Are willows suitable for flood defense? Quantifying mechanical properties of willow species

Abstract

The use of temporary willow wetland forests as nature-based flood defense requires reliable predictions of branch resistance to loads, due to either hydrodynamic forces or strong winds, while accounting for variations across Salix species and varieties, growth environments (e.g., local salinity), and shapes (e.g., shrubs vs. trees, natural vs. pollarded). The Salix genus comprises hundreds of species, each exhibiting high intraspecific diversity with numerous subspecies and genotypes. However, to assess their suitability for flood defense, it is crucial to understand their mechanical properties, which have been documented for only a few species and often using differing methodologies. Quantifying and understanding the variation in mechanical properties of willow branches can improve vegetation-wave models and their use in flood risk assessments. We studied 18 Salix species, varieties and hybrids (shrub and tree type, including the most common species of riverine wetland forests), various specimens of one species (S. alba) along a salinity gradient and two growth forms: natural and pollarded. Sampling took place in Belgium and the Netherlands. We conducted 3-point bending tests and quantified absolute and relative flexibility and strength (maximum load and extension, modulus of elasticity (MOE) and modulus of rupture (MOR)). For all species, the thicker the branch, the greater its mechanical strength. Tree type species were more flexible with lower modulus of elasticity and shrub type species stronger with higher modulus of rupture. S. alba was amongst the most flexible species and S. purpurea the strongest. We found a large variability for natural branches and far less variability for pollarded branches. The salinity gradient had no effect on mechanical properties. Our study showed that the type of species and variety as well as pollarding affect the mechanical branch properties. The values of tree mechanics found in this study can be used as input for models, and their application in nature-based flood defense solutions.