Substantial but highly variable diffusive methane emissions from eutrophic shallow lakes.

Lakes are critical natural methane (CH4) emitters, yet their emissions magnitude remains highly uncertain due to complex regulatory factors in human-dominated regions. To address this knowledge gap, the diffusive CH4 emissions and their drivers in heavily human-impacted shallow eutrophic lakes of China's eastern plain were investigated via integrating long-term field measurements and laboratory incubations with isotopic trajectories. Results showed that diffusive CH4 emissions were significant but exhibited substantial variability. The emissions were 2.1 times higher in the wet period (0.21 ± 0.13 mmol m-2 d-1) than in the dry period (0.10 ± 0.06 mmol m-2 d-1), consistent with greater sediment CH4 production rates, elevated water temperatures, higher chlorophyll-a (a proxy for algal biomass), and increased dissolved organic carbon. CH4 emissions span one order of magnitude (0.04 ± 0.01 to 0.40 ± 0.39 mmol m-2 d-1) across lakes, driven by strong gradients in autochthonous (algal-derived) and allochthonous (soil-derived) organic carbon inputs, and likely associated with highly variable sediment CH4 production modulated by eutrophication status. Further analysis revealed a seasonal shift in dominant drivers, chlorophyll-a (representing algal-derived carbon) dominated dry-period emissions, while dissolved organic carbon (primarily allochthonous such as soil-derived) regulated emissions in wet period. These observations were supported by incubations showing both algal- and soil-derived organic matter produced large CH4 via degradation. Random forest modeling identified chlorophyll-a as the top predictors of diffusive emissions, and the carbon isotope fractionation factor derived from incubations and field measurements suggested the dominant role of acetoclastic methanogenic pathway in CH4 production of algal-dominated lakes. Our findings likely provide a mechanistic framework for estimating CH4 emissions and predicting their future variability in human-impacted shallow lakes.