Improving anomaly selection by statistical estimation of background variations in regional geochemical prospecting

Abstract

The first part of this study, based on numerous data collected during the Mineral Resources Inventory Programme in France, investigates the influence of environmental factors on geochemical background in regional prospecting. Within the 30 lithological units selected in basement areas of the Massif Armoricain, Vosges and Massif Central, two main differentiation factors are commonly identified: (1) a dilution effect of trace elements by a barren siliceous phase related to various environmental parameters such as the nature of the substratum and overburden, or the type of material sample; (2) a coprecipitation effect of Zn, Ni, Co, Cu and P with Fe-Mn hydroxides, marked by a frequent association between these elements and, in Brittany, by their enrichment in stream sediments compared to the soils. Both factors appear to be closely related to the mineralogy of the samples, and are better characterized by multielement analytical data than by coded field observations. In the second part of the study, an example of Zn background estimation and correction is given for 1530 samples collected in the Nort-sur-Erdre district (Brittany). A simple technique of regression by local estimation, known as “neighbourhood regression”, has been applied on Zn in the chemical space of a suite of “background characteristic” elements defined previously. A reappraisal of the Zn anomalies is carried out on the residual component of Zn and, in order to give importance to the most diluted samples, this index is also normalized, either by the raw Zn value or by its local standard deviation. Synthetic mapping of these various indices helps in the final stages of anomaly interpretation and selection.