PREDICTING 3D GRAIN SIZE FROM 2D IMAGE ANALYSIS: AN EXPLORATION OF STEREOLOGICAL CORRECTIONS IN SILTY SEDIMENT

Analysis of grain size (a defining attribute of sedimentary rocks and sediments) is a powerful tool for interpreting sediment transport histories, depositional environments, and paleoclimate. For most modern sediment, carbonates, and rock samples with calcareous or hematite cement, chemical disaggregation procedures enable measurement of sediment grain size using techniques such as laser particle-size analysis (LPSA). However, well-lithified samples (especially those with silica cement) resist disaggregation, and call for alternative 2D methods for grain-size determinations. Whereas grain-size analysis using LPSA enables precise and repeatable measurements of particles on a volume-based distribution, 2D assessment, e.g. measurement in thin section, or by image-analysis techniques of 2D views measure cross-sectional diameter (or area) that is fine-biased unless the 2D cut bisects the grain perfectly. Stereological corrections enable 3D interpretation of 2D measurements, but are limited in geological application owing to the size and shape variability of grains in natural systems. Several stereological corrections have been attempted, but no consensus approach exists. In this study, we show grain-size results using both 2D image-analysis techniques (Cameca SX100 electron probe micro-analyzer) and traditional 3D methods (Malvern Mastersizer 3000 laser particle size analyzer) from the same silt samples in order to refine an empirical best-fit curve of 2D data (quartz only) to “true” 3D data. If we are able to find consistency using this approach (e.g. “the coarsest 45% of grains from 2D methods is the best representation of the 3D curve”), then this “predictive” technique for stereology can be applied more broadly, namely in studies where sample disaggregation for a 3D measurement is not possible.