Using Concatenated Profiles from High-Speed Laser Profile Scanners to Estimate Debris-Flow Characteristics: A Novel Approach Based on Particle Image Velocimetry

Abstract:

We use globally unique datasets from paired laser profile scanners to measure debris-flow height, velocity and discharge in two well-known debris-flow channels in Switzerland. Since 2011, these scanners have been scanning passing debris flows at rates of up to 75 Hz, acquiring millions of cross-bed profiles. The profiles can be concatenated through time, generating unique 2.5D representations of passing debris flows. Applying a large-scale Particle Image Velocimetry (PIV) approach to these datasets has proven successful to measure surface flow velocities. Flow height can also be estimated from the laser scanners, and thus a discharge estimate can be given. To account for changes to the channel bed due to erosion and deposition during the debris flow, we compute two flow height estimates using a pre-event as well as a post-event channel geometry in order to visualize discharge variability.
Velocity outliers need to be excluded to provide reliable estimates of peak discharge, and changes to the channel bed are assumed to be the largest source of uncertainty. However, the latter problem is inherent to all debris-flow discharge measurements, and we have found the new system to offer distinct advantages over the conventional system relying on geophones and a radar gauge. The wide scan angle of up to 190° renders the scanners insensitive to changes of the flow path, and the point density of roughly 20 points per meter offer unprecedented spatial coverage.
In addition, the geometries of the cross-bed profiles have been analyzed, revealing distinct changes of cross-flow convexity between the front and the tail of the flows in several cases. This is assumed to indicate changes of debris-flow mixtures, but further research is needed to better understand this signal.
We hope that our preliminary analysis and toolbox will facilitate working with these kinds of datasets so as to further improve debris-flow understanding, monitoring and modeling efforts in the future.