Search

Geochemical maps provide invaluable evidence to guide decisions on issues of mineral exploration, agriculture, and environmental health. However, the high cost of chemical analysis means that the ground sampling density will always be limited. Traditionally, geochemical maps have been produced through the interpolation of measured element concentrations between sample sites using models based on the spatial autocorrelation of data (e.g. semivariogram models for ordinary kriging). In their simplest form such models fail to consider potentially useful auxiliary information about the region and the accuracy of the maps may suffer as a result. In contrast, this study uses quantile regression forests (an elaboration of random forest) to investigate the potential of high resolution auxiliary information alone to support the generation of accurate and interpretable geochemical maps. This paper presents a summary of the performance of quantile regression forests in predicting element concentrations, loss on ignition and pH in the soils of south west England using high resolution remote sensing and geophysical survey data. Through stratified 10-fold cross validation we find the accuracy of quantile regression forests in predicting soil geochemistry in south west England to be a general improvement over that offered by ordinary kriging. Concentrations of immobile elements whose distributions are most tightly controlled by bedrock lithology are predicted with the greatest accuracy (e.g. Al with a cross-validated R 2 of 0.79), while concentrations of more mobile elements prove harder to predict. In addition to providing a high level of prediction accuracy, models built on high resolution auxiliary variables allow for informative, process based, interpretations to be made. In conclusion, this study has highlighted the ability to map and understand the surface environment with greater accuracy and detail than previously possible by combining information from multiple datasets. As the quality and coverage of remote sensing and geophysical surveys continue to improve, machine learning methods will provide a means to interpret the otherwise-uninterpretable. [ABSTRACT FROM AUTHOR]

The requirements for quantitative data in geological surveillance are ever increasing; traditional geological maps and 3D models are evolving into quantitative conceptual models based on a broad range of analytical measurements of surface and subsurface properties. The British Geological Survey's Geochemical Baseline Survey of the Environment (G-BASE) project provides one such source of data: national coverage of stream sediment, stream water and soil geochemistry. While this geochemical data is of obvious benefit to our understanding of the present day surface environment, the rich geological diversity of south west England provides an ideal setting in which to showcase the potential of such data to enhance our understanding of the underlying bedrock. In this paper we use compositional data analysis to extract lithostratigraphic information from G-BASE stream sediment data. We find that variations in G-BASE stream sediment geochemistry correspond very closely to mapped variations in bedrock geology. Geochemical variations between the 16 lithostratigraphic domains into which we classify the region provide insight into the distinct geological histories of the lithologies within each domain, particularly in relation to depositional environments and sediment provenance. [ABSTRACT FROM AUTHOR]