Reliability assessment and calibration of elemental signal values by XRF core scanning in Qinghai Lake

XRF core scanning has been extensively employed for semi-quantitative analysis of elements in various sediment types over the past three decades, particularly in lacustrine deposits due to its rapid, continuous, non-destructive, and high-resolution advantages. However, despite the susceptibility of element signal values obtained through XRF core scanning to instrument settings and core physical properties, there remains a scarcity of comprehensive evaluation regarding data reliability and calibration effects. In this study, a 2.39-m–long sedimentary core from Qinghai Hu (Lake) (QHH) was obtained for high-resolution scanning using an XRF core scanner. Physical and chemical characteristics in water content, grain size distribution, loss on ignition, and actual elemental composition were analyzed for each subsample. Moreover, the accuracy of element signal values and ratios by XRF core scanning and their influencing factors was effectively assessed, and the reliability of calibration results was simultaneously calibrated using internationally recognized models such as Normalized Median-scaled Calibration and Multivariate Log-ratio Calibration (MLC). Results demonstrate that the Zr signal values corresponded accurately to the actual contents in the sediment core sequence, while weak correlations were observed for Si and Ti, indicating their limited significance. Additionally, the presence of higher water content in the core sections significantly attenuated in signal intensity and fluctuation amplitude for elements of Al, Si, K, Ca, Ti, Fe and Mn. Reversely, dry core sections exhibited greater fluctuations in signals of above elements due to high-resolution scanning and variations in particle composition, thereby attenuating their correlations with actual concentrations. Trace elements of higher atomic weights, such as Rb, Sr, and Zr, demonstrated reduced susceptibility to the variations in water content and particle composition in terms of signal distributions. Finally, using the ratio between adjacent elements based on the XRF core scanning was proven a highly effective approach for quickly eliminating the consistent influence of multiple factors. Furthermore, the multivariate log-ratio calibration (MLC) model exhibited superior calibration effects on individual element signal values throughout the QHH core and within each core section. These findings not only offered valuable reference to the scientific application of high-resolution data acquired by XRF core scanning for lake sediments, but also established a foundation for the reconstruction of climate change and for comprehension of human-environment relationships in the northeastern Tibetan Plateau.