Abstract
While harmful cyanobacterial blooms (HCBs) are extensively characterized in eutrophic lakes, the ecological dynamics of connected river networks remain oversimplified, obscuring the mechanisms of community assembly and toxin distribution across the lake-river interface. This study investigated the spatial heterogeneity of HCBs and microcystins (MCs) in the Lake Taihu watershed, revealing a complex functional divergence between lotic and lentic ecosystems. The rivers functioned as primary nutrient sources, with Total Nitrogen (3.35 ± 1.52 mg·L(-1)) and Total Phosphorus (0.21 ± 0.22 mg·L(-1)) concentrations being 1.7-fold and 1.8-fold higher, respectively, than those in the lake during peak periods. Conversely, the lake acted as a biological sink, supporting a peak cyanobacterial density (3.32 × 10(7) cells·L(-1)) nearly 1.5 times that of the river network. Phytoplankton community analysis revealed distinct ecological niches: while the lentic lake environment was almost exclusively dominated by colonial Microcystis (>90% relative abundance), the lotic river networks harbored a diverse assemblage with significant contributions from filamentous Oscillatoria and Dolichospermum. Correspondingly, intracellular MC (IMC) in the lake (up to 14.5 μg·L(-1)) significantly exceeded riverine levels (generally <1.0 μg·L(-1)). Despite these compositional differences, toxin dynamics exhibited strong bidirectional coupling (r > 0.75, p < 0.01), suggesting a spillover effect where the lake determines the watershed's toxin burden during rivers outflow period. Redundancy Analysis (RDA) further elucidated that limnetic blooms were primarily regulated by water temperature and pH, whereas riverine communities were strictly constrained by hydrodynamic flow. Consequently, the health risk assessment revealed a highly heterogeneous landscape where, beyond the northern lake bays, specific semi-lentic river segments emerged as cryptic hotspots. These findings demonstrate that while nutrients fuel the system, hydrodynamic conditions act as the ultimate ecological filter determining the spatiotemporal distribution of cyanobacterial risks, necessitating an integrated approach to monitoring the entire lake-river continuum.