Ecological threshold responses of fish communities to lake eutrophication

Identifying ecological thresholds, which are critical points at which ecosystems undergo abrupt shifts in structure and function, has become a cornerstone of conservation and management strategies. This study introduces a novel framework, the "ecological threshold spectrum", by integrating exploration, inference, and validation approaches, to address the limitations of existing methods. The effectiveness of this analytical framework is demonstrated through a case study on fish communities in shallow eutrophic lakes located in the middle and lower reaches of the Yangtze River. We analyzed phosphorus loading as a key driver of fish community and identified three critical ranges of total phosphorus (TP) using Gradient Forest, Threshold Indicator Taxa Analysis and Boosted Regression Trees. The first range, TP < 0.08 mg L^-1, is characterized by submerged macrophyte dominance and high fish biodiversity. The second range, TP between 0.08 and 0.12 mg L^-1, represents a transitional phase with significant species and diversity loss. The third range, TP between 0.12 and 0.17 mg L^-1, is dominated by pollution-tolerant species and associated with severe turbidity. We illustrate how the different analytical methods complement one another by capturing both species-specific and community-level responses to environmental gradients, thereby enabling more robust threshold identifications based on multiple response variables. Our framework addresses uncertainties in ecological threshold detection by combining the strengths of multiple methods while accounting for data quality, noise, and species-specific masking effects. The findings emphasize the importance of TP control in maintaining ecological stability and inform targeted restoration strategies at different eutrophication stages. Overall, the ecological threshold spectrum framework offers new insights into state transitions and provides a robust, practical approach to sustainable management.