Sediment porosity is a key structural parameter governing hydraulic, geomorphological, ecological and geotechnical processes in fluvial systems. Despite its importance, commonly used porosity predictors rely almost exclusively on idealized, randomly packed grain arrangements and therefore fail to reproduce the porosity of natural sediments, which commonly exhibit internal structures such as imbrication and stratification. This dissertation quantitatively investigates how these two structural features modify pore architecture and porosity in granular riverbed sediments, with the overarching goal of improving porosity prediction for natural fluvial deposits. To analyse imbrication, a new CT-based workflow was developed that allows efficient, non-destructive determination of three-dimensional grain orientation within laboratory and field samples. A unified quantitative imbrication index was formulated, enabling comparison across datasets. Laboratory flume experiments with mono-sized ellipsoids demonstrated that increasing imbrication strength leads to a measurable but limited decrease in porosity (≈0.03 absolute) due to only slight enhancements in grain-scale organization. In contrast, field samples from the Buëch River showed that natural imbrication can reduce in-situ porosity by approximately 30% relative to the same material in a randomly packed state. This difference is attributed to the elimination of large, stable pores that arise from particle interlocking in natural sediments. The influence of stratification was investigated through a combination of non-destructive laboratory measurements and DEM-based numerical modelling. A novel experimental setup was developed to capture the vertical porosity distribution within stratified sphere packings, revealing characteristic transition layers in which porosity decreases and subsequently recovers. The experiments were used to calibrate and validate a DEM model, which enabled systematic exploration of particle-size ratios and depositional scenarios. From the combined laboratory and numerical datasets, empirical relationships were derived to predict transition-layer thickness, minimum porosity and average porosity as functions of grainsize ratio. These relationships for stratified systems were integrated with existing random-packing predictors to quantify the influence of stratification on packing porosity, revealing that stratification can increase porosity by up to 57% on average compared to a well-mixed sediment of identical grain composition. Finally, a morphodynamic modelling case study demonstrated that porosity reductions or increases induced by imbrication and stratification can substantially influence predicted bed-level changes, underscoring the importance of incorporating structural effects into sediment transport models. Overall, this research provides the first comprehensive quantification of porosity variations arising from imbrication and stratification in fluvial sediments, introduces new measurement and modelling tools, and establishes empirical predictors that significantly enhance the accuracy of porosity estimation in structured sediment deposits.
W.Y. Xu (Thu,) studied this question.