A Study of the Analytical Variation of Sampling and Analysis of Stream-Sediments from Mining and Milling Contaminated Areas

During an environmental study of stream-sediments, large variations in metal values were noticed in contaminated sediments taken from the same site at different times. The question arose as to whether these differences were due to variations in contaminate input or merely represented the heterogeneity of the contaminated sediments at the site. Large numbers of replicate samples and measurements could be utilized to help answer the questions, but this is time consuming and costly. Random error can be estimated by the use of a duplicate analyses scheme and much use has been made of duplicate analyses in geochemical surveys and prospecting. In the work reported here, use was made of duplicate analyses to study Cu, Pb, and Zn variation in environmental stream-sediment samples. Samples were collected from three areas: (1) adjacent to the La Bajada mine along the Santa Fe River, (2) adjacent to the Pecos mine along the Pecos River, and (3) downstream from the Alamitos mill located on a tributary of the Pecos River. Both rivers are located in northern New Mexico. The mines and the mill have been inactive for many years. ANOVA statistics were used to evaluate differences in Cu, Pb, and Zn values between contaminated sites that were immediately adjacent. The large variations in metal values noted on earlier trips at the sites on the Pecos River were not noted in this study.

Analytical error was the greatest contributor to total variance. Differences in the adjacent site means were most apparent in the smallest size fraction (<63 mm). But even these differences were shown to be mostly due to random error. The smallest size fraction contained the highest metal values at all three sites. Stream-sediments at the La Bajada mine site did not appear to be contaminated.

Keywords: statistical analysis; analysis of variance; anova; mining and milling contamination; stream-sediments; copper; lead; zinc

Questions and Comments on this abstract to: Lynn Brandvold