A Bottom-up Approach to Characterizing Least-Disturbed Stream Habitat

Many current efforts to forecast stream biotic conditions are based on top-down approaches using landscape-level anthropogenic disturbance data.  These data represent surrogate measures of in-stream habitat and water quality metrics, and are limited in their ability to describe the mechanistic response of streams’ physical, chemical, and biological characteristics to landscape alterations.  We developed a bottom-up approach for selecting and testing local habitat and water quality metrics’ effects on stream fish community structure using data collected from wadeable streams in Missouri between 2000 and 2013 (Central Plains Ecoregion n = 242, Ozark Ecoregion n = 361).  We used Canonical Correspondence Analysis paired with Threshold Indicator Taxa Analysis to identify species-specific and community-level physical habitat and water quality associations.  Our analysis revealed strong positive relationships between sensitive and lithophilic-spawning species to higher pH, dissolved oxygen, and greater percentages of coarse gravel substrates.  In contrast, wider, lower gradient streams with finer substrates and greater pool area consistently contained higher abundances of tolerant and omnivorous species.  Our approach extends current stream assessment methods by identifying potentially limiting habitat characteristics, and serves as a platform for developing a quantitative physical habitat index for Missouri streams.