Multi-element relationships and spatial structures of regional geochemical data from stream sediments, southwestern Quebec, Canada

Abstract

A regional geochemical survey, covering an area of 10000 km 2 in the Grenville Province of southwestern Quebec, was carried out by the Quebec Department of Energy and Resources. The fine (<177 μm) and heavy fraction of stream sediments were analyzed for more than 25 chemical elements by ICPES and by INAA. This study aims to identify the main geological processes affecting the geochemical data of both media, by considering the multi-element relationships and spatial structures of the geochemical data. Fine fraction and heavy mineral concentrates were subjected to: (1) principal component analysis (PCA) to determine the inter-element relationships controlling the background variations; and (2) variogram analysis to establish the spatial continuity of geological phenomena associated with the factors of the PCA. The stability of PCA results was tested by using a bootstrap replication technique. Detailed geochemical and mineralogical studies were conducted on different densimetric and granulometric fractions in order to characterize the geochemical behavior of each fraction, and to relate the inter-element associations to a specific mineralogy. Our study shows that the fine and heavy fractions indicate different geochemical information. The background levels and the multi-element associations differ between both media. A principal component model shows that the fine fraction is strongly influenced by surficial processes, explaining approximately 55% of the data variability. Two factors accounting for only 23% of the data variability in the fine fraction are related to lithologies. The heavy fraction reflects lithological characteristics. Three factors, accounting for 51% of the variation in the data, represent mineralogical features of the principal lithologic units. Two other factors reflect geochemical components that could be related to mineralization or lithologies of interest for prospecting. The variograms reveal the different spatial variation components of the geochemical signals. They show high nugget effect (more than 40%), indicating the erratic occurrence of phases hosting geochemical signals in the sampled media. They allow the spatial scales of geochemical components to be recognized, and the geological phenomena associated with regional trends to be differentiated from those acting at local scales. Knowledge of spatial variation components provides useful information for estimating suitable sampling intervals with regards to geochemical surveys, and aids in proposing solutions for reducing the nugget effect.