Landslides are a recurrent geohazard in Andean urban environments, where weak soils, intense seasonal rainfall, and unplanned urban expansion combine to increase slope vulnerability. In such settings, sustainable hillside management requires stabilization strategies that are both technically effective and environmentally compatible. This study evaluates the effect of root reinforcement by vetiver grass (Chrysopogon zizanioides) on slope stability in two representative soils from Loja, Ecuador: sandy silt (SM) and sandy clay (SC). A reduced-scale physical model with 30 days of root development was established, and consolidated–drained direct shear tests (ASTM D3080/D3080M-23) were performed to determine the shear strength parameters under bare and vetiver-reinforced conditions. These parameters were then incorporated into numerical slope stability analyses using Slide and PLAXIS 2D, considering three slope angles (30°, 45°, and 50°), six root-positioning configurations, and hydraulic conditions with and without a water table. Vetiver increased effective cohesion by 22.7% in sandy silt and 19.0% in sandy clay, while the internal friction angle increased by 21.8% and 12.2%, respectively. Across all modeled scenarios, vetiver produced a consistent improvement in the factor of safety. The most critical case, corresponding to sandy silt at 45° with a water table, increased from FS = 0.841 in the control condition to FS = 1.309 under the full-coverage configuration. Parametric sensitivity analysis yielded coefficients of variation between 4.97% and 7.03%, indicating a stable model response under controlled parameter perturbations. These findings support vetiver as an experimentally grounded and environmentally sustainable Nature-based Solution for slope stabilization and provide relevant evidence for sustainable management of hazard-prone urban hillsides in vulnerable Andean settings.

Plant roots can modify soil mechanical behaviour through anchorage, interlocking, and stress-transfer mechanisms; however, the magnitude of this effect is not uniform and depends on the soil matrix in which the root system develops. This study evaluates the biomechanical reinforcement induced by Chrysopogon zizanioides (L.) Roberty roots in seven contrasting Andean soil matrices classified as CL-ML, MH, CH, OH, OL, CL, and ML. Bare-soil and Vetiver-rooted specimens were compared across three depth intervals: 0.00–0.50 m, 0.50–1.50 m, and 1.50–2.00 m. Shear-strength parameters were determined through direct shear tests under controlled normal stresses of 50, 100, and 200 kPa, and the relative amplifications of apparent cohesion (Δc) and internal friction angle (Δφ) were calculated. ANOVA, Tukey HSD post hoc tests, effect-size analysis, descriptive depth-dependent regression analyses, and shear-resistance sensitivity analysis were then applied to identify soil-dependent reinforcement patterns. The results showed that the mechanical contribution of Vetiver roots was primarily controlled by soil type. The highest relative cohesion amplification was obtained in OL soil, where Δc reached 100.0%, whereas the strongest friction-angle response was observed in CH soil, with Δφ reaching 20.1%. MH soil exhibited an intermediate and technically balanced response, combining moderate cohesion gains with relevant frictional improvement. One-way ANOVA screening indicated that soil type was the main differentiating factor controlling the Vetiver-induced response, while depth had a less dominant effect. Tukey HSD comparisons confirmed that OL exhibited a distinct cohesion-amplification pattern relative to the other soil matrices. These findings indicate that C. zizanioides root reinforcement should not be interpreted as a uniform plant effect but as a soil-matrix-dependent biomechanical response governed by root–soil interaction. The initial mechanical condition of the soil, rather than root content alone, determines whether Vetiver reinforcement is expressed primarily as a cohesion gain, a friction-angle gain, or a combined response. Because the statistical evidence is screening-level owing to the absence of replication within soil–depth cells, these patterns should support soil-specific interpretation rather than be applied as universal amplification factors.

Depth-resolved mechanical responses of rooted forest soils are difficult to interpret because root properties and soil state can vary together with sampling depth. This study tested whether a common within-soil depth ordering of apparent shear-strength parameters persists across seven contrasting Andean Unified Soil Classification System (USCS) classes sampled beneath young Eucalyptus globulus at Surface (0–50 cm), Middle (50–150 cm), and Deep (150–200 cm) strata. Direct shear tests at normal stresses of 50, 100, and 200 kPa were used to fit apparent Mohr–Coulomb parameters (capp and φapp), while dry root mass content (RC) was evaluated separately; soil class was the paired blocking unit. Mean capp increased by 38.7% from Surface to Deep, and all seven soils followed the same Surface < Middle < Deep ordering (Friedman p < 0.001; Kendall W = 1.000), whereas mean RC decreased by 16.2% (Friedman p = 0.0019). In the only approximately depth-matched rooted–bare comparison, Deep-rooted capp exceeded the ≈170 cm bare reference in every soil class by 15.64 ± 1.28 kPa (exact Wilcoxon p = 0.0156). The novelty is the cross-soil paired-block demonstration that the same apparent cohesion ordering persists despite large differences in soil class, while RC follows the opposite profile. The pattern is depth-associated rather than causal because density, water state, stress history, and unmeasured root traits were not controlled independently across strata.

Vegetation-based nature-based solutions can improve shallow slope stability, yet transferring root-scale measurements to slope-scale assessment remains difficult. This study reanalyzes archived mechanical and engineering data for Salix integra used on an ecological canal embankment in northern Jiangsu Province, China. The dataset comprises 36 peak shear stress increments from constrained single-root direct shear tests conducted at four normal stresses and nine root-diameter classes, together with ten bifurcation records used for FracRoot reconstruction. Peak shear stress increment, Δτ, ranged from 2.41 to 39.28 kPa. Normal stress was the strongest statistical predictor within the clamped-end test configuration (r = 0.849, p < 0.001), whereas root diameter showed no significant unadjusted effect. The preserved FracRoot relationships were Dmax = 0.820D₀ and Dmin = 0.563D₀. An archived deterministic double-wedge calculation reported a factor-of-safety change from 0.92 to 1.14 after inclusion of aggregate root resistance. Because the separate unreinforced peak values, complete shear records, direct field root-depth measurements, planting geometry, and segment-level root-intersection calculations were not preserved, the statistical analysis is restricted to archived increments and the stability result is interpreted as an illustrative scenario rather than a verified design value. The novelty lies in transparently linking archived root-scale testing, branching architecture, and local slope-stability assessment while defining the evidence required for reproducible field-scale validation.