2 m −2 h −1 ) was observed in ponds in the subarctic wetland region of the Hudson Bay Lowlands, Canada, where CO 2 emissions were attributed to degrading peat (Hamilton et al 1994). The duckweed reservoir amplitude was 2.6 times higher than that observed in a larger eutrophic reservoir (0.119 km 2 ) (Hounshell et al 2023). We observed the lowest amplitude in Sister, which was an order of magnitude lower (0.042 g CO 2 m −2 h −1 ) than the duckweed reservoir.…”

“…To characterize diel patterns in CO 2 diffusion, CH 4 diffusion, and CH 4 ebullition, we plotted and visually inspected the fluxes interpolated with sequential Gaussian simulation across the sampling intervals. We then ran models explaining the measured point estimates of diffusive fluxes as a function of environmental variables we expected to relate to diel variation in diffusion, including DO at the top and bottom of the water column, windspeed, light, and temperature (Natchimuthu et al 2016; Sieczko et al 2020; Rudberg et al 2021; Hounshell et al 2023). We recorded DO concentration 0.25 m below the water surface and 0.1 m from the sediment using a handheld meter (YSI Pro Plus) at two locations near the edge of the reservoir and two in the center during every sampling period.…”

“…As a consequence of their numeric abundance and positions in river networks as recipients of substantial terrestrially derived carbon inputs (Harvey and Schmadel 2021), these reservoirs have the capacity to cumulatively emit carbon dioxide (CO 2 ) and methane (CH 4 ) in comparable magnitudes to larger reservoirs (Grinham et al 2018; Ollivier et al 2019 a ), which have been the primary focus of research for the past two decades. Variation in the dominant pathways, times, and locations of emissions in natural lakes and larger reservoirs has been shown to impact inferences about the total magnitude of their emissions (Beaulieu et al 2016; Sieczko et al 2020; Hounshell et al 2023); however, this variation has been poorly characterized in small reservoirs, and prior efforts to estimate their emissions have frequently relied on measurements of one pathway of emissions, at one time during the day, and in few locations in the reservoir (Wang et al 2017; Webb et al 2019; Ollivier et al 2019 a ). Unaccounted‐for variation in emissions may lead to misestimation of the contribution of small reservoirs to landscape greenhouse gas (GHG) emissions, which may be reduced or mitigated through management actions.…”

“…Wind is a major driver of surface turbulence and gas exchange in reservoirs (Crusius and Wanninkhof 2003). Windspeeds are generally higher during the day, which can force higher rates of gas exchange and, thus, higher daytime diffusive emissions (Sieczko et al 2020; Hounshell et al 2023). Transfer processes can also impact patterns of CH 4 ebullition, with drops in hydrostatic pressure yielding high bubbling rates; however, few studies have captured sub‐daily measurements of CH 4 emissions to characterize the periodicity of bubbling events (but see Grinham et al 2011; Varadharajan and Hemond 2012; Sieczko et al 2020).…”