An html version of the abstract and table of contents are below. The full report is available in a 3.2 MB pdf. Links to the pdf.
Five external quality-assurance programs were operated by the U.S. Geological
Survey for the National Atmospheric Deposition Program/National Trends Network
(NADP/NTN) from 1997 through 1999: the intersite-comparison program, the blind-audit
program, the field-audit program, the interlaboratory-comparison program, and
the collocated-sampler program. The intersite-comparison program assessed the
precision and bias of pH and specific-conductance determinations made by NADP/NTN
site operators. In two 1997 intersite-comparison studies, 82.8 and 85.8 percent
of the pH determinations met the NADP/NTN accuracy goals whereas 96.8 and 92.4
percent of the specific-conductance determinations met the NADP/NTN accuracy
goals. The percentage of pH and specific-conductance determinations that met
the accuracy goals in 1998 were, for the most part, higher than in 1997. In
two 1998 studies, 90.9 and 90.2 percent of the pH determinations met the accuracy
goals compared to 94.7 and 95.6 percent of the specific-conductance measurements
meeting the accuracy goals. In one 1999 intersite-comparison study, 88.3 percent
and 98.0 percent of pH and specific-conductance determinations, respectively,
met the NADP/NTN accuracy goals.
The blind-audit program evaluates the effects of routine sample handling, processing,
and shipping on the chemistry of weekly precipitation samples. A portion of
the blind-audit sample subject to the normal onsite handling, and processing
of a weekly precipitation sample is referred to as the bucket portion, whereas
the portion receiving only minimal handling is referred to as the bottle portion.
Positive bias in regard to blind-audit results indicates that the bucket portion
has a higher concentration than the bottle portion. The paired t-test for the
1997 through 1999 blind-audit data indicates that routine sample handling, processing,
and shipping introduced a positive bias for calcium and chloride and a negative
bias for hydrogen ion. During 1997 through 1999, the median paired differences
between the bucket and bottle portions ranged from 0.00 milligrams per liter
for nitrate and ammonium to +0.010 milligram per liter for both chloride and
sulfate. The median paired difference between the bucket and bottle portions
for hydrogen ion was -1.086 microequivalents per liter, whereas for specific
conductance, the median paired difference between the bucket and bottle portions
was -0.200 microsiemen per centimeter during 1997 through 1999.
Surface-chemistry effects due to variable amounts of precipitation contacting
prewashed sample-collection and shipping-container surfaces were studied in
the blind-audit program by using three different sample volumes. The sample-collection
and shipping containers used for the blind-audit study were obtained from the
site operator's supply and could have been used for precipitation samples. Results
of a Kruskal-Wallis analysis of variance test of the relation between paired
blind-audit sample differences in units of concentration and sample volume were
statistically significant for magnesium, chloride, sulfate, and hydrogen ion
during 1997 through 1999. Before 1994, at least 5 of the 10 analytes displayed
a statistically significant difference between paired blind-audit differences
in units of concentration and sample volume, supporting the premise that the
chemical reactions between the 13-liter bucket shipping container (primarily
the butadiene o-ring lid of the shipping container) and the sample, which resulted
in an increasing loss of hydrogen ion with increasing volume Before 1994, have
been eliminated by the new 1-liter bottle sample-shipping protocol.
The field-audit program was designed to measure the effects of field exposure,
handling, and processing on the chemistry of NADP/NTN precipitation samples.
In the field-audit program, the site operator is instructed to process and submit
a quality-control sample following a standard 7-day, Tuesday-to-Tuesday sampling
period with no precipitation. The requirement of a full week without precipitation
prevents loss of the routine weekly precipitation-chemistry data collected at
the site. The Wilcoxon signed-rank test indicated that no bias was introduced
during routine sample handling, processing, and shipping for any of the analytes
during 1997 through 1999. However, the paired t-test indicated statistically
significant differences for all of the analytes except for sulfate and specific
conductance for the "wet" coded samples and all but sodium and specific
conductance for the "dry" coded samples during the same time period.
Similar to the blind-audit program, surface-chemistry effects due to variable
amounts of precipitation contacting prewashed sample-collection and shipping
container surfaces were also studied in the field-audit program by using three
different sample volumes. Results of a Kruskal-Wallis analysis of variance test
of the relation between paired field-audit sample differences in units of concentration
and sample volume were statistically significant for magnesium, chloride, and
sulfate during 1997 through 1999. These were the same analytes that were statistically
significant in the blind-audit program during the same time period. The blind-audit
program also showed a statistically significant difference for hydrogen ion
that the field-audit program did not show.
The two objectives of the interlaboratory-comparison program are: (1) to estimate
the analytical precision of participating laboratories, and (2) to determine
if statistically significant differences exist among the analytical results
of participating laboratories. In 1997, results of the Friedman test indicated
significant differences in analyte measurements among the five participating
laboratories for all analytes except potassium. In 1998, the Friedman test indicated
significant differences in analyte measurements among the five participating
laboratories for all 10 parameters. The Friedman test indicated there were statistically
significant differences in the results from the seven participating laboratories
for all analytes in 1999.
Bar graphs depicting the number of data points outside the control limits were
prepared for the Central Analytical Laboratory in 1997, 1998, and 1999. For
magnesium, potassium, chloride, pH, and specific conductance, the number of
data points outside the control limits decreased from 1997 to 1999. Sodium and
ammonium had an upward trend in the number of data points outside the control
limits during the same time period.
Intralaboratory bias was indicated for most laboratories in tests of certified
analyte concentrations from standard reference material samples and from Ultrapure
deionized-water samples. The Ultrapure deionized-water used by the U.S. Geological
Survey is comparable to American Society for Testing and Materials (ASTM) Type
I water. Comparing precision estimates assessed variability among laboratories.
Precision estimates for the cations at the 50th percentile exhibited less variability
than precision estimates at the 50th percentile for the anions, pH, and specific
conductance for the seven laboratories that participated in the interlaboratory-comparison
program between 1997 and 1999
In October 1988, the collocated-sampler program was established to provide
a method of estimating the overall precision of the precipitation-monitoring
system used by the NADP/NTN. Results from the collocated-sampler program indicated
that the median relative error, calculated from deposition amounts, exceeded
20 percent for potassium at all but two of the 1997-99 collocated sites. The
median relative error, calculated from concentration amounts, was less than
10 percent for sulfate, specific conductance, and sample volume for all of the
1997-99 collocated sites. The sites with the lowest median hydrogen-ion concentrations
had much higher median relative errors associated with their paired-sample analyses.
The median relative errors for hydrogen-ion concentration were less than 10
percent at three sites, and these same sites had median hydrogen-ion concentrations
of 18 microeqivalents/L or greater. For sites with median hydrogen-ion concentrations
less than 10 microequivalents/L, the median relative errors exceeded 12 percent.
Abstract
Introduction
Statistical Approach
Intersite-Comparison Program
Results for Intersite-Comparison Studies 39 through 43
Intersite-Comparison Study Followup Program
Blind-Audit Program
Data Analysis
Field-Audit Program
Data Analysis
Assessing Analytical Bias
Ultrapure Deionized-Water Sample Results
Interlaboratory-Comparison Program
Laboratory Precision
Interlaboratory Bias
Intralaboratory Bias
Certified Standard Reference Samples
Ultrapure Deionized-Water Samples
Collocated-Sampler Program
Comparison of Laboratory and Network Error
Summary
Intersite-Comparison Program
Blind-Audit Program
Field-Audit Program
Interlaboratory-Comparison Program
Collocated-Sampler Program
References Cited
1 jgordon@usgs.gov, U.S. Geological Survey, Denver, CO
2nlatysh@usgs.gov, U.S. Geological Survey, Denver, CO
If you have Adobe Acrobat Reader installed on your computer, you may view and/or print the PDF version of this report. If you do not have Acrobat Reader, you may download it here.
The pdf file: WRIR 03-4027
| USGS | Water | Biology | Geology | Mapping | WRI Pubs | Water Related Pubs |