Spatial variability of erodibility of fine sediments deposited in gravel river beds: from field measurements to 2D numerical models.

Purpose: Gravel-bed rivers can store significant amounts of fine sediments, in the gravel matrix or at the bar surface. The contribution of the latter to suspended sediment fluxes depends on their erodibility which is highly variable spatially. The sensitivity induced by this spatial variability on outputs of a 2D hydro-sedimentary numerical model was investigated and recommendations for in situ erodibility measurement strategy were provided. Methods: The spatial variability of fine sediment erodibility was determined using the Cohesive Strength Meter (CSM) device in a 1-km-long river reach of the Galabre River in the southern French Alps. A 2D hydro-sedimentary numerical model was built on the monitored reach displaying three deposit zones with distinct erodibility values. The sensitivity of the modeled eroded masses to sediment erodibility variability was assessed through ten distinct sediment erodibility settings and three schematic flood events, based on the in situ monitoring of the river. Results and discussion: The spatial variability of fine sediment deposit erodibility was significant. Marginal deposits were more resistant than superficial or water-saturated ones. The sensitivity of the modeled eroded mass to erodibility parameters was different depending on the set of measurements used. When considering the entire dataset, which exhaustively characterizes the fine sediment deposits, the numerical sensitivity was relatively low. On the other hand, when a partial set of measurements outside the quartiles was considered, the sensitivity was more significant leading to large differences in eroded masses between spatially distributed and spatially averaged settings. Using bootstrap sampling, we recommended making 15 to 20 measurements in marginal and superficial zones to adequately capture the distribution of erodibility. Conclusions: This work provided insight on the spatial variability of erodibility and the sensitivity induced in 2D numerical modeling of fine sediments. The proposed methodology could be applied to other environments (e.g., reservoirs, estuaries, or lowland rivers) in order to adapt the monitoring and numerical modeling strategies. [ABSTRACT FROM AUTHOR]