In Reply Refer To:
Mail Stop 412
February 26, 2009
Office of Water Quality Technical Memorandum 2009.03
Subject: Potential Bias in Alkaline Persulfate Analysis of Total Nitrogen in Whole Water Samples and Recommendations for Quantifying Bias in Whole Water Samples
The purpose of this memorandum is to alert users of a potential negative bias in whole-water total nitrogen concentrations determined by USGS alkaline persulfate digestion method I-4650-03 (NWQL lab code 2756; USGS WRIR 03-4174) and to offer suggestions for quantifying bias in river and stream water samples submitted to the National Water Quality Laboratory (NWQL).
A negative bias was discovered in the Office of Water Quality (OWQ) requested experiments that were done at the NWQL with assistance by the Branch of Quality Systems (BQS). Known amounts of three different sediment and soil standard reference materials (SRMs) obtained from the National Institute of Standards and Technology (http://ts.nist.gov/measurementservices/referencematerials/index.cfm) were added to dissolved-nutrient-fortified reagent water and samples were tested by three different methods; high temperature combustion (HTCO), Total Kjeldahl Nitrogen (TKN) and method I-4650-03. Analyses indicate a low bias in concentrations of total nitrogen (TN) in SRMs analyzed by method I-4650-03 over a wide range of environmentally relevant nitrogen and sediment concentrations. Further testing indicated that the bias in I-4650-03 is due to incomplete digestion of sediment or soil materials in SRMs by alkaline persulfate reagents.
During the development of method I-4650-03, reagent limitations with increasing carbon loading was fully tested and the results documented (USGS WRIR 03-4174). The effect of suspended sediment concentration on nitrogen recovery was not identified as an issue of concern at the time of method development and therefore, was not investigated.
The SRMs investigation was initiated to address historical evidence from National Stream Quality Accounting Network (NASQAN) samples that the total nitrogen in whole water samples was underestimated by the alkaline persulfate digestion method for samples with high suspended sediment concentrations. For five years at some NASQAN sites in the late 1990s, multiple methods were used on the same samples to determine total nitrogen. A working hypothesis was established that at some concentrations higher than 500 mg/L, the alkaline persulfate digestion method might become “reagent limited,” i.e., the persulfate reagent would be consumed before oxidation of all nitrogen species to nitrate was complete thereby causing low bias in total nitrogen concentration results.
We chose to evaluate this hypothesis using SRM sediment/soil samples with varying nitrogen content by weight that were prepared by BQS and analyzed by the NWQL. The results of these experiments revealed that some negative bias occurred at the lowest sediment concentration tested (100 mg/L). With the dissolved nitrogen held constant, the negative bias increased proportionately as sediment concentrations increased and the three SRMs demonstrated different bias behavior (see attached Phase 1 Data Summary).
The results suggest that some portion of nitrogen in these reference materials was recalcitrant to alkaline persulfate digestion. Further investigation revealed that the underestimated nitrogen was found in post alkaline persulfate digestion solid residues. The residues from alkaline persulfate digestion were analyzed by the HTCO method, thus revealing presence of the “missing” total nitrogen.
The results of SRM tests with alkaline Persulfate have led to field testing. In FY 2009, NAWQA, and NASQAN personnel are undertaking a field study using comparison sampling at select field sites. The objective of the field study is to determine the extent of this potential bias in river and stream samples. A complete evaluation and final recommendation await completion of data collection and analysis. NAWQA personnel expect that any observed negative bias will probably be site specific. Differences are expected to vary in magnitude and be related to the character and origin of suspended material at individual sites in the 2009 study. For additional details about this planned study, contact Charlie Crawford (cgcrawfo@usgs.gov), Surface Water Status and Trends Coordinator for NAWQA and NASQAN.
NWQL customers who rely on the alkaline persulfate digestion method for determining total nitrogen in whole-water samples may want to consider using alkaline persulfate digestion for total dissolved nitrogen (TDN, lab code 2754) and high-temperature combustion/oxidation for total particulate nitrogen (TPN, lab code 2607). In the absence of bias, the sum of alkaline persulfate digestion total dissolved nitrogen and TPN (lab code 2607) concentrations should equal alkaline persulfate digestion whole-water total nitrogen (lab code 2756) within the limits of sampling and laboratory errors. For questions pertaining to the data contained in the Phase 1 Data Summary, contact the NWQL at labhelp@usgs.gov.
If a project or Water Science Center decision is made to undertake comparability testing of method I-4650-03 with the HTCO method to determine bias in sample results, a sufficient number of paired samples should be collected to cover the expected range of suspended sediment concentrations characteristic of the site being sampled. If a consistent bias is found in samples from specific sites, adjustments could be made to historical data for long-term data analysis or, at a minimum, the magnitude of the bias could be estimated for previously collected data.
The results from total Kjeldahl nitrogen testing are not yet complete. A subsequent memorandum on the total Kjeldahl results will be issued when they are available.