Evaluation of DEM size and grid spacing for fluvial patch-scale roughness parameterisation

Abstract

Surface roughness is a term used in fluvial research without a unanimous definition. Clarification of the term and improved parameterisation is needed in future research. Improvements to the collection of topographic data, using photogrammetry, have provided accurate digital elevation models (DEMs) of field and laboratory gravel-bed patches of varied sediment size and surface structure. In this study, a moving-window process is used for analysing spatial variability within DEMs. Using this information, and in unison, the effect of DEM size and grid spacing are considered on an extensive range of roughness parameters, in order to provide insights for obtaining grain-roughness statistics. It is shown that DEM size influences the calculated roughness statistics, and the observation of plateaus in statistics for DEM window sizes above 16 × D50A in both directions (where D50A is the median grain size of the bed-surface material) suggests this as a minimum DEM size for grain-scale roughness analysis. Further, the DEM grid spacing should be 1 mm or below, in order to adequately capture grain roughness, as coarser resolutions failed to detect particle imbrication. Finally, variability in roughness parameters was evident due to natural spatial variation in gravel-bed microtopography, suggesting using a single roughness parameter is not appropriate to holistically describe the roughness of a gravel patch. As the importance of topographic information continues to grow in geomorphology, the results of this study will assist researchers in the collection and subsequent analysis of DEMs.