Willow root distribution on riverbanks for soil and water bioengineering design: From field measurements to soil reinforcement estimation

The spatial and temporal distribution of soil reinforcement is closely linked to tree root arrangements in riparian soils which remains demanding to measure. Estimations of root-related soil reinforcement are used to predict bank stability, for example following the implementation of riverbank stabilization structures. In riparian environments, a frequently encountered pioneer species is willow ( Salix L.). Previous studies looking at willows root reinforcement in riverbanks mainly focused on young cuttings or on vegetation mixes. Here, we analyzed more than 2000 willow root measurements taken from 11 long-established sectors at two riverbank sites with low cohesive characteristics. The statistics of (horizontal and vertical) root distribution are presented alongside information on the respective contributions of the different root diameters to stabilizing effects and discussion of what type of reinforcement they can provide in this setting is discussed. Field data is then compared to the 1D Tron-model-based estimation of vertical root biomass, linking root distribution to groundwater level. Finally, we estimate the modification on critical shear stress of riverbanks induced by willows roots for the sites examined. The proposed three-step method yields promising results for the two field sites. The outcomes are highly relevant for practical applications, providing a quantification of the spatial distribution of willow roots in riparian soils and an assessment of their potential contribution to bank reinforcement, thereby supporting future operational use. • The root area ratio increases with willow age and root diameter. • Root density decreases as the vertical and horizontal distance from the stem collar increases. • On our sites with alluvial materials, groundwater levels drive willow root distributions. • In alluvial banks, root measurements provide a good fit with the estimation given by the 1D Tron-model.