Abstract

This study investigates the impacts of sample preparation procedures on grain-size measurements to determine comparability of data collected using differing methodologies. Grain-size distributions of marine and terrestrial sediments contain important information about the depositional environment. For example, the “sortable-silt index” ( or mean grain size between 10 and 63 μm in marine sediments, is used as an indicator of flow speed and has been applied to the reconstruction of ocean current strength before the instrumental period. Similarly, the mean grain size of a sediment is used to classify it (e.g., silt versus sand). Accurate measurements of grain-size distributions often require chemical pre-treatments in order to remove sedimentary components of biogenic origin (e.g., shells), and multiple ways to perform these pre-treatments exist. This study tests whether the choice of pre-treatment introduces variability into grain-size distributions. We simulate multiple commonly used pre-treatments on a well-characterized internal standard (“Sillikers”) and compare the resulting mean size and sortable-silt index in each treatment group to untreated samples using ANOVA. Two instruments, a Coulter Counter Multisizer III and a Coulter LS 230 Laser Diffraction Analyzer are used. Results from the Multisizer III suggest that the choice of pre-treatment method does not significantly impact the final grain-size distributions but underlines the importance of replicates. Results from the laser sizer suggest that oven-drying leads to a small but statistically significant difference of ∼ 0.3 μm in the sortable-silt index, and drying samples via hot plate leads to another small but statistically significant difference of ∼ 0.29 μm. While it is unclear what causes these differences in the laser sizer data, they are smaller than the observed variations in sortable-silt index used to infer changes in current speed in a typical paleoclimate study. In conclusion, grain-size measurements are a robust tool for sediment analysis and are resistant to changes from differing pre-treatment methods tested here.