Modified ASDPS for Geochemical Data Processing

Abstract

Geochemical data is firstly processed using various smoothing approaches to reduce random and/or systematic errors in the survey and/or analysis. In theory, a moving window varying from 2 × 2 to 9 × 9 can be used to smoothen the data with either a linear model or curved surface model. However, how to determine the ’best’ outcome from many smoothened datasets still depends on the analyser’s experience. In this paper, we propose a slightly modified ASDPS scheme to determine the likely ’best’ outcome among many smoothened datasets. This is achieved by introducing a statistics-based selector to determine the ’best’ smoothened result. The examples of processing Pb and other elements demonstrate that this modified ASDPS scheme is useful for geochemical data processing. The proposed statistical selector quantifies the determination of smoothing processing of geochemical data. This selection model also reveals that for geochemical data smoothing the 3× 3 surface model produces the closest result to its original data. However, the 5 × 5 surface model would become the right choice if the result of the 3×3 surface model is considered too ’fine’ whereas that of the 5× 5 linear model is too ’coarse’.