The Tennessee River Valley is simultaneously one of the most species-rich and most imperiled drainage basins in North America for freshwater fishes. As extinction rates of North American freshwater fishes are projected to increase in the coming decades due to ongoing human impacts and climate change, more research aimed at improving the efficiency, economy, and effectiveness of conservation efforts in the Tennessee River system is needed. Trait-based approaches offer a powerful framework for addressing this need by shifting focus from taxonomic composition alone to the structural and functional organization of assemblages. By quantifying the diversity of ecological, morphological, and life-history traits in stream fish assemblages, trait-based approaches allow for the assessment of functional diversity, an important component of biodiversity which may have more direct implications for ecosystem function and resilience. In this study, we will examine how multiple environmental stressors influence both the functional diversity and functional redundancy of stream fish assemblages across the Tennessee River Valley using long-term fish monitoring data provided by the Tennessee Valley Authority. We will use random forest models, a machine-learning approach capable of modeling the complex, non-linear relationships between landscape- and site-level stressors (e.g., land use/land cover, habitat) and fish assemblage responses. Similar methodologies have been used to characterize fish assemblage responses to multiple stressors in other river systems, but to our knowledge, large-scale analyses of this nature have yet been applied to the Tennessee River Valley. Our analysis will address three conservation-relevant questions: i) what is the current status and spatial distribution of fish functional diversity in Tennessee Valley streams, ii) how do environmental stressors affect the functional diversity and redundancy of stream fish assemblages, and iii) how might projected future stressors alter functional diversity and redundancy. By leveraging machine-learning models within a multiple-stressor, trait-based framework, our study aims to identify key streams, traits, and stressors of management concern and provide actionable insights for regional freshwater conservation and management.