A drier year markedly enhances methane, but not carbon dioxide, emissions in a Mediterranean reservoir (Dataset).

Dataset for: A drier year markedly enhances methane, but not carbon dioxide, emissions in a Mediterranean reservoir by Ramón et al. (2026), accepted in Global Change Biology. This dataset contains continuous eddy-covariance measurements of methane (CH₄) and carbon dioxide (CO₂) fluxes collected from a floating platform in Cubillas Reservoir (Spain) between 2022 and 2024. The dataset also includes temporal decompositions of the flux signals into seasonal, multiday, and diurnal components, environmental predictor variables used to explain flux variability, and outputs from Random Forest models. The dataset was developed to investigate the environmental controls of reservoir greenhouse-gas emissions and to assess how hydrological variability influences CH₄ and CO₂ flux dynamics. Study site Site name: Cubillas Reservoir Variable Description Country Spain Region Andalusia Reservoir coordinates 37.276° N, 3.671° W Surface area 1.94 km² Primary uses Irrigation and recreation Eddy-covariance measurements Variable Description Measurement method Eddy Covariance (EC) CH₄ analyzer LI-7700 (LI-COR) CO₂/H₂O analyzer LI-7500DS (LI-COR) Wind measurements WindMaster 3D ultrasonic anemometer (Gill) Platform coordinates 37.278° N, 3.672° W Sampling period January 2022–December 2024 Temporal resolution 30 min Dataset structure 1_GapFilledECdata Gap-filled half-hourly CH₄ and CO₂ flux measurements derived from the eddy-covariance system. 2_WaveletTemporalDecomposition Wavelet-based decomposition of CH₄ and CO₂ flux time series into seasonal, multiday, and diurnal components. 3_Predictors Meteorological, hydrological, physical, and biogeochemical variables used as predictors in Random Forest analyses. 4_RandomForest Outputs of Random Forest models, including model performance metrics, predicted values, and variable importance estimates.