The objectives of the present study were twofold: (1) to identify spatial sedimentation and erosion patterns developing within patches of epibenthic structures (i.e. physical structures that protrude from the sediments, originating either from animals or plants) as a consequence of biophysical interactions; and (2) to assess the relevance of hydrodynamic flume studies for the long-term sediment dynamics in the field. We addressed these objectives by using patches of well-defined artificial structures (bamboo canes) for which we could easily monitor the long-term sediment dynamics in the field, measure the hydrodynamic effects in detail in the flume, and simulate the field and flume set-up with a commercially available hydrodynamic model. Two-year monitoring in the field showed that sedimentation was much larger in the high-density patches than the low-density ones. Within the high-density patches, comparable spatial patterns emerged at different field sites: erosion at the front and the side of the patches, and sedimentation more down-stream within the patches. The low-density patches showed no such patterns, and were generally characterised by some small-scale erosion directly around individual bamboo canes. Sedimentation and erosion in the field was well explained by the patterns in bed shear stress that were derived from our flume measurements. The 3D hydrodynamic modelling facilitated up-scaling of the flume results to the field, but failed to simulate accurately the effects at the leading edge. We conclude that: (A) field observations on sedimentation revealed interesting spatial patterns, but could not elucidate underlying processes; (B) detailed hydrodynamic measurements in a flume can elucidate these underlying processes, provided that appropriate scaling is being used; (C) flume studies are by definition not able to capture all spatial scales that are relevant for estuarine landscape formation and will always cause some flow artefacts; (D) hydrodynamic modelling offers a valuable tool to upscale flume observations, even though present models are not yet capable of fully reproducing all detailed spatial patterns; and (E) spatial heterogeneity is very important when looking at small-scale patches. There is a need for more spatially explicit and scale-dependent knowledge on bio-physical interactions.