Can the Stream Quantification Tool (SQT) Protocol Predict the Biotic Condition of Streams in the Southeast Piedmont (USA)?

Abstract

In some states, the Stream Quantification Tool (SQT) has been adopted to quantify functional change of stream mitigation efforts. However, the ability of the SQT protocol to predict biological function and uphold the premise of the Stream Functions Pyramid (Pyramid) remains untested. Macroinvertebrate community metrics in 34 headwater streams in Piedmont, North Carolina (NC, USA) were related to NC SQT protocol (version 3.0) factors and other variables relevant to ecological function. Three statistical models, including stepwise, lasso, and ridge regression were used to predict the NC Biotic Index (NCBI) and Ephemeroptera, Plecoptera, and Trichoptera (EPT) richness using two datasets: 21 SQT variables and the SQT variables plus 13 additional watershed, hydraulic, geomorphic, and physicochemical variables. Cross-validation revealed that stepwise and ridge outperformed lasso, and that the SQT variables can reasonably predict biology metrics (R2 0.53–0.64). Additional variables improved prediction (R2 0.70–0.88), suggesting that the SQT protocol is lacking metrics important to macroinvertebrates. Results moderately support the Pyramid: highly predictive ridge models included metrics from all levels, while highly predictive stepwise models included metrics from higher levels, and not watershed hydrology. Reach-scale metrics were more important than watershed hydrology, providing encouragement for projects limited by watershed condition.