Abstract |
As is typical of environmental studies, the Eastern Lake Survey generated data that reflect both properties of the environment and properties of the sampling and measurement procedures. Thus, environmental conclusions cannot be drawn from these data without assessment of the sampling and measurement error. The paper shows how the properties of ion chromatography affect the nitrate and sulfate measurements in the survey. The investigation, which is based on measurements of routine-duplicate samples, field blanks, and field audit samples, is interesting because of its complexity. The measurement error has a within-day component and a day-to-day component. The variance of the within-day component increases linearly with concentration except near zero where the algorithm used to interpret the chromatographs seems to behave poorly. The day-to-day component depends on the calibration procedure. As an example, this investigation illustrates the strengths and weaknesses of some common designs for uncertainty assessment. |