Sources and preparation of data for assessing trends in concentrations of pesticides in streams of the United States, 1992–2010

This report updates a previously published water-quality dataset of 44 commonly used pesticides and 8 pesticide degradates suitable for a national assessment of trends in pesticide concentrations in streams of the United States. Water-quality samples collected from January 1992 through September 2010 at stream-water sites of the U.S. Geological Survey (USGS) National Water-Quality Assessment (NAWQA) Program and the National Stream Quality Accounting Network (NASQAN) were compiled, reviewed, selected, and prepared for trend analysis. The principal steps in data review for trend analysis were to (1) identify analytical schedule, (2) verify sample-level coding, (3) exclude inappropriate samples or results, (4) review pesticide detections per sample, (5) review high pesticide concentrations, and (6) review the spatial and temporal extent of NAWQA pesticide data and selection of analytical methods for trend analysis. The principal steps in data preparation for trend analysis were to (1) select stream-water sites for trend analysis, (2) round concentrations to a consistent level of precision for the concentration range, (3) identify routine reporting levels used to report nondetections unaffected by matrix interference, (4) reassign the concentration value for routine nondetections to the maximum value of the long-term method detection level (maxLT-MDL), (5) adjust concentrations to compensate for temporal changes in bias of recovery of the gas chromatography/mass spectrometry (GCMS) analytical method, and (6) identify samples considered inappropriate for trend analysis. Samples analyzed at the USGS National Water Quality Laboratory (NWQL) by the GCMS analytical method were the most extensive in time and space and, consequently, were selected for trend analysis. Stream-water sites with 3 or more water years of data with six or more samples per year were selected for pesticide trend analysis. The selection criteria described in the report produced a dataset of 21,988 pesticide samples at 212 stream-water sites. Only 21,144 pesticide samples, however, are considered appropriate for trend analysis.