Estimating Small Scale River Channel Roughness Using a Through-water Photo-based technique

Abstract:

Channel roughness is critical to the understanding of fluvial geomorphology and hydrology due to its connection with the transportation of sediment and effect on flow discharge. Due to manual measurement methods being costly and time consuming, and traditional visual observation methods being subjective, we have explored the use of a close-range remote sensing approach, based on through-water photography to estimate channel characteristics. Previous similar photo-based measurements have focused on estimating water depth by correcting data from stereo image pairs for refraction at the water surface. Here, we extend this approach to multi-image data sets, and implement refraction correction for data from commonly used ‘structure from motion’ based software. The accuracy of applied corrections is assessed in a laboratory setting using a gravel surface submerged at a range of water depths. We demonstrate the approach in the field by photographing cross sections to produce high density point clouds and hence digital elevation models of the stream bed. Correcting submerged regions for refraction effects allows channel characteristics such as wetted perimeter and water depth to be estimated. We explore the use of parameters thus estimated for deriving coefficients of channel flow resistance such as Manning’s coefficient.