Estimation of sediment texture from spectral induced polarisation data using cluster and principal component analysis

Abstract

Spectral induced polarisation data are usually interpreted with simple models in order to derive petrophysical relationships between electrical and sedimentological properties, such as texture, clay content, and permeability. The aim of this work is to explore the value of spectral induced polarisation in addition to conventional direct‐current resistivity measurements for determining textural properties of saturated samples collected from alluvial deposits. For this, an advanced data processing approach that combines cluster and principal component analysis was developed and applied to integral parameters derived from Debye decomposition of spectral induced polarisation data. This data processing procedure allowed identifying groups of samples with a similar spectral induced polarisation response and to derive a characteristic grain‐size distribution for each group of samples. The method to estimate the grain‐size distribution from spectral induced polarisation data was successfully validated using independent sediment samples. The remaining uncertainty in the estimation of sediment texture from spectral induced polarisation data was attributed to the effect of pore size distribution and mineralogy, which were not considered in the present work but can be added in the future within the same conceptual workflow.

Integrated Methodology to Link Geochemical and Geophysical-Lab Data in a Geophysical Investigation of a Slag Heap for Resource Quantification

“…Therefore, here we assume that the 22 samples collected at different positions (x, y, z) represent ground truth data enough to capture the physical and chemical properties of the slag heap overall. Note that in the literature we can find sampling to be challenging and often samples are collected at surface, at several sites and/or at few locations (x, y) (Inzoli et al, 2016;Martin et al, 2021;Florsch et al, 2011). Few available samples may not represent the lateral nor vertical heterogeneity of the waste deposit, and consequently the resulting classification of the field data may only be valid in the vicinity of the samples positions.…”