Recovery of Incomplete Rare Earth Element Datasets – A New Approach to REE Data Evaluation

Rare earth elements (REEs), a group of elements with similar physical properties and coherent geochemical behaviour, are widely used as proxies for many biogeochemical processes. Interpretation of REE data is primarily based on distribution patterns in normalised graphs. Therefore, missing REEs can alter the appearance of the REE patterns and, consequently, the interpretation. Data for certain REEs may be missing for various reasons, e.g., they could not be measured (neutron activation analysis, isotope dilution techniques), the measurement was near to or below the limits of quantification or certain REEs were used as spikes. To address this, we introduce a novel method that leverages REEs' characteristically smooth distribution patterns to reconstruct missing REE data. Our approach provides accurate and precise REE data (<10% deviation) well within common analytical uncertainties of modern analytical techniques (e.g., ICP-MS). The accuracy and precision were determined using a method verification dataset of >13,000 mafic and ultramafic rock samples. The re-modelled REE data can be used to interpret the overall pattern and to quantitatively determine anomalies, one of the most important tools in REE research. Furthermore, our method offers new opportunities for REE data handling and processing, enabling researchers to assess the usability and reliability of REE data, whether self-produced or data from journal publications and data repositories. We implemented our method into our software tool GeoArmadillo, which facilitates geochemical data processing, evaluation and assessment.