Evaluation of nine USGS reference materials for quality control through Univariate Data Analysis System, UDASys3

Mauricio Rosales-Rivera,Lorena Díaz-González,Surendra P. Verma

DOI: https://doi.org/10.1007/s12517-018-4220-0

2019-01-01

Arabian Journal of Geosciences

Abstract:Data quality in any science plays a fundamental role to achieve valid inferences from experimental data. Inter-laboratory geochemical data for nine geochemical reference materials (GRMs) issued from the United States Geological Survey (USGS) were compiled from numerous literature sources. These data were processed in the third version of UDASYS (UDASys3; Univariate Data Analysis System), which applies automatically a refined statistical procedure to obtain both central tendency and dispersion parameters for univariate statistical data arrays and generates brief as well as extended reports. We present improved working values for central tendency and dispersion (total uncertainty) parameters for 10 major elements (SiO2 to P2O5), 14 rare earth elements (REE; La to Lu), and 42 (B to Zr; Ac to W) trace elements, along with LOI, CO2, H2O+, Cl, F, and S for these GRMs. Because the total uncertainty values of the mean reported in this work are generally lower than the literature uncertainties, the present statistical values are superior to those reported in all previous compilations. This implies that our statistical information will be more useful for instrumental calibration and quality control. An example is presented of the instrumental X-ray florescence spectrometric calibration, in which both sets of GRM concentrations and their 99% uncertainties (literature as well as those obtained from UDASys3) were used for the calibration of 10 major elements. The results show that UDASys3 provided generally more reliable calibration regression equations (higher linear correlation coefficient and lesser uncertainty in intercept and slope values) than the literature concentration estimates.

geosciences, multidisciplinary