Grain size is one of the most fundamental properties of sediments. It is frequently used in paleoclimate, paleoceanographic, and paleoenvironmental research as a proxy for river discharge, current and wind strength, and to identify mass flow deposits. Measuring grain size is, however, time-consuming and destructive. Given the strong influence of grain size on sediment inorganic geochemistry, single-element variations measured by, for example, X-ray fluorescence core scanning are increasingly used to estimate grain-size variations at high resolution in sediment cores. This approach is however limited to a narrow grain-size range since individual elements only monotonically relate to grain size over a narrow size range. Here, we present a simple, code-free, multielement method based on partial least squares regression to predict sediment mean grain size from inorganic geochemical data over the range of sizes commonly encountered in sedimentary basins (clay to sand). The method was first tested using river sediment samples separated in 11 grain-size fractions, and it was later successfully applied to two sediment cores from the Chilean fjords. Our method only requires measuring grain size on a limited number (around 10) of selected training samples, and it allows to predict mean grain size at X-ray fluorescence core scanner resolution. This method has the potential to be applied to any lake or marine sediment core, provided sediment provenance, weathering, and diagenesis remain relatively stable through time, and we anticipate that it will result in a significant increase in the resolution of sediment proxy records of climate and environmental change. Plain Language Summary Sediment grain size is one of the best indicators of past transport conditions. Here we present a method to predict mean sediment grain size from inorganic geochemical measurements. This method can be applied to geochemical measurements obtained by X-ray Fluorescence core scanning to generate grain-size profiles quickly and at high resolution, and we anticipate that it will result in much improved paleoclimate and paleoenvironmental reconstructions. |
