Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/32116849

EffectivenessofPalmandSimulatedGeotextilesinReducingRun-offandInter-rillErosiononMediumandSteepSlopes

ARTICLEinSOILUSEANDMANAGEMENT·AUGUST2007

ImpactFactor:1.47·DOI:10.1111/j.1475-2743.2007.00098.x·Source:OAI

CITATIONS

36

READS

36

4AUTHORS:

ToonSmets

FlemishInstituteforTechnologicalResearch

19PUBLICATIONS236CITATIONS

SEEPROFILE

J.Poesen

UniversityofLeuven

641PUBLICATIONS19,376CITATIONS

SEEPROFILE

MichaelFullen

UniversityofWolverhampton

184PUBLICATIONS1,575CITATIONS

SEEPROFILE

ColinA.Booth

UniversityoftheWestofEngland,Bristol

125PUBLICATIONS658CITATIONS

SEEPROFILE

Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,

lettingyouaccessandreadthemimmediately.

Availablefrom:MichaelFullen

Retrievedon:05February2016

Effectiveness of palm and simulated geotextiles in reducingrun-off and inter-rill erosion on medium and steep slopes

T. Smets1

, J . Poesen1

, M. A. Fullen2

& C. A. Booth3

1Physical and Regional Geography Research Group, K. U. Leuven, Geo-Institute, Celestijnenlaan 200E, B-3001 Heverlee, Belgium,2School of Applied Sciences, The University of Wolverhampton, Wulfruna Street, Wolverhampton WV1 1SB, UK, and 3School of

Engineering and the Built Environment, The University of Wolverhampton, Wulfruna Street, Wolverhampton WV1 1SB, UK

Abstract

Palm-leaf geotextiles could be an effective and cheap soil conservation method with enormous global

potential. However, there are very few data on the effectiveness of palm geotextiles in reducing soil

erosion by water. This study investigates the effectiveness of two types of palm geotextiles and the

effect of geotextile mesh size on infiltration, run-off and inter-rill erosion rate and soil surface rough-

ness on a medium and steep slope. A well-defined protocol was developed to conduct laboratory

experiments. Rainfall was simulated for 90 min with an intensity (I) of 45 and 67 mm h)1 on an inter-

rill erosion plot, filled with an erodible sandy loam and having slope gradients (S) of 15 and 45%.

Two palm-leaf geotextiles (Borassus aethiopum and Brazilian Buriti Palm) and three simulated geo-

textiles (polyethylene tarpaulin) with different mesh sizes (1 · 1, 5 · 5 and 12 · 12 cm) were tested on

a simulated fine tilth. Calculated k values from the Horton infiltration equation ranged from 0.025 to

0.145 and decreased linearly on both slopes with geotextile cover. Geotextiles are more effective in

reducing the run-off coefficient on a medium slope (15%) compared with that on a steep slope (45%),

ranging from 76.4 to 17.9%. Mean b values from the mulch cover equation equalled 0.024 for a 15%

slope and 0.045 for a 45% slope, indicating a higher effectiveness of geotextiles in reducing total inter-

rill soil loss on gentler slopes compared with commonly used mulches. Erosion-induced soil surface

roughness at the end of each experiment increased linearly with geotextile cover percentage and this

increase was not significantly different between the two slope gradients.

Keywords: Geotextiles, infiltration rates, inter-rill erosion, laboratory experiments, simulated rainfall

Introduction

On steep slopes intense rainfall can cause significant soil ero-

sion. The establishment of a vegetative cover, which can pro-

vide long-term protection of slopes, is counteracted by the

erosive forces of rain and run-off. Seeds and seedlings can be

damaged or even washed away, so vegetation growth on

steep erodible slopes is limited. During this period of high

erosion risk, geotextiles or erosion control mats are a poss-

ible temporary replacement for vegetative cover and can

offer immediate soil protection on steep slopes (Hann &

Morgan, 2006). Geotextiles are defined as ‘permeable textiles

used in conjunction with soil, foundation, rock, earth or any

geotechnical engineering-related material, as an integral part

of a man-made project’ (John, 1987). They are currently used

in various applications, including separating two distinct

materials, filtration, drainage, reinforcement of weak soils

and other materials and surface soil erosion control which is

considered in this study. Erosion control geotextiles are made

from natural or synthetic materials, including jute, coir, sisal,

cereal straw, nylon, palm leafs, polypropylene, polyester and

polyethylene (Rickson, 2006). In this study, we focus on

palm geotextiles for erosion control. Preliminary investiga-

tions suggest palm-leaf geotextiles could be an effective and

cheap soil conservation method with enormous global poten-

tial (Booth et al., 2005; Davies et al., 2006). However, there

are few data on the effectiveness of palm geotextiles to

reduce soil erosion by water. In many of the existing field

and laboratory studies as reviewed by Sutherland (1998a,b),

results are difficult to interpret because very few treatment

replications were conducted and often influencing variablesCorrespondence: T. Smets. E-mail: toon.smets@geo.kuleuven.be

Received January 2007; accepted after revision May 2007

Soil Use and Management, September 2007, 23, 306–316 doi: 10.1111/j.1473-2743.2007.00098.x

306 ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science

were not controlled between experiments. Even so, there is

little understanding of how geotextiles interact with the var-

ious hydrological (infiltration and run-off production) and

erosion processes (splash, inter-rill and rill erosion) on a

range of slopes and under different rainfall intensities. This

is mainly attributed to the lack of standardized protocols to

assess the erosion control effectiveness of geotextiles.

The objectives of this study were therefore to: (i) develop a

methodology to assess the effectiveness of different types of

geotextiles in controlling water erosion; (ii) assess the effect-

iveness of two types of palm geotextiles in increasing infiltra-

tion rates and reducing inter-rill run-off and erosion rates on

a medium and steep slope gradient and under two rainfall

intensities; and (iii) investigate the effect of palm geotextile

mesh size on increasing infiltration rates and reducing inter-

rill run-off and erosion rates. The hypothesis tested in this

study is that geotextiles applied to the soil surface act as

mulch elements in increasing infiltration rates and reducing

inter-rill run-off and erosion rates.

Materials and methods

Experimental design

Inter-rill erosion plot. All experiments were conducted using a

rectangular inter-rill erosion plot, 1.25 · 1.78 m2 (Poesen

et al., 1990) (Figure 1). The erosion plot had a central test

area of 0.60 · 0.94 m2, surrounded by a buffer area, identic-

ally treated as the test area to compensate for losses of water

and sediment due to splash from the central test area. The

development of rills was limited because of the limited size

of the test area. On the bottom of the erosion plot, wet

cloths were placed on top of perforated plates. Run-off and

percolated water from the test area were collected at the

downslope end or below the erosion plot. Given that soil

erosion can be severe on medium and steep slopes, experi-

ments were conducted on 15 and 45% slopes.

Soil type. The soil used in the experiments was Tertiary sandy

loam, erodible subsoil, quarried in central Belgium (Bierbeek)

and often found in road cuttings and on construction

sites. The soil has 13% clay (<0.002 mm), 24% silt

(0.002–0.063 mm) and 63% sand (0.063–2 mm) and 0.18%

organic matter. The inter-rill erosion plot was filled in the test

area to a depth of 0.10 m and in the buffer area to 0.05 m, so a

soil volume of 0.14 m3 was needed for each experiment.

Simulated rainfall. Rainfall was simulated for 90 min by a

single-nozzle, continuous-spray system (Poesen et al., 1990;

Poesen & Lavee, 1991). Water pressure in the system could

be adjusted by a pressure regulator. Experiments were con-

ducted with a water pressure of 0.59 and 0.71 bar at the noz-

zle outlet, yielding a mean rainfall intensity over the test area

of 45 and 67 mm h)1, respectively. Because fall-height equal-

led 3.25 m, and using the drop-size distribution and the cal-

culated drop-fall velocity (Laws, 1941), the simulated rainfall

produced a kinetic energy at the soil surface of 15.2 and

15.8 J m)2 mm)1, respectively. This equals to approximately

60% of the energy of natural rainfall with similar intensities

(Salles et al., 2002).

Geotextiles. Experiments were conducted with simulated geo-

textiles and two types of palm geotextiles: the Borassus aethi-

opum geotextile (Davies et al., 2006), manufactured in

Gambia and Brazilian Buriti Palm geotextiles (Guerra et al.,

2005), manufactured in Brazil (Figure 2). To investigate the

effects of geotextile mesh size on increasing infiltration rates

and reducing inter-rill run-off and erosion rates, simulated

geotextiles with three different mesh sizes were constructed

by cutting, weaving and stapling 1.5-cm-wide strips similar

to the Borassus and Buriti geotextiles from reinforced poly-

ethylene sheets using mesh sizes of 0.01 · 0.01, 0.05 · 0.05

RunoffCollecting

system

Buffer area

Test area

0.94 m

0.60 m

Figure 1 Inter-rill erosion plot used in this study.

(a) (b)

Figure 2 Palm-leaf geotextiles (a) Buriti and (b) Borassus geotextiles;

total length of scale bar equals 60 cm.

Effectiveness of erosion control palm geotextiles 307

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

and 0.12 · 0.12 m (Figure 3). The cover percentage, mesh

size and thickness of the Borassus, Buriti and simulated geo-

textiles are given in Table 1.

Preparation of experiments

To assess the effectiveness of geotextiles in increasing infiltra-

tion and reducing inter-rill run-off and erosion, a standard

procedure was applied. Prior to each experiment, the soil

was air dried (room temperature, initial gravimetric moisture

content: 9.0 ± 1.5%) for 4 days and sieved using a sieve

with circular holes of 2.7-cm diameter. The erosion plot was

filled with the sieved soil and the soil surface flattened using

a board, simulating a fine tilth similar to a fine seedbed, with

a bulk density of 1.20 ± 0.04 g cm)3. Next, the soil surface

of the test and buffer area was covered with a geotextile.

Measurement procedure

Before the start of each experiment, soil samples in the test

area were taken to determine the initial gravimetric soil mois-

ture (%) and bulk density (kg m)3). Rainfall intensity

(mm h)1) over the test area was measured before and after

each experiment for 3 min using a rain gauge. Each experi-

ment lasted for 90 min which was sufficient to reach steady-

state conditions. During the experiments, percolation and

run-off samples including sediment were taken for 1 min with

an interval of 2 min. The run-off samples were oven-dried at

105 �C for 24 h. After each experiment, the geotextiles were

removed and the erosion-induced soil surface roughness was

measured using the chain method (Saleh, 1993). A 1-m-long

chain with link lengths of 0.003 m was used. Chain roughness

(CR, %) was calculated using the equation:

CR ¼ 100 1� L2

L1

� �ð1Þ

where L1 is the full length of the chain (m) and L2 is the

horizontal distance between chain ends when placed on the

soil surface (m) (Saleh, 1993).

For each combination of slope gradient and rainfall

intensity, a control experiment (bare soil surface) was

conducted and for each combination of cover, slope

gradient and rainfall intensity, two to three replications

were conducted. In total, 60 rainfall experiments were

conducted (Table 2).

Data processing

Inter-rill soil loss (ISL, kg m)2) was determined by weighing

the oven-dried run-off samples. Run-off volume (mm) was cal-

culated by subtracting the mass of the ISL samples from the

mass of the run-off samples (water density: 1.0 g cm)3). Total

inter-rill soil loss (total ISL, kg m)2) was calculated by integra-

ting the ISL samples over the duration of each experiment.

Next, inter-rill sediment concentration (ISC, kg m)3) was cal-

culated. Because rainfall intensity was kept constant during

the experiments, infiltration rate (mm h)1) was calculated by

subtracting the run-off rate, determined by the run-off volume

and the sampling time, from the rainfall intensity. ANOVA

tests were conducted on the results of the experiments (infiltra-

tion rate, run-off rate and total ISL) using the SAS statistical

program (version 9.1, SAS Institute Inc., 2002).

Table 2 Outline of the inter-rill laboratory experiments

Rainfall intensity (I) 45 and 67 mm h)1

Duration 90 min

Slope gradient (S) 15 and 45%

Soil type (texture) Sandy loam

Cover type Bare, Borassus, Buriti and three

simulated geotextiles

Replications Two to three

Total number of

experiments

60

0.6 m

1 by 1 cm

(C = 84%)

5 by 5 cm

(C = 43%)

12 by 12 cm

(C = 22%)

Figure 3 Simulated geotextiles with different

mesh sizes (C = surface cover, %).

Table 1 Characteristics of Borassus, Buriti and simulated geotextiles

Geotextile Cover (%) Mesh size (m · m) Thickness (m)

Borassus 43 0.05 · 0.05 0.016

Buriti 42 0.04 · 0.04 0.013

Simulated

0.01 · 0.01 m2 84 0.01 · 0.01 0.001

0.05 · 0.05 m2 43 0.05 · 0.05 0.001

0.12 · 0.12 m2 22 0.12 · 0.12 0.001

308 T. Smets et al.

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

Results

Infiltration and run-off

The infiltration rates for the bare inter-rill erosion plots and

the plots covered with Borassus and Buriti geotextiles on the

two tested slope gradients are shown in Figure 4, for the 45-

mm h)1 rainfall intensity, and Figure 5 for the 67-mm h)1

rainfall intensity. For both slope gradients and rainfall inten-

sities, the infiltration rate was higher and decreased more

slowly over time on inter-rill plots covered with palm geo-

textiles compared with bare soil treatments. Once ponding

occurs, the change in infiltration rate over time can be

predicted by the following equation (Horton, 1933):

ft ¼ fc þ ðf0 � fcÞe�kt ð2Þ

where ft is the infiltration rate at some time t (mm h)1), fc is

the final infiltration rate (mm h)1); f0 is the initial infiltration

rate (mm h)1), k is a constant representing the rate of

decrease in ft and t is the time (h).

By calculating k values for all treatments, the effect of cover

percentage, slope gradient and rainfall intensity on the rate of

decrease in infiltration rate can be quantified. The k values

ranged from 0.025 (Borassus, I = 45 mm h)1, S = 15%) to

0.145 (bare soil, I = 67 mm h)1, S = 15%) (Figures 6 and

7). For a 67-mm h)1 rainfall intensity, k values for all treat-

ments are higher compared with that of a 45-mm h)1 rainfall

intensity, indicating a faster decrease in infiltration rate over

time during a 67-mm h)1 rainfall intensity. With an increasing

geotextile cover percentage, k values decrease linearly. How-

I = 67 mm h–1

t (min)0 20 40 60 80

10

20

30

40

50

60

70

Bare (S=15%) Bare (S=45%) Borassus (S=15%) Borassus (S=45%) Buriti (S=15%) Buriti (S=45%) mean rainfall intensity

f t (m

m h

–1)

Figure 5 Infiltration rate (ft) versus time (t) for bare inter-rill plots

and plots covered with Borassus and Buriti geotextiles (with a

different cover %) on two slope gradients (S) and during a rainfall

intensity (I) of 67 mm h)1. Plotted data represent the mean of two

to three replications.

I = 67 mm h–1

C (%)

0 20 40 60 80 100

k

0,02

0,04

0,06

0,08

0,10

0,12

0,14

0,16

0,18

sim.geotext. (S=15%)sim.geotext. (S=45%)Borassus (S=15%)Borassus (S=45%)Buriti (S=15%)Buriti (S=45%)S=15%; R2 = 0.80;n=9; P=0.108S=45%; R2 = 0.92;n=9; P=0.042

Figure 7 Calculated k values in the Horton infiltration equation

versus cover percentage (C) for bare inter-rill plots and plots covered

with Borassus, Buriti and simulated geotextiles on two slope

gradients (S) and during a rainfall intensity (I) of 67 mm h)1 (error

bars represent 1 SD).

I = 45 mm h–1

t (min)0 20 40 60 80

f t (m

m h

–1)

0

10

20

30

40

50

Bare (S=15%) Bare (S=45%)Borassus (S=15%)Borassus (S=45%)Buriti (S=15%)Buriti (S=45%)mean rainfall intensity

Figure 4 Infiltration rate (ft) versus time (t) for bare inter-rill plots

and plots covered with Borassus and Buriti geotextiles (with a differ-

ent cover %) on two slope gradients (S) and during a rainfall inten-

sity (I) of 45 mm h)1. Plotted data represent the mean of two to

three replications.

I = 45 mm h–1

C (%)0 20 40 60 80 100

k

0,02

0,03

0,04

0,05

0,06

0,07

0,08

0,09

sim.geotext. (S=15%)sim.geotext. (S=45%)Borassus (S=15%)Borassus (S=45%)

Buriti (S=15%)Buriti (S=45%)S=15%; R2 = 0.73; n=9; P=0.144S=45%; R2 = 0.63; n=9; P=0.203

Figure 6 Calculated k values in the Horton infiltration equation

versus cover percentage (C) for bare inter-rill plots and plots covered

with Borassus, Buriti and simulated geotextiles on two slope

gradients (S) and during a rainfall intensity (I) of 45 mm h)1 (error

bars represent 1 SD).

Effectiveness of erosion control palm geotextiles 309

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

ever, this linear decrease is steeper for the 15% slope than for

the 45% slope. This indicates that simulated geotextiles are

more effective in increasing infiltration rate on a medium

slope (15%) than on a steep slope (45%).

Total run-off depth after 90 min of simulated rain for all

treatments is given in Table 3. Run-off coefficients (RC, %),

calculated as the ratio of total run-off volume to the total

rain volume, from bare inter-rill plots and plots covered with

simulated, Borassus and Buriti geotextiles on the two tested

slope gradients are shown in Figure 8 for the 45-mm h)1

rainfall intensity and in Figure 9 for the 67-mm h)1 rainfall

intensity. Total run-off volume decreased with increasing sur-

face cover and slope gradient. For a rainfall intensity of

45 mm h)1 Borassus geotextiles reduced run-off volume com-

pared with the bare soil by 39 and 68% and Buriti geotex-

tiles to 27 and 51% on a slope of 45 and 15%, respectively.

RC ranged from 76.4% (bare soil, I = 45 mm h)1,

S = 15%) to 17.9% (simulated geotextiles: 0.12 · 0.12 m2,

22% cover, I = 45 mm h)1, S = 15%) and decreased with

an increasing surface cover and slope gradient. Hence, relat-

ively less run-off occurs on a steep slope (45%) than on a

medium slope (15%). For the simulated and the Borassus

geotextiles, the ratio between RC for the 45% slope and the

15% slope is higher for the 45-mm h)1 rainfall intensity than

for the 67-mm h)1 rainfall intensity, indicating a larger effect

of slope gradient on RC during a 45-mm h)1 rain than dur-

ing a 67-mm h)1 event. Significance and probability (P) val-

ues of the effect of slope gradient on run-off volume for

both rainfall intensities and all treatments are calculated

from an ANOVA (Table 4). Borassus and Buriti geotextiles

are more effective in reducing total run-off depth compared

with the simulated geotextile with a comparable cover

Table 3 Total run-off depth after 90 min for bare inter-rill plots and plots covered with simulated, Buriti and Borassus geotextiles (with a differ-

ent cover %) on two slope gradients (S) and under two rainfall intensities (I) (±1 SD)

Treatments (cover, %)

Total run-off depth (mm)

45-mm h)1 rainfall intensity 67-mm h)1 rainfall intensity

15% slope gradient 45% slope gradient 15% slope gradient 45% slope gradient

Control (0) 51.6 ± 2.7 36.1 ± 9.3 65.6 ± 1.6 51.6 ± 2.6

12 · 12 cm2 (22) 44.1 ± 0.2 38.4 ± 4.9 54.3 ± 4.1 50.7 ± 2.8

5 · 5 cm2 (43) 44.6 ± 8.2 28.0 ± 1.9 55.2 ± 5.3 43.2 ± 6.5

1 · 1 cm2 (84) 12.1 ± 4.3 21.4 ± 6.2 20.9 ± 2.9 21.2 ± 6.3

Buriti (42) 25.2 ± 3.1 26.4 ± 1.3 31.1 ± 3.5 35.6 ± 4.0

Borassus (43) 16.4 ± 5.5 22.2 ± 8.9 47.9 ± 6.2 35.5 ± 3.5

C (%)0 20 40 60 80 100

RC

(%

)

0

20

40

60

80sim.geotext. (S=15%)sim.geotext. (S=45%)Borassus (S=15%) Borassus (S=45%) Buriti (S=15%) Buriti (S=45%)

I = 67 mm h–1

Figure 9 Run-off coefficient (RC) versus cover percentage (C) for

bare inter-rill plots and plots covered with Borassus, Buriti and

simulated geotextiles on two slope gradients (S) and during a rainfall

intensity (I) of 67 mm h)1 (error bars represent 1 SD). Plotted data

represent the mean of two to three replications.I = 45 mm h–1

C (%)0 20 40 60 80 100

RC

(%

)

0

20

40

60

80

100

sim.geotext.(S=15%)sim.geotext.(S=45%)Borassus (S=15%) Borassus (S=45%) Buriti (S=15%) Buriti (S=45%)

Figure 8 Run-off coefficient (RC) versus cover percentage (C) for

bare inter-rill plots and plots covered with Borassus, Buriti and

simulated geotextiles on two slope gradients (S) and during a rainfall

intensity (I) of 45 mm h)1 (error bars represent 1 SD). Plotted data

represent the mean of two to three replications.

310 T. Smets et al.

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

percentage (0.05 · 0.05 m, 43% cover). This can most prob-

ably be attributed to the larger thickness of the Borassus and

Buriti geotextiles (Table 1). Because of this thickness, a phys-

ical barrier for overland flow is formed. More water can

therefore infiltrate with the Borassus and Buriti geotextiles

compared with that of the simulated geotextiles.

Inter-rill soil loss

Total inter-rill soil loss (total ISL, kg m)2) after 90 min of

simulated rain on the plots covered with simulated, Borassus

and Buriti geotextiles for the two tested slope gradients and

rainfall intensities are given in Table 5. The highest total ISL

(5.16 kg m)2) was measured on a bare soil inter-rill plot with

a 45% slope during a rainfall intensity of 67 mm h)1; the

smallest total ISL (0.02 kg m)2) was measured on a plot cov-

ered with simulated geotextiles with mesh sizes of

0.01 · 0.01 m2 (84% cover) on a 15% slope during a rainfall

intensity of 45 mm h)1. For all combinations of slope gradi-

ent and rainfall intensity, total ISL decreased with an increas-

ing surface cover percentage. Both Borassus and Buriti

geotextiles reduced total ISL more than the simulated geotex-

tiles with comparable surface cover.

Relative total ISL (ISLrel, expressed by total ISL of a

covered soil surface relative to that of a bare soil surface) is

plotted in Figure 10 against surface cover percentage for the

simulated, Borassus and Buriti geotextiles for the two tested

slope gradients for the 45-mm h)1 rainfall intensity and in

Figure 11 for the 67-mm h)1 rainfall intensity. For both rain-

fall intensities the Borassus, Buriti and simulated geotextiles

are more effective in reducing ISLrel on a 15% slope

compared with that on a 45% slope.

In RUSLE2.0 (Revised Universal Soil Loss Equation,

Version 2.0; Foster, 2005), the effect of surface cover on

inter-rill and rill erosion is determined by the ‘ground cover

subfactor’ of the C-factor (or cover management factor). The

RUSLE2.0 equation for the ground cover subfactor (gc) is

given by:

gc ¼ exp½�bfgnð0:24=RaÞ0:08� ð3Þ

where b is a coefficient (%)1) describing the relative

effectiveness of the ground (surface) cover in reducing

erosion, fgn is the net soil surface ground cover (%), Ra is

the adjusted roughness used to compute the soil surface

roughness subfactor (inches) and 0.24 is the assumed

adjusted soil surface roughness value (inches) for unit plot

conditions.

The calculated b values from the bare inter-rill erosion plots

and the plots covered with simulated geotextiles (increasing

cover percentage) on the two tested slope gradients and during

two rainfall intensities are given in Table 6 (Ra was assumed to

be 0.24 for all combinations of cover, slope gradient and rain-

fall intensity). The b values on a 45% slope are almost half the

b values on a 15% slope. Therefore, the simulated geotextiles

are much more effective in reducing ISLrel on a 15% slope

compared with that on a 45% slope for both rainfall intensi-

ties. Significance and probability (P) values of the effect of

slope gradient on total ISL for both rainfall intensities and all

treatments are calculated from an ANOVA (Table 7). For a

rainfall intensity of 67 mm h)1, total ISL is significantly differ-

ent on both slope gradients for all treatments. For the bare

inter-rill plots and the plots covered with the simulated

Table 5 Total inter-rill soil loss (total ISL) after 90 min for bare inter-rill plots and plots covered with simulated, Buriti and Borassus geotextiles

(with a different cover %) on two slope gradients (S) and under two rainfall intensities (I) (±1 SD)

Treatments (cover, %)

Total ISL (kg m)2)

45-mm h)1 rainfall intensity 67-mm h)1 rainfall intensity

15% slope gradient 45% slope gradient 15% slope gradient 45% slope gradient

Control (0) 1.46 ± 0.01 1.71 ± 0.32 1.99 ± 0.11 5.16 ± 0.59

12 · 12 cm2 (22) 0.63 ± 0.08 1.51 ± 0.37 1.37 ± 0.14 3.57 ± 0.30

5 · 5 cm2 (43) 0.32 ± 0.08 0.65 ± 0.13 0.41 ± 0.05 2.61 ± 0.41

1 · 1 cm2 (84) 0.02 ± 0.01 0.23 ± 0.13 0.04 ± 0.01 0.56 ± 0.25

Buriti (42) 0.13 ± 0.04 0.46 ± 0.25 0.29 ± 0.17 0.95 ± 0.47

Borassus (43) 0.07 ± 0.03 0.32 ± 0.21 0.19 ± 0.04 0.79 ± 0.21

Table 4 Probability values (P) and significance of the effect of slope

gradient (S) on run-off volume for both rainfall intensities (I) and all

treatments

Treatment (cover, %)

Rainfall intensity (I)

45 mm h)1 67 mm h)1

Control (0) 0.047* 0.025*

12 · 12 cm2 (22) 0.047* 0.444 ns

5 · 5 cm2 (43) 0.018* 0.047*

1 · 1 cm2 (84) 0.160 ns 0.903 ns

Buriti (42) 0.647 ns 0.090 ns

Borassus (43) 0.214 ns 0.043*

Significance of effect: not significant (ns); significant at *P < 0.05

and **P < 0.01.

Effectiveness of erosion control palm geotextiles 311

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

geotextiles, the ratio of total ISL for the 45% slope to the 15%

slope is higher for the 67-mm h)1 rainfall intensity than for the

45-mm h)1 rainfall intensity. The ratios of total ISL on the

45% slope to the 15% slope for both Borassus and Buriti

geotextiles are not significantly different for the 67- and

45-mm h)1 rainfall intensities.

Soil surface roughness

Soil surface roughness (SR) is an important parameter influ-

encing water erosion. SR determines the water volume that

can be stored on the surface as depression storage and affects

run-off velocity and erosivity. SR includes depressions where

sediment deposition occurs and soil heights of large and

stable soil aggregates that are resistant to soil detachment.

SR increases infiltration rate and slows run-off rate which

reduces total run-off and erosion (Foster, 2005). Soil surface

cover, slope gradient and rainfall intensity can have a consid-

erable effect on SR. In this study, SR is represented by chain

roughness (CR). The effect of slope gradient and geotextile

cover percentage on CR is shown in Figures 12 and 13 for a

rainfall intensity of 45 and 67 mm h)1 respectively. CR ran-

ged from 0.6% (bare soil, I = 67 mm h)1, S = 15%) to

22.7% (simulated geotextile: 0.01 · 0.01 m2, 84% cover,

I = 67 mm h)1, S = 45%). CR increased linearly with

increasing cover percentage of geotextiles and higher CR val-

ues were measured on the 45% slope compared with that on

the 15% slope. However, the linear increase in CR with

cover percentage is not significantly different for both slope

gradients. For all the treatments the ratios between CR for

the 45% slope and the 15% slope gradient are higher for the

67-mm h)1 rainfall intensity than for the 45-mm h)1 rainfall

intensity. During the highest rainfall intensity, slope gradient

C (%)0 20 40 60 80 100

ISL re

l (%

)

0

20

40

60

80

100

120

sim.geotext. (S=15%) sim.geotext. (S=45%) Borassus (S=15%) Borassus (S=45%) Buriti (S=15%) Buriti (S=45%)

I = 67 mm h–1

Figure 11 Relative total inter-rill soil loss (ISLrel) versus cover per-

centage (C) for bare inter-rill plots and plots covered with Borassus,

Buriti and simulated geotextiles on two slope gradients (S) and dur-

ing a rainfall intensity (I) of 67 mm h)1 (error bars represent 1 SD).

Plotted data represent the mean of two to three replications.

C (%)0 20 40 60 80 100

ISL re

l (%

)

0

20

40

60

80

100

120

sim.geotext. (S=15%) sim.geotext. (S=45%) Borassus (S=15%) Borassus (S=45%) Buriti (S=15%) Buriti (S=45%)

I = 45 mm h–1

Figure 10 Relative inter-rill soil loss (ISLrel) versus cover percentage

(C) for bare inter-rill plots and plots covered with Borassus, Buriti

and simulated geotextiles on two slope gradients (S) and during a

rainfall intensity (I) of 45 mm h)1 (error bars represent 1 SD). Plot-

ted data represent the mean of two to three replications.

Table 6 b values in the RUSLE2.0 ground cover equation (Foster,

2005) for inter-rill plots covered with simulated geotextiles (with a

different cover %) on two slope gradients (S) and under two rainfall

intensities (I); n is the number of data points used to calculate b

values

45-mm h)1 rainfall

intensity

67-mm h)1 rainfall

intensity

15% slope

gradient

45% slope

gradient

15% slope

gradient

45% slope

gradient

b value 0.048 0.023 0.042 0.024

r2 0.96 0.94 0.95 0.93

n 9 9 9 9

Table 7 Probability values (P) and significance of the effect of slope

gradient (S) on total inter-rill soil loss (total ISL) for both rainfall

intensities (I) and all treatments

Treatment

(cover, %)

45-mm h)1

rainfall intensity

67-mm h)1

rainfall intensity

Control (0) 0.309 ns 0.013*

12 · 12 cm2 (22) 0.061 ns 0.001**

5 · 5 cm2 (43) 0.095 ns 0.004**

1 · 1 cm2 (84) 0.082 ns 0.046*

Buriti (42) 0.159 ns 0.099 ns

Borassus (43) 0.183 ns 0.043*

Significance of effect: not significant (ns); significant at *P < 0.05

and **P < 0.01.

312 T. Smets et al.

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

has a larger effect on CR compared with that of the

45-mm h)1 rainfall intensity.

Discussion

Effects of surface cover, slope gradient and rainfall inten-

sity on infiltration, run-off, inter-rill soil loss and surface

roughness

Infiltration and run-off. All geotextile treatments increase the

infiltration rate compared with bare soil treatments on both

tested slope gradients (15 and 45%) and rainfall intensities

(45 and 67 mm h)1). However, calculated k values from the

Horton infiltration equation indicate that for all treatments,

infiltration rate decreases faster under a 67-mm h)1 rainfall

intensity than under a 45-mm h)1 rainfall intensity and simu-

lated geotextiles are more effective in increasing infiltration

rate on a medium slope (15%) than on a steep slope (45%).

Previous studies indicate surface cover, slope gradient and

rainfall intensity play major roles in determining infiltration

rate (e.g. Poesen, 1984; de Figueiredo & Poesen, 1998;

Janeau et al., 2003; Assouline & Ben-Hur, 2006). A greater

decrease in infiltration rate during higher rainfall intensity is

also observed by Assouline & Ben-Hur (2006) when simula-

ted rainfall intensities of 24 and 60 mm h)1 were applied to

a sandy soil on five slope gradients, ranging from 5 to 25%.

They also state that the final infiltration rate increases with

increasing slope gradient. On plots covered with various

straw cover percentages ranging from 0 to 100% and having

mean slope gradients of 6–8%; Li et al. (2001) applied a

122-mm h)1 simulated rainfall intensity and found a linear

increase in the final infiltration rate with increasing cover

percentage, a trend also observed in this study. However, on

a steep slope (45%), the results of this study indicate that a

linear increase in final infiltration rate with increasing cover

percentage was less steep compared with that on a medium

slope (15%). In other studies, the observed trend is that infil-

tration rate increases as surface cover (mulch, residue cover

and rock fragments) increases (Bradford & Huang, 1994;

de Figueiredo & Poesen, 1998). Infiltration rate also increases

with an increasing slope gradient (Poesen, 1984; Janeau

et al., 2003) and this effect is more pronounced for higher

rainfall intensities (Assouline & Ben-Hur, 2006). On steep

slopes, the development of a surface seal is obstructed by

intense rain-wash because more fine soil particles, susceptible

to be washed in, are transported by overland flow. Thus, the

final thickness and permeability of the surface seal layer is

reduced (Poesen, 1986).

The results of this study indicate that with increasing geo-

textile cover (decreasing mesh size), the run-off coefficient

decreases linearly. For most treatments, the run-off coeffi-

cient is significantly larger on the 15% slope than on the

45% slope, indicating less run-off is occurring on a steep

slope (45%) than on a medium slope (15%). However, for

the simulated and the Borassus geotextiles, the ratio between

the run-off coefficients for the 15% slope and the 45% slope

is higher for the 45-mm h)1 rainfall intensity than for the

67-mm h)1 rainfall intensity. This is in contrast to the trend

observed by Assouline & Ben-Hur (2006) who state that the

difference in cumulative run-off between a 5% slope and a

25% slope is smaller for the 24-mm h)1 rainfall than for the

60-mm h)1 rainfall.

Inter-rill soil loss. The results from this study indicate

Borassus, Buriti and simulated geotextiles with different cover

C (%)0 20 40 60 80 100

CR

(%

)

0

5

10

15

20

25

sim.geotext (S=15%) sim.geotext. (S=45%) Borassus (S=15%)

Borassus (S=45%)

Buriti (S=15%) Buriti (S=45%)

I = 45 mm h–1

Figure 12 Chain roughness (CR) versus cover percentage (C) for

bare inter-rill plots and plots covered with Borassus, Buriti and

simulated geotextiles on two slope gradients (S) and during a rainfall

intensity (I) of 45 mm h)1 (error bars represent 1 SD).

C (%)0 20 40 60 80 100

CR

(%

)

0

5

10

15

20

25

sim.geotext. (S=15%) sim.geotext. (S=45%) Borassus (S=15%) Borassus (S=45%) Buriti (S=15%) Buriti (S=45%)

I = 67 mm h–1

Figure 13 Chain roughness (CR) versus cover percentage (C) for

bare inter-rill plots and plots covered with Borassus, Buriti and

simulated geotextiles on two slope gradients (S) and during a rainfall

intensity (I) of 67 mm h)1 (error bars represent 1 SD).

Effectiveness of erosion control palm geotextiles 313

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

percentages to be very effective in reducing total ISL on the

two tested slope gradients (15 and 45%) and under two rainfall

intensities (45 and 67 mm h)1). Calculated b values of the

mulch cover equation indicate the relative effectiveness of

geotextiles in reducing total ISL and range from 0.023 to

0.048. However, on a 15% slope, b values are significantly

higher than on a 45% slope. Rainfall intensity has a smaller

effect on the b values than slope gradient. Previous studies

indicate that covering the soil surface with a mulch is highly

effective in reducing inter-rill soil loss. Soil surface cover

influences inter-rill erosion in several ways. An increasing soil

surface cover decreases the amount of bare soil exposed to

rainfall detachment, restricts the movement of soil by splash,

slows the mean flow velocity appreciably and significantly

decreases run-off volume (e.g. Lattanzi et al., 1974; Khan

et al., 1988; McGregor et al., 1988). Previous research shows

that b values derived from measured erosion data range from

approximately 0.025 to >0.1 (Foster, 2005). However,

depending on the combinations of rainfall intensities and slope

gradients investigated in previous studies, total ISL either

increases or is not correlated with slope gradient (Assouline &

Ben-Hur, 2006). Lattanzi et al. (1974) studied the effects of

mulch rate and slope steepness on inter-rill erosion using a

0.61 · 0.61-m inter-rill erosion plot and simulating rainstorms

with an intensity of 64 mm h)1 on slopes ranging from 2 to

20%. They found that the relative effectiveness of straw mulch

in reducing inter-rill erosion decreased with increasing slope

steepness. However, the differences in effectiveness of a cover

on different slope gradients are less pronounced than those

observed in this study (15 < S < 45%). Straw mulch is

mostly applied on croplands where slope gradients rarely

exceed 20%. On the other hand, geotextiles can be installed

on steep slopes (>20%) because they are anchored to the soil

surface. However, this study indicates that on a very steep

slope (45%), the effectiveness of geotextiles in reducing total

ISL is about half of that on a medium slope (15%).

For a medium slope gradient (12–15%), geotextiles are

more effective in reducing ISLrel (b = 0.042–0.048) com-

pared with mulch elements (b = 0.025, Foster, 2005). This

can be explained by the fact that geotextiles create bare soil

patches which are completely surrounded by fixed strips

which in turn disconnect the bare patches (Figure 2 and 3).

By contrast, ordinary mulches also create bare soil patches

but these are not completely surrounded by mulch elements

leading to a pattern of connected bare patches. Hence,

sediment connectivity is smaller with geotextiles compared

with that in commonly used mulch elements.

Some studies indicate that the effect of slope gradient on

total ISL is dependent on rainfall intensity (e.g. Lattanzi

et al., 1974; Fox & Bryan, 1999; Assouline & Ben-Hur,

2006). For the bare inter-rill plots and the plots covered

with the simulated geotextiles, the effect of slope gradient on

total ISL is larger for the 67-mm h)1 rainfall intensity than

for the 45-mm h)1 rainfall intensity. This corresponds to the

trend observed by Assouline & Ben-Hur (2006) who studied

the effect of rainfall intensity and slope gradient on the

dynamics of inter-rill erosion during surface seal formation.

A sandy loam was exposed to 70 mm of rainfall at two

intensities (24 and 60 mm h)1) and five slope gradients ran-

ging from 5 to 25%. They concluded that the ratio between

the cumulative soil loss for the 25% slope and the 5% slope

was 2.0 for the 24-mm h)1 rainfall and 5.1 for the

60-mm h)1 rainfall. In this study, the ratio between the total

ISL for the 45% slope to the 15% slope was for all treat-

ments an average of 4.2 for the 45-mm h)1 rainfall and 5.5

for the 67 mm h)1 rainfall.

Soil surface roughness. Chain roughness (CR, %) increased

linearly with increasing surface cover of geotextiles and lar-

ger CR values were measured on the 45% slope than the

15% slope. However, the effectiveness of the simulated geo-

textiles in increasing CR is independent of slope gradient

and rainfall intensity. The CR values measured in this study

are of the same order as the values observed by Merrill

(1998) who compared field research measurements made with

the chain set with micro-topographic data collected with a

laser scanner.

Comparison with soil erosion models

In this study, b values, indicating the relative effectiveness of

the soil surface cover in reducing total ISL, were calculated

for the two tested slope gradients (15 and 45%) and rainfall

intensities (45 and 67 mm h)1) (Table 6). In RUSLE2.0 the

effect of cover on inter-rill and rill erosion is determined by

the ‘Ground cover subfactor’ of the C-factor (Foster, 2005).

To describe the relative effectiveness of surface cover in redu-

cing only inter-rill erosion, a constant b value of 0.025 is

assumed in RUSLE2.0 (Foster, 2005). The relation between

surface cover and the predicted ISLrel according to

RUSLE2.0 and the observed ISLrel from this study is shown

in Figure 14. The observed data from this study confirm the

exponential trend as predicted by RUSLE2.0. However,

depending on the slope gradient and the simulated rainfall

intensity, the b values in this study range from 0.048 to 0.023

(Table 6). On a medium slope (15%), simulated geotextiles

are far more effective in reducing ISLrel than on a steep slope

(45%). The ground cover subfactor equation of the

RUSLE2.0 predicts the observed trend very well; however,

no effect of slope gradient is incorporated in this equation.

The effect of slope gradient on the b value is larger com-

pared with the effect of rainfall intensity. The constant b

value of 0.025 used in RUSLE2.0 for inter-rill erosion is

derived from data observed by Lattanzi et al. (1974) and

McGregor et al. (1988). However, in these studies b values

were determined from inter-rill plots with slopes ranging

from only 2 to 20%.

314 T. Smets et al.

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

Conclusions

In the last two decades, much expansion has occurred within

the erosion control industry, including an increase in the

number, variety and application of erosion control geo-

textiles (Ziegler et al., 1997; Sutherland, 1998b). However,

there are no standardized protocols available and only

limited data exist on the effectiveness of geotextiles in

increasing infiltration and in reducing run-off, rill and inter-

rill erosion (Theisen, 1992; Allen, 1996; Ziegler et al., 1997;

Sutherland, 1998a,b,c; Ziegler & Sutherland, 1998; Rickson,

2006). To assess the effectiveness of different types of geo-

textiles in controlling water erosion, a well-defined protocol

was developed in this study. Both Borassus and Buriti geo-

textiles significantly (a=0.05; n = 12) increased the infiltra-

tion rate and decreased run-off volume and total inter-rill

soil loss compared with a bare soil treatment on a 15% slope

and a 45% slope and during 45- and 67-mm h)1 rainfall

intensities. Calculated k values from the Horton infiltration

equation, ranged from 0.025 to 0.145 indicating that for all

combinations of cover and slope gradient, infiltration rate

decreased faster during a 67-mm h)1 rainfall intensity than

during a 45-mm h)1 rainfall intensity and simulated

geotextiles were more effective in increasing infiltration rate

on a medium slope (15%) than on a steep slope (45%). The

tested geotextiles were more effective in reducing total inter-

rill soil loss on a medium slope (15%) than on a steep slope

(45%). Calculated b values as derived from the ground cover

equation in RUSLE2.0 (Foster, 2005) were significantly

higher on the 15% slope (0.045 ± 0.003) compared with that

on the 45% slope (0.024 ± 0.001). In RUSLE2.0, a constant

b value of 0.025 is assumed to describe the relative effective-

ness of surface cover in reducing only inter-rill erosion

(Foster, 2005). However, this study indicates, depending on

the slope gradient, the b value for inter-rill erosion can

significantly change. Induced soil surface roughness at the

end of each experiment, measured as chain roughness (CR,

%) (Saleh, 1993), ranged from 0.6 to 22.7%. CR increased

linearly with increasing geotextile cover and higher CR

values were measured on the 45% slope compared with that

on the 15% slope. Overall, the geotextiles tested in this study

behaved similar to mulch. However, on a medium slope

(15%), geotextiles are more effective in reducing relative

inter-rill soil loss compared with mulch elements which are

attributed to a smaller sediment connectivity between the

bare soil patches for a surface covered with geotextiles. Care

should be taken when extrapolating results on the effective-

ness of geotextiles in increasing infiltration rates and

reducing inter-rill run-off and erosion rates from low and

medium slopes to steep slopes.

Acknowledgements

Many thanks are due to Jeroen Monsieur for his help during

the experiments. This research is part of the BORASSUS

Project: ‘The environmental and socio-economic contribution

of palm-leaf geotextiles to sustainable development and soil

conservation’ (contract number: INCO-CT-2005-510745),

funded by the European Commission (EC), Specific Targeted

Research Projects (FP6-STREPs) for Developing Countries

(INCO-DEV) Programme. This Project is dedicated to the

memory of Dr Kathy Davies who was the initiator and inspi-

ration for palm-leaf geotextile research.

References

Allen, S.R. 1996. Evaluation and standardization of rolled erosion

control products. Geotextiles and Geomembranes, 14, 207–221.

Assouline, S. & Ben-Hur, M. 2006. Effects of rainfall intensity and

slope gradient on the dynamics of interrill erosion during soil sur-

face sealing. Catena, 66, 211–220.

Booth, C.A., Davies, K. & Fullen, M.A. 2005. Environmental and

socio-economic contributions of palm-leaf geotextiles to sustaina-

ble development and soil conservation. In: Ecosystems and sustain-

able development V (eds E. Tiezzi, C.A. Brebbia, S.E. Jorgensen &

D. Almorza-Gomar), pp. 649–658. WIT Press, Southampton, UK.

Bradford, J.M. & Huang, Chi-hua. 1994. Interrill soil erosion as

affected by tillage and residue cover. Soil & Tillage Research, 31,

353–361.

Davies, K., Fullen, M.A. & Booth, C.A. 2006. A pilot project on the

potential contribution of palm-mat geotextiles to soil conservation.

Earth Surface Processes and Landforms, 31, 561–569.

C (%)0 20 40 60 80 100

ISL re

l (%

)

0

20

40

60

80

100

sim.geotext. (I=45 mm h–1 , S=15%), b=0.048, R2=0.96; n=9; P=0.001sim.geotext. (I=67 mm h–1 , S=15%), b=0.042, R2=0.95; n=9; P=0.029sim.geotext. (I=45 mm h–1 , S=45%), b=0.023, R2=0.94; n=9; P=0.054sim.geotext. (I=67 mm h–1 , S=45%), b=0.024, R2=0.93; n=9; P=0.016RUSLE2.0 , b=0.025

Figure 14 Relative total inter-rill soil loss (ISLrel) versus cover per-

centage (C) for two rainfall intensities (I) and two slope gradients (S)

according to this study and according to the ground cover equation

(3) from RUSLE2.0 (for inter-rill erosion) (Foster, 2005); b is a coef-

ficient that describes the relative effectiveness of surface cover for

reducing inter-rill erosion.

Effectiveness of erosion control palm geotextiles 315

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316

de Figueiredo, T. & Poesen, J. 1998. Effects of surface rock fragment

characteristics on interrill runoff and erosion of a silty loam soil.

Soil & Tillage Research, 46, 81–95.

Foster, G.R. 2005. Revised universal soil loss equation, Version 2.0

(RUSLE2.0): science documentation. USDA-Agricultural

Research Service, Washington, DC.

Fox, D.M. & Bryan, R.B. 1999. The relationship of soil loss by

interrill erosion to slope gradient. Catena, 38, 211–222.

Guerra, A., Marcal, M., Polivanov, H., Sathler, R., Mendonca, J.,

Guerra, T., Bezerra, F., Furtado, M., Lima, N., Souza, U.,

Feitosa, A., Davies, K., Fullen, M.A. & Booth, C.A. 2005.

Environmental management and health risks of soil erosion gullies

in Sao Luıs (Brazil) and their potential remediation using palm-

leaf geotextiles, In: Environmental health risk III (eds C. A

Brebbia, V. Popov & D. Fayzieva), pp. 459–467. WIT Press,

Southampton, UK.

Hann, M.J. & Morgan, R.P.C. 2006. Evaluating erosion control

measures for biorestoration between the time of soil reinstatement

and vegetation establishment. Earth Surface Processes and Land-

forms, 31, 589–597.

Horton, R.E. 1933. The role of infiltration in the hydrological cycle.

American Geophysical Union, Transactions, 14, 446–460.

Janeau, J.L., Bricquet, J.P., Planchon, O. & Valentin, C. 2003. Soil

crusting and infiltration on steep slopes in northern Thailand.

European Journal of Soil Science, 54, 1–11.

John, M.W.M. 1987. Geotextiles. Blackie and Son, Glasgow.

Khan, M.J., Monke, E.J. & Foster, G.R. 1988. Mulch cover and

canopy effect on soil loss. Transactions of the American Society of

Agricultural Engineers, 31, 706–711.

Lattanzi, A.R., Meyer, L.D. & Baumgardner, M.F. 1974. Influences

of mulch rate and slope steepness on interrill erosion. Soil Science

Society of America Journal, 38, 946–950.

Laws, J. 1941. Measurements of the fall-velocity of waterdrops and

raindrops. Transactions, Geophysical Union, 22, 709–721.

Li, Y., Tullberg, J.N. & Freebairn, D.M. 2001. Traffic and residue

cover on infiltration. Australian Journal of Soil Research, 39,

239–247.

McGregor, K.C., Bengtson, R.L. & Mutchler, C.K. 1988. Effects of

surface straw on interrill runoff and erosion of Grenada silt loam

soil. Transactions of the American Society of Agricultural Engin-

eers, 31, 111–116.

Merrill, S.D. 1998. Comments on the chain method for measuring

soil surface roughness: use of the chain set. Soil Science Society of

America Journal, 62, 1147–1149.

Poesen, J. 1984. The influence of slope angle on infiltration rate and

Hortonian overland flow. Zeitschrift fur Geomorphologie, Supple-

ment Band, 49, 117–131.

Poesen, J. 1986. Surface sealing as influenced by slope angle and

position of simulated stones in the top layer of loose sediments.

Earth Surface Processes and Landforms, 11, 1–10.

Poesen, J. & Lavee, H. 1991. Effects of size and incorporation of

synthetic mulch on runoff and sediment yield from interrills in a

laboratory study with simulated rainfall. Soil & Tillage Research,

21, 209–223.

Poesen, J., Ingelmo-Sanchez, F. & Mucher, H. 1990. The hydrologi-

cal response of soil surfaces to rainfall as affected by cover and

deposition of rock fragments in the top layer. Earth Surface Pro-

cesses and Landforms, 15, 653–671.

Rickson, R.J. 2006. Controlling sediment at source: an evaluation of

erosion control geotextiles. Earth Surface Processes and Land-

forms, 31, 550–560.

Saleh, A. 1993. Soil roughness measurement: chain method. Journal

of Soil and Water Conservation, 48, 527–529.

Salles, C., Poesen, J. & Sempere-Torres, D. 2002. Kinetic energy of

rain and its functional relationship with intensity. Journal of

Hydrology, 257, 256–270.

SAS Institute Inc. 2002. SAS Version 9.1. SAS Institute Inc., Cary,

NC.

Sutherland, R.A. 1998a. Rolled erosion control systems for hillslope

surface protection: a critical review, synthesis and analysis of

available data. I. Background and formative years. Land Degrada-

tion & Development, 9, 465–486.

Sutherland, R.A. 1998b. Rolled erosion control systems for hillslope

surface protection: a critical review, synthesis and analysis of

available data. II. The post-1990 period. Land Degradation &

Development, 9, 487–511.

Sutherland, R.A. 1998c. A critical assessment of the research con-

ducted at the hydraulics and erosion control laboratory – a focus

on rolled erosion control systems applied to hillslopes. Geotextiles

and Geomembranes, 16, 87–118.

Theisen, M.S. 1992. The role of geosynthetics in erosion and sedi-

ment control: an overview. Geotextiles and Geomembranes, 11,

535–550.

Ziegler, A.D. & Sutherland, R.A. 1998. Reduction in interrill sedi-

ment transport by rolled erosion control systems. Soil & Tillage

Research, 45, 265–278.

Ziegler, A.D., Sutherland, R.A. & Tran, L.T. 1997. Influence of

rolled erosion control systems on temporal rainsplash response – a

laboratory rainfall simulation experiment. Land Degradation &

Development, 8, 139–157.

316 T. Smets et al.

ª 2007 The Authors. Journal compilation ª 2007 British Society of Soil Science, Soil Use and Management, 23, 306–316