Data from: Rates and drivers of carbon emissions from hydropower reservoirs in the southeastern United States

Reservoirs are a significant source of carbon (C) to the atmosphere, but their emission rates vary in space and time. Here, we compared C emissions via diffusion and ebullition pathways at several stations in six large hydropower reservoirs in the southeastern US that were previously sampled in summer 2012. We used similar methodology to compare emissions rates, dominant pathways, and predictors of emissions between the two sampling campaigns. We found that carbon dioxide (CO2) diffusion was the dominant flux pathway during 2012 and 2022, with only three exceptions where methane (CH4) diffusion or CH4 ebullition dominated. However, the direction of the CO2 diffusive flux shifted between 2012 and 2022, where all but three stations across all reservoirs emitted CO2 in summer 2012, but every station sequestered CO2 in summer 2022. Next, we explored drivers of variation and found that indicators of greater algal production were associated with CO2 sequestration (i.e., negative CO2 flux), including surface chlorophyll-a concentration, surface dissolved oxygen saturation, and pH. We also linked CH4 diffusion rates with water temperature as a likely driver of CH4 production. However, we found limited evidence for predictors of CH4 ebullition, likely due to extremely high variability within and across reservoirs, with rates ranging from 0 to 739 mg C m-2 day-1. Additional sampling campaigns outside the summer season highlighted the importance of seasonal phenology in primary production on the direction of CO2 diffusive fluxes, which shifted to positive CO2 fluxes by the end of August as productivity decreased. Our results demonstrate the importance of capturing CO2 sequestration in field and modelling measurements and understanding the seasonal drivers of these estimates when considering upscaling. Measuring C emissions from multiple pathways in reservoirs and understanding their spatiotemporal responses and variability is vital to reducing uncertainties in global upscaling efforts.