Spatial and Temporal Variability of Diffusive CO2 and CH4 Fluxes From the Amazonian Reservoir Petit‐Saut (French Guiana) Reveals the Importance of Allochthonous Inputs for Long‐Term C Emissions

Variability in greenhouse gas emissions from reservoirs creates uncertainty in global estimates of C emissions from reservoirs. This study examines the temporal and spatial variability in CO2 and CH4 surface water concentrations and diffusive fluxes from an Amazonian reservoir using an original data set combining both a high temporal (1 central site × 22 years) and spatial (44 sites × 1 season) resolution monitoring. The gas concentrations at the central site decreased over time and suggested reduced bioavailability of C in the initial flooded soil but exhibited strong seasonal variation. Not accounting for this variability may result in uncertainties in estimates of annual concentrations (ranging from −68.9% to +260% for CH4 and from −71.5% to +156% for CO2) and thus in estimates of diffusive gas emissions. Gas concentrations and diffusive fluxes exhibited high spatial variability in the reservoir, 24 years after impoundment. In particular, diffusive fluxes were higher in littoral and transitional areas than in open areas, suggesting a large contribution of allochthonous C to current gaseous emissions. Not accounting for this spatial variability in diffusive fluxes may underestimate the total emissions expressed in CO2 equivalents from the whole reservoir by 50.7%. Our study stresses the importance of well‐resolved temporal and spatial monitoring to provide reliable estimated of C emissions and a comprehensive understanding of the processes involved; both of these inputs are needed to support decision‐making for developing energy strategies. Plain Language Summary While touted as a low carbon source of electricity, some tropical hydropower reservoirs were reported to emit substantial quantities of dioxide carbon and methane. Tropical reservoirs, especially those located in the lowland reaches of Amazon basin, are the greatest reservoir emitters on the planet. Despite this, hundreds of dams are set to be constructed in these areas. In the context of global warming and the expanding growth of hydroelectricity, the quantification of greenhouse gas emissions from tropical reservoirs is recognized as a priority. Yet emissions vary greatly in time and space. Such variability has to be addressed to provide reliable estimates of carbon emissions from reservoirs and to guide energy decision‐making. With long‐term monitoring of a 24 year old Amazonian reservoir, we reported high gaseous concentrations in the first years following impoundment due to the mineralizati