# High-resolution XRF-CS/ICP-MS mineral element data calibration and potential applications in sub-Antarctic peat records

**Authors:** François De Vleeschouwer, Stephen J. Roberts, Gaël Le Roux, Thomas Bishop, Sarah J. Davies, Angela Gallego-Sala, Charlotte Green, Bianca Perren, Krystyna M. Saunders, Alex Whittle, Anjali L. Dhunna, Dominic A. Hodgson

PMC · DOI: 10.1038/s41598-026-41047-8 · Scientific Reports · 2026-02-26

## TL;DR

This paper introduces a new method to calibrate XRF data for mineral elements in peat cores, enabling detailed environmental studies in sub-Antarctic regions.

## Contribution

A novel multivariate calibration protocol for XRF-CS data using PLS models to improve quantitative analysis of mineral elements in peat records.

## Key findings

- A four-element PLS model (Ca, Ti, Sr, Zr) provides reliable predicted XRF-CS concentrations for Ti and Zr.
- Calibrated XRF-CS concentrations align well with ICP-MS measurements across five sub-Antarctic peatland sites.
- The calibration protocol expands the potential for high-resolution palaeoenvironmental research at decadal–centennial timescales.

## Abstract

Peatlands are important environmental archives and mineral dust trapped in peat cores from multiple sites can be used to track past changes in hemispheric and global wind circulation patterns. X-ray Fluorescence Core Scanning (XRF-CS) can rapidly geochemically characterise minerals deposited in peat at sub-millimetre-resolution, but calibration is needed to obtain quantitative data. Here, we present a unique calibration of > 14,000 contiguous mm-scale XRF-CS measurements depth-matched to 268 interval-based cm-scale Inductively Coupled Plasma Mass Spectrometry (ICP-MS) quantitative measurements from five peat records located on the west coasts of four sub-Antarctic islands impacted by the Southern Hemisphere Westerly Winds. Of eight calibration models tested, a four-element (Ca, Ti, Sr, Zr) multivariate partial least squares (PLS) model optimised for the widely used dust flux elements Ti and Zr accounts for covariance and provides the most reliable predicted XRF-CS concentrations for Ti (R2CV = 0.76, RMSEPboot = 2203 ± 705 mg kg− 1, R2pred. = 0.87, RMSEpred. = 2136 mg kg− 1, P < 0.0001). Predictions for Zr are indicative due to low Zr concentrations, but calibrated Ti and Zr XRF-CS concentrations align well with ICP-MS Ti and Zr concentrations for all five peatland sites. Our multivariate calibration protocol expands the scope of quantitative high-resolution palaeoenvironmental and geochemical research potentially to decadal–centennial timescales for peat records from, and beyond, the sub-Antarctic islands.

The online version contains supplementary material available at 10.1038/s41598-026-41047-8.

## Linked entities

- **Chemicals:** Ca (PubChem CID 271), Ti (PubChem CID 23963), Sr (PubChem CID 104798), Zr (PubChem CID 23995)

## Full-text entities

- **Chemicals:** XRF (-), CS (MESH:D002586)

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12988183/full.md

## Figures

4 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12988183/full.md

## References

1 references — full list in the complete paper: https://tomesphere.com/paper/PMC12988183/full.md

---
Source: https://tomesphere.com/paper/PMC12988183