Salinity regulates the fate of dissolved organic matter in ice-covered lakes: coupling physical fractionation with microbial transformation

Lakes in cold and arid regions play a critical role in the global carbon cycle. However, the mechanisms governing dissolved organic matter (DOM) dynamics during freeze-thaw cycles under increasing salinization remain poorly understood, particularly the synergistic effects of physicochemical properties and microbial communities. This study investigated four typical lakes in China's cold and arid regions as a salinity gradient experimental platform. By integrating three-dimensional excitation-emission matrix spectroscopy (3DEEMs), Ultraviolet-visible (UV–Vis) spectroscopy, and 16S rRNA sequencing, we systematically revealed the multi-scale distribution patterns of DOM at the ice-water interface and its coupling with microbial communities. Our results demonstrate that: (1) The freezing process exerted significant selective effects on DOM from different sources, with ice DOM showing higher autochthonous characteristics and lower humification, while molecular weight and humification degree increased with ice depth, revealing vertical fractionation. (2) Salinity was a key environmental factor regulating DOM composition and fractionation efficiency, as evidenced by decreasing DOM aromaticity and humification with increasing salinity, where high-salinity conditions weakened physical fractionation efficiency by altering ice physical structure through brine channel formation. (3) Salinity, pH, and ice thickness were critical drivers of microbial community differences between ice and water systems, with the unique ice habitat shaping communities dominated by Proteobacteria, Actinobacteria, Bacteroidota, and Cyanobacteria, which finely regulated DOM composition and transformation in ice. This study provides the first systematic elucidation of the biogeochemical mechanisms governing DOM partitioning during freeze-thaw cycles in cold-arid lakes, highlighting the crucial influence of salinization on DOM migration and transformation during ice cover periods, thereby offering a fundamental theoretical basis for accurately modeling carbon cycling processes and assessing ecological risks under global change.