The precision of wet atmospheric deposition data from national atmospheric deposition program/national trends network sites determined with collocated samplers

Abstract

Abstract A collocated, wet-deposition sampler program has been operated since October 1988 by the U.S. Geological Survey to estimate the overall sampling precision of wet atmospheric deposition data collected at selected sites in the National Atmospheric Deposition Program and National Trends Network (NADP/NTN). A duplicate set of wet-deposition sampling instruments was installed adjacent to existing sampling instruments at four different NADP/NTN sites for each year of the study. Wet-deposition samples from collocated sites were collected and analysed using standard NADP/NTN procedures. Laboratory analyses included determinations of pH, specific conductance, and concentrations of major cations and anions. The estimates of precision included all variability in the data-collection system, from the point of sample collection through storage in the NADP, /NTN database. Sampling precision was determined from the absolute value of differences in the analytical results for the paired samples in terms of median relative and absolute difference. The median relative difference for Mg 2+ , Na + , K 4 + and NH concentration and deposition was quite variable between sites and exceeded 10% at most sites. Relative error for analytes whose concentrations typically approached laboratory method detection limits were greater than for analytes that did not typically approach detection limits. The median relative difference for S0 4 2− and NO 3 − concentration, specific conductance, and sample volume at all sites was less than 7%. Precision for H + concentration and deposition ranged from less than 10% at sites with typically high levels of H + concentration to greater than 30% at sites with low H + concentration. Median difference for analyte concentration and deposition was typically 1.5-2-times greater for samples collected during the winter than during other seasons at two northern sites. Likewise, the median relative difference in sample volume for winter samples was more than double the annual median relative difference at the two northern sites. Bias accounted for less than 25% of the collocated variability in analyte concentration and deposition from weekly collocated precipitation samples at most sites.