Catena 118 (2014) 168–178

Contents lists available at ScienceDirect

Catena
journal homepage: www.elsevier.com/locate/catena

Evaluation of erosion control geotextiles on steep slopes. Part 1: Effects

on runoff and soil loss
Jesús Álvarez-Mozos a,⁎, Eguzki Abad a, Rafael Giménez a, Miguel A. Campo a, Mikel Goñi a, Maider Arive a,
Javier Casalí a, Javier Díez a, Ignacio Diego b
a
Public University of Navarre, Dep. of Projects and Rural Engineering, Los Tejos Building, Arrosadia Campus, 31006 Pamplona, Spain
b
Huesker Geosintéticos S.A., P.I., Talluntxe II, Calle O, 31110 Noáin, Spain

a r t i c l e i n f o a b s t r a c t

Available online 20 June 2013 Geotextiles have been demonstrated to be effective in reducing erosion and subsequent slope degradation
processes. However, most studies devoted to assessing the effectiveness of those elements have considered
Keywords: small or intermediate slope gradients (normally ≤ 30°), whereas real engineering works often lead to steeper
Erosion control slopes needing protection. In addition, although a large amount of laboratory studies exist on this topic, there
Engineering slopes are few articles reporting field experiments which consider large plots, real precipitation and prolonged
Geotextiles
study periods. This paper is the first of two in which the performance of erosion control geotextiles on
Runoff
Soil loss
steep slopes (45° and 60°) has been assessed. In this paper, the influence of geotextiles on runoff and soil
loss reduction was evaluated. The second paper focuses on the effects of geotextiles on the establishment
and growth of vegetation on the slope. The research was carried out on an experimental embankment
built in Pamplona (Spain) following standard construction procedures in order to resemble real engineering
slopes. Two biological geotextiles (jute net and coir blanket) and a synthetic polyester geogrid, installed in
two positions (surface-laid and buried), were evaluated and compared with a vegetated hydroseeded control
plot. After each significant rain event, runoff and soil loss amounts produced on each plot were recorded. The
results showed that coir and jute geotextiles produced 2–3 times larger runoff volumes than the control plots
on both the 45° and 60° slopes, whereas the synthetic geogrid did not give significantly different runoff rates
from the control (at p b 0.05). Geotextiles were more effective in reducing soil loss at 45° than at 60°. The
most effective treatment was the surface-laid geogrid, with a median Soil Loss Reduction Effectiveness
(SLRE) of 76% and 53% for the 45° and the 60° slopes, respectively. At 45° the coir blanket showed lower
soil losses than the control (median SLRE of 61%), but at 60° its protection action was more irregular (median
SLRE of only 8%). The jute net and the buried geogrid produced lower erosion rates (median soil loss of
3.2 g m−2 and 2.1 g m−2, respectively) than the control (median soil loss of 3.6 g m−2) at 45°, but at 60°
the erosion rates observed were similar to those of the control. In short, the surface-laid geogrid produced
the lowest soil loss rates. However, when buried, the geogrid did not effectively control erosion, at least
not that of the upper soil layer. Thus, its surface installation seems to be a better option. In cases where burial
is preferred for aesthetic reasons (i.e. hiding the geogrid), a higher soil loss must be assumed.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction One of the most used control measures is the placement of

Civil engineering projects often lead to disturbed, bare soils with
steep slopes, which are very sensitive to runoff and soil loss pro-
cesses. The degradation of those slopes causes severe impacts
onsite (e.g. reduced soil structure and water holding capacity
resulting in lower fertility) (Cerdá, 2007) and offsite (e.g. malfunc-
tion of gutters and drains affecting road conservation and safety,
turbidity and siltation of water courses) (Morgan, 2005; Owens
et al., 2005), so that the design and implementation of erosion
control measures is crucial in any project of this kind.
⁎ Corresponding author. Tel.: +34 948169235; fax: +34 948169644.
E-mail address: jesus.alvarez@unavarra.es (J. Álvarez-Mozos).
geotextiles over sloping surfaces (Rickson, 1995). Geotextiles are
nets or mats which protect the soil and reduce the detachment
and transport capacity of rainfall and overland flow. Ideally, they
should also promote vegetation growth, due to their enhanced soil
and water holding capacity. Their protection mechanism is there-
fore expected to be twofold: first, through their direct protection
of the soil surface against eroding agents, and second, through ben-
efits in the development of a dense protective vegetation cover.
Several physical properties of geotextiles have proven to be
vital for soil protection. Specifically, the fraction of open area (or
the relative size of apertures), moisture sorption depth, mat thick-
ness, hydraulic roughness and tensile strength are key properties
controlling the performance of such materials in each particular

0341-8162/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.catena.2013.05.018
 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178
setting (Rickson, 2006). Additionally, the erosion reduction capac-
ity of a geotextile will also vary depending on slope length and
gradient, soil type and local precipitation regime (Bhattacharyya
et al., 2010).
Geotextiles are normally made of permeable materials which can
be either biological or synthetic. The former are mainly made from
natural fibres (i.e., jute, coir or palm). Synthetic geotextiles are nor-
mally made of polymers (i.e., polyester, polypropylene or polyolefin).
The cost of synthetic geotextiles is significantly higher than that of bi-
ological materials. However, the performance of a geotextile is related
to its physical properties (mentioned above) which are much more
important than the geotextile material itself (Ziegler et al., 1997).
Synthetic geotextiles offer stronger reinforcing action (i.e. higher
tensile strength) and last longer (≥ 20 years) than natural fibres
(~ 2–5 years) (Li and Khanna, 2008). In turn, the degradation of bi-
ological geotextiles (contributing organic matter and nutrients to
the soil) could be beneficial for the development of vegetation on
slopes, which may play a key role as a long term erosion protection
measure (Fullen et al., 2011).
Several comprehensive reviews focusing on the influence of dif-
ferent factors and geotextile properties on their performance have
been published previously (Bhattacharyya et al., 2010; Rickson,
2006; Sutherland, 1998a, 1998b). Bhattacharyya et al. (2010) consid-
ered that three separate erosion types (rainsplash, interrill and rill–
interrill erosion) depended on the length of slopes to be protected,
and they summarised the results of many laboratory and field exper-
iments. Most experiments reported by Bhattacharyya et al. (2010)
confirmed the effectiveness of geotextiles for reducing runoff, obtaining
similar results with both synthetic and biological geotextiles. Their ana-
lyses also highlighted the role of geotextiles in reducing the impact of
splash erosion behaving as a physical barrier, which dissipated the ki-
netic energy of raindrops and their erosivity. Accordingly, the area
cover fraction was identified as a key property for erosion reduction.
Characteristics of slopes to be protected are also crucial for the ad-
equate performance of geotextiles. As already mentioned, slope length
and gradient are particularly important properties which influence,
and even sometimes determine, the type of erosion process occurring.
Long and steep slopes usually lead to higher energy runoff with a stron-
ger sediment detachment and transport capacity (Smets et al., 2007).
However, steeper slopes do not necessarily have lower infiltration
rates since other factors (such as surface roughness and sealing) can af-
fect infiltration on them (Assouline and Ben-Hur, 2006). The protective
action of a geotextile might also vary depending on the length and gra-
dient of the slope to be protected. Some authors observed better soil
contact (and hence more effective protection) on short slopes than on
long ones, where void spaces could cause preferential flow paths under-
neath the geotextile (Chen et al., 2011). In short, care should be taken
when extrapolating results on the effectiveness of geotextiles from
low and medium slopes to steep slopes (Chen et al., 2011; Ogbobe et
al., 1998; Smets et al., 2007).
Most experiments assessing the protective action of geotextiles
have been carried out in laboratory settings using rainfall simulators
and considering reduced plot dimensions and slope gradients
(Bhattacharyya et al., 2010). Outdoor experiments on this topic are
scarce, particularly those considering long time periods. Although
laboratory experiments provide valuable information on the process-
es and factors involved, the conditions tested in these settings might
not entirely represent real field conditions (Smets et al., 2011). In fact,
soil properties (moisture, texture, structure, depth, and profile) and
rainfall characteristics (intensity, duration, depth, kinetic energy,
and water quality) can be significantly different (Smets et al., 2011).
As a result, it is necessary to further extend research activities in
this topic to more realistic conditions, where longer and steeper
slopes are examined under natural precipitation.
In this study, the performance of erosion control geotextiles on
steep slopes (45° and 60°) is evaluated under natural conditions.
169
With this aim, an experimental embankment was constructed follow-
ing standard construction practices. The focus of the study is to assess
the influence of geotextiles on runoff and soil loss reduction.
2. Materials and methods
2.1. Experimental setting
The experiment was carried out in the experimental fields of the
School of Agricultural Engineering of the Public University of Navarre
in Pamplona (Spain). Its geographical coordinates are: 42°47′39″N,
1°37′52″W, and it is 435 m above sea level. The climate is humid sub-
Mediterranean with annual precipitation of ~800 mm and a mean tem-
perature of ~12 °C. The precipitation regime is strongly seasonal, with
dry summers with short, intense convective storms and longer
front-like events during the rest of the year. The soils are silty–clay–
loam in texture (sand = 13.8%; silt = 53.9%; clay = 32.3% and organic
matter = 1.8%) with significant silt and low organic matter proportions
and, hence, they are highly erodible.
An experimental embankment was constructed following stan-
dard construction practices. The use of an outdoor experimental set-
ting makes it possible to reproduce an engineered slope with fixed
characteristics (gradient, length or soil type) more realistically and
is the most convenient choice for medium to long term experiments
with natural precipitation events and vegetation dynamics. However,
one drawback is that the continuous measurement of hydrological
variables (infiltration, overland flow and sediment concentration) is
more difficult than in laboratory plots.
The material used in the earthwork operations was the same soil
present on the site, which was mechanically compacted (layer by
layer) using a smooth steel-wheel roller, in order to reproduce ‘real’
earthworks as far as possible. The final mean bulk density at the soil
surface was 1.57 Mg m−3 (n = 20, SD = 0.1), which corresponds to
a significantly compacted soil. The local undisturbed soil had a mean
bulk density of ~1.34 Mg m−3. The orientation of the embankment
was east and its total width was 50 m. It was divided into two sections
with different slope gradients, 45° and 60°, and an intermediate transi-
tion area. Five different geotextile treatments were created on each sec-
tion (each treatment had three replicates) making a total of 30 plots,
each 1.3 m wide. The embankment reached a height of ~4.5 m above
its base. The first 1.5 m was required to install flow collecting gutters
and tanks and the remaining 3 m were used for the experiment,
resulting in a final length for the experimental plots of 4.1–4.6 m for
the 45° slope and 3.4–3.8 m for the 60° slope. A detailed topographic
survey was used to obtain the exact dimensions of each plot (Table 1).
The different sizes of the 60° plots (~7.5 m2) and 45° plots (~13 m2)
need to be taken into account when comparing the results of the two
datasets. The top of the embankment was sealed with plastic and a
crossfall of 3–5% was placed towards the rear in order to avoid overland
flow into the slope and to reduce infiltration through the embankment.
The experimental slope was hydroseeded on 11 May 2009 using a
seed mixture with species adapted to local climate conditions, mainly
grasses (~ 70%) and legumes (~ 25%) (Alvarez-Mozos et al., submitted
for publication). The seeding rate was 70 g m−2. Due to logistic con-
straints, the seeding was performed outside the recommended
seeding period. Therefore, in order to ensure adequate germination,
a sprinkler irrigation system was installed, which was switched on
Table 1
Plot size (m2) for each treatment and slope gradient.
Jute net Coir blanket Geogrida Geogrid (B)b Control
60° slope 7.86 7.53 7.34 7.66 7.57
45° slope 13.05 13.58 13.79 12.84 12.95
a
Surface laid.
b
Buried.
170 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178

Fig. 1. Experimental slope 1 month after hydroseeding.

for 2 min every 2 h during daytime and kept on until 1 July 2009,
when the establishment of vegetation (in most plots) was consid-
ered to be successful. Control plots, with no geotextile cover, were
also created on both 45° and 60° slopes. The control plots were
also hydroseeded. Fig. 1 shows the experimental slope 1 month
after hydroseeding.
2.2. Geotextiles evaluated
The four treatments tested on both 45° and 60° slopes consisted of
geotextile products currently available on the market, and comprised
both biological (jute net and coir blanket) and synthetic (polyester)
materials. The most important properties of the tested geotextiles
are summarised in Table 2.
Jute is a widespread material in the geotextile industry, which can
be marketed in products with different characteristics. Its main
properties are its high water sorption capacity and mass, which en-
hance its attachment to the soil when wet (drapability) (Chen et
al., 2011). In addition, it is a very flexible material with a moderate
tensile strength. It is highly biodegradable and its action normally
lasts for 1–2 years (Mitchell et al., 2003). The coir (coconut fibre) is
also a very common biological material, which is lighter and has a
lower water sorption capacity than jute (Sutherland and Ziegler,
2007). Coir geotextiles usually have a very close random fibre struc-
ture that accounts for its superior durability compared to other nat-
ural fibres (Lekha, 2004). As a result, its open area fraction is usually
very small.
The synthetic material tested here is a three-dimensional polyes-
ter geogrid (Fortrac® 3D). This geogrid has a 3D traverse structure,
which is supposed to provide enhanced soil holding ability. In spite
of its light weight, it is a very strong material with a very low
deformability. Fortrac® 3D is marketed in different products with in-
creasing tensile strengths. In the present study, the 30 kN m−1 product,
which has a maximum elongation of 12.5% and a mesh size of
1.8 × 0.8 cm, was used. Its durability is much longer than that of
the biological geotextiles and can reach up to 20 years. The geogrid
was installed in two positions: buried and surface-laid. The buried
geogrid was covered with a 2–3 cm layer of loose silty soil prior to
hydroseeding.
2.3. Data collection
The experiment started on 1 June 2009. Runoff and sediment
concentration measurements were acquired after each significant
precipitation event (Fig. 2) (i.e., an event producing measurable
runoff), or occasionally after a succession of smaller events. Runoff
depth was measured directly in each tank and sediment concentra-
tion was determined by taking a 500 ml sample from each plot.
Prior to sample acquisitions, the water contained in each tank was
homogenised and after the measurement was completed, the
tanks were emptied and cleaned using pressure water. Runoff sam-
ples were oven dried (at 110 °C) until the water evaporated and
only dry sediment remained (normally 7–10 days). In summary,
the variables measured for each plot and event were: runoff
(mm), sediment concentration (g l− 1) and soil loss (g m− 2). The
three replications made for each treatment were averaged in order
to obtain a mean value.
The experimental period lasted for 21 months, but the 60° slope
suffered stability problems after heavy rainfall in November 2009
and finally collapsed, so the data acquisition had to be stopped on
this slope 7 months after the initiation of the experiment. In total,
six series of measurements were acquired on the 60° slope and 16
on the 45° slope. The first two series were acquired when the irriga-
tion system was on, the rest under natural precipitation.
Runoff and soil loss measurements were tested for normality using
the Shapiro–Wilk test. The Shapiro–Wilk tests yielded p-values b 0.05
indicating that normality could not be assumed. The distributions
obtained were skewed towards higher runoff or soil loss values. As a re-
sult, the differences between each treatment and the control plot were
evaluated using the Mann–Whitney non-parametric test, which does
not require normally distributed populations.
In order to compare runoff and soil loss values observed in
each treatment with those of the control plots, ratios between
treatments and control (Ogbobe et al., 1998) and effectiveness in-
dexes (Sutherland, 1998a) were proposed. In this experiment, the

Table 2
Main properties of tested geotextiles.

Treatment Roll size Thickness Mass per area Open area Water holding capacity Tensile strength
(m × m) (mm) (g m−2) (%) (mm) (kN m−1)

1 — Jute net 1.22 × 68.6 5 500 60 3.1 21

2 — Coir blanket 2 × 33 8 271 7 0.84 3
3 — Geogrid 4.5 × 100 10 300 50 – 30
4 — Geogrid (B) 4.5 × 100 10 300 100 – 30
 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178 171

Fig. 2. Experimental period and data acquisition on the 45° and 60° slopes.

runoff reduction effectiveness (RRE, %) and soil loss reduction ef-
fectiveness (SLRE, %) were used (Sutherland, 1998a):
 
Rcontrol −Ri
RREi ¼ ⋅100
Rcontrol
 
SLcontrol −SLi
SLREi ¼ ⋅100
SLcontrol
where, RREi and SLREi are runoff and sediment loss reduction ef-
fectiveness of a treatment i, and R and SL are runoff and soil loss,
respectively. A positive effectiveness indicates that the geotextile re-
duces runoff or soil loss, whereas a negative value corresponds to
geotextiles that produce more runoff or soil loss than the control. In
the literature, effectiveness has been calculated either by using individ-
ual runoff (or soil loss) measurements observed during an experiment
and computing their mean (or median) (Sutherland, 1998a; Sutherland
and Ziegler, 2007), or by employing total accumulated values at the end
of the experiment (Bhattacharyya et al., 2009; Mitchell et al., 2003).
Even if the effectiveness concept is the same, the influence of individual
extreme events is much greater when total values are used. From a statis-
tical viewpoint, the mean (or median) effectiveness computed using a se-
ries of individual values might be more correct. In any case, both
ð1Þ
effectiveness values are computed and discussed in this study.
An automatic weather station was located close to the experi-
mental slope (b 200 m), where the main meteorological variables
ð2Þ
(temperature, precipitation, wind speed and direction, relative hu-
midity and incoming solar radiation) were measured every 10 min.
Several precipitation variables were calculated for the time lapse be-
tween two measurements. In particular, the accumulated precipitation
depth (Pac, mm), rainfall duration (tR, h), intensity (I30, mm h−1) and
erosivity (EI30, MJ mm ha−1 h−1) were calculated for each period
and for the main erosive event in the period (Table 3). Additionally,
the wetness condition of each period was estimated using an Anteced-
ent Precipitation Index (API7, mm), which considers precipitation
recorded during the seven previous days (Heggen, 2001).
Table 3 reports the main precipitation characteristics recorded be-
tween every two measurements. Precipitation recorded during the

Table 3
Precipitation characteristics recorded in each measurement period. Accumulated precipitation, duration, intensity and erosivity of the main event and the accumulated character-
istics of the period are given.

Date Main event Period

Pac tR I30 EI30 ΣPac ΣtR ΣI30 ΣEI30 API7

(mm) (h) (mm h−1) (MJ mm (mm) (h) (mm h−1) (MJ mm (mm)
ha−1 h−1) ha−1 h−1)

9-Jun-2009 9.20 8.33 6.40 10.56 19.40 9.17 14.80 17.82 5.38
22-Jun-2009 26.20 3.33 31.20 171.72 28.80 3.33 31.20 171.72 9.18
11-Aug-2009 9.20 1.33 10.80 17.64 13.80 1.33 10.80 17.64 1.39
22-Sep-2009 36.00 34.33 6.00 23.33 48.80 37.17 17.60 36.96 3.27
23-Oct-2009 19.60 6.00 5.20 12.03 38.40 13.17 16.40 20.80 3.97
9-Nov-2009 72.20 30.83 11.20 103.00 118.00 54.17 25.60 123.03 14.58
7-Dec-2009 29.80 19.67 5.60 19.24 57.20 32.50 19.20 28.62 17.44
30-Dec-2009 11.80 7.50 7.20 10.02 41.60 18.00 8.80 11.24 5.91
19-Jan-2010 45.60 27.67 4.80 23.90 76.40 42.17 14.80 36.44 14.85
9-Feb-2010 29.80 31.33 4.00 12.50 54.00 49.33 7.60 14.59 9.74
2-Mar-2010 11.00 16.83 2.00 2.10 33.20 16.83 2.00 2.10 7.40
12-May-2010 31.40 6.83 18.40 98.68 114.40 45.67 57.20 149.91 6.03
22-Jun-2010 18.60 12.17 4.00 8.03 50.20 19.67 7.60 10.21 4.77
27-Jul-2010 26.40 4.50 24.40 133.14 45.20 12.17 36.80 143.76 5.36
21-Oct-2010 5.20 0.50 10.00 10.40 58.20 16.50 24.40 30.23 2.65
25-Nov-2010 19.60 10.17 8.00 19.93 96.00 58.00 26.00 42.71 9.62
172 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178
experiment was variable. In general, larger and longer precipitation
events were observed in winter and spring than in summer, when
the events were more intense.
3. Results
3.1. Influence of geotextiles on runoff production
Biological geotextiles (coir and jute) systematically produced
more runoff than geogrid and control (i.e. hydroseeded vegetation)
plots on both slopes (Fig. 3). The first two measurements were ac-
quired in the hydroseeding establishment phase, when vegetation
development was poor and the soil surface was wet due to sprinkler
irrigation. As a result, runoff volumes recorded on these first dates
were notable, even if precipitation was low. This was especially true
for the coir and jute treatments. Apart from those first measurements,
the events recorded between November 2009 and March 2010 pro-
duced the most runoff in all treatments.
The runoff volumes measured on the 60° slope treatments were clear-
ly higher than those observed at 45°, but this affirmation cannot be statis-
tically validated due to the short measurement series on the 60° slope.
Table 4 summarises the main statistics of measured runoff rates.
The total runoff volume, minimum and maximum values, median
and inter-quartile range (IQR) are given for each treatment. The me-
dian and IQR were used instead of the mean and standard deviation,
since runoff data were not normally distributed. In skewed distribu-
tions the median (2nd quartile) is a more realistic statistical location
measure than the mean. In these cases, the dispersion is better repre-
sented by the IQR, which is calculated as the difference between the
Fig. 3. Accumulated runoff measurement series in treatments on both 45° and 60° slopes.
3rd and 1st quartiles, and hence reflects the range where the 50% cen-
tral values are located for each treatment. It can be observed that run-
off from jute and coir plots had a greater variability (e.g. on the 45°
slope, the IQR was 3–4 times larger in jute and coir than in geogrid
and control plots), whereas geogrid treatments and the control
were less dispersed and generally had less runoff.
On the 45° slope, jute and coir treatments produced total runoff
volumes of > 100 mm, with median values of 9.43 and 7.36 mm, re-
spectively (Table 4). Their variability was also high with IQR values
of > 7 mm. On the other hand, both geogrid treatments and the con-
trol showed very similar values in terms of total runoff (~ 54 mm),
median runoff (~ 3 mm) and IQR (~ 1.5–2 mm). The Mann–Whitney
test confirmed these observations and detected significant differences
(at p b 0.05) between the control and both jute and coir plots. The
median runoff volumes measured were 2.5–3 times higher for coir
and jute than the control. On the contrary, the geogrid treatments
were not statistically different (p b 0.05) from the control in terms
of runoff.
On the 60° slope, jute and coir plots also produced 2–3 times larg-
er runoff volumes than the control, both in terms of total (~ 114 and
156 mm for jute and coir, respectively, versus 48 mm for the control)
and median (~ 17 and 22 mm for jute and coir, respectively, versus
7 mm for the control) runoff. In this case, their variability was not
as large, when compared with the control. The surface-laid geogrid
gave lower runoff values (~ 4 mm) than the control, whereas the bur-
ied treatment produced more runoff (~ 11 mm). In any case, statisti-
cally significant differences (at p b 0.05) were only found between
the coir plot and control. This might be a consequence of the small
size of the 60° dataset (n = 6 events).
 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178 173

Table 4
Runoff data collected for different treatments (no. of observations on 45° slope = 16, on 60° slope = 6). IQR stands for inter quartile range and RRE stands for runoff reduction
effectiveness.

45° slope 60° slope

Jute⁎ Coir⁎ Geogrid Geogrid (B) Control Jute Coir⁎ Geogrid Geogrid (B) Control

Total runoff (mm) 143.14 134.46 53.95 53.78 53.75 113.99 156.84 39.78 59.27 47.82
Range (mm) 1.76–16.09 2.14–21.72 0.44–7.76 0.31–7.94 0.46–9.03 2.04–40.33 3.72–59.36 2.04–14.03 1.17–15.4 0.53–14.4
Median (mm) 9.43 7.36 3.15 3.27 2.90 17.49 22.11 5.11 10.77 7.13
IQR (mm) 9.08 7.40 1.94 2.08 1.54 8.87 5.11 4.22 5.12 6.90
Overall RRE (%) −166.31 −150.16 −0.37 −0.06 – −138.37 −227.98 16.81 −23.94 –
Median RRE (%) −160.98 −114.64 1.07 −6.33 – −155.35 −211.46 19.05 −20.81 –
⁎ Significantly different than control at p b 0.05.

Fig. 4 plots runoff volumes of each treatment against those
obtained from the control. Most points in the geogrid treatments fol-
low the 1:1 line, showing identical behaviour to the control. The
points of the jute and coir plots are all well above the 1:1 line, show-
ing a systematic increase in runoff production compared with the
control. This was also reflected by the effectiveness values (RRE)
(Table 4). Both overall (total) RRE and median RRE are given. The dis-
tribution of RRE values obtained for each treatment is represented
with boxplots in Fig. 5. Biological geotextiles gave very low negative
RRE values, whereas geogrid treatments also had negative values
but closer to 0% at 45°, and at 60° positive values in the surface-laid
geogrid, and moderately negative values in the buried geogrid. RRE
had a very low variability in the geogrid treatment (Fig. 5). Geogrid
(B) was less effective and its dispersion was still low. On the other
hand, jute and coir geotextiles displayed highly dispersed RRE distri-
butions. This can be explained by the enhanced runoff production of
these treatments in wet periods (leading to very low RRE values),
whereas in drier periods runoff is closer to that of the control (RRE
values close to 0).
3.2. Influence of geotextiles on soil loss
The series of accumulated soil loss values (Fig. 6) illustrate the im-
portance of certain precipitation events in the occurrence of erosion.
In particular, the first two precipitation events (where vegetation
was poorly developed and soil wetness was high) produced very
high soil loss rates, especially on the control and geogrid (B) plots.
But the most extreme erosion rates were measured in May 2010,
when a series of events resulted in very significant soil losses in the
control and, to a lesser extent, in the geogrid (B) and jute plots.
These results highlight the crucial role of geotextiles when particular-
ly erosive events occur. In these major events, erosion rates were al-
ways highest in control plots.
The very high variability of measured soil loss rates is represented
in the large dispersion of data (Table 5). Distributions were again not
normal (as confirmed by the Shapiro–Wilk test), with long tails to-
wards high soil loss values. The control and geogrid (B) plots showed
the highest dispersion, while geogrid and, especially, coir and jute,
had a lower variability (Table 5).
On the 45° slope, differences in total soil loss amounts were very
important with a maximum value of 384 g m−2 in the control plot,
followed by geogrid (B) (142 g m−2), jute (120 g m−2), coir
(40 g m−2) and geogrid (30 g m−2). However, differences in terms
of median values were not so high, particularly between the control

Fig. 4. Runoff rates from different treatments versus those from control plots. No. of ob-
servations is 16 for treatments on the 45° slope and 6 on the 60° slope. The 1:1 line is Fig. 5. Boxplots showing the distribution of runoff reduction effectiveness (RRE)
plotted as a reference. obtained for the different treatments on both 45° and 60° slopes.
174 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178

Fig. 6. Accumulated soil loss measurement series from treatments on both 45° and 60° slopes.

(3.6 g m−2), jute (3.2 g m−2)and geogrid (B) (2.1 g m−2) treat-
ments. Surface-laid geogrid and coir gave clearly lower soil loss
rates (0.6 g m−2 and 1.5 g m−2, respectively) than the control, and
their differences were statistically significant at p b 0.05.
The results of the 60° slope do not show coherence between
total and median soil loss values. Although the total soil loss was
clearly larger in the control (87 g m− 2) than in all the treatments
(jute ~ 35 g m− 2, coir ~ 39 g m− 2, geogrid ~ 20 g m− 2, geogrid (B)
~ 34 g m− 2), median values showed a different picture with coir,
geogrid (B) and jute having higher median values (6.7 g m− 2,
6.4 g m− 2 and 6.3 g m− 2, respectively) than the control (4.7 g m−2).
The geogrid was the only treatment with a median soil loss below
the control (2.7 g m− 2). However, the large variability of the con-
trol (IQR ~ 6.4 g m− 2) and the small size of the sample analysed
(n = 6 events) did not permit the detection of significant differ-
ences (at p b 0.05) between treatments and the control.
Fig. 7 shows that, for most events, the control plot had higher
erosion rates than any treatment on the 45° slope (points below
the 1:1 line), with geogrid and coir showing the lowest erosion
rates. Differences between treatments and control were largest for
high soil loss rates (e.g. for the event with highest soil loss rates
control ~ 300 g m− 2, jute and geogrid (B) ~ 80 g m− 2 and coir and
geogrid ~ 10 g m− 2). On the 60° slope these differences are not
that clear, and all treatments, except for the geogrid, had, in some
points, higher soil loss rates than the control, especially when soil
loss rates were low.
The effectiveness values reported in Table 5 confirm these find-
ings. The highest SLRE values were obtained for geogrid and coir
treatments at 45° (median SLRE of 76% and 61%, respectively) and
geogrid at 60° (median SLRE of 53%). At 45° all SLRE values were pos-
itive, representing the beneficial effect, in general terms, of
geotextiles in protecting the slope. But SLRE values clearly differed

Table 5
Soil loss data collected for different treatments (no. of observations on 45° slope = 16, on 60° slope = 6). IQR stands for inter quartile range and SLRE stands for soil loss reduction
effectiveness.

45° slope 60° slope

Jute Coir⁎ Geogrid⁎ Geogrid (B) Control Jute Coir Geogrid Geogrid (B) Control

Total soil loss (g m−2) 120.34 39.81 29.75 141.96 384.44 34.77 38.69 19.72 33.68 87.30
Range (g m−2) 0.45–67.87 0.48–9.68 0.22–10.82 0.44–88.25 0.58–291.93 2.29–9.1 3.22–9.14 0.79–8.13 2.31–8.3 0.85–65.4
Median (g m−2) 3.17 1.49 0.63 2.15 3.58 6.31 6.75 2.69 6.43 4.67
IQR (g m−2) 2.15 2.18 1.91 3.37 4.01 4.03 3.72 0.73 3.96 6.42
Overall SLRE (%) 68.70 89.64 92.26 63.07 – 60.17 55.68 77.41 61.42 –
Median SLRE (%) 20.28 61.51 76.11 41.19 – −19.49 8.34 53.15 −37.95 –
⁎ Significantly different than control at p b 0.05.
 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178
Fig. 7. Soil loss rates from different treatments versus those from control plots. No. of
observations is 16 for treatments on the 45° slope and 6 on the 60° slope. The 1:1
line is plotted as a reference.
between treatments, depending on whether overall or median SLRE
was considered (Table 5). Median SLRE values were more conserva-
tive (i.e. lower) than total SLRE, which was more sensitive to the in-
fluence of extreme events (particularly on the control plots). From a
statistical viewpoint, the use of median SLRE seems to be more rigor-
ous. On the 60° slope, efficiencies were positive when total SLRE was
considered (with values above 55%). But when median SLRE was
used, geogrid and coir were the only treatments achieving positive ef-
fectiveness values (median SLRE of 53% and 8%, respectively). Again,
the small size of the dataset analysed did not permit us to extract
very sound conclusions on the 60° slope treatment.
Boxplots in Fig. 8 illustrate a higher variability of SLRE on the
60° slope. All treatments, except for the geogrid, had a very high
dispersion, with boxplot limits (25% and 75% percentiles) ranging
from b − 50% to > 50%. The geogrid had positive SLRE values of be-
tween 50% and 80%. On the 45° slope, variability was lower and
most of the treatments usually had positive SLRE values. Again,
geogrid had the lowest variability, and the highest SLRE values,
followed by coir, geogrid (B) and jute.
3.3. Correlation analysis
In order to correctly interpret the results, correlations between mea-
sured runoff and soil loss rates and several rainfall variables were ex-
plored. The correlation analysis shows Pearson correlation coefficients
and slopes of the linear model fitted between selected precipitation var-
iables and runoff and soil loss, respectively (Tables 6 and 7). The precip-
itation variables analysed were: rainfall depth (Pac), duration (tR),
intensity (I30) and erosivity (EI30). These were calculated for all run-
off/soil loss measurements in two ways: (1) considering only the
most significant event in the period prior to a measurement, and (2)
in accumulated terms, considering all precipitations recorded in the pe-
riod prior to a measurement. In addition, the wetness conditions of the
175
Fig. 8. Boxplots showing the distribution of soil loss reduction effectiveness (SLRE)
obtained for the different treatments on both 45° and 60° slopes.
slope on each measurement time interval were evaluated using mean
API7.
Runoff was strongly correlated with precipitation depth and dura-
tion of the main rainfall event (Table 6). Accumulated precipitation in
the period was also positively correlated with runoff, as was API7. The
slope between runoff and Pac was steeper in the biological geotextile
treatments (coir and jute). This can be interpreted as a less stable per-
formance of these materials in terms of runoff production, with rela-
tively higher runoff rates expected as precipitation increased. At 60°,
results were similar, but significant correlations were also observed
between runoff and rainfall intensity and erosivity. Thus, it seems
that as the slope gradient increases, the infiltration capacity is more
controlled by storm intensity.
Soil loss rates were positively correlated with rainfall intensity and
erosivity, and no clear associations were observed between rainfall
amounts and soil loss rates. The wetness (API7) and runoff volumes
of each event (not shown) did not have a clear relation to observed
soil loss rates. Therefore, the hypothesis here might be that erosion
was mostly caused by processes which were not runoff driven. In
other words, concentrated flow erosion (rills) and interrill erosion
do not seem to be the main factors responsible for the soil loss
observed in this experiment. In fact, rainsplash as well as gravity-
driven soil loss processes might account for most of the soil loss
observed. The relation between rainfall intensity and soil loss was
stronger in the control and geogrid (B) treatment (at 45°), whereas
geogrid and coir (materials with a larger area coverage) showed a
weaker relationship, probably indicating decreased rainsplash im-
pacts (Rickson, 2006). The slopes of the linear regressions obtained
were higher in the control and geogrid (B), indicating lower protec-
tion as precipitation intensity increased.
4. Discussion
4.1. Influence of geotextiles on runoff production
The placement of geotextiles on bare soils is generally expected
to decrease runoff volumes (Sutherland, 1998a). However, the re-
sults of this experiment show a negative effectiveness of coir and
jute geotextiles in reducing runoff when compared to a vegetated
hydroseeded control plot. Similar findings have been reported in
176 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178

Table 6
Correlation analysis between runoff and precipitation characteristics for each treatment (no. of observations on 45° slope = 16, on 60° slope = 6). The first number indicates the
slope of the linear regression model fitted and the second the Pearson's correlation coefficient.

Runoff 45° slope 60° slope

Jute Coir Geogrid Geogrid (B) Control Jute Coir Geogrid Geogrid (B) Control

Main event Pac 0.18/0.62⁎ 0.23/0.73⁎⁎ 0.1/0.84⁎⁎⁎ 0.1/0.89⁎⁎⁎ 0.12/0.9⁎⁎⁎ 0.17/0.89⁎ 0.48/0.92⁎ 0.16/0.89⁎ 0.11/0.52 0.71/0.74
tR 0.26/0.6⁎ 0.24/0.52⁎ 0.11/0.61⁎ 0.1/0.6⁎ 0.13/0.68⁎⁎ 0.16/0.62 0.54/0.6 0.15/0.51 0.12/0.34 0.75/0.43
I30 −0.03/−0.04 0.05/0.08 0.00/0.00 0.01/0.06 −0.02/−0.06 0.3/0.25 0.33/0.13 0.18/0.41 0.25/0.48 0.24/0.55
R 0.02/0.18 0.03/0.33 0.01/0.27 0.01/0.34 0.01/0.24 0.07/0.62 0.12/0.52 0.05/0.75 0.05/0.7 0.14/0.83⁎
Period ΣPac 0.03/0.2 0.06/0.32 0.04/0.64⁎⁎ 0.05/0.72⁎⁎ 0.05/0.67⁎⁎ 0.09/0.87⁎ 0.29/0.93⁎⁎ 0.1/0.84⁎ 0.06/0.45 0.44/0.65
ΣtR 0.08/0.3 0.1/0.34 0.07/0.62⁎⁎ 0.07/0.69⁎⁎ 0.08/0.68⁎⁎ 0.13/0.76 0.45/0.79 0.14/0.68 0.09/0.37 0.68/0.52
ΣI30 0/0.01 0.04/0.11 0.02/0.15 0.03/0.23 0.02/0.11 0.66/0.76 1.29/0.67 0.48/0.83⁎ 0.57/0.85⁎ 1.63/0.94⁎⁎
ΣR 0.02/0.17 0.03/0.32 0.01/0.3 0.01/0.38 0.01/0.27 0.07/0.7 0.14/0.61 0.05/0.82⁎ 0.06/0.74 0.17/0.88⁎
API7 0.58/0.54⁎ 0.69/0.6⁎ 0.38/0.88⁎⁎⁎ 0.33/0.8⁎⁎⁎ 0.4/0.87⁎⁎⁎ 0.99/0.95⁎⁎ 2.51/0.95⁎⁎ 0.88/0.98⁎⁎⁎ 0.81/0.77 3.58/0.9⁎
⁎ Significant at p b 0.05.
⁎⁎ Significant at p b 0.01.
⁎⁎⁎ Significant at p b 0.001.

the literature. For instance, in their review, Bhattacharyya et al.
(2010) mentioned at least 17 studies in which geotextile cover
(either biological or synthetic) was found to be ineffective in reduc-
ing runoff. Laboratory overland flow simulations also demonstrated
that geotextiles (either biological or synthetic) had little effect on
reducing runoff volume (Rickson, 2006). In particular, geotextiles
with dense covers (similar to the coir blanket evaluated here)
were found to be especially ineffective in reducing runoff (Davies
et al., 2006; Mitchell et al., 2003). Similarly, Bhattacharyya et al.
(2011a) observed that plots completely covered with Borassus
palm-mats produced higher runoff rates than palm-mats placed in
buffer strips (area coverage 10%), and even higher than those of
bare control plots. Permanent grassed plots had least total runoff
and most runoff reduction effectiveness. In this experiment, jute,
and especially coir plots, only achieved reduced vegetation develop-
ment, whereas a very dense vegetation cover developed on the control
plots (Alvarez-Mozos et al., submitted for publication). Therefore, vege-
tation cover appears as a more effective runoff reduction measure than
biological geotextile emplacement. Vegetation elements store pre-
cipitation in leaves and stems, increase surface roughness, reduce
soil moisture through transpiration and increase aggregate stability,
thus promoting infiltration (Fullen and Booth, 2006).
The influence of geotextiles on runoff generation is also related to
slope gradient. Smets et al. (2007) found that palm-leaf geotextiles
were more effective in reducing runoff coefficients on a medium
slope (15°) compared to a steep slope (45°). Similarly, Ogbobe et al.
(1998) observed that runoff velocities increased linearly with in-
creasing slope gradients from 15° to 45°. On gentle or moderate
slopes, biological geotextiles might absorb rainfall water and slow
runoff generation, whereas on steep slopes water can slip through
the geotextile fibres and create superficial flow paths without infil-
trating into the soil. The contact between geotextile and soil has
been found to decrease as the slope gradient increases (Chen et al.,
2011). This might further limit infiltration of rain water into the soil
and promote runoff flow downslope through the geotextile fibres.
Experiments analysing the effectiveness of geotextiles have fre-
quently been carried out on plots tilled to create an erodible environ-
ment, leading to soils with low bulk densities (Bhattacharyya et al.,
2011a; Rickson, 2006; Sutherland, 1998b; Sutherland and Ziegler,
2007). On the contrary, in this study, the experimental slope was
aimed at reproducing real engineered slopes and thus, soils were
strongly compacted. In these soils, infiltration rates are low and
geotextiles might not be as effective in reducing runoff as in loose
soils. Therefore, vegetation might play a relatively more important
role in promoting infiltration.
Both geogrid treatments (surface laid and buried) resulted in sim-
ilar runoff rates to those of the control (i.e. hydroseeded vegetation)
plots. When surface laid, the geogrid should retard surface runoff
due to its transverse 3-D yarns and increased thickness, which are
properties directly correlated with the hydraulic roughness of the
slope (Rickson, 2006; Smets and Poesen, 2009). On the other hand,
when buried, the loose upper soil layer should provide a higher infil-
tration rate. However, those mechanisms seem to be minor compared
to the infiltration-promoting role of vegetation. Therefore, for these
steep slopes, the geogrid's role in promoting vegetation development
appears as a much more effective runoff reduction mechanism than
its physical properties.
4.2. Influence of geotextiles on soil loss
At 45°, all geotextiles produced mostly lower soil loss rates than
the control. But the protection mechanism seems to be different on
the one hand in coir and jute geotextiles, and on the other hand in
the geogrid. As mentioned above, both biological products resulted

Table 7
Correlation analysis between soil loss and precipitation characteristics for each treatment (no. of observations on 45° slope = 16, on 60° slope = 6). The first number indicates the
slope of the linear regression model fitted and the second the Pearson's correlation coefficient.

Soil loss 45° slope 60° slope

Jute Coir Geogrid Geogrid (B) Control Jute Coir Geogrid Geogrid (B) Control

Main event Pac 0.14/0.14 0.02/0.14 0.01/0.07 0.11/0.08 0.45/0.1 −0.06/0.15 0.02/−0.26 −0.01/−0.09 0.05/0.5 −0.03/−0.06
tR −0.1/−0.07 −0.01/−0.06 −0.08/−0.32 −0.44/−0.24 −1.2/−0.19 −0.57/0.46 0.09/0.06 −0.06/−0.34 0.09/0.52 0.01/−0.33
I30 0.54/0.27 0.11/0.36 0.22/0.63⁎⁎ 1.13/0.42 3.17/0.36 2.42/0.26 0.07/0.42 0.24/0.95⁎ 0.13/0.5 0.1/0.95⁎
R 0.09/0.28 0.02/0.34 0.03/0.61⁎ 0.17/0.41 0.49/0.36 0.32/0.3 0.01/0.23 0.03/0.81 0.03/0.66 0.01/0.84⁎⁎
Period ΣPac 0.25/0.47 0.02/0.23 0.03/0.28 0.3/0.42 1.1/0.47 −0.13/0.01 0/−0.43 −0.02/−0.25 0.02/0.35 −0.03/−0.2
ΣtR 0.24/0.27 0.01/0.07 0/−0.01 0.19/0.17 0.86/0.22 −0.43/0.22 0.03/−0.24 −0.05/−0.39 0.05/0.41 −0.03/−0.36
ΣI30 0.85/0.7⁎⁎ 0.11/0.63⁎⁎ 0.16/0.79⁎⁎⁎ 1.23/0.78⁎⁎⁎ 3.97/0.75⁎⁎ 2.57/0.39 0.14/0.16 0.22/0.67 0.25/0.74 0.05/0.77
ΣR 0.13/0.44 0.02/0.45 0.03/0.7⁎⁎ 0.21/0.56⁎ 0.64/0.51⁎ 0.3/0.32 0.01/0.19 0.03/0.75 0.03/0.69 0.01/0.79
API7 −0.43/−0.12 −0.15/−0.29 −0.07/−0.11 −0.46/−0.1 −1.22/−0.08 1.65/0.17 0.09/−0.26 0.14/0.27 0.3/0.57 −0.13/0.32
⁎ Significant at p b 0.05.
⁎⁎ Significant at p b 0.01.
⁎⁎⁎ Significant at p b 0.001.
 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178
in higher runoff rates than the control. But the sediment concentra-
tions measured in these treatments were lower than the control.
This supports the hypothesis that runoff mostly flowed over the fi-
bres of these biological geotextiles without reaching the soil
(Bhattacharyya et al., 2011b).
Dense geotextiles, such as the coir blanket evaluated, have been
found to be very successful in reducing rainsplash erosion (Rickson,
2006). But, as slope inclination increases, surface runoff might play
a more important role in soil erosion than rainsplash. On steep slopes,
runoff has been found to scour under the mats (particularly when
mats have high coverage ratios) flushing soil out (Chen et al., 2011).
In this experiment, no evidence of concentrated flow erosion was no-
ticed under the coir blankets at the end of the experiment. At 45°,
some soil loss was measured in this treatment, but it was generally
much lower than that of the control. However, at 60° erosion was
higher than in the control plot in several points, suggesting runoff
soil detachment below the blanket.
The jute plots produced higher erosion rates than the coir mats.
The open area fraction of the jute net was higher (Table 2) and, con-
sequently, its protection against rainsplash and runoff was expected
to be lower (Rickson, 2006). Similar findings were reported by Chen
et al. (2011), who found a 28% erosion reduction efficiency for jute
plots at 45°. However, jute geotextiles were successful in reducing
erosion on moderate slopes (Mitchell et al., 2003). Their effectiveness
has been normally attributed to their great drapability (adherence to
the soil when wet). But jute fibres degrade quite rapidly and, when
placed on steep slopes, their reduced tensile strength and lack of
transverse structures can cause a lower erosion reduction capacity.
The effectiveness of jute decreased as rainfall persisted and became
more intense (Chen et al., 2011; Sutherland, 1998a). Biological prod-
ucts with random fibres (similar to the coir blanket) have normally
been found to be more effective than open weave materials (similar
to the jute net) (Sutherland and Ziegler, 2007).
The effectiveness of geotextiles in reducing soil erosion decreased
as the slope gradient increased (Chen et al., 2011; Smets et al., 2007).
On steep slopes the structural properties of the material are crucial
for adequate erosion protection (Chen et al., 2011). The geogrid eval-
uated in this experiment has 3D transverse structures and a higher
tensile strength than the biological products tested. These properties
might be more important than the percentage area cover for the re-
duction of erosion under such steep slopes. Therefore, the geogrid
seems to be more suitable than the coir and jute products for
protecting steep slopes. The placement of the geogrid is vital; when
buried, the soil loss rates observed were much larger than when
surface-laid. Similar findings have been reported (Rickson, 2006;
Sutherland, 1998a). Even if manufacturers frequently recommend
buried placement, the resulting unconsolidated superficial soil is ex-
tremely sensitive to rainsplash and runoff erosion processes. In this
study, total soil loss was lower in the buried geogrid than in the con-
trol, but when median values were compared no real differences
were evident (Table 5).
5. Conclusions
The performance of several geotextiles (coir blanket, jute net and
a synthetic geogrid) were evaluated on steep slopes (45° and 60°) in
terms of runoff and soil loss. Coir and jute geotextiles resulted in in-
creased runoff volumes (2–3 times larger) compared to the control
(i.e. hydroseeded vegetated) plots. This produced negative effective-
ness values (RRE b − 100%), suggesting that rain may scour through
the fibres of these materials without infiltrating into the soil. Instead,
geogrid treatments produced very similar runoff volumes to the con-
trol (with median runoff values of ~ 3 mm). The well-developed veg-
etation cover in the control plot appears as a key factor in promoting
infiltration and reducing runoff.
177
Regarding soil loss, the best performance was achieved by the
geogrid, which produced very low and consistent erosion rates, followed
by the coir (median soil loss rates ~0.6 g m−2 and 1.5 g m−2, respec-
tively, on the 45° slope). The low erosion rates of the geogrid can be at-
tributed to its 3D transverse structures, higher tensile strength and the
well-developed vegetation canopy. However, when buried, the geogrid
did not effectively control erosion (median soil loss rate ~2.1 g m−2 on
the 45° slope), at least that of the upper soil layer. Thus, its surface instal-
lation seems to be a better option. In cases where the buried installation
is preferred for aesthetic reasons, a higher soil loss must be assumed.
References
Alvarez-Mozos, J., Abad, E., Goñi, M., Giménez, R., Campo, M.A., Díez, J., Casalí, J., Arive,
M., Diego, I., 2013. Evaluation of erosion control geotextiles on steep slopes. Part
2: influence on the establishment of vegetation. Catena (submitted for
publication).
Assouline, S., Ben-Hur, A., 2006. Effects of rainfall intensity and slope gradient on the
dynamics of interrill erosion during soil surface sealing. Catena 66, 211–220.
Bhattacharyya, R., Fullen, M.A., Davies, K., Booth, C.A., 2009. Utilizing palm-leaf
geotextile mats to conserve loamy sand soil in the United Kingdom. Agriculture,
Ecosystems & Environment 130, 50–58.
Bhattacharyya, R., Smets, T., Fullen, M.A., Poesen, J., Booth, C.A., 2010. Effectiveness of
geotextiles in reducing runoff and soil loss: a synthesis. Catena 81, 184–195.
Bhattacharyya, R., Fullen, M.A., Booth, C.A., 2011a. Using palm-mat geotextiles on an ar-
able soil for water erosion control in the UK. Earth Surface Processes and Land-
forms 36, 933–945.
Bhattacharyya, R., Fullen, M.A., Booth, C.A., Kertesz, A., Toth, A., Szalai, Z., Jakab, G.,
Kozma, K., Jankauskas, B., Jankauskasiene, G., Buhmann, C., Paterson, G.,
Mulibana, E., Nell, J.P., van der Merwe, G.M.E., Guerra, A.J.T., Mendonca, J.K.S.,
Guerra, T.T., Sathler, R., Bezerra, J.F.R., Peres, S.M., Yi, Z., Yongmei, L., Li, T.,
Panomtarachichigul, A., Peukrai, S., Thu, D.C., Cuong, T.H., Toan, T.T., 2011b. Effec-
tiveness of biological geotextiles for soil and water conservation in different
agro-environments. Land Degradation and Development 22, 495–504.
Cerdá, A., 2007. Soil water erosion on road embankments in eastern Spain. Science of
the Total Environment 378, 151–155.
Chen, S.C., Chang, K.T., Wang, S.H., Lin, J.Y., 2011. The efficiency of artificial materials
used for erosion control on steep slopes. Earth Sciences and Environment 62,
197–206.
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 Land-
forms 31, 561–569.
Fullen, M.A., Booth, C.A., 2006. Grass ley set-aside and soil organic matter dynamics
on sandy soils in Shropshire, UK. Earth Surface Processes and Landforms 31,
570–578.
Fullen, M.A., Subedi, M., Booth, C.A., Sarsby, R.W., Davies, K., Bhattacharyya, R., Kugan,
R., Luckhurst, D.A., Chan, K., Black, A.W., Townrow, D., James, T., Poesen, J., Smets,
T., Kertesz, A., Toth, A., Szalai, Z., Jakab, G., Jankauskas, B., Jankauskiene, G.,
Buhmann, C., Paterson, G., Mulibana, E., Nell, J.P., van der Merwe, G.M.E., Guerra,
A.J.T., Mendonca, J.K.S., Guerra, T.T., Sathler, R., Bezerra, J.F.R., Peres, S.M., Yi, Z., Li,
Y.M., Li, T., Panomtaranichagul, M., Peukrai, S., Thu, D.C., Cuong, T.H., Toan, T.T.,
Jonsyn-Ellis, F., Sylva, J.Z., Cole, A., Mulholland, B., Dearlove, M., Corkill, C.,
Tomlinson, P., 2011. Utilising biological geotextiles: introduction to the Borassus
Project and global perspectives. Land Degradation and Development 22,
453–462.
Heggen, R.J., 2001. Normalized antecedent precipitation index. Journal of Hydrologic
Engineering 6, 377–381.
Lekha, K.R., 2004. Field instrumentation and monitoring of soil erosion in coir geotextile
stabilised slopes — a case study. Geotextiles and Geomembranes 22, 399–413.
Li, M.H., Khanna, S., 2008. Aging of rolled erosion control products for channel erosion
control. Geosynthetics International 15, 224–231.
Mitchell, D.J., Barton, A.P., Fullen, M.A., Hocking, T.J., Zhi, W.B., Yi, Z., 2003. Field studies
of the effects of jute geotextiles on runoff and erosion in Shropshire, UK. Soil Use
and Management 19, 182–184.
Morgan, R.P.C., 2005. Soil Erosion and Conservation, Third ed. Blackwell Publishing, Ox-
ford (304 pp.).
Ogbobe, O., Essien, K.S., Adebayo, A., 1998. A study of biodegradable geotextiles used
for erosion control. Geosynthetics International 5, 545–553.
Owens, P.N., Batalla, R.J., Collins, A.J., Gomez, B., Hicks, D.M., Horowitz, A.J., Kondolf,
G.M., Marden, M., Page, M.J., Peacock, D.H., Petticrew, E.L., Salomons, W.,
Trustrum, N.A., 2005. Fine-grained sediment in river systems: environmental sig-
nificance and management issues. River Research and Applications 21, 693–717.
Rickson, R.J., 1995. Simulated vegetation and geotextiles. In: Morgan, R.P.C., Rickson,
R.J. (Eds.), Slope Stabilization and Erosion Control: A Bioengineering Approach.
Taylor & Francis, Abingdon, UK.
Rickson, R.J., 2006. Controlling sediment at source: an evaluation of erosion control
geotextiles. Earth Surface Processes and Landforms 31, 550–560.
Smets, T., Poesen, J., 2009. Impacts of soil tilth on the effectiveness of biological geotextiles
in reducing runoff and interrill erosion. Soil and Tillage Research 103, 356–363.
Smets, T., Poesen, J., Fullen, M.A., Booth, C.A., 2007. Effectiveness of palm and simulated
geotextiles in reducing run-off and inter-rill erosion on medium and steep slopes.
Soil Use and Management 23, 306–316.
178 J. Álvarez-Mozos et al. / Catena 118 (2014) 168–178
Smets, T., Poesen, J., Bhattacharyya, R., Fullen, M.A., Subedi, M., Booth, C.A., Kertesz, A.,
Szalai, Z., Toth, A., Jankauskas, B., Jankauskiene, G., Guerra, A.J.T., Bezerra, J.F.R., Yi,
Z., Panomtaranichagul, M., Buhmann, C., Paterson, G., 2011. Evaluation of biological
geotextiles for reducing runoff and soil loss under various environmental condi-
tions using laboratory and field data. Land Degradation and Development 22,
480–494.
Sutherland, R.A., 1998a. Rolled erosion control systems for hillslope surface protection:
a critical review, synthesis and analysis of available data. I. Background and forma-
tive years. Land Degradation and 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 and Development 9, 487–511.
Sutherland, R.A., Ziegler, A.D., 2007. Effectiveness of coir-based rolled erosion control
systems in reducing sediment transport from hillslopes. Applied Geography 27,
150–164.
Ziegler, A.D., Sutherland, R.A., Tran, L.T., 1997. Influence of rolled erosion control sys-
tems on temporal rainsplash response — a laboratory rainfall simulation experi-
ment. Land Degradation and Development 8, 139–157.