thumbnail

Comparison of Surrogate Models To Estimate Pesticide Concentrations at Six U.S. Geological Survey National Water Quality Network Sites During Water Years 2013–18

Scientific Investigations Report 2022-5109
National Water Quality Program
By: , , , and 
https://doi.org/10.3133/sir20225109

Links

Abstract

During water years 2013–18, the U.S. Geological Survey National Water-Quality Assessment Project sampled the National Water Quality Network for Rivers and Streams year-round and reported on 221 pesticides at 72 sites across the United States. Pesticides are difficult to measure, their concentrations often represent discrete snapshots in time, and capturing peak concentrations is expensive. Three types of regression models were developed to estimate concentrations for two selected pesticides at each of six National Water Quality Network for Rivers and Streams sites. The regression models used continuously measured streamflow and water-quality properties (differing combinations of pH, specific conductance, turbidity, and water temperature); discrete water-quality samples analyzed for atrazine, azoxystrobin, bentazon, bromacil, imidacloprid, simazine, and triclopyr; and time as an additional explanatory variable for seasonality.

The modeling approaches included (1) a standard regression that included surrogates (differing combinations of pH, specific conductance, turbidity, and water temperature) and periodic functions (sine-cosine) of pesticide application use as predictor variables; (2) the seasonal wave with flow adjustment model that included a seasonal component and flow anomalies but excluded surrogates; and (3) the seasonal wave with flow adjustment model that included a seasonal component, flow anomalies, and surrogates. Models were evaluated using three measures of model performance: generalized coefficient of determination (generalized R2), Akaike’s Information Criteria, and scale (the estimated standard deviation of the tobit regression error term). Because of low observation numbers, results from this study can be considered a pilot effort with the possibility that some models are overfit.

In all cases, estimated pesticide concentrations modeled with base SEAWAVE-Q were better than the standard surrogate regression models; all 39 generalized R2 values increased by 3–56 percent (median of 25 percent) when compared to the standard surrogate regression models, and all Akaike’s Information Criteria and scale values decreased. The addition of surrogate variables such as pH, specific conductance, turbidity, and water temperature to the base SEAWAVE-Q model to improve estimates of pesticide concentrations resulted in only modest improvements; generalized R2 values increased by only 0–10 percent (median of 3 percent). In some instances, combinations of the surrogates produced more appreciative improvements in model results, but in those instances, we hypothesize that the surrogates correlated with some unknown measure that directly relates to pesticide transport.

Suggested Citation

Covert, S.A., Bunch, A.R., Crawford, C.G., and Oelsner, G.P., 2023, Comparison of surrogate models to estimate pesticide concentrations at six U.S. Geological Survey National Water Quality Network sites during water years 2013–18: U.S. Geological Survey Scientific Investigations Report 2022–5109, 17 p., https://doi.org/10.3133/sir20225109.

ISSN: 2328-0328 (online)

Study Area

Table of Contents

  • Abstract
  • Introduction
  • Purpose and Scope
  • Study Design and Methods
  • Results
  • Summary and Conclusions
  • Acknowledgments
  • References Cited
Publication type Report
Publication Subtype USGS Numbered Series
Title Comparison of surrogate models to estimate pesticide concentrations at six U.S. Geological Survey National Water Quality Network sites during water years 2013–18
Series title Scientific Investigations Report
Series number 2022-5109
DOI 10.3133/sir20225109
Year Published 2023
Language English
Publisher U.S. Geological Survey
Publisher location Reston, VA
Contributing office(s) Ohio-Kentucky-Indiana Water Science Center
Description Report: v, 17 p.; Data Release
Country United States
State Indiana, Kansas, Missouri, North Dakota, Oregon
City De Soto, Durham, Grand Forks, Hazelton, Herman, Sedgwick
Other Geospatial Fanno Creek, Kansas River, Little Arkansas River, Missouri River, Red River of the North, White River
Online Only (Y/N) Y
Additional Online Files (Y/N) N
Google Analytic Metrics Metrics page
Additional publication details