קולוקוויום בחוג לגאופיזיקה: New Perspective on River Dynamics Across Scales

Zoom: https://tau-ac-il.zoom.us/j/85032703038

Abstract:

Rivers are dynamic systems that evolve across a wide range of spatial and temporal scales. At the smallest scales, turbulence controls sediment entrainment and transport; during floods, erosion and deposition reshape river channels; and over longer timescales, these processes drive landscape evolution. Bridging these scales remains a fundamental challenge in fluvial geomorphology because the relevant processes are difficult to observe continuously and across large spatial extents. In this talk, I will present three research projects that use geophysical observations, including topo-bathymetric LiDAR and passive seismic monitoring, to investigate river dynamics across scales.

First, I will present a new workflow for processing airborne topo-bathymetric LiDAR data in challenging river environments (Nativ et al., 2026; EGUsphere). The approach combines full-waveform reanalysis, supervised Random Forest classification (3DMASC), and unsupervised classification of deep LiDAR echoes to distinguish river-bed points from water-column noise and reconstruct continuous bathymetric surfaces. The workflow is demonstrated using a unique dataset acquired along 28 km of the Ardèche River (southern France), where fast-flowing, turbid waters confined within a bedrock canyon were surveyed in October 2021, producing approximately 1.1 billion 3D points. The final high-resolution topo-bathymetric product comprises 42 million classified points, including 18 million bathymetric points, with interpolation limited to a few deep pools (>4.5 m) and a median bathymetric density of 21.7 pts/m². Compared to standard airborne LiDAR processing, the proposed workflow increased depth penetration by 55% (2.9 to 4.5 m), bathymetric coverage from 70% to 86%, and the proportion of channel sections with near-complete coverage (≥95%) from 32% to 87%.

In the second part, I will present a complementary dataset acquired 17 days after the initial survey, following a major flood with a peak discharge of ~2,000 m³/s (approximately a 10-year recurrence event). This rare before-and-after dataset provides a unique opportunity to investigate how large floods reshape river morphology through erosion and sediment deposition. To quantify the hydrodynamic drivers of geomorphic change, I simulated the complete flood hydrograph using HEC-RAS at 1-m resolution. Initial analyses reveal that direct grid-to-grid comparisons between flood forces and erosion–deposition patterns fail to explain the observed morphological response. I therefore developed an object-based Random Forest framework that analyzes alternating sequences of erosional and depositional features (EDFs), enabling the relative influence of channel morphology, flood hydraulics, EDF geometry, and upstream sediment supply to be quantified across scales.

Finally, I will demonstrate how high-frequency seismic monitoring (>100 Hz) can reveal river processes that are otherwise difficult to observe directly, including turbulence and sediment transport. I will present a newly developed framework that combines passive seismic observations with Kolmogorov's phenomenological theory of turbulence to classify turbulent flow conditions and bedload transport. This approach offers a new pathway for monitoring river dynamics continuously and non-invasively, bridging the gap between flow-scale processes and longer-term geomorphic change.

​

מארגן האירוע: ד"ר ליאור רובננקו