A surrogate regression approach for computing continuous loads for the tributary nutrient and sediment monitoring program on the Great Lakes

• DOI: 10.1016/j.jglr.2017.10.003
• Publication Date: 2018-01-05T21:30:08Z
• Authors: Dale M. Robertson; Laura E. Hubbard; David L. Lorenz; Daniel J. Sullivan
• Keywords: Seasonality; Turbidity

Water quality (WQ) in many Great Lake tributaries has been degraded (increased nutrient and sediment concentrations) due to changes their watersheds, resulting downstream eutrophication. As part of the Lakes Quality Agreement, specific goals were established for loading constituents (e.g., phosphorus). In 2010, Restoration Initiative was launched identify problem areas, accelerate restoration efforts, track progress. 2011, U.S. Geological Survey a monitoring program on 30 lakes, representing ~ 46% draining area spectrum land uses. Discrete measurements nutrients suspended sediment, continuous flow WQ surrogates (turbidity, temperature, conductance, pH, dissolved oxygen) are being collected these document estimate (5-min) loading. To loadings, two regression models developed each constituent site: one using seasonality factor; flow, seasonality, surrogates. Variables included final chosen from explanatory variables that worked "best" all sites. computing loads, when surrogate data unavailable short periods, loads computed models. Prediction intervals calculated results both These provide better understanding short-term variability long-term affecting environmental health than traditional techniques employ only parameters.