A Drier Year Markedly Enhances Methane, but Not Carbon Dioxide, Emissions in a Mediterranean Reservoir.

Reservoirs are increasingly recognized as important sources of greenhouse gases (GHGs), yet the temporal scales and environmental drivers governing CO2 and CH4 flux variability remain poorly constrained. This is particularly the case for Mediterranean systems, which are subject to strong hydrological variability. Here, we present and analyze year-round, high-frequency eddy covariance measurements of CO2 and CH4 fluxes in a Mediterranean reservoir over two years with contrasting hydrological conditions, explicitly evaluating variability at diurnal, multiday, and seasonal timescales. The reservoir acted as a net source of both gases in both years, but CH4 fluxes exhibited substantially greater temporal variability than CO2 fluxes. The seasonal component dominated CH4 flux variability (> 75%), with annual areal fluxes and reservoir-wide emissions 44% higher during the drier year due to an earlier onset and longer duration of the high-emission summer period. Near-sediment temperature and water-column depth were the primary predictors of seasonal CH4 emissions, reflecting temperature controls on methanogenesis and hydrological regulation of ebullition. In contrast, CO2 flux variability was dominated by diel processes, with wind speed emerging as the main predictor of short-term emissions, while water depth and surface temperature regulated seasonal variability. For both gases, chlorophyll-a was an important predictor at multiday timescales. Our results demonstrate that the temporal variability of GHG fluxes from Mediterranean reservoirs is highly sensitive to hydrologically mediated changes in lake water levels, particularly for methane. Under climate change, GHG emissions could increase significantly in Mediterranean regions, and potentially in other seasonally dry regions worldwide, as a result of projected trends towards stronger eutrophication and increases in drought frequency and duration. Our study underscores the value of long-term, high-frequency observations to improve emission estimates and to better represent reservoir processes in regional and global carbon cycle assessments.