Final Report Summary - BIOEPIC SLOPE (Use of BioEngineered Plant-Integrated Cover (BioEPIC) to Enhance Slope Performance)
During the project period, all four objectives defined in the proposal were achieved:
(1) To systematically characterise soil-water-plant interaction by identifying how soil state (e.g. void ratio) and root characteristics (e.g. root length distribution) affect transpiration-induced suction in laboratory;
(2) To parameterise root-water uptake by establishing physically-based relationship with hydraulic properties;
(3) To investigate long-term (at least 20 years) deformation characteristics (serviceability limit state) of BioEPIC slope under different scenarios of seasonal climate conditions from physical model tests; and
(4) To investigate stability and failure mechanism(s) (ultimate limit state) of BioEPIC slope under different scenarios of intense rainfall due to the climate change from physical model tests
The following specific research tasks has been conducted to address all four objectives. A synopsis of each task, including key findings, is given below.
1. Topic: Correlating hydrologic reinforcement of vegetated soil with plant traits – This aims to quantify and compare the hydrological reinforcement induced by ten woody species widespread in Europe; and then to examine and identify measureable plant traits that may be associated with the hydrologic reinforcement of soil. It is revealed that C. avellana, I aquifolium and U. europaeus, exhibited contrasting hydrologic reinforcement down the soil profile. While I. aqufolium provided hydrologic reinforcement mainly in shallow soil, U. europaeus induced greater hydrologic reinforcement in the entire soil profile down to 0.7 m depth. During winter, evergreen species had a much slower water uptake rate and smaller increase in matric suction compared with summer. Despite their slow increase in matric suction during winter, the magnitude of suction preserved was much higher than the value recorded in deciduous soil columns, thus providing greater increases in soil strength. As far as fast establishment and hydro-mechanical reinforcement are concerned, the nitrogen-fixing U. europaeus may be a more suitable candidate for soil eco-engineering purposes than the “resources saver” I. aquifolium.
2. Topic: Root biomechanical properties – This aims to measure and quantify root biomechanical properties during their early stage establishment, which represents a particularly challenging period for these plants; analyse and compare the tensile strength-diameter relationship of the ten woody species. The root tensile strength-diameter relationships of the ten species highlighted three different trends. The commonly-quoted negative power law in the literature was applicable for the strength-diameter data of only two out of the ten species tested, E. europaeus and U. europaeus. Indiscriminate use of the negative power law model to describe root tensile strength-diameter relation could overestimate root reinforcement, especially in the range of very small diameter roots where high root tensile strength would tend to be predicted by the model. Caution should be taken when this root biomechanical model is assumed for different taxa or for the same species that grows under different environmental conditions or during the challenging establishment period.
3. Topic: Effects of plant roots on soil-water retention and induced suction in vegetated soil – This task is to compare the effects of the two hydrological mechanisms, namely (a) change of soil water retention ability induced by roots and (b) plant evapotranspiration (ET) and the associated root-water uptake, on suction induced in vegetated soil, and then to quantify their relative importance. The calculation shows that for a prolonged 24-h rainfall with a return period of 10 years, mechanical root reinforcement is effective to stabilise the shallow soil of up to 0.5 m depth generally, where most of the root biomass exists. Hydrological reinforcement considering both the effects (a) and (b) provides much significant effects of soil stabilisation in deeper depths (i.e. 1–2 m), where slip failure is normally of major concern. The presence of roots in a vegetated slope preserves higher suction, hence higher shear strength, after rainfall, as compared to a bare slope. Reduction in saturated permeability (ks) due to the presence of intact roots is the most predominant hydrological effects. In contrast, increase in ks due to the presence of dying/decaying roots could be detrimental to slope stability at 1–2 m depth due to the reduced ability of the vegetated slope to preserve suction. Other effects, in particular the root water uptake through ET during rainfall, are minimal. Their contribution to slope stabilisation could be practically negligible.
4. Topic: A new and simple water retention model for root-permeated soils – This task aims to quantify and model the root effects on both soil water retention curve (SWRC) and soil hydraulic conductivity function (SHCF) of soils using the instantaneous profile method. Four types of vegetated soil, namely bare, grass-only, tree-only and mixed tree-grass soils, were subjected to a controlled drying-wetting cycle in a plant room. Plant roots affect the air-entry value (AEV), saturated permeability (ks) and reduction rate of unsaturated hydraulic conductivity (with respect to suction) most significantly, but it does not affect the reduction rate of volumetric water content much. When planted with single species, the AEV of silty sand increased, while saturated ks and reduction rate of unsaturated permeability as suction decreased. However, under the mixed planting conditions, opposite results are found. A new void ratio function is developed to model the observed decrease and increase in soil void ratio due to root occupancy (upon growth) and root shrinkage (upon decay), respectively, in an unsaturated vegetated coarse-grained soil. The function requires two root parameters, namely root volume ratio and root decay ratio, both of which can be readily measured through root excavation and image-based analysis. The new function is incorporated into a void ratio-dependent SWRC model for predicting SWRC of vegetated soils. Similarly, the same function can be combined with the Kozeny-Carman equation for predicting ks. The model prediction is then compared with a set of new field test data and an existing laboratory dataset for a silty sand vegetated with a tree under the family Schefflera. Good agreements are obtained between the measurements and predictions.
5. Topic: Centrifuge modelling and testing of the combined mechanical and hydrological effects of plant roots on slope hydrology and stability
Task 1: In order to simulate transpiration-induced suction in a geotechnical centrifuge, a novel root system that enables suction to be induced and controlled at high-g are developed and verified. This new root system consists of a high air-entry value (AEV) porous filter, cellulose acetate, which has scaled mechanical properties, including tensile strength, elastic modulus, and axial rigidity, similar to living roots. This filter is fully saturated with de-aired water and it is connected to an airtight chamber for controlling vacuum pressures. The function of the water-saturated porous filter is to maintain hydraulic gradient between soil and the root system for any vacuum pressure lower than the AEV of the filter. Any reduction of soil moisture due to applied vacuum hence induces suction. Suctions induced by the new root system were verified to be consistent with those induced by a living tree (Schefflera heptaphylla) at 1-g and that retained by vegetation in the field. Both vertical and horizontal influence zones of suction of the living tree were captured. For centrifuge tests carried out at 15-g, suctions of up to 25 kPa can be simulated.
Task 2: In this task, a series of centrifuge pull-out tests were conducted using the newly developed plant root models developed in Task 1 for testing the effects of transpiration of root pull-out resistance. Three idealised and simplified root geometries were considered, namely tap-, heart- and plate-shaped. All tests were carried out under identical rainfall conditions at high-g, where the stress state of the soil and root dimensions can be modelled more closely to field conditions. The test results revealed that, after a rainfall event, pore water pressure retained by the tap- and heart-shaped roots (which have longer root depths) was much lower than that retained by the plate-shaped root. The presence of soil suction enhanced the pull-out resistance significantly due to increased tendency of constraint dilatancy upon soil–root interface shearing. Among the three root geometries, the tap- and heart-shaped roots were identified to be more favourable in resisting pull-out because they consisted of a vertical taproot component that effectively mobilised soil–root interface friction against pull-out.
Task 3: After development plant root models in Task 1 and characterizing the mechanical properties of these root models in Task 2, this task aims to quantify both the hydrological and mechanical effects of root geometry on the stability of shallow slopes. Centrifuge tests were conducted to measure soil suction in slope models supported by root models. These root models exhibit three different representative geometries (i.e. tap, heart and plate) and could simulate the effects of transpiration. The measured suction was then back-analysed through a series of finite element seepage-stability analyses to determine the factor of safety (FOS). It is revealed that after a rainfall event with a return period of 1000 years, the slope supported by heart-shaped roots retained the highest suction within the root depth, and thus, this type of root provided the greatest stabilisation effects. The FOS of the slope supported by the heart-shaped roots, through both mechanical reinforcement and transpiration-induced suction, is 16 and 28 % higher than that supported by the tap- and plate-shaped roots, respectively.
Influences of transpiration-induced suction and mechanical reinforcement of different root geometries (i.e. tap and heart-shaped) to the stability of slopes of different gradients were also investigated by subjecting them to an intense rainfall with an intensity of 70 mm/h (corresponding to a return period of 1000 years),. The same root models were used to simulate transpiration-induced suction in the centrifuge. Transient seepage analyses were performed using a finite-element software, SEEP/W, to back-analyse the suction responses due to transpiration and rainfall observed in the centrifuge. Subsequently, the back-analysed suction was used to assess the FOS of the slopes using another software, SLOPE/W. It is revealed that heart-shaped roots provided greater stabilization effects to a 60° clayey sand slope than tap-shaped roots. The heart-shaped roots induced higher suction, leading to 14% reduction of rainfall infiltration and 6% increase in shear strength. Although transpiration-induced suction in a 45° slope was reduced to zero after the rainfall, mechanical root reinforcement was found to be sufficient to maintain slope stability.
The outcome of all research tasks have been synthesized mainly in the forms of journal articles, 24 in total, and oral presentation. The full list of publications is given in Section 2 of this report.
(1) To systematically characterise soil-water-plant interaction by identifying how soil state (e.g. void ratio) and root characteristics (e.g. root length distribution) affect transpiration-induced suction in laboratory;
(2) To parameterise root-water uptake by establishing physically-based relationship with hydraulic properties;
(3) To investigate long-term (at least 20 years) deformation characteristics (serviceability limit state) of BioEPIC slope under different scenarios of seasonal climate conditions from physical model tests; and
(4) To investigate stability and failure mechanism(s) (ultimate limit state) of BioEPIC slope under different scenarios of intense rainfall due to the climate change from physical model tests
The following specific research tasks has been conducted to address all four objectives. A synopsis of each task, including key findings, is given below.
1. Topic: Correlating hydrologic reinforcement of vegetated soil with plant traits – This aims to quantify and compare the hydrological reinforcement induced by ten woody species widespread in Europe; and then to examine and identify measureable plant traits that may be associated with the hydrologic reinforcement of soil. It is revealed that C. avellana, I aquifolium and U. europaeus, exhibited contrasting hydrologic reinforcement down the soil profile. While I. aqufolium provided hydrologic reinforcement mainly in shallow soil, U. europaeus induced greater hydrologic reinforcement in the entire soil profile down to 0.7 m depth. During winter, evergreen species had a much slower water uptake rate and smaller increase in matric suction compared with summer. Despite their slow increase in matric suction during winter, the magnitude of suction preserved was much higher than the value recorded in deciduous soil columns, thus providing greater increases in soil strength. As far as fast establishment and hydro-mechanical reinforcement are concerned, the nitrogen-fixing U. europaeus may be a more suitable candidate for soil eco-engineering purposes than the “resources saver” I. aquifolium.
2. Topic: Root biomechanical properties – This aims to measure and quantify root biomechanical properties during their early stage establishment, which represents a particularly challenging period for these plants; analyse and compare the tensile strength-diameter relationship of the ten woody species. The root tensile strength-diameter relationships of the ten species highlighted three different trends. The commonly-quoted negative power law in the literature was applicable for the strength-diameter data of only two out of the ten species tested, E. europaeus and U. europaeus. Indiscriminate use of the negative power law model to describe root tensile strength-diameter relation could overestimate root reinforcement, especially in the range of very small diameter roots where high root tensile strength would tend to be predicted by the model. Caution should be taken when this root biomechanical model is assumed for different taxa or for the same species that grows under different environmental conditions or during the challenging establishment period.
3. Topic: Effects of plant roots on soil-water retention and induced suction in vegetated soil – This task is to compare the effects of the two hydrological mechanisms, namely (a) change of soil water retention ability induced by roots and (b) plant evapotranspiration (ET) and the associated root-water uptake, on suction induced in vegetated soil, and then to quantify their relative importance. The calculation shows that for a prolonged 24-h rainfall with a return period of 10 years, mechanical root reinforcement is effective to stabilise the shallow soil of up to 0.5 m depth generally, where most of the root biomass exists. Hydrological reinforcement considering both the effects (a) and (b) provides much significant effects of soil stabilisation in deeper depths (i.e. 1–2 m), where slip failure is normally of major concern. The presence of roots in a vegetated slope preserves higher suction, hence higher shear strength, after rainfall, as compared to a bare slope. Reduction in saturated permeability (ks) due to the presence of intact roots is the most predominant hydrological effects. In contrast, increase in ks due to the presence of dying/decaying roots could be detrimental to slope stability at 1–2 m depth due to the reduced ability of the vegetated slope to preserve suction. Other effects, in particular the root water uptake through ET during rainfall, are minimal. Their contribution to slope stabilisation could be practically negligible.
4. Topic: A new and simple water retention model for root-permeated soils – This task aims to quantify and model the root effects on both soil water retention curve (SWRC) and soil hydraulic conductivity function (SHCF) of soils using the instantaneous profile method. Four types of vegetated soil, namely bare, grass-only, tree-only and mixed tree-grass soils, were subjected to a controlled drying-wetting cycle in a plant room. Plant roots affect the air-entry value (AEV), saturated permeability (ks) and reduction rate of unsaturated hydraulic conductivity (with respect to suction) most significantly, but it does not affect the reduction rate of volumetric water content much. When planted with single species, the AEV of silty sand increased, while saturated ks and reduction rate of unsaturated permeability as suction decreased. However, under the mixed planting conditions, opposite results are found. A new void ratio function is developed to model the observed decrease and increase in soil void ratio due to root occupancy (upon growth) and root shrinkage (upon decay), respectively, in an unsaturated vegetated coarse-grained soil. The function requires two root parameters, namely root volume ratio and root decay ratio, both of which can be readily measured through root excavation and image-based analysis. The new function is incorporated into a void ratio-dependent SWRC model for predicting SWRC of vegetated soils. Similarly, the same function can be combined with the Kozeny-Carman equation for predicting ks. The model prediction is then compared with a set of new field test data and an existing laboratory dataset for a silty sand vegetated with a tree under the family Schefflera. Good agreements are obtained between the measurements and predictions.
5. Topic: Centrifuge modelling and testing of the combined mechanical and hydrological effects of plant roots on slope hydrology and stability
Task 1: In order to simulate transpiration-induced suction in a geotechnical centrifuge, a novel root system that enables suction to be induced and controlled at high-g are developed and verified. This new root system consists of a high air-entry value (AEV) porous filter, cellulose acetate, which has scaled mechanical properties, including tensile strength, elastic modulus, and axial rigidity, similar to living roots. This filter is fully saturated with de-aired water and it is connected to an airtight chamber for controlling vacuum pressures. The function of the water-saturated porous filter is to maintain hydraulic gradient between soil and the root system for any vacuum pressure lower than the AEV of the filter. Any reduction of soil moisture due to applied vacuum hence induces suction. Suctions induced by the new root system were verified to be consistent with those induced by a living tree (Schefflera heptaphylla) at 1-g and that retained by vegetation in the field. Both vertical and horizontal influence zones of suction of the living tree were captured. For centrifuge tests carried out at 15-g, suctions of up to 25 kPa can be simulated.
Task 2: In this task, a series of centrifuge pull-out tests were conducted using the newly developed plant root models developed in Task 1 for testing the effects of transpiration of root pull-out resistance. Three idealised and simplified root geometries were considered, namely tap-, heart- and plate-shaped. All tests were carried out under identical rainfall conditions at high-g, where the stress state of the soil and root dimensions can be modelled more closely to field conditions. The test results revealed that, after a rainfall event, pore water pressure retained by the tap- and heart-shaped roots (which have longer root depths) was much lower than that retained by the plate-shaped root. The presence of soil suction enhanced the pull-out resistance significantly due to increased tendency of constraint dilatancy upon soil–root interface shearing. Among the three root geometries, the tap- and heart-shaped roots were identified to be more favourable in resisting pull-out because they consisted of a vertical taproot component that effectively mobilised soil–root interface friction against pull-out.
Task 3: After development plant root models in Task 1 and characterizing the mechanical properties of these root models in Task 2, this task aims to quantify both the hydrological and mechanical effects of root geometry on the stability of shallow slopes. Centrifuge tests were conducted to measure soil suction in slope models supported by root models. These root models exhibit three different representative geometries (i.e. tap, heart and plate) and could simulate the effects of transpiration. The measured suction was then back-analysed through a series of finite element seepage-stability analyses to determine the factor of safety (FOS). It is revealed that after a rainfall event with a return period of 1000 years, the slope supported by heart-shaped roots retained the highest suction within the root depth, and thus, this type of root provided the greatest stabilisation effects. The FOS of the slope supported by the heart-shaped roots, through both mechanical reinforcement and transpiration-induced suction, is 16 and 28 % higher than that supported by the tap- and plate-shaped roots, respectively.
Influences of transpiration-induced suction and mechanical reinforcement of different root geometries (i.e. tap and heart-shaped) to the stability of slopes of different gradients were also investigated by subjecting them to an intense rainfall with an intensity of 70 mm/h (corresponding to a return period of 1000 years),. The same root models were used to simulate transpiration-induced suction in the centrifuge. Transient seepage analyses were performed using a finite-element software, SEEP/W, to back-analyse the suction responses due to transpiration and rainfall observed in the centrifuge. Subsequently, the back-analysed suction was used to assess the FOS of the slopes using another software, SLOPE/W. It is revealed that heart-shaped roots provided greater stabilization effects to a 60° clayey sand slope than tap-shaped roots. The heart-shaped roots induced higher suction, leading to 14% reduction of rainfall infiltration and 6% increase in shear strength. Although transpiration-induced suction in a 45° slope was reduced to zero after the rainfall, mechanical root reinforcement was found to be sufficient to maintain slope stability.
The outcome of all research tasks have been synthesized mainly in the forms of journal articles, 24 in total, and oral presentation. The full list of publications is given in Section 2 of this report.