Quality of Nutrient Data from Streams and Ground Water Sampled During Water Years 1992-2001
By David K. Mueller and Cindy J. Titus
USGS Scientific Investigations Report 2005-5106, 27 p., 8 figs.
This document is available in pdf format:
SIR2005-5106 (1.61 MB)
(Requires Adobe
Acrobat Reader)
The citation for this report, in USGS format, is as follows:
Mueller, David K. and Titus, Cindy J., 2005, Quality of Nutrient Data
from Streams and Ground Water Sampled During Water Years 1992-2001: U.S.
Geological Survey Scientific Investigations Report 2005–5106, 27
p.
Abstract
Proper interpretation of water-quality data requires consideration of the effects that bias and variability might have on measured constituent concentrations. In this report, methods are described to estimate the bias due to contamination of samples in the field or laboratory and the variability due to sample collection, processing, shipment, and analysis. Contamination can adversely affect interpretation of measured concentrations in comparison to standards or criteria. Variability can affect interpretation of small differences between individual measurements or mean concentrations. Contamination and variability are determined for nutrient data from quality-control samples (field blanks and replicates) collected as part of the National Water-Quality Assessment (NAWQA) Program during water years 1992–2001. Statistical methods are used to estimate the likelihood of contamination and variability in all samples. Results are presented for five nutrient analytes from stream samples and four nutrient analytes from ground-water samples. Ammonia contamination can add at least 0.04 milligram per liter in up to 5 percent of all samples. This could account for more than 22 percent of measured concentrations at the low range of aquatic-life criteria (0.18 milligram per liter). Orthophosphate contamination, at least 0.019 milligram per liter in up to 5 percent of all samples, could account for more than 38 percent of measured concentrations at the limit to avoid eutrophication (0.05 milligram per liter). Nitrite-plus-nitrate and Kjeldahl nitrogen contamination is less than 0.4 milligram per liter in 99 percent of all samples; thus there is no significant effect on measured concentrations of environmental significance. Sampling variability has little or no effect on reported concentrations of ammonia, nitrite-plus-nitrate, orthophosphate, or total phosphorus sampled after 1998. The potential errors due to sampling variability are greater for the Kjeldahl nitrogen analytes and for total phosphorus sampled before 1999. The uncertainty in a mean of 10 concentrations caused by sampling variability is within a small range (1 to 7 percent) for all nutrients. These results can be applied to interpretation of environmental data collected during water years 1992–2001 in 52 NAWQA study units.
Contents
Foreward
Abstract
Introduction
Purpose and Scope
Acknowledgments
Nutrients in Streams and Ground Water
Types of Quality-Control Samples
Compilation of Environmental and Quality-Control Data
Methods of Data Analysis
Methods Used to Determine Bias
Methods Used to Determine Variability
Quality of the Nutrient Data
Contamination Bias
Sampling Variability
Trends in Variability
Confidence Intervals
Implications for Interpretation of Environmental Data
Potential Effects of Contamination
Potential Effects of Sampling Variability
Effects on Measured Concentrations
Effects on Comparisons between Concentrations
Conclusions
References
Glossary of Data-Quality Terms
