In Reply Refer To: April 21, 1986 WGS-Mail Stop 412 OFFICE OF WATER QUALITY TECHNICAL MEMORANDUM 86.02 SUBJECT: PROGRAMS AND PLANS--Evaluation of Central Laboratories Nutrient Data Recent concern about the accuracy of nutrient data provided by the Atlanta Central Laboratory has led to an evaluation of data from both the Atlanta and Denver Laboratories. Data from the double- blind Quality Assurance Program (which is operated outside the Laboratory System) were statistically evaluated, and a special nutrient analysis study between the Denver and Atlanta Laboratories was conducted to determine whether WATSTORE files contain inaccurate nutrient analysis data and what corrective action might be taken. Results of these two independent studies indicate that: 1. Nutrient data produced by the Atlanta and Denver Central Laboratories have comparable means, but do not have comparable standard deviations. These data should not be expunged from WATSTORE. 2. Accuracy and precision of values reported for nitrite plus nitrate, orthophosphate, and phosphorus, are acceptable. 3. Precision of the TKN analysis (ammonia plus organic nitrogen) and ammonia analysis (NH4) is less than expected. 4. Investigators computing mass balances may need to reevaluate their data using Table 1 to assess error bounds. Evaluation of available information did not provide a definitive explanation of the problems causing the poor precision of determinations for ammonia and TKN. Data from the external double-blind Quality Assurance Program, evaluated for the years 1983, 1984, and 1985 (Table 1), indicate no bias between laboratories, but do indicate differences in precision. Samples for quality-assuring nutrient determinations are prepared in single batches, then split and submitted within the same month to both laboratories. Analysis of data from reanalyzed samples generally indicates that the nitrogen species in the reference samples are stable for at least 6 months. The pooled standard deviations shown in Table 1 should accurately represent the precision of reference sample data and be useful in assessing the minimum error bounds for dissolved and total nitrogen species determinations in hydrologic investigations. Table 1. Mean annual concentration range (mg/L), number of determinations, and pooled standard deviations (mg/L) for external QA samples. NO2 + NO3 NH4 TKN __________________ ____________________ _________________ Range N s.d. Range N s.d. Range N s.d. 1983 ATL 0.5-5.3 383 .280 0.16-2.67 383 .141 0.2-3.2 383 .346 DEN 0.5-4.5 459 .250 0.20-2.72 456 .148 0.6-3.9 459 .457 1984 ATL 0.3-4.3 377 .219 0.16-3.15 377 .234 0.3-4.7 377 .470 DEN 0.3-4.5 464 .179 0.16-3.53 464 .294 0.4-5.1 462 .272 1985 ATL 0.6-4.5 305 .388 0.04-1.69 306 .101 0.6-2.8 299 .505 DEN 0.5-4.4 467 .375 0.05-1.74 467 .047 0.4-3.2 466 .285 These two evaluations of the worth of nutrient data reported by the Central Laboratories System prompts us to reconsider the reporting level for ammonia and TKN analysis. It is clear that the currently-stated reporting limit of 0.1 mg/L is not consistent with the standard deviation of performance over the range of the method. A technical memorandum will be issued soon to restate the reporting limit. Copies of the two evaluation reports, "Quality Assurance Audit of the Atlanta Laboratory Nutrient Data" and "Nutrient Analysis Interlaboratory Study," may be requested from this office. David A. Rickert Distribution: A, B, FO, PO Key Word: This memorandum does not supersede any previous memorandum.