Large alpine deep lake as a source of greenhouse gases: A case study on Lake Fuxian in Southwestern China

Freshwater lakes are recognized as potential sources of greenhouse gases (GHGs) that contribute to global warming. However, the spatiotemporal patterns of GHG emissions have not been adequately quantified in large deep lakes, resulting in substantial uncertainties in the estimated GHG budgets in global lakes. In this study, the spatial and seasonal variability of diffusive GHG (CO2, CH4, and N2O) emissions from Lake Fuxian located on a plateau in Southwestern China were quantified. The results showed that the surface lake water was oversaturated with dissolved GHG concentrations, and the average concentrations were 24.25 μM CO2, 0.044 μM CH4, and 14.28 nM N2O, with diffusive emission rates of 8.82 mmol CO2 m-2 d-1, 31.94 μmol CH4 m-2 d-1, and 4.94 μmol N2O m-2 d-1, respectively. Diffusive CH4 flux exhibited high temporal and spatial variability similar to that in most lakes. In contrast, diffusive CO2 and N2O flux showed distinct seasonal variability and similar spatial patterns, emphasizing the necessity for increasing the temporal resolution in GHG flux measurements for integrated assessments. Water temperature and/or oxygen concentrations were crucial in regulating seasonal variability in GHG emissions. However, no limnological parameter was found to govern the spatial GHG patterns. The frequent advection mixing caused by wind-driven currents might be the reason for the low spatial heterogeneity in GHGs, in which the inconspicuous mechanism requires further research. It was recommended that at least 11 locations were needed for representative whole lake flux estimates at each sampling campaign. In addition, the maximum peak of CH4 in the oxycline from Lake Fuxian indicated that low CH4 oxidation occurred in oxic waters. Overall, this study suggests that, compared to other tropical and temperate lakes, this alpine deep lake is a minor CO2 and CH4 source, but a moderate N2O source, which are horizontally uniform.