Bioengineering of slope using plants has been recognised as an environmental-friendly solution for shallow slope stabilisation. It is evident that plants with different root geometries can provide different degrees of mechanical reinforcing effects and the amount of soil suction induced by transpiration. Although geotechnical centrifuge has been increasingly used to study the effects of vegetation on slope stability, existing studies mainly focus on the use of plant roots to stabilise soil slopes mechanically. Effects of induced soil suction, which can increase shear strength of soil and reduce its water permeability, are generally ignored. Furthermore, the influence of root geometry and its effects on the magnitude and distribution transpiration-induced suction and hence on slope stabi...[ Read more ]

Bioengineering of slope using plants has been recognised as an environmental-friendly solution for shallow slope stabilisation. It is evident that plants with different root geometries can provide different degrees of mechanical reinforcing effects and the amount of soil suction induced by transpiration. Although geotechnical centrifuge has been increasingly used to study the effects of vegetation on slope stability, existing studies mainly focus on the use of plant roots to stabilise soil slopes mechanically. Effects of induced soil suction, which can increase shear strength of soil and reduce its water permeability, are generally ignored. Furthermore, the influence of root geometry and its effects on the magnitude and distribution transpiration-induced suction and hence on slope stability is not considered. The principal objectives of this thesis are to develop a new modelling technique to simulate both mechanical and suction effects of plant roots in a geotechnical centrifuge and to utilise this new technique to quantify and compare the effects of different root geometry on slope stability with dual consideration of hydrological effects of transpiration on suction and mechanical effects of root reinforcement.

A series of centrifuge tests were conducted to investigate induced and retained soil suction and the stability of vegetated silty sand model slopes (45° and 60°) subjected to various rainfall events. To simulate both mechanical and hydrological contribution of vegetation, novel artificial roots with different geometries (i.e., tap, heart and plate) developed were installed in these model slopes. Each artificial root is made of porous material and connected to a vacuum source, which can create a total head gradient between the artificial root and the surrounding soil. By controlling total head difference, soil moisture can be reduced to increase soil suction in a model slope. The porous material has a high air-entry value and is water-saturated for maintaining hydraulic connection between the root and the surrounding soil. The tensile strength, interface friction angle and axial rigidity of each are also scaled according to those of real roots. Induced suction by artificial roots was compared with laboratory and field measurements. To determine the factor of safety (FOS) of vegetated model slopes, measured suction was then back-analysed through a series of finite element seepage analyses followed by stability analyses through strength reduction method.

Suctions induced by the new artificial roots are found to be consistent with those observed in laboratory and in the field. It was revealed that after an extreme rainfall event up to 1000-year return period (i.e., 70 mm/hr for 8 hours in prototype), 45° model slope remained stable, irrespective to different root geometries. It is evident that heart-shaped root consisting of a taproot component and two horizontal branches was the most effective one to improve the slope stability. One of the reasons is that it has a larger hydraulic contact area with the surrounding soil resulting in the highest retained suction within the root depth. Another reason is that the horizontal branches of heart-shaped root can provide 12% higher pull-out resistance than the tap-shaped root. The slope supported by heart-shaped roots, therefore, has 16% and 28% higher FOS than that supported by the tap- and plate-shaped roots, respectively. For 60° steep slope, slope failure happened after 3 hours of rainfall (20-year return period) in the model slope with tap-shaped roots, whereas the stability of slope supported by heart-shaped roots could be maintained up to 5 hours (100-year return period). Under the same rainfall return period, the heart-shaped root also provided smaller volume of failure mass and shorter runout distance as compared to the slope supported by tap-shaped roots. The heart-shaped root is thus the most favorable geometry for improving slope stability, among the three geometries considered.