External Quality-Assurance Results for the National Atmospheric Deposition Program/National Trends Network, 1995-96:

The blind audit program utilizes a paired sample design to evaluate the effects of routine sample handling, processing and shipping on the chemistry of weekly precipitation samples. The portion of the blind audit sample subject to all of the normal onsite handling and processing steps of a regular weekly precipitation sample is referred to as the bucket portion, whereas the portion receiving only minimal handling is referred to as the bottle portion. Throughout the report, the term positive bias in regard to blind-audit results indicates that the bucket portion had a higher concentration than the bottle portion. The paired t-test of 1995 blind-audit data indicated that routine sample handling, processing, and shipping introduced a very small positive bias (a=0.05) for hydrogen ion and specific conductance and a slight negative bias (a =0.05) for ammonium and sodium. In 1995, the median paired differences between the bucket and bottle portions ranged from -0.02 milligram per liter for both ammonium and nitrate to +0.002 milligram per liter for calcium. Although the paired t-test indicated a very small positive bias for hydrogen ion, the median paired difference between the bucket and bottle portions was 0.00 microequivalents per liter, whereas for specific conductance, the median paired difference between the bucket and bottle portions was 0.200 microsiemens per centimeter in 1995. The paired t-test of blind-audit results in 1996 indicated statistically significant bias for 6 of the 10 analytes. Only chloride, nitrate, hydrogen ion, and specific conductance were not biased in 1996. However, the magnitude of the bias in 1996 was very small and only of limited importance from the viewpoint of an analytical chemist or data user. The median paired differences between the bucket and bottle portions ranged from -0.02 milligram per liter for both ammonium and chloride to +0.006 milligram per liter for calcium. For hydrogen ion, the median paired difference between the bucket and bottle portions was ‑0.357 microequivalent per liter; for specific conductance, the median paired difference between the bucket and bottle portions was 0.00 microsiemens per centimeter in 1996. 

            Surface-chemistry effects due to different amounts of precipitation contacting the sample collection and shipping container surfaces were studied in the blind-audit program using three different sample volumes. The results of a hypothesis test of the relation between hydrogen-ion differences and sample volume were not statistically significant in either 1995 or 1996. This supports the premise that the chemical reactions between the 13-liter bucket shipping container and the sample, which resulted in an increasing loss of hydrogen ion with increasing volume prior to 1994, have been eliminated by the new 1-liter bottle sample-shipping protocol. 

            In the interlaboratory-comparison program in 1995, results of the Friedman test indicated significant (a =0.05) differences in analyte measurements for calcium, potassium, nitrate, sulfate, and hydrogen ion among the five laboratories. In 1996, the Friedman test indicated significant (a=0.05) differences in analyte measurements for potassium, ammonium, chloride, sulfate, nitrate, and hydrogen ion. Intralaboratory bias was indicated for most laboratories in tests of certified analyte concentrations from standard reference material samples and from Ultrapure deionized-water samples. Variability among laboratories was assesed by comparing precision estimates. Precision estimates for the cations at the 50^th percentile exhibited less variability than precision estimates at the 50^th percentile for the anions, pH, and specific conductance for the five laboratories that participated in the interlaboratory comparison program. 

            Results from the collocated-sampler program indicated the median relative error calculated from deposition amounts exceeded 20 percent for sodium and potassium at all of the 1995 collocated sites. In 1996 the median relative error calculated from deposition amounts also exceeded 20 percent at two of the four sites for sodium and potassium. In contrast, the median relative error for sulfate and nitrate deposition was less than 20 percent at all sites in 1995 and 1996, and at four of the eight sites, sulfate and nitrate median relative error was less than 10 percent. The median relative error for hydrogen-ion concentration and deposition ranged from 4.6 to 26.3 percent at the eight sites and was inversely proportional to the acidity of the precipitation at a given site. In those cases where the median laboratory error was greater then zero, median collocated-sampling error estimates in units of concentration were typically at two to four times larger than median laboratory error estimates for most analytes. In fact, the median laboratory error estimates for the laboratory that analyzes the samples for he NADP/NTN was 0.00 milligrams per liter for 4 of 10 analytes.

CONTENTS

Abstract
Introduction
Statistical Approach
Intersite-Comparison Program
  Results for Intersite-Comparison Studies 33 Through 36
    Intersite-Comparison Study Followup Program
Blind-Audit Program
    Data Analysis
Interlaboratory-Comparison Program
    Laboratory Precision
    Interlaboratory Bias
    Intralaboratory Bias
     National Institute of Standards and Technology Standard Reference Samples
     Ultrapure Deionized-Water Samples
Collocated-Sampler Program
Comparison of Laboratory and Network Error
Summary
References Cited

¹ jgordon@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.