Large-scale triaxial tests of vegetated soil at low confining stresses

Abstract

<p>The focus of geotechnical researchers and practitioners has recently been on the impact of vegetation on the mechanical behaviour of the soil as nature-based techniques against erosion and landslides. Although numerous laboratory studies have already been produced on this subject, there seems to be a lack of discussion on the significance of the results in relation to the representative elementary volume (<em>REV</em>) used. An excessive or scarce root/soil ratio can result in over- or underestimation of the strength of the soil specimen tested, respectively. In addition, a root/soil ratio very different from that which the plants have in-situ would risk making the laboratory results difficult to upscale to the slope or catchment level. To this end, the aim of this study is to present triaxial compression tests of large vegetated soil specimens (<em>h </em>= 400 mm <em>Φ </em>= 200 mm).</p><p>Silty sand was used and statically compacted at a dry density <em>ρ<sub>d</sub></em> = 1.60 Mg/m<sup>3</sup> and at a water content <em>w</em> = 15%. Samples were then thoroughly poured with water up to a high degree of saturation (<em>S<sub>r</sub></em> ≈ 0.95). Eight of them were seeded with <em>Cynodon dactilon, </em>maintaining fixed seeding spacing and density. Samples were irrigated for eight months to induce sprouting and root growth: during this period, matric suction was monitored. The same procedure was followed to prepare ten fallow specimens.</p><p>Prior to testing, samples were sealed and left in the darkness in a temperature/relative humidity-controlled room for 24 hours, to equalise the desired value of initial suction. An isotropic consolidation stress between 10 and 50 kPa was imposed prior to shearing at a vertical displacement rate of 0.016 mm/min. Matric suction was measured by a tensiometer and the water content was checked at the beginning and at the end of each test. Finally, soil samples were washed to retrieve the entire root architecture, to assess root volume and tensile strength. The resulting values of the root volume ratio of <em>Cynodon dactilon</em> were in good agreement with those observed in-situ in literature studies.</p><p>Generally, the higher the initial soil matric suction, the higher the strength observed in the tests, with vegetated soil systematically showing greater strength than the bare one at similar hydro-mechanical states. In fact, at low values of suction, additional resistance in vegetated soil was observed once reaching large shear deformations, whereas, in drier soils, root reinforcement was activated at smaller strains. Indeed, soil hydraulic state affected the root failure mechanism. In nearly saturated soil, the roots subjected to shearing/tension are free to stretch and slip whereas in slightly saturated soil they are firmly bonded within the matrix and so they experience a more immediate breakage.  </p><p>Despite the root reinforcement, the vegetated samples exhibited larger volume deformations upon shearing, due to the changes generated by root growth on soil fabric (fissures).</p><p>A shear strength criterion for partially saturated soils was used to interpret successfully the results, considering suction, degree of saturation, and soil microstructure. Roots predominantly increased the apparent cohesion of the soil, with minor changes on the friction angle.</p>