Grain-size distribution of volcaniclastic rocks 2: characterizing grain size and hydraulic sorting

Quantification of the grain size distribution of sediments allows interpretation of processes of transport and deposition. Jutzeler et al. (2012) developed a technique to determine grain size distribution of consolidated clastic rocks using functional stereology, allowing direct comparison between unconsolidated sediments and rocks. Here, we develop this technique to characterize hydraulic sorting and infer transport and deposition processes. We compare computed grain size and sorting of volcaniclastic rocks with field-based characteristics of volcaniclastic facies for which transport and depositional mechanisms have been inferred. We studied pumice-rich, subaqueous facies of volcaniclastic rocks from the Oligocene Ohanapecosh Formation (Ancestral Cascades, Washington, USA), Pliocene Dogashima Formation (Izu Peninsula, Honshu, Japan), Miocene Manukau Subgroup (Northland, New Zealand) and the Quaternary Sierra La Primavera caldera (Jalisco State, Mexico). These sequences differ in bed thickness, grading and abundance of matrix. We propose to evaluate grain size and sorting of volcaniclastic deposits by values of their modes, matrix proportion (< 2 mm; F-1) and D₁₆, instead of median diameter (D₅₀) and standard deviation parameters. F-1 and D₁₆ can be uniformly used to characterize and compare sieving and functional stereology data. Volcaniclastic deposits typically consist of mixtures of particles that vary greatly in density and porosity. Hydraulic sorting ratios can be used to test whether mixed clast populations of pumice and dense clasts are hydraulically sorted with each other, considering various types of transport underwater. Evaluation of this ratio for our samples shows that most studied volcaniclastic facies are deposited by settling from density currents, and that basal dense clast breccias are emplaced by shear rolling. These hydraulic sorting ratios can be applied to any type of clastic rocks, and indifferently on consolidated and unconsolidated samples.