Computer software for identifying compositional subpopulations in marine sediment geochemical data using threshold value analysis

The material of marine sediments is commonly derived from multiple sources or processes. Consequently, individual sediment samples can be viewed as mixtures of two or more distinct geochemical subpopulations. Certain quantitative procedures such as threshold value analysis are capable of isolating and identifying the subpopulations that are present within a suite of sediment samples through an analysis of bulk sediment compositional data. Thus, these methods are useful in mineral exploration programs because they facilitate the task of discriminating between samples containing background versus anomalous chemical signals. This study reports the development and testing of computer software (ProbabilityGrapher) for threshold value analysis using a probability graphing technique. This program is designed for microcomputers so that it can be readily applied in field situations where the rapid in situ reduction and interpretation of geochemical data is important to the planning and continuation of an exploration survey. The software can also be used in concert with the Q_—LM software package previously developed for Q‐mode factor analysis and linear modeling, thus providing a powerful combination of data reduction and interpretation tools for investigating sediment composition information.