<b>Long-term evaluation of pollution</b><b>dynamics</b><b>in West Lake Taihu, China: Enhanced WQI and TLI models for improved accuracy and management relevance</b>

Accurately assessing water quality and eutrophication remains a critical challenge in global lake pollution management. Traditional models, such as the Water Quality Index (WQI) and Trophic Level Index (TLI), often rely on expert judgment and outdated guidelines, limiting their adaptability to evolving pollution conditions. This study improves overall model performance by integrating Principal Component Analysis (PCA) and Factor Analysis (FA) for parameter weighting, and using Random Forest (RF) to validate weight rankings. West Lake Taihu (WTH), the largest natural pretreatment reservoir of Lake Taihu, was selected as the case study for assessing long-term (2008–2023) pollution dynamics. Results showed that total phosphorus (TP) and total nitrogen (TN) were the dominant factors influencing both water quality and eutrophication, replacing traditional dissolved oxygen (DO) and chlorophyll-a (Chl-a). RF and exceedance rate analyses confirmed TP (24.79%) and TN (19.51%) as the primary contributors to WQI, with exceedance rates &gt; 80%, while DO showed little influence (&lt; 9%) and an exceedance rate below 5%. For TLI, TN and TP contributed 31.83% and 30.12%, while Chl-a exhibited lower contribution (23.76%) and a distinct temporal mismatch. From 2008 to 2023, WQI increased by 27.25% and TLI decreased by 22.70%, indicating substantial improvements in both water quality and eutrophication. However, persistent seasonal and spatial differences disrupted WQI-TLI consistency in spatiotemporal trends. The enhanced WQI model provided more realistic assessments by reducing the influence of stable parameters and emphasizing frequently exceeding ones. The modified TLI model improved the identification of eutrophic zones and periods, enhancing the model’s targeting ability and correlating more strongly with algal density. Thus, both models demonstrated their superior accuracy and relevance for contemporary lake pollution management, compared with traditional approaches. This study proposes a robust, data-driven framework for lake pollution evaluation, supporting adaptive management in dynamic freshwater ecosystems.