Distinct seasonal variations of dissolved organic matter across two large freshwater lakes in China: Lability profiles and predictive modeling.

Biological lability of dissolved organic matter (DOM) is a crucial indicator of carbon cycle and contaminant attenuation in freshwater lakes. In this study, we employed a multi-stage plug-flow bioreactor and spectrofluorometric indices to characterize the seasonal variations in DOM composition and lability across Poyang Lake (PY) and Lake Taihu (TH), two large freshwater lakes in China with distinct hydrological seasonality. Our findings showed that the export of floodplain-derived organics and river-lake interaction led to a remarkable increase in terrestrial aromatic and humic-like DOM with high molecular weights and long turnover times in PY. Consequently, the labile fraction was extremely low (average LDOC% of 3%) during the rising-to-flood season (spring and summer). Conversely, autochthonous production in TH considerably enriched semi-labile (average SDOC% of 26%) and biodegradable DOM (average BDOC% of 34%) during the phytoplankton bloom to post-bloom season (summer and autumn). This was reflected by the accumulation of low-light-absorbing and protein-like components with high biological and fluorescence indices. In the dry and non-bloom season (winter), the better preservation of humic substances maintained the high molecular weight and humic degree of DOM in PY, while the decay of aquatic plants strengthened autochthonous production, resulting in a similar BDOC% of PY samples (23%-34%) to TH samples (18%-33%). We further applied partial least squares regression using DOM optical indices as predictive proxies, which generated a greater prediction strength for BDOC% (R = 0.80) compared to SDOC% (R = 0.57) and LDOC% (R = 0.28). The regression model identified aromaticity (SUVA) as the most effective and negative predictor and low molecular weight (A/A) as the highly and positively influential factor. Our study provides new evidence that the seasonality of DOM lability profiles is regulated by the trade-off between flow-related variation and phytoplankton production, and presents an approach to describe and predict DOM lability across freshwater lakes.