Triaxial test investigation of the reinforcement effect of Acacia dealbata roots on mountain red soil Authors: Xirui Dai, Zongheng xu, Hongchen Ye, Yun Zeng Published: 2025 ABSTRACT AbstractVegetation roots play a pivotal role in enhancing the shear resistance and stability of degraded mountain red soils, yet the quantitative effects of root density and orientation remain poorly constrained. In this study, unconsolidated undrained triaxial tests were conducted on red soil reinforced with Acacia dealbata roots to examine the influence of root area ratio (RAR), spatial arrangement, and confining pressure on stress–strain behaviour and shear strength. Shear strength parameters were derived from Mohr–Coulomb envelopes, and the increments in cohesion (Δc) and internal friction angle (Δφ) were analytically related to the increment of the major principal stress (Δσ₁). The results indicate that the presence of root systems enhanced both peak strength and the reinforcement index. The configuration with high RAR and horizontal root distribution achieved up to 64.80% higher deviator stress than the unreinforced control, whereas sparse vertical roots could weaken the composite under low confinement. Reinforcement efficiency was governed by the combined effects of root density and orientation, with dense, horizontally distributed networks most effectively mobilising interfacial friction and tensile resistance. Strength envelopes indicated simultaneous increases in apparent cohesion and internal friction angle, where Δc contributed nearly linearly to Δσ₁, while Δφ induced a quadratic amplification; their synergistic effect was further enhanced at higher confining pressures. This study provides robust experimental evidence and a quantitative framework to inform eco-engineering and slope stabilisation practices in mountain red soils, highlighting the importance of root system density and architecture in maximising vegetation-based reinforcement.Keywords: Root–soil composites, Triaxial shear test, Shear strength parameters, Root distribution and density, Mountain red soilSubject terms: Engineering, Solid Earth sciences Introduction Soil erosion represents a pervasive and critical ecological challenge in the mountainous regions of Southwest China, particularly within landscapes dominated by red soils. These soils are inherently susceptible to detachment due to their weak erosion resistance and the prevalence of steep slopes, conditions that markedly accelerate sediment loss under seasonal heavy rainfall events. Such processes not only exacerbate surface degradation but also destabilize slopes, thereby increasing the likelihood of secondary geological hazards such as landslides. Vegetation is widely recognized as an effective natural barrier to erosion1, playing a pivotal role in reducing soil loss and mitigating hazard occurrence2, with notable efficacy in red soil environments3,4. The mechanical properties of plant root systems confer both shear resistance and anchorage5, creating a root–soil composite that enhances slope stability6. Through reinforcement and interlocking, roots increase soil cohesion and shear strength while lowering the susceptibility to surface erosion7. Furthermore, vegetation intercepts rainfall, modulates soil moisture, and reduces the erosive energy of overland flow6. Understanding the reinforcement function of plant root systems within root–soil composites is therefore of substantial practical and ecological significance for mitigating erosion and associated hazards in mountainous red soil regions. Substantial advances have been achieved in elucidating the reinforcing effect of plant roots on soils, with research primarily focusing on the mechanical response of root–soil composites to variations in root content, moisture content, and environmental cycling. Experimental studies using triaxial compression have shown that shear strength is markedly influenced by root content, water content, and the number of wet–dry cycles8–10, while consolidated undrained triaxial tests and electron microscopy further demonstrate that root diameter and hydrological cycling significantly affect composite stability11. Beyond these physical factors, increasing attention has been directed to plant species and root morphological traits. For instance, direct shear tests with strain-gauge instrumentation reveal that root inclination governs the force–deformation mode12, while combined triaxial compression, wave velocity measurements, and FLAC3D modelling have been used to assess the influence of vegetation coverage and freeze–thaw cycles13. Integration of direct shear testing with X-ray CT scanning indicates that root mass and volume parameters alter shear strength by modifying internal structure14. The role of spatial distribution and arrangement patterns has also been clarified: triaxial tests demonstrate that primary versus lateral root layouts, as well as distribution angles, control the stress–strain relationship, pore water pressure development, and stress paths15, and that geometric features such as root diameter, length, density, and angle directly affect shear resistance16. Further studies link root content and distribution to cohesion, internal friction angle, and even fractal dimension of the composite17, and highlight the influence of root morphology and hierarchical structure on both mechanical performance and reinforcement mechanisms3. Complementary approaches have refined the understanding of root–soil interaction: mechanical parameters for unreinforced and reinforced soils have been derived from triaxial testing18, and dynamic behaviours under cyclic shear have been characterised using triaxial and dynamic simple shear apparatus, with discrete element modelling proposed to simulate contact mechanisms in reinforced soils19. Collectively, these findings underscore that the reinforcing capacity of plant roots depends not only on species but also on root arrangement, inclination, and structural attributes, offering critical insights for soil and water conservation strategies. As a distinctive zonal soil in the mountainous and plateau regions of Southwest China, red soil has undergone increasingly severe erosion and degradation in recent years under the combined influence of natural processes and human activities. Although vegetation–soil interactions have been widely studied, systematic research on the erosion mechanisms of mountain red soils and the reinforcing role of vegetation under local environmental conditions remains scarce. Experimental evidence directly linking root quantity, spatial distribution, and the mechanical behaviour of red soil composites under controlled loading conditions is particularly limited. In response to this research gap, this study uses red soil from the Chenggong area of Kunming, Yunnan Province, and incorporates roots of Acacia dealbata, a locally dominant and ecologically significant shrub, to construct root–soil composites. Indoor triaxial tests are conducted to assess the influence of root content and spatial distribution on shear strength parameters, thereby clarifying the mechanisms by which vegetation enhances soil stability. The results provide a scientific basis for improving the mechanical performance of red soils in Southwest China and for guiding targeted ecological restoration and slope protection strategies. Materials and methods Study area The sampling site was located at Songmao Reservoir in Chenggong District, Kunming City, Yunnan Province (The geographical coordinates are 102.88° east longitude and 24.88° north latitude.) (Fig. 1a, b). The sampling area can be a natural forest area. This area lies within the third-level new tectonic zone of the Kunming Depression in the Chuan–Dian fault block uplift area, with the Xiaojiang fault zone as the major structure, representing a typical strong earthquake-prone region in China. The reservoir is situated in a low mountain and hilly setting on a low-latitude plateau plain, where the dam area river is wide and gentle, relief is small, slopes are moderate, and the weathering crust is well preserved20. Regional river systems include the upper reaches of the Yangtze River, Red River, and Pearl River21. The subsurface is dominated by Quaternary alluvial and lacustrine deposits, overlaid by extensive deep zonal ancient soil, mountain red soil, with well-developed vegetation and high organic matter content. The Dianchi Lake Basin, where the reservoir is located, has been significantly altered by both natural processes and human activities. Severe deforestation has caused widespread vegetation loss, slope erosion, and soil and water degradation, prompting reforestation efforts with Acacia dealbata Link as a secondary forest species. Although Acacia dealbata is an invasive species and an aggressive invader in southern Europe, the Americas, and southern Africa22, known to reduce native understory plant abundance and alter ecosystem functions such as soil respiration and enzyme activity23, it has become dominant in the study area (Fig. 1c). Its abundant leaf litter and root turnover increase soil microbial content and contribute to soil and water conservation in the understory. This study examines how the roots of Acacia dealbata influences understory soil erosion, assessing restoration effectiveness, soil recovery, and ecosystem resilience22,24. Study area and sampling points. (a) (b) The location distribution of sampling points in Chenggong District, Yunnan Province, China (c) Sampling site environment with widespread Acacia dealbata growth (d) Acacia dealbata roots and mountain red soil at the sampling site. (Arcgis10.8. (https://desktop.arcgis.com/zh-cn/desktop/index.html) GS(2024)1158. ASTGTMN24E103. The data set is provided by Geospatial Data Cloud site, Computer Network Information). Characteristics of the tested soil and roots The roots of the vegetation are mainly distributed within the upper 60 cm of the soil, and soil samples for this study were collected from the mountain red soil beneath the Acacia dealbata forest within this depth range (Fig. 1d). Three particle size tests (sample 1–3) were conducted, and the average results are presented in Fig. 2a. Mountain red soil, typical of warm and humid climatic zones, has undergone prolonged and intense chemical weathering, leading to the decomposition of parent minerals and the formation of abundant clay minerals such as kaolinite. The d50 was 1.33 μm, with the mean particle size of 9.38 μm and a sorting coefficient of 2.47, indicating poor sorting and wide range of particle sizes. These properties favor the development of water- -permeable channels while maintaining adequate water retention and aeration. Such a balance enhances soil structure stability and pore connectivity, creating favorable conditions for roots growth and helping to mitigate soil erosion. Physical and mechanical characteristics of the tested soil and roots. (a) Particle size composition of the tested soil; (b) Tensile test results of the tested roots (tensile deformation-force relationship curve); (c) On-site root tensile test. To evaluate the tensile properties of the roots, the tests were conducted using the universal testing machine. Only results where fracture occurred within the effective test length were considered valid (Fig. 2c). The root system’s peripheral periderm contains fiber bundles that fail progressively rather than simultaneously, producing step-like features in the tensile deformation-force curve (Fig. 2b). Initial rupture of weaker bundles transfers stress to remaining bundles, which continue to bear the load until subsequent failure events occur. This sequential breakage increases deformation incrementally, culminating in peak tensile force when the internal xylem is ruptured. According to calculations derived from Fig. 2b, the mean tensile strength was 802.50 kPa, reflecting high load-bearing capacity. The progressive, multi-step deformation prior to failure indicates substantial ductility, which, in turn, contributes to soil reinforcement and mitigates water-induced erosion. Preparation and testing methods of root-soil composites To ensure the accuracy and reliability of the test results, strict control over sample uniformity was maintained. In this study, Acacia dealbata Link roots were standardized to a length of 8.00 cm, with diameters controlled within 1.00 ± 0.20 mm. Morphological variations such as branching, curvature, and diameter irregularities were excluded. Collected roots were thoroughly cleaned, air-dried to remove surface moisture, and stored in a constant-temperature and -humidity chamber to maintain uniform pre-conditioning. Following GB/T 50,123 − 201925, the mountain red soil was sieved through a 2 mm mesh, oven-dried, and adjusted to a moisture content of 30% before specimen preparation. To ensure consistency in density and minimize variability, all cylindrical specimens (5 cm diameter, 8 cm height) were compacted in four layers, with roughened interlayer surfaces to enhance bonding. To systematically investigate the reinforcement effects of roots across varying densities and orientations, comparative tests were designed using root area ratios (hereinafter referred to as RAR) of 0.08, 0.16, and 0.2426–28. Horizontally distributed root specimens (GT1 ~ GT3) included two, four, and six roots arranged in perpendicular pairs within the layers. Vertically distributed specimens (GT4 ~ GT6) were prepared by inserting varying numbers of roots from the top surface to the base (Fig. 3a). Unreinforced specimens were designated as GT0 and served as the control group. Root–soil composite schematic and test arrangement. (a) Schematic diagram of the root–soil composite; (b) Triaxial test system; (c) Morphology of the partially failed triaxial specimen of the composite after loading. To minimize potential result randomness, unconsolidated undrained (UU) triaxial shear tests were conducted under four different confining pressures of 100 kPa, 200 kPa, 300 kPa, and 400 kPa (Fig. 3b). The testing instrument used was a stress-controlled triaxial shear permeameter (SLB-1), which allowed equal stress control and applied a stable axial load. The shear rate of the test instrument was 0.80 mm/min. After the sample failed failure under shear, the test continued until the axial strain exceeded 15%, at which point the test was concluded. Shear strength and its parameters, cohesion (c) and internal friction angle (φ), were determined based on the obtained stress–strain curves. Analytical methods Root effects on soil strength evaluation In the stress-strain curve, the failure load is defined by the failure mode, with 15% axial strain indicating failure for hardening and peak stress for softening. To assess the impact of root number and distribution on soil failure load and strength, this study introduces the root reinforcement coefficient m and growth rate of cohesion and internal friction angle29. The root reinforcement coefficient m is defined as the ratio of shear strength in root-reinforced soil to that in unreinforced soil, calculated as: where (σ1-σ3)p and (σ1-σ3)q represent the failure loads of root-reinforced and unreinforced soils, respectively. If m > 1, it indicates that roots enhance the soil’s shear strength. The value of m allows analysis of how different root distributions and densities affect the shear strength of mountain red soil. To quantify the effects of root reinforcement on cohesion and internal friction angle, the growth rates of these parameters are introduced: where c and φ represent the cohesion and internal friction angle of unreinforced soil, and c′ and φ′ are those of the root-reinforced soil. kc and kφ reflect the improvements in cohesion and internal friction angle due to root reinforcement. Mohr stress circle analysis and strength envelopes The Mohr stress circle method is employed to analyze the ultimate failure states of unreinforced soil and root-soil composites. As shown in Fig. 4, Circle Ⅰ represents the Mohr stress circle for the unreinforced soil, while circle Ⅱ represents the Mohr stress circle for the root-soil composite. The same confining pressure σ3 is used for all samples, but the major principal stress σ1 at failure varies across the different composite states. Comparing Circle Ⅰ and Ⅱ, distinct strength envelopes can be plotted. Experimental results indicated that root inclusion enhanced both cohesion and internal friction angle, thereby increasing the failure load. Accordingly, in Fig. 4, the composite parameters (c′ > c, φ′ > φ, σ1gt > σ₁) reflect the improved shear strength of the root–soil composites. Model diagram of Mohr stress circle and strength envelope of triaxial test of root-reinforced and unreinforced soil. During the test, a constant confining pressure σ3 was applied to both the unreinforced soil and the root-soil composite to achieve a new limit equilibrium state for each specimen. The major principal stresses at failure were denoted as for the unreinforced soil and for the root-soil composite. According to the Mohr-Coulomb failure criterion, the following relations hold: where c and φ represent the cohesion and internal friction angle of the unreinforced soil, respectively, while Δc and Δφ denote the increases in cohesion and internal friction angle of the root-soil composite. To assess the effect of root reinforcement on failure load (σ1 − σ3), The difference between the composite and the unreinforced soil can be calculated as follows: . Assume, , the expression can be expanded as: The expression (6) provides a detailed relationship for calculating the increment in the major principal stress Δσ1 due to the root reinforcement. By further analyzing this equation, the influence of root-induced changes in cohesion and internal friction angle on the overall strength of the root-soil composite can be evaluated. Results and analysis Stress-strain behaviours The failure modes of the samples encompass shear, dilation, and contraction9. In this study, shear and dilation were the predominant failure modes, with shear failure being the most prevalent (Fig. 3c). During the unloading phase at the end of the test, a rebound phenomenon occurred in the vertically placed roots, causing localized damage at the top of the sample. Due to the high clay content in the mountain red soil, all samples retained their structural integrity throughout the test. Figure 5 illustrates the relationship between axial strain and principal stress difference across different root-reinforced soil configurations. The triaxial stress–strain responses qualitatively demonstrate the reinforcement and stress-modulating effects of varying root quantities and spatial arrangements. At the initial stage of loading (small axial strain), the stress level remains low, and the contribution of roots to the composite strength is limited. As the strain increases, roots begin to deform in coordination with the surrounding soil, enabling their tensile strength to be progressively mobilized. During this phase, the principal stress difference increases in a non-uniform linear fashion. With further straining, the reinforcing effect of the roots approaches its maximum, reflected either by a distinct stress peak or continued strain hardening. These trends indicate that the reinforcement effect of root–soil composites becomes significant only after the soil undergoes sufficient axial deformation. Relationship between principal stress difference and axial strain for various root-soil composites (a) GT0; (b) GT1; (c) GT2; (d) GT3; (b) GT4; (c) GT5; (d) GT6. The influence of root reinforcement under different confining pressures is further illustrated in Fig. 6. Under low confining pressure (100 kPa, Fig. 6a, b), the principal stress difference of composites is substantially higher than that of the unreinforced specimens when the axial strain is below 2%. This suggests that root reinforcement is particularly effective in loose soils under low confinement. Nevertheless, the overall peak stress remains lower than that observed under medium and high confining pressures, with the maximum deviator stress reaching only 325.70 kPa. At medium confining pressures (200 kPa and 300 kPa, Fig. 6c ~ f), the stress–strain curves of root-reinforced and unreinforced soils exhibit little difference during the initial loading stage. However, as axial strain increases, the divergence in principal stress difference becomes more apparent. This implies that under moderate confinement, root reinforcement may initially disturb the soil fabric, and its beneficial effects only emerge after substantial deformation. At high confining pressure (400 kPa, Fig. 6g, h), the roots continue to enhance the shear strength and load-bearing capacity of the soil. The closer interlocking between soil particles and root surfaces results in stronger root–soil interaction, thereby significantly improving the composite strength. Relationship between principal stress difference and axial strain at identical confining pressure (a) (b) 100 kPa; (c) (d) 200 kPa; (e) (f) 300 kPa; (g) (h) 400 kPa. Failure load and shear strength parameters The variation of the peak principal stress difference of the composite material under different confining pressures (Table 1) indicates. From the analysis of the results of the maximum principal stress difference in the experiment, it can be concluded that the larger the RAR, the greater the horizontal root system can enhance the shear bearing capacity of the soil. Under the same RAR conditions, the tensile performance of GT1 to GT3 is always superior to that of GT4 to GT6. With the increase of RAR, the principal stress difference also shows a continuous upward trend. Under a confining pressure of 400 kPa, the peak value of GT3 in the horizontally arranged root system was 532.70 kPa, which was 64.80% higher than that in the blank sample GT0 and 28.00% greater than that in the vertically arranged root system GT6. Compared with GT0, all root reinforcement specimens except GT4 showed significant strength improvement at all pressure levels, and the enhancement was particularly obvious under high confining pressure conditions, indicating that the reinforcement potential of roots in compact-structured soil is greater. The peak principal stress difference of the horizontal root system is higher than that of the vertical root system in reinforcing the soil, which means that the horizontal root system can enable the soil to withstand a greater stress difference without damage. This indicates that the spatial arrangement of the root system plays a key role in mobilizing the tensile capacity. Table 1Peak principal stress difference at different confining pressures.Confining pressuresGT0 (kPa)GT1 (kPa)GT2 (kPa)GT3 (kPa)GT4 (kPa)GT5 (kPa)GT6 (kPa)100 kPa241.20280.30304.40325.70174.10287.20312.30200 kPa268.90320.50347.30364.30259.40312.60335.60300 kPa288.80340.70352.30390.50284.50352.10381.20400 kPa323.30373.90405.90532.70323.50380.80415.00 Peak principal stress difference at different confining pressures. The variation of the root reinforcement coefficient m shown in Fig. 7 further confirms this trend. Under all confining pressures, the root system reinforcement coefficient m is greater than 1, indicating that the presence of roots significantly enhances the shear strength of the composite soil. Under the same RAR conditions, the m values of the horizontal arrangement groups (GT1 to GT3) are all higher than those of the vertical arrangement groups (GT4 to GT6). With the increase of RAR, the root reinforcement coefficient m also shows a continuous upward trend. At 400 kPa, the m value of GT3 in the horizontally arranged root system was the highest, at 1.65, and the reinforcement effect was the most obvious, which was much greater than the m of GT6 in the vertically arranged root system under the same confining pressure. In conclusion, these findings emphasize the synergistic influence of root content, spatial arrangement and confining pressure on the reinforcing performance of composite materials. In the test configuration, a high RAR combined with a horizontal root distribution is the most effective in enhancing the shear strength of the composite material. It can be expressed from both the principal stress difference and the root reinforcement coefficient m that the horizontally arranged root system has a better reinforcement effect than the vertically arranged root system. Variation of root reinforcement coefficient. Mohr’s circles and the corresponding strength envelopes were plotted with the principal stresses (σ₁, σ₃) on the x-axis and shear stress (τ) on the y-axis. As shown in Fig. 8, the specimens tested under a confining pressure of σ₃ = 100 kPa are presented. The shear strength envelopes of the composites, as illustrated in Fig. 8, display higher intercepts than the control group (GT0), indicating that roots substantially enhance soil cohesion. The nonparallel nature of the envelopes suggests that the strengthening effect of roots is also extends to an increase in the internal friction angle. Notably, the envelopes of GT4 ~ GT6 are consistently lower than those of GT1 ~ GT3, underscoring the influence of root spatial configuration on reinforcement performance. The apparent cohesion principle16, which attributes the improvement of soil shear strength to the combined effects of intrinsic soil resistance, root tensile strength, and root–soil interface friction, provides a sound explanation for these findings. The results of this study align well with this theoretical framework. Mohr stress circle and strength envelope of the specimens (a) GT0, GT1, GT2, GT3 (σ3 = 100kP) (b) GT0, GT4, GT5, GT6 (σ3 = 100kP). The shear strength parameters, cohesion (c) and internal friction angle (φ), together with their respective growth rates derived from the strength envelopes, are shown in Fig. 9. Figure 9a indicates that, with the exception of GT4, all root-reinforced specimens exhibited appreciable increases in both c and φ relative to the control (GT0). GT3 achieved the highest values, with cohesion of 124.50 kPa and φ of 8.70°, representing increases of 20.57% and 31.42%, respectively (Fig. 9b). These results demonstrate that high root content in a horizontal arrangement provides the most effective reinforcement of soil strength. Substantial improvements were also observed for GT2 and GT6, suggesting that an increase in RAR consistently enhances the mechanical performance of the soil matrix. In contrast, GT4 (low root content with vertical orientation) deviated from this trend, displaying lower c and φ than other reinforced specimens, with the growth rate of φ even becoming negative (− 1.86%). This implies that under conditions of low confining pressure and insufficient root density, vertical orientation may weaken the composite structure. Although the improvement observed in GT5 was less pronounced than in GT3, it still surpassed that of GT4, indicating that increased root density can partly compensate for the disadvantages associated with vertical root distribution. Shear strength parameters and their growth rates of different specimens (a) cohesion (c) and internal friction angle (φ); (b) their growth rates. Strength increments related to cohesion and friction angle Based on the experimental results, the cohesion and internal friction angle of the GT0 were determined as c = 94.74 kPa and φ = 6.91°, respectively. Assuming Δc remains constant, Eq. (6) can be expressed as: Δσ₁= d + bΔφ + a(Δφ)2, where Δσ₁ denotes the increment of the major principal stress, Δφ represents the increment of the internal friction angle, and a, b, and d are incremental coefficients that vary with confining pressure σ₃ (100, 200, 300, and 400 kPa). Based on the Δc values calculated from Fig. 9a, the relationships between Δσ₁ and Δφ (ranging from 0 to 1.80°, with the maximum experimental Δφ being 1.79°) were plotted under different confining pressures, as illustrated in Fig. 10a ~ d. These curves clearly reveal how the reinforcement effect of roots translates into changes in major principal stress through modifications of the internal friction angle. Relationship between major principal stress increment Δσ1 and internal friction angle increment Δφ, and cohesion increment Δc. (a) (e) σ3 = 100 kPa; (b) (f) σ3 = 200 kPa; (c) (g) σ3 = 300 kPa; (d) (h) σ3 = 400 kPa. As Δφ increases, the growth rate of Δσ₁ exhibits a quadratic trend, indicating that the strengthening effect of roots via changes in internal friction angle is not a simple linear superposition but rather involves a distinct nonlinear amplification mechanism. Overall, composites with horizontally oriented roots consistently exhibit greater Δσ₁ than those with vertically oriented roots. Furthermore, for a given root arrangement, an increase in root quantity further enhances the reinforcing effect. Of particular note, the GT4 (Δc = − 1.76 kPa) deviates markedly from this general trend, with Δσ₁ not only substantially lower than those of other composites but even becoming negative under low Δφ conditions. This suggests that under unfavorable arrangements and insufficient root content, the presence of roots may undermine, rather than enhance, the overall performance of the composite. This observation is consistent with the experimental findings discussed earlier. Assuming Δφ remains constant, Eq. (6) can be simplified as: Δσ₁= e + fΔc, where Δσ₁ denotes the increment of major principal stress in the composites, Δc represents the cohesion increment, and e and f are incremental coefficients that vary with confining pressure σ₃. Based on the Δφ values calculated from Fig. 9a, the relationship curves between Δσ₁ and Δc (ranging from 0 ~ 30.00 kPa, with the maximum experimental Δc being 29.76 kPa) can be further derived under different confining pressures, as illustrated in Fig. 10e ~ h. The results indicate that Δσ₁ exhibits a nearly linear increase with Δc, while the rate of increase is strongly modulated by the magnitude of Δφ; larger Δφ values correspond to higher Δσ₁. This finding suggests that although cohesion enhances composite strength in a nearly proportional manner, its contribution is substantially amplified by the frictional resistance induced by root reinforcement. In addition, an increase in confining pressure further elevates the overall level of Δσ₁, implying that under higher confinement the root–soil interfacial friction and interlocking effects are more effectively mobilised, thereby strengthening the synergy between Δc and Δφ. It is noteworthy that the GT4 (Δφ = 0.20°) shows a marked deviation from the general trend, with Δσ₁ consistently lower than that of other composites at the same Δc. When both Δc and Δφ are relatively small, Δσ₁ even takes negative values, indicating that insufficient root content or unfavourable root configuration may lead to adverse effects on the composite strength. However, once Δc reaches a critical threshold (approximately 3 kPa), this negative effect diminishes and is progressively replaced by a positive reinforcement effect, thereby highlighting the existence of a transition from weakening to strengthening in the composites’ behaviour. As shown in Fig. 10a ~ h, cohesion exhibits an approximately linear strengthening effect on the composite material, while the internal friction angle contributes a nonlinear amplification response. These two factors act synergistically, and their combined reinforcement becomes increasingly evident with rising confining pressure. Discussion Mechanisms governing the mobilisation of root-induced strength The experimental findings can be further interpreted through the Wu–Waldron model30,31, which attributes the reinforcement effect of roots primarily to an increase in the equivalent cohesion of the soil. This strengthening mechanism is manifested through increments in both cohesion and internal friction angle, as illustrated in Fig. 11. In this framework, roots are idealised as rigid inclusions intersecting the shear plane, where A denotes the soil area influenced by a single root, T is the tensile resistance of the root, θ is the shear deformation angle, x is the displacement along the shear plane, and H represents the thickness of the shear zone (Fig. 11a). When the composite is subjected to shear displacement, relative movement tends to occur between roots and soil particles, which is resisted by the frictional forces at the root–soil interface. Shear mechanical model of root–soil composites (a) vertically arranged roots; (b) horizontally arranged roots (the model is modified based on the Waldron model30. The observed distinction between vertical and horizontal root arrangements can also be rationalised within this model. Vertically oriented roots predominantly experience tensile elongation during shearing. Despite their inherent tensile strength and ductility, composite failure is largely governed by interfacial slippage rather than root rupture. Moreover, only a limited number of vertical roots intersect the shear plane, resulting in a relatively weak contribution to shear resistance, consistent with the lower reinforcement efficiency observed in GT4–GT6 (Fig. 11a). Notably, experimental data show that the m value of GT4 at 100 kPa is only 0.72, and remains below 1 even under higher confining pressures. This suggests that under low confinement, the inclusion of a small number of vertically oriented roots may compromise the integrity of the soil matrix, particularly in fine-grained clay, due to insufficient restraint and a low root area ratio (RAR). When root density is limited, vertical configurations tend to weaken rather than enhance the overall mechanical performance of the composite material. In contrast, horizontally distributed roots are directly mobilised by vertical confinement. The bidirectional, multilayer arrangement employed in this study induces substantial lateral frictional resistance (Fig. 11b), effectively constraining particle displacement. This force-bearing mechanism underpins the superior shear strength and overall reinforcement efficiency exhibited by the horizontally rooted composites (GT1–GT3). The influence of root quantity further clarifies the mechanism. Increasing root content systematically enhances the mobilization of interfacial interlocking, frictional, and adhesive forces, thereby improving the structural integrity of the composite. These observations are in line with the Wu–Waldron model, which highlights the controlling role of root length density and root diameter. In the present study, the dense Acacia dealbata root network demonstrably increased the equivalent cohesion of the red soil, confirming the model’s applicability and offering mechanistic insight into the observed strength gains. Effect of root density and orientation on reinforcement performance The test results demonstrate that both root quantity and spatial configuration exert a decisive influence on the efficiency of soil reinforcement. An increase in the RAR systematically enhanced the peak principal stress difference and the associated shear strength parameters. For instance, GT3, mobilised the strongest reinforcement effect, achieving a 64.80% improvement in peak strength compared with the unreinforced control. In contrast, GT4, characterised by low root content and vertical orientation, not only failed to improve the composite strength but, under certain conditions, even exhibited adverse effects. This indicates the existence of a threshold root density below which the reinforcing contribution of roots cannot be effectively mobilised. These contrasting behaviours can be explained by the mechanisms of root–soil interaction. A higher number of roots provides greater interfacial contact and stronger interlocking with the soil skeleton, thereby enhancing the mobilisation of both tensile resistance and frictional forces. Conversely, when root density is insufficient, the lack of interfacial resistance and the occurrence of local stress concentrations may compromise the integrity of the composite, resulting in reduced strength. Equally important is the spatial orientation of roots: horizontally distributed roots are more readily engaged during shear deformation, mobilising lateral restraint and frictional resistance at an early stage, whereas vertically aligned roots primarily rely on axial tension, which becomes fully mobilised only under larger deformations and higher strains. This interpretation aligns well with the predictions of the Wu–Waldron model, which highlights root density, diameter, and orientation as key determinants of reinforcement. Dense networks of horizontally arranged roots are able to mobilise interfacial friction and tensile resistance synergistically, leading to concurrent increases in both apparent cohesion and internal friction angle, and thus producing a nonlinear amplification of shear strength. By contrast, sparse or unfavourably oriented roots may act as weak inclusions, undermining rather than enhancing composite performance. When considered as a whole, these findings emphasise that reinforcement efficiency is not a simple additive function of root abundance, but emerges from the interplay between root quantity and spatial distribution. This insight is consistent with earlier observations by Gray and Sotir32 and Genet et al.33, who reported that both root density and orientation strongly influence the stabilising effect of vegetation on soil. From an applied perspective, the results suggest that eco-engineering and slope stabilisation practices should prioritise not only sufficient root density but also favourable spatial orientations, thereby ensuring that the tensile and frictional capacities of roots are jointly mobilised to maximise the reinforcing effect. Conclusion This study systematically evaluated the reinforcement effects of Acacia dealbata roots on mountain red soil through unconsolidated undrained triaxial tests, with particular focus on the influence of root content, spatial distribution, and confining pressure. The main findings are summarized as follows: Root reinforcement markedly altered the stress–strain response of mountain red soil. Whereas unreinforced soil predominantly exhibited dilation failure, root–soil composites were governed by shear failure. Both root quantity and spatial configuration were shown to be decisive in controlling shear strength. An increase in RAR consistently enhanced the peak principal stress difference and reinforcement coefficient m, with GT3 (high RAR, horizontal distribution) achieving the highest strength gain (64.80% greater than the control). Conversely, insufficient and vertically aligned roots (e.g. GT4) not only failed to reinforce the soil but also induced adverse effects under low confinement. Root reinforcement markedly altered the stress–strain response of mountain red soil. Whereas unreinforced soil predominantly exhibited dilation failure, root–soil composites were governed by shear failure. Both root quantity and spatial configuration were shown to be decisive in controlling shear strength. An increase in RAR consistently enhanced the peak principal stress difference and reinforcement coefficient m, with GT3 (high RAR, horizontal distribution) achieving the highest strength gain (64.80% greater than the control). Conversely, insufficient and vertically aligned roots (e.g. GT4) not only failed to reinforce the soil but also induced adverse effects under low confinement. (2)Shear strength envelopes confirmed that root inclusion improved both apparent cohesion and internal friction angle. While cohesion increments contributed linearly to strength, increments in the internal friction angle produced a nonlinear amplification effect. The combined action of these two mechanisms was more pronounced under higher confining pressures. Shear strength envelopes confirmed that root inclusion improved both apparent cohesion and internal friction angle. While cohesion increments contributed linearly to strength, increments in the internal friction angle produced a nonlinear amplification effect. The combined action of these two mechanisms was more pronounced under higher confining pressures. (3)The reinforcement efficiency arose from the interaction between density and orientation. Dense, horizontally arranged root networks mobilised interfacial friction and tensile resistance most effectively, thereby delivering superior shear strength improvements. These insights emphasise that eco-engineering and slope stabilisation strategies should prioritise root systems with adequate density and favourable spatial orientation to maximise reinforcement in mountain red soil. The reinforcement efficiency arose from the interaction between density and orientation. Dense, horizontally arranged root networks mobilised interfacial friction and tensile resistance most effectively, thereby delivering superior shear strength improvements. These insights emphasise that eco-engineering and slope stabilisation strategies should prioritise root systems with adequate density and favourable spatial orientation to maximise reinforcement in mountain red soil. Author contributions Xirui Dai was responsible for writing the original draft. Zongheng Xu was responsible for the review and editing of the manuscript, as well as providing guidance on figures preparation. Hongchen Ye and Yun Zeng conducted the experimental operations. All authors have read and agreed to the published version of the manuscript. Funding This research was financially supported by Yunnan Natural Science Foundation Project (202401AT070119). Data availability All data generated or analyzed during this study are fully included in this published article and its supplementary information files. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data generated or analyzed during this study are fully included in this published article and its supplementary information files.