AbstractSediment samples tend to dry out during storage and are, therefore, stored refrigerated at about 4°C after wrapping in plastic foil. During XRF core scanning however, the samples must be taken out of their cover, increasing the risk of drying and formation of desiccation cracks on the surface. Because scan times can often amount to several hours and at highest resolution may take over a day to complete, the core will progressively dry out during scanning. With this study we aim to increase our understanding of how this slow drying of the samples during scanning and storage influences the XRF signal because of changes in water content, sediment surface topography, and the development of small, but slowly expanding cracks in the sediment core. Results show that the desiccation of samples during scanning and storage influence the XRF measurements in several ways. Most importantly, slow desiccation of the cores results in both a general lowering of the sample surface, and a shortening of the core due to shrinkage. Larger distance between sediment surface and detector leads to increased noise levels and poor reproducibility for many elements, while the shrinking of cores may shift individual data points between runs, resulting in poor reproducibility and offsets between datasets obtained at different times. Moreover, the loss of light elements, such as hydrogen and oxygen, can influence the matrix effect, especially for organic‐rich sediment. Because the XRF signals of individual elements are affected to different degrees, these changes may induce artificial shifts and biases in many elemental ratios commonly used for paleoenvironmental reconstruction.