资源环境科技发展态势分析平台

Steep streams with massive sediment supply are among the most complex systems to study, even in the laboratory. Their shallow sediment-laden flows create self-adjusting bed geometries that evolve rapidly. Often, morphological changes and flow processes cannot be dissociated. Because these very shallow and unstable flows cannot be equipped with measurement sensors, image analysis techniques, such as photogrammetry (e.g., structure-from-motion, SfM) and large-scale particle image velocimetry (LSPIV), are interesting options for capturing the characteristics of these systems. The present work describes a complete procedure using both techniques to measure spatially distributed surface velocity and bed properties (deposit patterns, channel slope, and local roughness). The velocity data are used to assess the local flow directions along which the channel slope and roughness are extracted from the SfM digital elevation models. Fergusons variable power equation friction law, having been previously validated by comparison with approximately 100 local flow depth measurements, was used in a second step with the collected data to reconstruct a complete mapping of the depth-averaged flows, thereby enabling a comprehensive analysis of the hydrogeomorphic system where shallow water equations apply. The assumptions, details, use of the friction law with roughness standard deviation rather than diameter as parameter and limitations of the procedure as well as possible sources of errors are discussed here, along with possibilities for improvements. This affordable and simple-to-implement procedure can provide a large amount of data, allowing for a more comprehensive analysis of complex hydraulic systems.

Plain Language Summary Steep-slope stream hydraulics are complicated and insufficiently known because fast, active, mobile, and profoundly correlated with massive sediment transport processes. Capturing data on such flows is complicated, although necessary to develop and validate models. Here we propose a combination of two image analysis methods (SfM and LSPIV) and an inverse procedure using a friction law to acquire thousands of flow observations. This affordable and simple to use method gives promising perspectives for other application in the still widely used flume experiments.