Model methodology for estimating pesticide concentration extremes based on sparse monitoring data

Scientific Investigations Report 2017-5159

National Water Quality Program



This report describes a new methodology for using sparse (weekly or less frequent observations) and potentially highly censored pesticide monitoring data to simulate daily pesticide concentrations and associated quantities used for acute and chronic exposure assessments, such as the annual maximum daily concentration. The new methodology is based on a statistical model that expresses log-transformed daily pesticide concentration in terms of a seasonal wave, flow-related variability, long-term trend, and serially correlated errors. Methods are described for estimating the model parameters, generating conditional simulations of daily pesticide concentration given sparse (weekly or less frequent) and potentially highly censored observations, and estimating concentration extremes based on the conditional simulations. The model can be applied to datasets with as few as 3 years of record, as few as 30 total observations, and as few as 10 uncensored observations. The model was applied to atrazine, carbaryl, chlorpyrifos, and fipronil data for U.S. Geological Survey pesticide sampling sites with sufficient data for applying the model. A total of 112 sites were analyzed for atrazine, 38 for carbaryl, 34 for chlorpyrifos, and 33 for fipronil. The results are summarized in this report; and, R functions, described in this report and provided in an accompanying model archive, can be used to fit the model parameters and generate conditional simulations of daily concentrations for use in investigations involving pesticide exposure risk and uncertainty.

Suggested Citation

Vecchia, A.V., 2018, Model methodology for estimating pesticide concentration extremes based on sparse monitoring data: U.S. Geological Survey Scientific Investigations Report 2017–5159, 47 p.,

ISSN: 2328-0328 (online)

Table of Contents

  • Foreword
  • Acknowledgments
  • Abstract
  • Introduction
  • Purpose and Scope
  • Model Methodology
  • Examples of SEAWAVE–QEX Model Results
  • Model Testing
  • Model Assumptions and Limitations
  • Data Preparation and Screening
  • SEAWAVE–QEX Model Applications
  • Summary and Conclusions
  • References Cited
  • Appendix. Description of R Functions and Model Archive for Running SEAWAVE–QEX
  • References Cited

Additional publication details

Publication type:
Publication Subtype:
USGS Numbered Series
Model methodology for estimating pesticide concentration extremes based on sparse monitoring data
Series title:
Scientific Investigations Report
Series number:
Year Published:
U.S. Geological Survey
Publisher location:
Reston, VA
Contributing office(s):
North Dakota Water Science Center, Dakota Water Science Center
Report: viii, 47 p.; Appendix; Data release
Online Only (Y/N):
Additional Online Files (Y/N):