This thesis aims towards a more physically-based incorporation of fluvial sedimentology into process models, through the development and application of new techniques to measure and model fluvial sedimentology. The two new techniques presented in this thesis are Discrete Element Models (DEMs) and Terrestrial Laser Scanning (TLS). Chapter 2 presents a methodology for creating static DEM sediment beds, with are used in the thesis as a model of fluvial sedimentology. Many of the problems with incorporating fluvial sedimentology relate to the difficulties of measuring it, and therefore can be overcome through the use of DEM sediment beds. In order to assess these beds, a suite of field data was collected from the River Feshie. As part of this suite, TLS was used to capture high-resolution data of fluvial gravel surfaces, and a methodology for this is presented in Chapter 3. In Chapter 4, the DEM beds are utilised to produce a physically-based model of bedload grain size distribution (GSD). This model is designed to be sufficiently simple to be implemented within a reduced complexity framework such as might be used to model reach-scale dynamics. Chapter 5 assesses the use of DEM beds in the model formulation by determining the similarity between the DEM beds and fluvial gravels, through a comparison of the grain entrainment properties of the DEM beds and the field data. Chapter 6 continues this comparison by comparing the structure and roughness of the DEM beds and field data, and considers how these data could be used to parameterise roughness in a flow velocity equation.