Sources of Uncertainty in Stream Nutrient Sampling Below a Point Source

Author

Alexandria Hayden Campbell, Utah State University

Date of Award

5-2015

Degree Type

Thesis

Degree Name

Departmental Honors

Department

Biology

Abstract

The goal of this study was to determine what aspects of sampling and sample storage could lead to uncertainty when taking samples in a stream below a point source. Sources of uncertainty studied were the locations where the samples were taken to assess if nutrients were adequately mixed within a cross-section, different filtration techniques, dilution errors, analytical uncertainty, and freezing time. Bootstrapping analyses were used to determine whether mixing and dilution errors led to uncertainty, while one-way ANOVAs were used to evaluate filtration techniques and storage time. Sample spikes to determine percent recovery of nutrients and repeat sample analyses are routinely performed as part of the lab quality assurance/quality control plan (QA/QC), and are used here to evaluate analytical uncertainty. Comparison of coefficients of variation (CV) of samples collected within a cross section at four locations, above, at, and below a point source, revealed that mixing of nutrients within a cross section appeared to be different at the different locations. The filtration devices analyzed were an electric pump and a manual syringe. These two devices gave statistically similar results in ammonium, nitrate, and soluble reactive phosphorus concentrations (p > 0.05). Dilution error was determined by comparing seven diluted samples with the original sample with which they were made. Dilutions proved to have the highest uncertainty relative to other treatments. The diluted samples were consistently higher than the original sample for all nutrients and were more variable than lab QA/QC duplicates for ammonium and soluble reactive phosphorus. Analytical uncertainty was found to be less than uncertainty associated with sample collection and storage except for unanticipated protocol failure. For this study, QA/QC data beyond 20% were considered fails, and the samples required reanalysis. In most cases the percent recovery of spiked samples was within 20% and coefficients of variation of samples repeatedly analyzed were much less than 20%. However, ammonium, total nitrogen, and total phosphorus incurred the most failures. Freezing samples appeared to be an adequate storage method. Samples frozen for 12 weeks showed statistically significant declines in total nitrogen (TN) and total phosphorus (TP) concentrations (p < 0.05), however these declines were less than 9% of the initial values. This is within the range of variation seen for analytical duplicates.

Recommended Citation

Campbell, Alexandria Hayden, "Sources of Uncertainty in Stream Nutrient Sampling Below a Point Source" (2015). Undergraduate Honors Capstone Projects. 574.
https://digitalcommons.usu.edu/honors/574