Effects of vegetation on the angle of internal friction of ... · Effects of vegetation on the...

61For. Snow Landsc. Res. 82, 1: 61–77 (2009)

Effects of vegetation on the angle of internal friction of a moraine

Frank Graf1, Martin Frei2 and Albert Böll3

1 WSL Institute for Snow and Avalanche Research SLF, CH-7260 Davos Dorf, [email protected]

2 Forest Department Grisons, CH-7000 Chur, Switzerland. [email protected] WSL Swiss Federal Institute for Forest, Snow and Landscape Research, CH-8903 Birmensdorf,

Switzerland. [email protected]

AbstractVegetation clearly affects soil strength, but how to take these effects into account in conventionalcalculations of slope stability is still an unsolved problem. To quantify the important influence ofplant roots on shear strength of a moraine, we performed isotropic, consolidated-undrained triax-ial compression tests with different confining pressures (σ3' = 50, 75, 100 kPa). Three differenttypes of samples were tested: A) planted soil and B) pure soil at low dry unit weight (γ ≈ 15.5kN/m3) as well as C) pure compacted soil at γ ≈ 19 kN/m3. The planted samples were preparedwith alder seeds (Alnus incana). For each sample type, failure lines (kf lines) were calculated usingthe peak shear strength points of the corresponding p'-q paths. Robust statistics were performedto fit the failure lines and to test for significance.No differences were found in the cohesion (c') of the different soils. However, there was a signifi-cant difference in the angle of internal friction (Φ') of about 5° between the samples of pure soilat low dry unit weight and those of both compacted and planted soil. The vegetation effect is thusapparent as an increase in the angle of internal friction Φ' in planted soil compared to pure soil atthe same dry unit weight. This finding can also be considered as a virtual increase in soil density(from γ ≈ 15.5 to γ ≈ 19 kN/m3).

Keywords: triaxial compression tests, plants, slope stability, robust statistics

1 Introduction

During the last decade there has been a pronounced increase in the number of catastrophicevents including shallow landslides and erosion processes after heavy rainstorms, particularlyin mountainous regions, which has raised public awareness of the hazard (GÄRTNER 2003;RICKLI et al. 2004; BEZZOLA and HEGG 2007). Slope instability is thus a major concern forall those responsible for the protection of human lives and infrastructure against naturalhazards. Consequently, methods to protect slopes against processes of erosion and sliding,and to stabilise those already affected are needed, as are reliable models and ways of esti-mating and predicting slope stability, and of calculating the factor of safety against failure.

For protecting slopes, joint technical and biological measures are particularly suitable asthey combine environmental compatibility with cost effectiveness and durability (BÖLL

et al. this issue). However, compared to purely technical constructions, these nowadayscalled eco-engineering methods face a major disadvantage. Until today, no very satisfactoryways of quantifying the biological effects on soil stability have been developed (BÖLL andGRAF 2001; CAZZUFFI and CRIPPA 2005). The contributions of technical constructions tosoil strength can be directly fitted into conventional slope stability and security calculations,but this is not yet appropriately possible for biological measures.

The public has recently become much more aware of natural hazards in general, anddemand for security has increased. Thus, precise information about the effects of plants on

62 Frank Graf et al.

slope stability is needed, and the development of methods to provide evidence for vegetationeffects is an urgent objective. In particular, we need to find ways of including the influence ofplants in conventional models for estimating the stability of natural slopes and embank-ments (JANBU 1954), which are generally based on the Mohr-Coulomb failure criterion (τf =c'+ σ'tanφ', in LANG et al. 1996). For this purpose, the vegetation effects need to be assignedto shear strength, i.e. to the angle of internal friction Φ' or to the cohesion c'.

To calculate and model vegetation effects on soil stability, suitable measurement tech-niques are necessary to properly address the relevant parameters. During the early historyof soil mechanics, the direct shear test was the most popular approach, but it has some considerable disadvantages. To overcome some of its most serious limitations, the triaxialcompression apparatus was developed in the 1930s (CASAGRANDE 1936; BARNES 1995). Thetriaxial compression test is more demanding and time consuming than the direct shear test,but it is also much more versatile. Several improvements have made it the appropriatechoice, today, for experimental investigations of complex stress paths (LANG et al. 1996).

Nevertheless, direct shear tests are still very popular because they are simpler to applyand can be used to conduct experiments in the field. Accordingly, this method has beenwidely applied in investigations of soil reinforced by roots (WALDRON 1977; WU et al. 1979;STYCZEN and MORGAN 1995; WU and WATSON 1998; CAZZUFFI and CRIPPA 2005), andtends to be the first approach used to quantify vegetation effects on soil stability. Evenrecently proposed approaches using fibre bundle models adapted for estimating themechanical effects of vegetation on soil stability are finally based on direct shear tests(POLLEN and SIMON 2005; POLLEN 2007).

WU (1984) proposed, probably for methodical reasons, implementing the effects of plantroots on soil stability as an additive constant of the cohesion c' in the Mohr-Coulomb failurecriterion (Mohr-Coulomb: τf = c'+ σ'tanΦ'; Wu: sr = c'+ σ'tanΦ'+ cr'; with sr as the shearstrength τf and r for root). This approach has the advantages that, as a plain strength value,the additional cohesion cr' mobilised by roots, may be measured relatively simply by directshear tests. Nevertheless, the soil stability conditions and, correspondingly, the effects ofroots in the near-surface zone of the soil are not satisfactorily described by the cohesion c'.The stress-dependent term σ'tanΦ' represents the proper characteristics much better.During a simple shear test in which the only stresses measured are the normal and shearstresses on horizontal planes, it is not possible to define the stress state completely. As aresult, the failure stresses measured may not represent the true strength of the soil. Theresults of laboratory tests on Cowden Till and on Blue London Clay (ATKINSON et al. 1991)showed that the strengths measured in simple direct shear tests differed from those meas-ured in triaxial compression tests. The conventional interpretation of direct shear tests leadsto a false cohesion intercept with friction angles smaller than those measured in triaxialcompression tests (ATKINSON et al. 1991).

Recent experimental investigations on fibre reinforcement in sand yielded controversialfindings, depending on the method applied. Using direct shear tests, YETIMOGLU andSALBAS (2003) found no improvement in the shear strength of the composite compared topure sand, and OPERSTEIN and FRYDMAN (2000) reported an essentially constant angle ofinternal friction of soil reinforced by roots, but an increase in the apparent cohesion withincreasing cross-sectional area and tensile strength of the roots. Accordingly, they inter -preted the general increase in shear strength of the composite as the result of an increase incohesion.

However, analyses based on triaxial compression tests revealed an increase in the angleof internal friction of a composite (fibre reinforced sand) compared to the untreated granularmatrix (STAUFFER and HOLTZ 1995; CONSOLI et al. 2002). The addition of fibres to cohesion-less pure sand yielded an increase in the angle of internal friction without any change in thecohesion but when added to cemented sand (with cohesion), the increase in the angle of

63For. Snow Landsc. Res. 82, 1 (2009)

internal friction went along with a decrease in cohesion (CONSOLI et al. 2002). Furthermore,it was found that the reinforcement effect generally correlates positively with the fibreaspect ratio, and, if the aspect ratio and concentration of fibres are kept constant, the composite strength is positively correlated with the length of the fibres (MICHALOWSKI andCERMÁK 2003).

The triaxial compression test is better suited than the direct shear test at representingprocesses and characteristics of the superficial soil layers reasonably well. There is no rotation of the principal stresses, and, although stress concentrations still exist, they are significantly less. Normal stress is applied in three dimensions, and the area of shearing doesnot change during the test procedure. Furthermore, the failure plane can occur anywhere,and the stress paths can be controlled reasonably well. This means that complex stress pathsin the field can be more effectively modelled in the laboratory. In particular, if undrainedshear strength and the effective stress parameters of low-permeability material are needed,the triaxial compression test (consolidated undrained, with pore water pressure measure-ments) is by far more adequate (LAMBE and WHITMAN 1979; HOLTZ and KOVACS 1981;BARNES 1995; HEAD 1998).

The present study was undertaken to examine the influence of roots on the shearstrength of moraine in a recent slide area. The focus is on the triaxial testing of moraine witha programme including pure soil and planted soil samples with low dry unit weight (γ ≈ 15.5kN/m³) as well as on compacted pure soil samples with high dry unit weight (γ ≈ 19 kN/m³).Consolidated undrained triaxial testing was performed at different confining pressures toassess the effects of roots and dry unit weight on soil stability. The concept of “virtual density”(BÖLL and GRAF 2001) is proposed as a practical description of the shear strength of plantedsoils. “Virtual density” represents the higher dry unit weight of pure soil with a shearstrength comparable to planted soil at lower dry unit weight.

2 Material and methods

2.1 Soil analysis and sample preparation

The soil investigated is a moraine of the subalpine landslide area “Schwandrübi” in CentralSwitzerland. The grain size distribution was analysed and the soil material classified (ASTMD422 2000, ASTM C136 2001, ASTM D2487 2002). Furthermore, the maximum dry unitweight at optimum water content, the liquid and plastic limits as well as the plasticity indexwere determined (ASTM D698 2000, ASTM D4318 2000). The coefficient of permeabilitywas assessed with the oedometer method (LANG et al. 1996).

From oven-dried soil material (24 h at 105° C), the fractions ≤ 10 mm were used to preparethe soil samples for the triaxial compression tests. The soil material was moistened to a watercontent of 6 % and tamped into PVC-plastic tubes (diameter: 70 mm; height: 140 mm).Subsequently, the specimens were dynamically compacted in three layers to the given dryunit weight and three different treatments were applied: A) soil (γ ≈ 15.5 kN/m³) plantedwith Alnus incana (L.) Moench, B) untreated soil with a dry unit weight of γ ≈ 15.5 kN/m³,and C) untreated compacted soil with a dry unit weight of γ ≈ 19 kN/m³. For treatment A) 15alder seeds were applied to each sample and reduced to three seedlings after four weeks ofgrowing. The samples were arranged completely randomly and maintained in a greenhousefor 20 weeks with 16 h of daylight daily and temperatures of 17° C during the day and 10° Cat night. Totally, 15 samples were finally evaluated, with five from each treatment.

After the testing procedure, the cleaned plant roots were analyzed with a flatbed scannerand the root length was determined with the software WinRhizo® (2000). The root lengthper sample volume [cm/cm3] was used as an indicator for plant growth.


2.2 Triaxial compression tests

Isotropic consolidated-undrained triaxial compression tests were performed to determinethe shearing resistance of the soil samples (ASTM D 4767-95, 2000). Following FREI et al.(2003), the specimens were surrounded with a filter-paper drain and enclosed within a double latex membrane. The specimens were saturated with de-aired water under a pressureof 20 kPa. The initial degree of saturation was improved by applying a back pressure of 1200kPa following the recommendations of BISHOP and HENKEL (1957). The B-value waschecked according to DAY (2001). All samples had B-values greater than 0.98 and wereassumed to be fully saturated. For each of the three treatments (A, B, C), compression testswere performed with effective confining pressures of σ3' = 50, 75, and 100 kPa with one ortwo specimens (Table 1). Axial stress, axial strain and pore water pressure were recordedduring the shearing process. The axial strain rate was 0.015 % per minute. All tests werestopped once the loading piston had covered a distance of 20 mm.

Table 1. Confining pressures σ3' [kPa] and corresponding number of samples [n] tested for each of thethree different treatments: A (soil planted with Alnus incana with γ ≈ 15.5 kN/m3), B (untreated soilwith γ ≈ 15.5 kN/m3), and C (untreated compacted soil with γ ≈ 19 kN/m3).

TreatmentA B C

confining pressure σ3' [kPa]: 50 75 100 50 75 100 50 75 100number of samples [n]: 2 1 2 2 1 2 2 1 2

dry unit weight γ [kN/m3] 15.3 15.6 18.9

The shearing strength was determined according to the deviator stress (σ1–σ3) at the yield-strength point in the stress-strain diagram for the highly compacted soil samples. This corresponds to the “peak” in the stress-strain diagram and to the local maximum in the p'-qdiagram (ZHU and ANDERSON 1998). For the planted and the pure soil samples at low dryunit weight, the stress-strain values were determined at “phase transformation”, which characterises more ultimate sample behaviour (NEGUSSEY et al. 1988). In the p'-q diagram,this state is determined by the inflexion point of the path (ZHU and ANDERSON 1998;SPRINGMAN et al. 2003).

2.3 Statistical analysis

All statistical calculations were performed with the software R 2.5.0 (2007). For each of thetreatments A, B, and C, failure lines (kf lines) were calculated, based on the peak shearstrength points of the corresponding p'-q paths, by performing all combinations of 3 out of 5[C(5,3)] without repetition. This resulted in 10 kf lines for each treatment (Equation 1).

[1]

where C (binomial coefficient, also known as the “choose function”) is the number of all k-combinations from a set of n elements without repetitions, with k = 3 (confining pressure)and n = 5 (number of samples); see also Table 1.

10)!35(!3

!5;

)!(!!

),( =−−

=⎟⎟⎠

⎞⎜⎜⎝

⎛=

knkn

k

nknC

⎫⎪⎪⎬⎪⎪⎭ ⎫⎪⎪⎬⎪⎪⎭⎫⎪⎪⎬⎪⎪⎭


To fit the failure lines, robust linear statistics were performed based on an MM-estimator(R-function: rlm) as proposed by VENABLES and RIPLEY (2002). Confidence intervals (1-α)for the difference between 20 % trimmed means (WILCOXON 2005; function trimpb2) werecalculated to compare the failure lines (intercept = cohesion [c'] and slope = angle of internalfriction [Φ']) of the three treatments A, B, and C. The stress-strain data of the triaxial com-pression tests were subjected to local polynomial regression fitting (R-function:smooth.spline) after HASTIE and TIBSHIRANI (1990).

Additionally, robust Kruskal-Wallis rank sum tests (R-function: kruskal.test) and pair-wise comparisons using Wilcoxon rank sum tests (R-function: pairwise.wilcox.test) were performed with correction for multiple testing according to HOLM (1979).

In addition to the conventional mean values and standard deviations, a robust locationparameter and measure of scale are given where appropriate, with the median and the medianabsolute deviation (mad), respectively.

3 Results

3.1 Soil and samples

The grain size distribution of the moraine, including the coarse parts up to 63 mm, is: 75.3 %gravel, 15.5 % sand, 7.1 % silt, and 2.1 % clay. The corresponding values of the soil materialused for the triaxial compression tests (grain size ≤ 10 mm) are: 40.4 % gravel, 37.1 % sand,17.0 % silt, and 5.0 % clay (Fig. 1). According to ASTM D2487 (2000), the moraine is classified as a clayey gravel with sand (GC-CL).

particle size [mm]

perc

ent p

assi

ng [%

]

0.001 0.01 0.1 1 10 1000

10

20

30

40

50

60

70

80

90

100clay silt sand gravel

max. particle size: 10 mm max. particle size: 63 mm

Fig. 1. Grain size distribution of the “Schwandrübi” soil material: � = distribution curve of the moraineinvestigated, including the coarse grains up to 63 mm; � = distribution curve of the material used for triaxial compression tests, including grains ≤ 10 mm.


Further properties of the moraine include: the liquid limit (wL: 21.5 %), plastic limit (wP: 12.9 %), the plasticity index (IP: 8.6 %), the specific gravity (Ds: 26.9 kN/m3), the maximumdry unit weight (γ: 21.9 kN/m3) with optimum water content (w: 7.9 %), and the coefficientof permeability (k: 6 x 10–10 m/s).

During the 20 weeks in the greenhouse, changes in the initial dry unit weight occurred.These were more pronounced in the planted soil (Δγ: 1.84 kN/m3 ± 0.67) and pure soil (Δγ:1.86 kN/m3 ± 0.43) samples of low dry unit weight than in the compacted soil samples (Δγ:0.6 kN/m3 ± 0.99), although the differences were not significant.

The comparison of the dry unit weights before and after the consolidation process of thethree different treatments revealed significant differences in the values before consolidation(Kruskal-Wallis p-value: 0.0087) but not afterwards (Kruskal-Wallis p-value: 0.1111). Thepair-wise comparisons of the dry unit weights before consolidation showed that those of thecompacted soil samples were significantly higher than those of both the planted soil samples(p-value: 0.033) and the pure soil samples at low dry unit weight (p-value: 0.033) (Fig. 2,Table 2). The mean values and corresponding standard deviations of the three treatmentswere 15.26 kN/m3 ± 0.43 (A), 15.66 kN/m3 ± 0.64 (B), 18.85 kN/m3 ± 0.76 (C) before consoli-dation, and 19.98 kN/m3 ± 0.45 (A), 20.60 kN/m3 ± 0.75 (B), 20.70 kN/m3 ± 0.42 (C) afterconsolidation.

Within the plants in treatment A, the root length per soil volume [cm/cm3] varied con -siderably, with a minimum of 0.61 cm/cm3, a maximum of 2.17 cm/cm3, and a mean value of1.26 cm/cm3 ± 0.71 (Table 2).

15

16

17

18

19

20

21

before consolidation

dry

unit

wei

ght

γ [k

N/m

3 ]

planted pure compact

15

16

17

18

19

20

21

after consolidation


dry

unit

wei

ght

γ [k

N/m

3 ]

Fig. 2. Dry volume weight γ [kN/m3] of the samples of planted, pure, and compacted soil before and afterthe consolidation process.


3.2 Triaxial compression tests

The yielding at axial strain at failure εf of the individual specimens in treatment A (plantedsoil) varied within the range of 2.3 % and 7.6 %, in treatment B (pure soil at low dry unitweight) between 2.4 % and 6 %, and in treatment C (compacted soil) within 1 % and 2.2 %(Table 2). No significant difference in εf was detected between the treatments A and B (p-value: 1.0), but both have significantly higher values compared to treatment C (p-values:0.024 each).

The deviator stresses (q) and, in good approximation, the mean effective stresses (p') ofthe treatments A and B indicate a positive correlation with the confining pressure (Figs. 3,4). This is, however, not the case for treatment C. There, the specimen tested at 75 kPa by farexceeds the others in both the mean effective stress (p') and the deviator stress (q).Furthermore, one of the two specimens tested at 100 kPa has a lower q-value than one ofthose tested at 50 kPa (Table 2, Fig. 5).

Table 2. Properties of the individual samples used for the triaxial compression tests. pld = planted soilsamples (treatment A); sld = pure soil samples (treatment B); shd = pure, compacted soil samples (treat-ment C); V0 = volume of the samples after the greenhouse period; γbc = dry unit weight before con -solidation; γac = dry unit weight after consolidation; σ3 = confining pressure; p' = mean effective stress(σ1' + σ3')/2; q = deviator stress (σ1–σ3)/2; εf = strain at failure; root = root length per volume.

sample V0 γbc γac σ3 p' q εf root

pld 1 494 14.5 20.1 50 38.14 19.39 2.29 0.61pld 2 497 15.4 19.5 50 27.03 16.21 5.13 1.88pld 3 488 15.5 20.7 75 35.99 23.63 4.94 2.17pld 4 451 15.5 19.8 100 51.26 31.81 2.98 0.82pld 5 455 15.4 19.8 100 44.47 27.21 7.60 0.81

sld 1 467 16.2 21.0 50 26.80 15.35 5.84 —sld 2 467 15.2 19.3 50 33.96 17.85 2.84 —sld 3 478 15.2 20.7 75 38.52 23.21 4.47 —sld 4 439 16.5 21.2 100 66.07 36.92 2.40 —sld 5 481 15.0 20.8 100 55.91 32.58 5.96 —

shd 1 528 19.1 20.5 50 41.87 25.22 2.20 —shd 2 542 19.2 20.1 50 59.18 36.86 1.74 —shd 3 537 19.4 20.9 75 90.72 58.25 1.75 —shd 4 524 19.0 20.8 100 75.10 41.07 1.02 —shd 5 429 17.5 21.2 100 62.97 30.82 1.07 —

dim. [cm3] [kN/m3] [kN/m3] [kPa] [kPa] [kPa] [%] [cm/cm3]

The variance of the 10 failure lines (kf lines) of each treatment, represented by robustregression lines, is least in soil at low dry unit weight, slightly more in planted soil, and mostin compacted soil (Figs. 3–5). Concerning the treatments A and B, the comparison of thetrimmed means of each of the 10 transformed regression lines reveals a significant differ-ence in the slope, i.e. in the angle of internal friction [Φ'], and no significant difference in theintercept, i.e. in the cohesion [c'] (Table 3, Fig. 6).


Table 3. Comparison of intercept (cohesion c' [kPa]) and slope (angle of internal friction Φ' [°]) ofuntreated soil at low dry unit weight (B) and planted soil (A), based on 1-α confidence intervals forthe difference between 20 % trimmed means (WILCOXON 2005); est. diff. = estimated difference.

lower limit upper limit ø planted ø untreated est. diff. p-value dim.

intercept [c']: –2.4894 1.4011 –1.5635 –0.9269 –0.4503 0.6345 [kPa]slope [tan(Φ')]: 0.0146 0.0974 0.6766 0.6220 0.0583 0.0076 [°]

No significant differences were detected however, between the treatments A and C orbetween B and C neither in the slope [tanΦ'] with a p-value of 0.9599 for A and C and oneof 0.8607 for B and C nor in the intercept [c'] with a p-value of 0.6979 for A and C and one of0.6997 for B and C. The range of dispersion of intercepts [c'] and slopes [tanΦ'] was rathernarrow for treatments A and B and remained within the same array but it was considerablywider for treatment C (Fig. 6).

The estimated angle of internal friction [Φ'] and cohesion [c'] for the three treatmentswere 39.39° and –0.86 kPa for A, 34.25° and –0.04 kPa for B, and 40.14° and –3.89 kPa for C,respectively. The plants’ effect on the soil is reflected by an increase of 5° in the angle ofinternal friction, compared to pure soil samples at the same dry unit weight (γ ≈ 15.5 kN/m3),if we ignore the cohesion, which is virtually zero. Furthermore, the planted samples (A) atlow dry unit weight (γ ≈ 15.5 kN/m3) mobilise the same angle of internal friction of about 40°as the compacted soil samples (C) at a dry unit weight of γ ≈ 19 kN/m3 (Fig. 7).

0 20 40 60 80 100

0

10

20

30

40

50

60

ε−, p'q−diagramγ of untreated soil: ø=15.6 kN/m3

ε [%] ; p'−value [kPa]

q−va

lue

[kP

a]

(0%) (2%) (4%) (6%) (8%) (10%)

⎡⎣⎢

σ1−σ3

2⎤⎦⎥

⎡

⎣⎢σ1'+σ3'

2

⎤

⎦⎥

Fig. 3. Untreated soil at γ ≈ 15.5 kN/m³. To improve legibility, the stress-strain data are not shown com-pletely. Black solid lines = stress-strain lines; dashed and solid blue curves = specimens tested at 50 kPa;solid red curve = specimen tested at 75 kPa; dashed and solid green curves = specimens tested at 100 kPa confining pressure; violet dotted lines = kf lines; � = inflexion points.


806040200 100

0

10

20

30

40

50

60

ε−, p'q−diagramγ of planted soil: ø=15.3 kN/m3


q−va

lue

[kP

a]

(0%) (2%) (4%) (6%) (8%) (10%)

⎡⎣⎢

σ1−σ3

2⎤⎦⎥

⎡

⎣⎢σ1'+σ3'

2

⎤

⎦⎥

Fig. 4. Planted soil at γ ≈ 15.5 kN/m³. For legend, see Fig. 3.

0 20 40 60 80 100

0

10

20

30

40

50

60

ε−, p'q−diagramγ of compacted soil: ø=18.9 kN/m³


q−va

lue

[kP

a]

(0%) (2%) (4%) (6%) (8%) (10%)

⎡

⎣⎢σ1'+σ3'

2

⎤

⎦⎥

⎡⎣⎢

σ1−σ3

2⎤⎦⎥

Fig. 5. Untreated soil at γ ≈ 19 kN/m³. For legend, see Fig. 3.


−5

0

5

10

soil treatments

inte

rcep

t [kP

a]


0.40

0.45

0.50

0.55

0.60

0.65

0.70

soil treatments

slop

e


Fig. 6. Comparison of intercept and slope values of the regression lines of each of the three treatments.

0 10 20 30 40 50

0

10

20

30

40

50

σ ’ [kPa]

τ [kP

a]

shear angle: 39.39°intercept: −0.86 kPa



Fig. 7. Comparison of cohesion c' [kPa] and angle of internal friction Φ' [°] of the three different treat-ments; planted, pure, and compacted soil. Dotted black = compacted soil (γ ≈ 19 kN/m3); dashed lightgrey = pure soil (γ ≈ 15.5 kN/m3); and solid dark grey = planted soil (γ ≈ 15.5 kN/m3). The dash-dottedblack line represents the curve of the compacted soil displaced parallel to the intercept of the curve ofthe planted soil.


4 Discussion

The shear tests were performed with confining pressures (σ3') of 50, 75 and 100 kPa, repre-senting the stress conditions in the superficial soil area, where root growth mainly occurs.Although even lower confining pressures would have been desirable, the value of 50 kPamarked the limit with respect to the practicability of shear tests given the laboratory equipment. However, the chosen conditions are well in accordance with those of otherresearchers investigating the strength of either soils reinforced by the roots of grass species(KIRSTEN and FOURIE 1999), or cemented and uncemented sand reinforced by fibres(CONSOLI et al. 2002; MICHALOWSKI and CERMÁK 2003) and 3D inclusions (ZHANG et al.2006).

The isotropic consolidated undrained triaxial compression tests clearly indicate anincrease in the shear strength of the moraine investigated (clayey gravel with sand, GC-CL)with the presence of alder roots (Alnus incana). The gain in soil stability due to plant roots isexclusively based on the increase of the angle of internal friction Φ' from 34.3° to 39.4° with-out any noticeable change in apparent cohesion c'. The fact that some slight negative cohe-sion c' was found may be attributed to some extent to the statistical evaluation procedure.This observation, however, is also consistent with the results of investigations of glacial till,where slightly positive or negative values were found (TIKA et al. 1996; IVERSON et al. 1998),and of fault gouge (MARONE et al. 1990; MARONE 1998; SCHOLZ 2002) and of the craterwalls of volcanoes (MOON et al. 2005).

RATHBUN et al. (2008) conducted double direct shear experiments with different glacialtill and obtained negative cohesion values as low as c' = –6.8 kPa, which is well below thevalues found in this study, namely –0.86 kPa for planted soil (γ ≈ 15.5 kN/m3), –0.04 kPa foruntreated loose soil (γ ≈ 15.5 kN/m3), and –3.89 kPa for untreated compacted soil (γ ≈ 19.0kN/m3). ATKINSON et al. (1990) conducted cylinder dispersion tests on a number of differentsoils to investigate true cohesion and dispersion in soils and found that reconstituted soilsmay exhibit positive, zero or negative cohesion. Moreover, for a given soil, the cohesioncharacteristics seemed to be governed by the chemistry of the pore water. Three characteristicbehaviours of samples were distinguished, including non-dispersive with zero cohesion, non-dispersive with positive cohesion, and dispersive with negative cohesion. Comparablebehaviours were found in soil aggregate stability tests (FREI et al. 2003; GRAF et al. 2006)conducted with the same moraine material used in the present study.

The soil-reinforcing effects of the plants are likely also to be attributable to the higherstability of unplanted soil with a higher dry unit weight, which in this case corresponded toan increase from γ ≈ 15.5 kN/m3 to γ ≈ 19 kN/m3.

Most investigations of the influence of vegetation on slope stability to date have dependedon direct shear tests. The increase in shear strength because of the roots reported in all thestudies, has normally been attributed to an increase in the apparent cohesion c' (WALDRON

1977; SHEWBRIDGE and SITAR 1989; WU and WATSON 1998; OPERSTEIN and FRYDMAN

2000; CAZZUFFI and CRIPPA 2005; WU 2007). An exception is the study of KIRSTEN andFOURIE (1999), who investigated the influence of three different grass species on the shearstrength of frictional (pure) sand and clayey sand applying drained triaxial compressiontests. In contrast to our results, they found not only an increase in the angle of internal friction Φ', but also on the apparent cohesion c'. The effect was more pronounced for theclayey sand than for the pure sand. However, there were some inconsistencies in their experiments. They added trace elements, fertiliser, organic compost, and vermiculite in theplanted pure sand, but not in the clayey sand specimens and possibly not in the unreinforcedcontrol samples with pure sand. Consequently, the root growth in the pure sand was twicethat in the clayey sand specimens. The mean values of the shear strength parameters weretaken from experiments with different confining pressures, with σ1 = 100 and 200 kPa for


both types of unreinforced soil, with σ1 = 20, 40, 100, and 170 kPa for the planted pure sand,and, with σ1 = 20, 80, and 160 kPa for the planted clayey sand. Furthermore, they applied apower function to describe the failure envelope of the unreinforced sand to allow for zerocohesion, as they expected no cohesion for this material, but used linear fits for the plantedtreatments. As a consequence, it is difficult to compare the effects of the different plants onsoil stability both in their work and with the findings from our study.

Various investigations on peaty soil have been performed using triaxial compressiontests, showing that the behaviour of the peaty soil is essentially frictional, with high angles of internal friction (Φ') and relatively small cohesion (c') intercepts (ADAMS 1965; EDIL

and DHOWIAN 1981; YAMAGUCHI et al. 1985; FARRELL and HEBIB 1998; COLA andCORTELLAZZO 2005). The high angles of internal friction are due to not entirely decom-posed fibres intersecting the failure plane. This indicates that shearing resistance depends onthe mutual orientation of fibres and failure plane. Undrained triaxial compression tests con-ducted by YAMAGUCHI et al. (1985) revealed significant differences in the angle of internalfriction of samples taken vertically (Φ' = 52°) and horizontally (Φ' = 35°).

According to WU et al. (1988), for roots to reinforce soil most effectively the vertical andhorizontal growth should take up the stress applied.

From the methodical point of view, FARRELL and HEBIB (1998) found that either the ringshear or direct shear test will most probably determine the shear resistance of the matrix,without the reinforcing effects of the fibres, a phenomenon already reported in LANDVA andLA ROCHELLE (1983) who used a ring shear device to test Canadian Radforth peat. Possiblymethodologically induced as well are the results of YETIMOGLU and SALBAS (2003) whoanalysed randomly distributed fibre-reinforced sands by applying direct shear tests, resultingin no discernible effect on shear strength, neither concerning the angle of internal friction Φ'nor the cohesion c'.

Findings from research on granular composites, reinforced by continuous filaments/fibres, result in similar method based conclusions (MOREL and GOURC 1997; CONSOLI et al.2002; MICHALOWSKI and CERMÁK 2003). Consequently, in recent years the triaxial com -pression test has become the method of choice as it yields a more promising description offibre-reinforced soils than tests based on a “direct shear” concept (SAWICKI 1983; DE

BUHAN et al. 1989; MICHALOWSKI and ZHAO 1993, 1996; MICHALOWSKI 1997; CONSOLI

et al. 2002; MICHALOWSKI and CERMÁK 2002, 2003; GHIASSIAN et al. 2004; ZHANG et al.2006; TANG et al. 2007). Investigations of a clayey soil reinforced with polypropylene fibreconducted by TANG et al. (2007) revealed an increase in both the angle of internal friction Φ'and the apparent cohesion c'. Similar findings were made by CHEN and LOEHR (2008) withOttawa sand (Grade F-75). Research on silty sand and sandy soil in general (PARK and TAN

2005) – our moraine material can be attributed to the latter – found that the inclusion ofshort fibres usually yields a distinct increase in the angle of internal friction, which is onlysometimes accompanied by a change in the apparent cohesion (MICHALOWSKI 1997;CONSOLI et al. 2002; MICHALOWSKI and CERMÁK 2002, 2003; ZHANG et al. 2006).

As found in the peat investigations, fibres orientated in the direction of maximum specimenextension contribute most to soil strength. In axisymmetric compression, as performed inour study, the maximum extension occurs in the horizontal plane. Conclusively, the con -tribution of the horizontal fibres is the largest (MICHALOWSKI and CERMÁK 2002). In ourstudy, ductile sample deformation has been observed during triaxial compression tests,resulting in barrelling and, accordingly, in increasing specimen diameter (HEAD 1994). Theapparent dependence of the angle of internal friction from the orientation of the reinforcingfibres indicates that in our study the root effect on the shear strength of soil is probably evenunderestimated as roots grew predominantly vertically, induced by the narrow diameter(70 mm) of the specimens. It further demonstrates that there may be pronounced discrep -


ancies between data gathered in the field and data obtained from laboratory experiments(GYSSELS et al. 2005).

The experimental tests of MICHALOWSKI (1997) showed that the angle of internal frictionΦ' of the composite is larger than that of the granular matrix, whereas the applied modelpredicted changes only in the apparent cohesion c'. The reason for this is that the contributionof the filaments to the composite strength was included as a dissipation term independent ofthe composite stress state. Investigations of fine and coarse sand revealed that at small fibreconcentrations (0.5 % Vol), the increase in the composite shear strength is larger for finethan for coarse sand. However, with increased fibre concentration, coarse sand benefitsmore than fine sand does.

The micromechanical behaviour of the fibre/matrix interface depends on the bindingmaterial properties in the soil, the normal stress around the fibre body, the effective contactarea and the fibre surface roughness. The interface roughness is known to play an importantrole in reinforced soil systems (TANG et al. 2007). The strain in a coarse matrix with a highfibre concentration (≈ 2 %) involves bending of the fibres to accommodate the change in therelative configuration of grains during the deformation process, which, in turn, enhances theinteraction between grains and fibres. Furthermore, the reinforcing effect is positively corre-lated to the fibre aspect ratio given by the fibre length divided by its diameter, and, if thisratio is kept constant, with the length of the fibres. The reinforcement effect is more effectivethe larger the fibres are compared to the size of the grains (MICHALOWSKI and CERMÁK

2003). Our findings from this study and GRAF et al. (2006) are similar, and we observed thatshear strength and soil aggregate stability were positively correlated with the root length persoil volume.

In clayey soils, the “bridge” effect of fibres can efficiently impede the further developmentof tension cracks and deformation of the soil. Bond strength and friction at the interfaceseem to be the dominant mechanisms controlling the reinforcement benefit. In fibre-reinforced soil, interactions that occur at the interface between the fibre surface and the claygrains play key roles in the soil-mechanical behaviour. Furthermore, fibre-sliding resistancewas observed to be strongly dependent on their surface roughness (FROST and HAN 1999;TAGNIT-HAMOU et al. 2005; TANG et al. 2007).

In our study of the moraine, we found, as in fibre-reinforced sand, that the planted speci-mens with the highest root length density (2.17 cm roots/cm3 soil) mobilised the highestshear strength. That is, the angle of internal friction (Φ' = 41.03°, c' = 0) was 1.64° above thecalculated mean value, which was derived from the 10 kf-lines of the planted specimens. Thevariance in the kf-lines of the planted specimens was higher than that of the pure soil at thesame dry unit weight, which might be explained by the apparent differences in root lengthdensity. It is assumed that there exists a positive correlation between root length density andramification with lateral roots. Therefore, a pronounced increase in the complexity of theroots’ 3D-structure is expected. Clearly, the higher complexity allows for an even moreeffective sand-fibre interaction and, consequently, further increases shear strength. In theirstudies on sand-reinforced with 3D inclusions, ZHANG et al. (2006) found confirming results,i.e. a positive correlation between the shear strength and the complexity of the 3D structuresadded.

The reinforcing effects of plant roots on soil behaviour may be explained in a similar wayto the interaction processes found between sand grains and artificial fibres. The results ofour investigations of the effects of vegetation on the stability of a moraine fit quite well withthose found in studies of soil reinforced with fibres. The isotropic consolidated undrainedtriaxial compression tests on moraine revealed a higher shear strength due to plant roots,reflected by a significant increase in the angle of internal friction Φ'.

As discussed above, the angle of internal friction Φ' of planted soil was higher than thatof pure soil at the same dry unit weight, and equal to the angle of internal friction of pure


soil samples at higher dry unit weight. This confirms the concept of the “virtual density” proposed by BÖLL and GRAF (2001). They recommend a new approach to quantify the bio-logical effects on soil stability within eco-engineering measures, i.e. to quantify biologicalinterventions on the stability of slopes affected by erosion and superficial sliding processescharacterised by loose mixtures of grains with only small proportions of fine fractions andorganic matter. A positive correlation is known to exist between the dry unit weight γ andthe angle of internal friction Φ' of pure soil. The “virtual density” approach simply explainsthe increase in soil stability due to vegetation as a virtual increase in the dry unit weight γ,and assigns the shear strength parameter Φ' of pure soil at the corresponding higher dry unitweight to the planted soil at lower dry unit weight. In this study, the planted soil with a dryunit weight of γ ≈ 15.5 kN/m3 may be regarded as unplanted compacted soil at γ ≈ 19 kN/m3.It is thus assigned the angle of internal friction of the unplanted compacted soil with Φ' =40.1°, which is substantially the same as the effective value for the planted samples (Φ' =39.4°).

The testing procedure we applied with triaxial compression tests as well as the robust statistical methods proved appropriate for assessing the effects of vegetation on soil stabilityand yielded comprehensible results. However, the operation sequence is time consumingand demanding, so the number of replications was rather low. Further tests are needed toverify our results, particularly with different soil types. As eco-engineering measures are notrestricted to one plant species only, future investigations should quantify the simultaneouseffects of different plant species on soil stability, including important symbiotic micro-organi sms that naturally occur together. Within such a community, the diversity of soil-rein-forcing structures is considerably higher and, stabilisation processes are probably accelerated.Another step to cope better with natural conditions is to adapt the specimen size and/ortheir preparation procedure to allow more distinctive root growth in horizontal direction aswell. Possible sources of data for this purpose are either undisturbed natural samples orinstalling large containers containing plants and taking samples by extracting drilling coresat appropriate size after the designated growing period.

AcknowledgementsWe thank Prof. Sarah Springman for her challenging discussions on the meaning of the data andtwo anonymous reviewers for their helpful suggestions and comments as well as Silvia Dingwallfor correcting the English. The PhD project of co-author Martin Frei, who provided the data, wasfunded by the Swiss Federal Office for the Environment FOEN.

5 References

ADAMS, J.I., 1965: The engineering behavior of a Canadian Muskeg, In: Proceeding of the VIInternational Conference on Soil Mechanics and Foundation Engineering, Montreal, Vol. 1.3–7.

ASTM, D4318-00, 2000: Standard Test Method for Liquid Limit, Plastic Limit, and Plasticity Indexof Soils. Ann. b. ASTM standards, sect. four: constr. Volume 04.08: 580-593.ASTM, C136-01,2001: Standard Test Method for Sieve Analysis of Fine and Coarse Aggregates. Ann. b. ASTMstandards, sect. four: constr. Volume 04.02: 85–89.

ASTM, D2487-00, 2000: Standard Practice for Classification of Soils for Engineering Purpose(Unified Soil Classification System). Ann. b. ASTM standards, sect. four: constr. Volume 04.08:238–248.

ASTM, D422-63, 2000: Standard Test Methods for Particle-Size Analysis of Soils. Ann. b. ASTMstandards, sect. four: constr. Volume 04.08: 10–17.


ASTM, D4767-95, 2000: Standard test methods for consolidated undrained triaxial compressiontests for cohesive soils, Ann. b. ASTM standards, sect. four: constr., Volume 04.08: 882–891.

ASTM, D698-00a, 2000: Standard Test Method for Laboratory Compaction Characteristics of SoilUsing Standard Effort (12400 ft-lbf/ft3 [600 kN-m/m3]). Ann. b. ASTM standards, sect. four:constr. Volume 04.08: 81–91.

ATKINSON, J.H.; CHARLES, J.A.; MHACH, H.K., 1990: Examination of erosion resistance of clays inembankment dams. Q. J. Eng. Geol. Hydrogeol. 23: 103–108.

ATKINSON, J.H.; LAU, W.H.W.; POWELL, J.J.M., 1991: Measurement of soil strength in simple sheartests. Can. Geotech. J. 28: 255–262.

BARNES, G. 1995: Soil mechanics. Principles and practice. New York, Macmillan.BEZZOLA, G.R.; HEGG, C. (eds) 2007: Ereignisanalyse Hochwasser 2005. Teil 1 – Prozesse,

Schäden und erste Einordnung. Bern, Bundesamt für Umwelt BAFU, Birmensdorf,Eidgenössische Forschungsanstalt WSL. 215 pp.

BISHOP, A.W.; HENKEL, D.J., 1957: The Measurement of Soil Properties in the Triaxial Test.Edward Arnold LTD. 190 pp.

BÖLL, A.; BURRI, K.; GERBER, W.; GRAF, F., 2009: Long-term studies on joint technical and bio-logical measures. For. Snow Landsc. Res. 82, 1: 9–32.

BÖLL, A.; GRAF, F., 2001: Nachweis von Vegetationswirkungen bei oberflächennahen Boden -bewegungen – Grundlagen eines neuen Ansatzes. Schweiz. Z. Forstwes. 152: 1–11.

CASAGRANDE, A., 1936: Determination of the preconsolidation load and its practical significance.In: Proceedings of the International Conference on Soil Mechanics and FoundationEngineering, vol. III. Cambridge, MA, Harvard University. 60–64.

CAZZUFFI, D.; CRIPPA, E., 2005: Shear strength behaviour of cohesive soils reinforced with vegetation. Proc. 16th International Conference on Soil Mechanics and GeotechnicalEngineering, Osaka. 2493–2498.

CHEN, C.W., LOEHR, J.E., 2008: Undrained and drained triaxial tests of fibre-reinforced sand. In:LI, G.; CHEN, Y.; TANG, X. (eds) Geosynthetics in Civil and Environmental Engineering. Proc.4th Aisan Regional Conference on Geosynthetics, Shanghai, China. 114–120.

COLA, S.; CORTELLAZZO, G., 2005: The shear strength behavior of two peaty soils. Geotech. Geol.Eng. 23: 679–695.

CONSOLI, N.C.; MONTARDO, J.P.; PRIETTO, P.D.M.; PASA, G.S., 2002: Engineering behavior of asand reinforced with plastic waste. J. Geotech. Geoenviron. Eng. 128: 462-472.

DAY, R.W., 2001: Soil testing manual. Ohio, McGraw-Hill. 618 pp.DE BUHAN, P.; MANGIAVACCHI, R.; NOVA, R.; PELLEGRINI, G.; SALENÇON, J., 1989: Yield design of

reinforced earth walls by homogenization method. Géotechnique 39: 189–201.EDIL, T.B.; DHOWIAN, A.W., 1981: At-Rest Lateral Pressure of Peat Soils. Journal of the

Geotechnical Engineering Division, ASCE 107: 201–217.FARRELL, E.R.; HEBIB, S., 1998: The determination of the geotechnical parameters of organic

soils, In: Yanagisawa, E.; Moroto, N.; Mitachi, T. (eds) Proceedings of the InternationalSymposium on Problematic Soils, IS-TOHOKU, Sendai, Vol. 1, A.A. Balkema, Rotterdam, TheNetherlands. 33–36.

FREI, M.; BÖLL, A.; GRAF, F.; HEINIMANN, H.R.; SPRINGMAN, S., 2003: Quantification of the influence of vegetation on soil stability. In: LEE, C.; THAM, L. (eds) Slope Engineering: Proc.Int. Conf., Dept. of Civil Engineering, University of Hong Kong. 872–877.

FROST, J.D.; HAN, J., 1999: Behavior of interfaces between fiber-reinforced polymers and sands. J.Geotech. Geoenviron. Eng. 125: 633–640.

GÄRTNER, H., 2003: The applicability of roots in Dendrogeomorphology. In: SCHLESER, G.;WINIGER, M.; BRÄUNING, A.; GÄRTNER, H.; HELLE, G.; JANSMA, E.; NEUWIRTH, B.; TREYDTE,K. (eds) TRACE – Tree Rings in Archaeology, Climatology and Ecology, Vol. 1. 120–124.

GHIASSIAN, H.; POOREBRAHIM, G.; GRAY, D.H., 2004: Soil reinforcement with recycled carpetwaste. Waste Manage. Res. 2: 108–114.

GRAF, F.; FREI, M.; SCHWARZ, M.; BÖLL, A., 2006: Use and importance of mycorrhiza in site-specific restoration. Proceedings of “Soil Bioengineering: ecological restoration with nativeplants and seed material”, Gumpenstein, 5–7 September 2006.

GYSSELS, G.; POESSEN, J.; BOCHET, E.; LI, Y., 2005: Impact of plant roots on the resistance of soilsto erosoin by water: a review. Prog. Phys. Geogr. 29: 189–217.


HASTIE, T.J.; TIBSHIRANI, R.J., 1990: Generalized additive models. Chapman and Hall, London.335 pp.

HEAD, K.H., 1994: Permeability, Shear Strength and Compressibility Tests. London, PentechPress. 441 pp.

HEAD, K.H., 1998: Effective Stress Tests, Volume 3, Manual of Soil Laboratory Testing, 2nd

Edition, John Wiley & Sons.HOLM, S., 1979: A simple sequentially rejective multiple test procedure. Scand. J. Stat. 6: 65–70.HOLTZ, W.G.; KOVACS, W.G., 1981: An introduction to geotechnical engineering. NJ, Prentice-Hall,

Engelwood Cliffs.IVERSEN, N.R.; HOOYER, T.S.; BAKER, R.W., 1998: Ring-shear studies of till deformation:

Coulomb-plastic behaviour and distributed strain in glacier beds. J. Glaciol. 44: 634–642.JANBU, N., 1954: Application of the composite slip surfaces for stability analysis. Proc. European

Conference on Stability of Earth Slopes, Stockholm Sweden, Nr. 3: 43–49.KIRSTEN, L.; FOURIE, L., 1999: The strength of root reinforced soil. Proc. Third Conference on

Environmental Engineering, Kwa Maritane, May 1999. 10 pp.LAMBE, T.; WHITMAN, V., 1979: Soil mechanics, SI Version. Chichester, John Wiley & Sons. 553 pp.LANG, H.-J.; HUDER, J.; AMANN P., 1996: Bodenmechanik und Grundbau (6. Aufl.). Berlin,

Springer. 320 pp.LANDVA, A.O.; LA ROCHELLE, P., 1983: Compressibility and Shear Characteristics of Radforth

Peats. In: JARETT, P.M. (ed), Testing of Peats and Organic Soils, ASTM STP 820, AmericanSociety for Testing and Materials. 157–191.

MARONE, C., 1998: Laboratory-derived friction constitutive laws and their application to seismicfaulting, Annu. Rev. Earth Planet. Sci. 26, 643–696, doi:10.1146/annurev.earth.26.1.643.

MARONE, C.; RALEIGH, C.B.; SCHOLZ, C.H., 1990: Frictional behaviour and constitutive modelingof simulated fault gouge, J. Geophys. Res. 95(B5), 7007–7025, doi:10.1029/JB095iB05p07007.

MICHALOWSKI, R.L., 1997: Limit stress for granular composites reinforced with continuous filaments. J. Eng. Mech. 123: 852–859.

MICHALOWSKI, R.L., ZHAO, A., 1996: Failure of fibre-reinforced granular soils. J. Geotech.Geoenviron. Eng. 122:226-234.

MICHALOWSKI, R.L.; CERMÁK, J., 2002: Strength anisotropy of fiber reinforced sand. Comput.Geotech. 29: 279–299.

MICHALOWSKI, R.L.; CERMÁK, J., 2003: Triaxial compression of sand reinforced fibres. J. Geotech.Geoenviron. Eng. 129: 125–136

MICHALOWSKI, R.L.; ZHAO, A., 1993: Failure criteria for homogenized reinforced soils and appli-cation in limit analysis of slopes. Proc. Geosynthetics 93, Industrial Fabrics Assoc. Int. Vol 1,Minneapolis, Minn. 443–453.

MOON, V.; BRADSHAW, J.; SMITH, R.; DE LANGE, W., 2005: Geotechnical characterisation of stra-tocone crater wall sequences, White Island Volcano, New Zealand. Eng. Geol. 81, 2: 146–178.

MOREL, J.C.; GOURC, J.P., 1997: Mechanical behaviour of sand reinforced with mesh elements.Geosynthetics International, 4: 481–508.

NEGUSSEY, D.; WIJEWICKREME, W.K.D.; VAID, Y.P., 1988: Constant-volume friction angle of granularmaterials. Can. Geotech. J. 25: 50–55.

OPERSTEIN, V.; FRYDMAN, S., 2000: The influence of vegetation on soil strength. GroundImprovement 4: 81–89.

PARK, T.; TAN, S.A., 2005: Enhanced performance of reinforced soil walls by the inclusion of shortfiber. Geotext. Geomembr. 23: 348–361.

POLLEN, N., 2007: Temporal and spatial variability in root reinforcement of streambanks:Accouting for soil shear strength and moisture. Catena 69: 197–205.

POLLEN, N.; SIMON A., 2005: Estimating the mechanical effects of riparian vegetation on streambank stability using a fibre bundle model. Water Resour. Res. 41, W07025, doi: 10.1029/2004WR003801.

R 2.5.0, 2007: R Development Core Team. R: A language and environment for statistical comput-ing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URLhttp://www.R-project.org .

RATHBUN, A.P.; MARONE, C.; ALLEY, R.B.; ANANDAKRISHNAN, S., 2008: Laboratory study of thefrictional rheology of sheared till. J. Geophys. Res. 113, F02020, doi:10.1029/2007JF000815.


RICKLI, C.; BUCHER, H-U.; BÖLL, A.; RAETZO, H., 2004: Untersuchungen zu oberflächennahenRutschungen des Jahres 2002 im Napfgebiet und in der Region Appenzell. Bull. angew. Geol.9: 37–49.

SAWICKI, A., 1983: Plastic limit behavior of reinforced earth. J. Geotech. Eng. 109: 1000–1005.SCHOLZ, C.H., 2002: The Mechanics of Earthquakes and Faulting, 2nd ed. Cambridge Univ. Press,

Cambridge, U.K.SHEWBRIDGE, S.E.; SITAR, N., 1989: Deformation Characteristics of Reinforced Sand in Direct

Shear. J. Geotech. Eng. 115: 1134–1147.SPRINGMAN, S.M.; JOMMI, C.; TEYSSEIRE, P., 2003: Instabilities on moraine slopes induced by loss

of suction: a case history. Géotechnique 53: 3–10.STAUFFER, S.D.; HOLTZ, R.D., 1995: Stress-strain and strength behavior of staple fiber and contin-

uous filament-reinforced sand. Transp. Res. Rec. 1474. Transp. Res. Board, Washington D.C.82–95.

STYCZEN, M.E.; MORGAN, R.P.C., 1995: Engineering properties of vegetation. In: Morgan, R.P.C.;Rickson, R.M. (eds) Slope stabilization and erosion control: A bioengineering approach.London – New York, Spoon. 5–58.

TAGNIT-HAMOU, A.; VANHOVE, Y.; PETROV, N., 2005: Microstructural analysis of the bond mech -anism between polyolefin fibers and cement pastes. Cem. Concr. Res. 35: 364–370.

TANG, C.; SHI, B.; GAO, W.; CHEN, F.; CAI, W., 2007: Strength and mechanical behavoir of shortpolypropylene fiber reinforced and cement stabilized clayey soil. Geotext. Geomembr. 25:194–202.

TIKA, T.E.; VAUGHAN, P.R.; LEMOS, L., 1996: Fast shearing of preexisting shear zones in soil.Geotechnique 46: 197– 233.

VENABLES, W.N.; RIPLEY, B.D., 2002: Modern Applied Statistics with S. Fourth Edition. New York,Springer. 495 p.

WALDRON, L.J., 1977: The shear resistance of root-permeated homogeneous and stratified soil.Soil Sci. Soc. Am. J. 41: 843–849.

WILCOX, R.R., 2005: Introduction to Robust Estimation & Hypothesis Testing. New York,Academic Press. 609 pp.

WinRhizo, 2000: Régent Instruments Inc., 4040 rue Blain, Quebec, Qc G2B 5C3, Canada,www.regent.qc.ca

WU, T.H., 1984: Effect of Vegetation on Slope Stability, Transportation Research Report 965:37–46.

WU, T.H., 2007: Root reinforcement: analyses and experiments. In: STOKES, A.; SPANOS, I.;NORRIS, J.E.; CAMMERAAT, E. (eds) Eco- and Ground Bio-Engineering: The Use of Vegetationto Improve Slope Stability. Berlin, Springer Verlag. 21–30.

WU, T.H.; MCKINNELL, W.P. III; SWANSTON, D.N., 1979: Strength of tree roots and landslides onPrince of Wales Island, Alaska. Can. Geotech. J. 16: 19–33.

WU, T.H.; MCOMBER, R.M; ERB, R.T.; BEAL, P.E., 1988: Study of Soil-Root Interaction. Journal ofGeotechnical Engineering, 114: 1351–1375.

WU, T.H.; WATSON, A., 1998: In-situ shear tests of soil blocks with roots. Can. Geotech. J. 35:579–590.

YAMAGUCHI, H.; OHIRA, Y.; KOGURE K.; MORI, S., 1985: Undrained shear characteristics of nor-mally consolidated peat under triaxial compression and extension conditions. Soils Found. 25:1–18.

YETIMOGLU, T.; SALBAS, O., 2003: A study on shear strength of sands reinforced with randomlydistributed discrete fibers. Geotext. Geomembr. 21: 103–110.

ZHANG, M.X.; JAVADI, A.A.; MIN, X., 2006: Triaxial tests of sand reinforced with 3D inclusions.Geotext. Geomembr. 24: 201–209.

ZHU, J.H.; ANDERSON, S.A., 1998: Determination of shear strength of Hawaiian residual soil subjected to rainfall-induced landslides. Géotechnique 48: 73–82.

Revised version accepted May 11, 2009

Effects of vegetation on the angle of internal friction of ... · Effects of vegetation on the...

Documents

Transcript of Effects of vegetation on the angle of internal friction of ... · Effects of vegetation on the...