-
Fs
FDU
a
ARRAA
KPRS
1
btarcwa1f
0h
Ecological Engineering 61P (2013) 633 645
Contents lists available at ScienceDirect
Ecological Engineering
journa l h om epage: www.elsev ier .com/ locate /eco leng
orest protection and protection forest: Tree root degradation
over hydrologicalhallow landslides triggering
ederico Preti
ipartimento di Economia, Ingegneria, Scienze e Tecnologie
Agrarie e Forestali, Sezione di Ingegneria dei Biosistemi Agrari e
Forestali,niversit di Firenze, via S. Bonaventura 13, 50145
Firenze, Italy
r t i c l e i n f o
rticle history:eceived 21 July 2012eceived in revised form 19
October 2012ccepted 12 November 2012vailable online 20 December
2012
eywords:rotection forestoot degradationhallow landslides
triggering
a b s t r a c t
The potential use of protection forests to combat shallow slope
instabilities is becoming increasinglyimportant and considerable,
especially in the light of the recent landslides and debris/mud
flows in regionstriggered by rainfalls with increased intensity.
Tree vegetation has been constantly subjected to silvicul-tural
activity both in exclusively productive forest areas and in more
conservative ones meant to contrasthydrogeological risk. It is
important to quantify the root system dynamics in order to
correctly evaluatethe impact of wood felling or plants death on
slope stability. Based on field investigation (on experimentalplots
and 29 occurred landslides) and numerical modelling (on slope
stability and root distribution), theaim of this work is to
determine the effects of the evolution of the mechanical
characteristics of rootsystems (and consequently on landslide
probability). The paper investigates variations over time in
thehazard of rainfall-triggered landslides as a result of root
degradation after forest cutting (or death). Thecase under study is
related to experimental investigations aimed at determining the
tensile strength andelasticity of root samples of trees dead within
a decade, which correspond to decreasing values of soilcohesion
(root reinforcement). Two kinds of samples were taken into account:
living beech roots fromprotected wood areas to determine the
current characteristics and roots from dead beeches (felled
inprevious years and at present in degradation) to analyse the
evolution of root mechanical characteristics.To analyse the
stability of representative slopes, we calculated the return time
associated with a rainfallevent, which in saturated conditions
would lead to the attainment of the limit value of the safety
factor
and the associated hazard for different rainfall durations
during a fixed period of time. Information aboutthe increasing risk
of collapse with the degradation of root system was obtained and
compared withlandslides occurrence in forested slopes of the study
area. The results of the present paper show thatsuch slopes may
remain stable if they are covered with intact protective
vegetation, but they will becomeunstable if the conditions of the
forest deteriorate or after a wooded area dies off: within a decade
of tree
rotec
ssfrfaeaT
death the root system of p
. Introduction
Forest cover reduces shallow landslide hazard by influencingoth
hydrological and geo-mechanical factors which contributeo slope
stability. From a hydrological perspective, forest coverffects the
soil moisture regime as it increases both transpirationates during
interstorm periods and the evaporation rates byanopy interception,
as well as it enhances the formation ofell drained soil surface
horizons (OLoughlin, 1974; Waldron
nd Dakessian, 1981; Ziemer, 1981; Watson and OLoughlin,985;
Alila et al., 2009; Preti et al., 2011). As a consequence,
therequency of occurrence of high soil water potentials in
shallow
Tel.: +39 3209223758.E-mail address: [email protected]
stspl22
925-8574/$ see front matter 2012 Elsevier B.V. All rights
reserved.ttp://dx.doi.org/10.1016/j.ecoleng.2012.11.009tion forests
loses most of its soil-stabilising function. 2012 Elsevier B.V. All
rights reserved.
oils is significantly reduced, with a favourable effect to the
slopetability during rainstorms. From a geo-mechanical
perspective,orest cover also reinforces the soils explored and
bounded by itsoot system, which improves the slope stability
independentlyrom the actual soil water content (e.g. Selby, 1993;
Nilaweerand Nutalaya, 1999; Abernethy and Rutherfurd, 2001;
Schmidtt al., 2001; Simon and Collison, 2002; Frei et al., 2003;
Graynd Barker, 2004; Fournier et al., 2006; Reubens et al.,
2007).his geo-mechanical effect is generally parameterised in
slopetability models by an additional apparent cohesion provided
byhe root system to the soil. However, the geometry of the
rootystem is the results of eco-hydrological processes, i.e. of the
soil
lant atmosphere interactions, which in turn are influenced
byocal climatic regimes and soil hydraulic properties (Laio et
al.,006; De Beats et al., 2008; Preti et al., 2010; Giadrossich et
al.,012).
dx.doi.org/10.1016/j.ecoleng.2012.11.009http://www.sciencedirect.com/science/journal/09258574http://www.elsevier.com/locate/ecolengmailto:[email protected]/10.1016/j.ecoleng.2012.11.009
-
6 neerin
ssitars2aplarrhalos
emWeodPbMa(mrddmrpl2o
amdptoda
maA
iTtltFset
el
2
mpTclibbtmtav
uswpjofttflbftlapiott
3
3
bOao2lbams
1ico
34 F. Preti / Ecological Engi
It is important to consider the entire living cycle of the
forestystem when one has to analyse the role of the forest cover
onlope stability. Particularly from a forest management
perspective,t is important to assess the effect of the death of the
forest cover onhe slope stability, as it can occur after a sudden
perturbation, suchs tree cut, wildfire and plant disease attack.
After tree death, theeduction of overload due to biomass reduction
generally is notignificant for the slope stability (e.g. Preti,
2006; Schwarz et al.,010a,b). The favourable effects on the soil
water potential regimere lost immediately (e.g. Selby, 1993). The
roots is subjected to arogressive decomposition, eventually causing
gaps in the inter-ocking root system of neighbouring individual
trees (Burroughsnd Thomas, 1977), which inevitably implies a
reduction of theoot tensile strength and of the apparent cohesion
provided by theoot system to the soil, with a progressive increase
of the landslideazard over time (Ziemer, 1981; Sidle et al., 2005;
Sidle, 1992; Sidlend Ochiai, 2006; Ammann et al., 2009). This
scenario is particu-arly critical in places where the timing of the
natural renovationf the forest, with a fully developed root system,
starting from newpruces and pioneer flora, encompass several
hydrological years.
Detailed studies of rotting processes in tree roots and
theirffects on the strength of the roots have so far been carried
outainly in North America, New Zealand and Asia (OLoughlin
andatson, 1979; Ziemer, 1981; Ekanayake et al., 1997; Watson
t al., 1997, 1999). All these studies have been motivated by
thebservation of increased occurrence of shallow landslides
aftereforestation. In Alaska, roots of Tsuga heterophylla [Raf.]
Sarg. eicea sitchensis [Bong.] Carr showed a reduction in root
strengthy a 2025% only three years after the trees death (Ziemer,
1981).oreover, the frequency of landslides increased by 3.8 times
after
large-scale decline of Chamaecyparis nootkatensis [D.Don]
SpachJohnson and Wilcock, 2002). Little information is available on
theonthly degradation rate of root tensile strength of a decaying
dead
oot (Schmidt et al., 2001). Ammann et al. (2009) examined
theecrease in root tensile strength of decomposing stumps, after
theying-off or cutting-down of red firs (Picea abies Karst.) in
Swissountain forests. Tensile strength and elasticity was measured
on
oot segments sampled 8 to 12 years after tree death and com-ared
with living ones, although these were taken from living treesocated
at altitudes different from the dead ones (Ammann et al.,009).
Genet et al. (2006, 2008) have shown the effect of altituden the
characteristics and resistance of the root system.The reinforcement
provided by the roots to the soil can be
ssessed in various ways (Sidle and Ochiai, 2006): (1)
laboratoryeasurements of tensile or cutting strength on single
roots withifferent diameters; (2) laboratory direct shear tests on
soil sam-les with roots; (3) in situ measurements using cut boxes
appliedo the soil horizon explored by roots; (4) laboratory
measurementsf cutting strength opposed by a roots column; (5)
uprooting evi-ences of stumps or plants; (6) back analysis on
collapsed slopesfter storm events.The last approach implies the
application of a hydrological
odel for predicting the spatial distribution of the soil
moisturefter the storm event (e.g. Borga et al., 1998; Chirico et
al., 2003a,b;ndriola et al., 2009).In this study we examine the
strength reduction of roots and
ts influence on the slope stability of a beech forest in
Northernuscany (Italy). The study is structured as follows. We
first presenthe experimental plots in the study area and the
recently occurredandslides. Then we illustrate the model employed
for predictinghe root system geometry from local climatic and
pedological data.
ollowing we illustrate the root sample design and laboratory
mea-urements of tensile strength on sample roots. Finally a model
forstimating the strength decrease of a cut or senescent root
overime and its effect on slope stability when combined with
rainfall
fmrp
g 61P (2013) 633 645
vents of different scale is presented and discussed, by
examiningandslide events occurred in the study area.
. Study sites
The study area is the Northern Toscana (Fig. 1a). The
experi-ental plots are located within the Serchio River watershed,
in theortion of the Apennine range separating Emilia Romagna
fromoscana in Central Italy (Casone di Profecchia Fig. 1b). The
geologi-al setting consists of sandstones layers averagely inward
dipping,ocally covered by debris subjected to runoff erosion. The
bedrocks sub-outcropping. The top soil consists in a transient high
forest ofeech, showing standards of relevant size in its upper
portion. Theeech basal area per hectare G is equal to 33 m2/ha in
the inves-igated area. The data examined in this study have been
collectedainly in experimental plots located in the Province of
Lucca, in
he Castiglione Garfagnana district, at an altitude between
1290nd 1533 m a.s.l., with an average slope of 1525% and peak
slopealues up to 4550% (Fig. 1c).In December 2009, one month after
our field investigation, an
nusual hydrological event occurred in northern Tuscany. A
wide-pread and heavy snowfall occurred on December 18th and
19th,ith air temperature significantly under the average value of
theeriod. Following, around Christmas time a heavy rainfall
occurred,oined with a sudden temperature increase. The rainfall
height wasf 10-years return time, but the total available water at
the sur-ace was much higher, as the sudden temperature increase
meltedhe snow accumulated in the previous week. The combination
ofhese two events is infrequent in central Italy and caused
relevantoods, especially along the Serchio river, and over 800
slope insta-ilities, such as shallow land-slides, debris-flows and
localized treealling in the northern Apennines of Tuscany. Fig. 1d
illustrateshe locations of 29 investigated landslides (red
triangles). Particu-arly, Crespole (near Pescia in Fig. 1d)
landslide has very similar soilnd slope characteristics of the
representative one for the presentaper, at an average altitude of
690 m a.s.l., along a north-west fac-ng slope, with an average
slope angle of s = 36.80. The slope wasriginally covered by trees
which have been clear cut 9 years priorhe landslide occurred and
during field surveys we could verify thathe coppices were highly
degraded.
. Materials and methods
.1. Data collection
Beech (Fagus sylvatica L.) sample roots have been collected
fromoth living and dead beech plants, felled in 2008, 2006 and
2004.ther sample roots dead in 2002 have been collected in
nearbyreas at same elevations, assuming that they are
representativef roots with similar characteristics, following Genet
et al. (2006,008). Other samples have been taken in the upper,
middle and theower part of the natural beech range of the Serchio
Valley. Livingeech roots have been taken from areas at elevations
of 800, 1450nd 1600 m a.s.l., corresponding to the minimum,
intermediate andaximum limits of the elevation distribution of
beechwoods in the
tudy region.The beech roots have been carefully dug out to a
depth of
.0 m by hand and they were identified morphologically follow-ng
Kutschera and Lichtenegger (2002). Provided that beech rootsan
expand over large extents, even larger that the coppice spacing,ne
can be easily get confused in the identification of the coppice
rom which the sampled root is originated. In order to avoid
anyistake, we traced each collected sample back to one of the
main
oots which can be clearly attributed to the corresponding
cop-ice. The root samples have been collected with a minimum
length
-
F. Preti / Ecological Engineering 61P (2013) 633 645 635
Fig. 1. Study area localization (a and b) and example of beech
vegetation on slope (c); 29 other landslide sites (red triangles
(d)). (For interpretation of the references to colori
obtadw
vwdtDdtlata
T
u
M
wiDeMd2
M
wda
n this figure caption, the reader is referred to the web version
of the article.)
f 4 cm and a diameter ranging from 1.9 to 9.5 mm, excluding
theark. The roots were kept in closed bags before the tests to
maintainheir initial moisture content. A total of 276 samples was
collectednd tested. Root water content has been measured with the
ovenrying method and has been expressed as the ratio of the mass
ofater to the mass of the solid material.For tensile strength
measurements, we used the the Amsler uni-
ersal testing machine, with maximum load set at 40 kN,
equippedith a pressure transducer for loads recording and a
potentiometricisplacement transducer for elongation measurements,
connectedo a computer for data acquisition (Preti and Giadrossich,
2009).eformation measurements were performed by fixing the
trans-ucer directly on the roots with two light clamps adapted to
griphe root. The results of the root tensile tests were used to
calcu-ate the tensile strength at the moment of maximal tensile
forcepplied. Tensile strength Tr (N/mm2) is calculated as the ratio
ofhe maximum tension force F (N) to the initial root
cross-sectional
rea A0 (mm2) (DIN 52188, 1979 in Preti and Giadrossich,
2009).
r = FA0
(1)
r
fm
We calculated the Modulus of Elasticity (MoE or Youngs mod-lus)
expressed as follows:
oE =
= FD0A0x
(2)
here and are, respectively, the stress and strain of the root,
Fs the pullout force difference in the near-linear elastic range
(N),0 is the initial distance between the two clamps (mm), x is
thelongation in the near-linear elastic range (mm). The
experimentaloE data values can be expressed by a non-linear
function of theiameter d as follows (Operstein and Frydman, 2000;
Fan and Su,008; Schwarz et al., 2010a,b):
oE = cd1 (3)here c is a parameter depending on species and the
degree ofegradation. Root tortuosity may affect the apparent MoE
because
tortuous root may stretch with reduced stress transmission to
the
oot tissue (Commandeur and Pyles, 1991; Schwarz et al.,
2010a,b).
The following geotechnical properties have been measuredrom soil
sample measurements at Casone di Profecchia experi-ental plots:
angle of internal friction = 30; saturated specific
-
6 neerin
ws
tc(pCa(E
GsG
3
qtslFpcsvib
wt
ls
c
wsaseiote
t
imos
R
ws
f
c
wcc
c
c
dte
efededbtf
R
b
dfg
b
woeea
o
r
T
wia
mvse
36 F. Preti / Ecological Engi
eight sat = 2.2 kN/m3; submerged specific weight = 1.2 kN/m3;oil
porosity = 0.35, water content at field capacity = 0.28.
We characterized the 29 landslide slope mechanical parame-ers
too, according to the field surveys, as follows: soil cohesions = 0
kPa (coefficient of variation C.V. = 0), friction angle = 29.96
C.V. = 0.192), specific weight of saturated soil sat = 2.2
kN/m3, soilorosity = 0.35, slope angle 36.80 (C.V. = 0.218), root
cohesionv at 1 mt of depth = 3.52 kPa estimated as in Eq. (8) (C.V.
= 1.159),verage rooting depth b = 0.83 m (C.V. = 0.958) estimated
as in Eq.10), unit tensile strenght of roots Tr(1) = Tr1 = 41 MPa
(according toq. (11)).The measured tree basal area at Casone di
Profecchia site is
= 33 m2/ha, while the average value at the abandoned
landslideites is G = 17 m2/ha and at Crespole site (near the
landslide scarp)
= 10 m2/ha.
.2. Root reinforcement
For the slope stability analysis of vegetated slopes, we have
touantify root reinforcement, as increase of shear strength or
addi-ional soil cohesion Cv. Additional soil cohesion due to root
tensiletrength can be calculated using different models on
different rootoadings and breakage assumptions (Ji et al., 2012).
Particularlyiber or Root Bundle Models could be implemented if all
necessaryarameters are available (Schwarz et al., 2010a, 2010b).
The modelhoice was dictated by the need to keep the numerical model
asimple as possible, with limited number of parameters. A
modifiedersion of the well-known Wu (1976) and Waldron (1977)
theoret-cal model (Preti, 2006; Schwarz et al., 2010a,b; Ji et al.,
2012) haseen here used considering the MohrCoulomb failure
criterion as:
sr = cv + cs + tan (4)here cv is the additional cohesion due to
the presence of roots, cs
he bare soil cohesion, the normal stress on the shear plane, and
the bare soil friction angle.When soil mass movement occurs, such
as translational shallow
andslides, shear forces are developed and roots crossing the
slipurface are mobilized in tension.
Resulting additional cohesion cv is thus expressed as:
v = kktR (5)here tR is the mobilized root tensile strength per
unit area ofoil and k is a correction factor used to account for
the roots thatre not oriented perpendicular to the slipping
surface, usually con-idered equal to 1.2 (ranging from 1.0 to 1.3,
Waldron, 1977; Wut al., 1979). The parameter k = 0.4 is an
empirical correction factorntroduced by Preti (2006) in order to
correct the bias due to theverestimation of the cohesion values
with WM. Ji et al., 2012 foundhat this simply model revision
provided the most conservative cvstimates as compared with other
six models.The mobilized root tensile strength
R = Tr RAR (6)s defined as the product of the mean tensile
strength of roots (Tr)ultiplied by the root area ratio (RAR), i.e.
the relative surface ratiof the soil profile occupied by roots. In
our case, for each sampleite, RAR is calculated using the following
equation:
AR =nArj (7)j=1A
here n is the number of root sections observed in a vertical
cross-ectional area A, Arj is the cross-sectional area of j-th
root.
t
ht
g 61P (2013) 633 645
In order to account for the variability of the root strength
asunction of the root size, the last equation is rewritten as
follows:
v = kkn
j=1Trj
ArjA
(8a)
here N is the number of diameter classes, j the current
diameterlass, Trj and Arj the mean root tensile strength and the
mean rootross-sectional area respectively in the class j.
For technical purposes, root cohesion at z soil depth can be
cal-ulated as follows:
v(z) = kkTrRAR(z) (8b)
The root area ratio can be also computed as function of the
soilepth, RAR(z), by computing the ratio of the sum of root cross
sec-ions at given depth z (Ar(z)) and the total vertical extent of
soilxplored by the root system A = Ars.The variation of the root
density Ar(z) over depth can be mod-
lized by a negative exponential function with two parametersor
the applications of technical interest: Ar0 is the area of
rootsxtrapolated at the initial depth and b is the average rooting
depthepending on hydrological and pedological characteristics
(Pretit al., 2010). The scaling factor Ar0 is specie-specific and
plant-age-ependent; in fact, the basal stump area and the
above-groundiomass are different from specie to specie, increasing
over timeo a mature state. As a consequence, RAR(z) can be assessed
by theollowing formula:
AR(z) = Ar0Ars
e(1/b)z = RAR0e(1/b)z (9)
RAR0 is the root area ratio at the surface and can be estimatedy
the tree basal area (RAR0 = G/10.000, Preti et al., 2010).The
parameter b represents the average rooting depth and it is
erived from the long-period soil water balance in the
followingorm (in water controlled eco-systems and during the
vegetativerowing season; Preti and Giadrossich, 2009; Preti et al.,
2010):
= (fc w)(1 0/Tp)
(10)
here 0 is the mean frequency of rain events (n. events/day)n
growing vegetative phase; is the mean intensity of rainfallvents
(mm/event) on growing vegetative phase; Tp is the
potentialvapotranspiration ratio (mm/day) on growing vegetative
phase;
is the effective soil porosity; fc and w is the moisture
contentst field capacity and the wilting point, respectively.The
calculated b values are in good agreement with all the
bserved root profiles.The expected value of the mean root
tensile strength for a given
oot diameter can be expressed as follows
r = Tr1d (11)
here Tr1 is the tensile strength for diameter of unit length and
ds the root diameter. The parameters Tr1 and can be obtained by
linear regression model after logarithmic transformation.In 3
plants sampled at 800, 1450 and 1645 m a.s.l., Tr1 is esti-
ated equal to 42.6 N/m2 while is estimated equal to 0.51.
Thesealues are comparable with other authors values for the
samepecies (e.g. Tr = 41, 57d0,98 at 1100 m a.s.l. in Bischetti et
al., 2005,ven if in a different diameter range of thinner roots
1.78 1.19 mm
han the ones tested in the present study).
Based on the field surveys, in the study area the following
dataave been found and they have been used for the model
simula-ions:
-
F. Preti / Ecological Engineerin
0
0,5
1
1,5
2
2,5
0 1 3 5 7
degradation years
Ten
sile s
tren
gth
rati
o T
r/T
r2009
Fig. 2. Box-plot of tensile strength ratio [Y-axis, Tr values
divided by Tr(2009)]ofq
-
-
4
4
b
T
wT
dttTtTd
F3
ver degradation time [X-axis: 0 = 2009, 1 = 2008, 3 = 2006, 5 =
2004 and 7 = 2002]or all samples. The plotted values are: minimum
and maximum, first quartile, thirduartile, median. Regression line
Tr(DY)/Tr(2009) = 0.1168 DY + 1.0322 R2 = 0.9958.
root cohesion at 0 depth cv(0) = 178.8 kPa as maximum value
cor-responding to the tree basal area of the slope under
examinationG = 33 m2/ha as in the experimental site, neglecting the
reduc-tion coefficient k in Eqs. (5) and (8) and considering Tr =
45 kPa
(according to literature data).
root cohesion at 0 depth cv(0) = 28.1 kPa, corresponding to
youngforest case with tree basal area G = 5.2 m2/ha, considering
onlystandard plants and/or newly-sprouting spruces and pioneer
ave
(a)
(b)
ig. 3. (a) Tr (d) regression power lines for different
degradations years (Eq. (11)). (b) Unit = 2006, 5 = 2004 and 7 =
2002]: decay rate = 9.1%; annulment (intercept of X-axis) after 1g
61P (2013) 633 645 637
flora and the same hypotesis as above for k and Tr or,
alter-natively, to a tree basal area G = 17.5 m2/ha (representative
ofthe investigated landslide areas) with conservative k = 0.4 andTr
= 33.4 kPa (average value of measured data).
. Results and discussion
.1. Root reinforcement decay after tree death
The decay of root tensile strength over time can be representedy
the following linear function:
rDY = Tr0[1 (DY DY0)] (12)here Tr0 is the root tensile strength
at the reference year DY0 whilerDY = Tr(DY) is the tensile strenght
at year DY, after the root death.In Fig. 2 the Box-plot of the
relative tensile strength by
ividing for the 2009 value over time Tr(DY)/Tr(2009)
(degrada-ion years DY) is presented, showing the expected reduction
inensile strenght with respect to the reference year DY0 = 2009.he
plotted values are: minimum and maximum, first quartile,hird
quartile, median. The median values have regression
liner(DY)/Tr(2009) = 0.1168 DY + 1.0322 with R2 = 0.9958. The
yearlyecreasing rate is 11% and the total decay time is ca. 9
years.
The regression power curves Tr(d) (Eq. (11)) for several
years
ppear to have similar trends: by analysing their
log-transformedalues (Fig. 3a), we found that the angular
coefficients are almostqual, resulting in the parallelism of the
curves, and that the Y-axis
tensile strength (d = 1 mm) [Mpa] over degradation time.
[X-axis: 0= 2009, 1 = 2008,1 years R2 = 0.9957.
-
638 F. Preti / Ecological Engineering 61P (2013) 633 645
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
0 2 4 6 8 10Degradation years
Te
ns
ile
str
en
gth
ra
tio
diameter 4 mm
New Zeland Beech
(Sidle & Ochiai, 2006)
y=-0.0962x+1
R2=0.9435
y=-0.1214x+1
R2=0.9713
y=-0.0969x+1
R2=0.9055
F 9) oni r studo
itd9(a
d(
d4smauoctsa
s(tlcW
M
ca2o(2m(l
SlooabdlRewaidad3
is attributed to variations in tissue density and cellulose
contentand are affected by branching order (Genet et al., 2005;
Wang andYen, 1974; Schwarz et al., 2010a,b). In our results, both
Tr (Fig. 2)
0
0,5
1
1,5
2
2,5
3
0 1 3 5 7Degradation time [year]
Ela
sti
cit
y r
ati
o (
Mo
E/M
oE
2009) ig. 4. Tensile strength ratio [Y-axis, Tr values divided
by the maximum or the Tr(200n samples grouped by a diameter
threshold (< and >4 mm) and comparison with othef X-axis)
after 10.4 and 8.3 years, respectively.
ntercepts Tr1 (root resistance for d = 1 mm as the interpolation
inhe power regression curves Tr(d) for each degradation year)
areecreasing over time (Fig. 3b): an annual linear decrease of
about.1% can be observed with a complete annulment after ca. 11
yearsregression equation Tr(d) = 4.4528 DY + 49.066 characterized
by
very high R2 = 0.9957).Eq. (11) and Eq. (12) imply that roots of
different diameters
egrade in a similar way, consistently with findings by
Ziemer1981) and Ammann et al. (2009).
Results in Fig. 4 show that the dimensionless tensile
strengthecreases after cutting also for diameters ranging between
1.9 and
mm (almost exactly as in Sidle and Ochiai, 2006 for the
samepecies) and for those between 4 and 9.5 mm (being 4 mm theedian
diameter), reaching an almost null value after more or less
decade with a decreasing rate of ca. 10% and significant R2
val-es (0.900.97). These results are in agreement with other
authorsnes too as e.g. Ziemer (1981) and to Ammann et al. (2009),
even ifomparisons have to be made paying attention due to the fact
thatensile strength is affected by differences in the selection of
rootamples (altitudes, diameter and curvature), root moisture
contentnd testing methods.Thus, if we assume that RAR is constant
over time, the apparent
oil cohesion will also decrease linearly over time, according to
Eqs.8) and (11). The hypothesis that RAR is constant over time
implieshat the hillslope should be affected by external events,
such asandslides, which may affect the root mass distribution as
naturalonsequence of forest adaptation to disturbing forces
(Johnson andilcock, 2002; Saklas and Sidle, 2004; Preti et al.,
2010)A similar analysis can be done also for MoE:
oEDY = MoE0[1 (DY DY0)] (13)
Knowledge of the elasticity is important for quantifying
pro-esses inducing mechanical activation of the rootsoil
interfacend its shear strength (Mickovski et al., 2007, 2008;
Schwarz et al.,010a,b), but there is a lack of information
concerning MoE valuesr the slope of stressstrain relationship
during root tensile testsCommandeur and Pyles, 1991; Operstein and
Frydman, 2000; Tosi,
007; Fan and Su, 2008; Ammann et al., 2009). In Fig. 5 all
elasticityeasurements have been analysed over the diameter using
Eq. (2b)
the coefficient c assuming value equal to 893 in our study case
foriving beech, while e.g. is equal to 696 for different plant
species as
FMatD
e] over degradation time [X-axis: 0 = 2009, 1 = 2008, 3 = 2006,
5 = 2004 and 7 = 2002]ies (Sidle and Ochiai, 2006): degradation
rate = ca. 10 and 12%; annulment (intercept
esbania cannabina, Medicago sativa, Rosmarinus officinalis,
Pistaciaentiscus and Cistus in Schwarz et al., 2010a,b). In Fig. 5
the Box-plotf MoE measurements is presented: a roughly 4% linear
decreasef the median values can be observed, with extrapolated null
valuefter ca. 25 years (R2 = 0.7). The unit MoE (interpolation for
d = 1 mmy the hyperbolic regression curves MoE(d) in Eq. (3) for
each degra-ation year) over degradation time is shown in Fig. 6
with an annualinear decrease of about 4.8% with annulment after
20.9 years and2 = 0.81. (Apparent) MoE could be lower if we also
consider rootlongation of tortuous roots (Schwarz et al., 2010a,b):
in our testse avoided this problem measuring directly on
stretched-out roots,fter by excluding very tortuous roots. For each
year of degradation,n Fig. 7 the decreasing of average MoE values
is analysed for twoata-sets, grouped by a 4 mm threshold (median
diameter value):
lower starting value for larger root diametres (64%) and a
smallerecreasing rate (ca. 3% and 6%, with intercept of X-axis
equal to ca.3 and 18 years, respectively), contrarily to the Tr
case (Fig. 4).The dependency of stressstrain relationship on root
diameterig. 5. Box-plot of Module of Elasticity ratio [Y-axis, MoE
values divided byoE(2009)] over degradation time [X-axis: 0 = 2009,
1 = 2008, 3 = 2006, 5 = 2004nd 7 = 2002]. The values are minimum
and maximum, first quartile, third quar-ile, median. Regression
line MoE(DY)/MoE(2009) = 0.0304 DY + 0.769 R2 = 0.7075.egradation
rate = 4%; annulment (intercept of X-axis) after 25 years.
-
F. Preti / Ecological Engineering 61P (2013) 633 645 639
y = -46,649x + 976,05
R2 = 0,8105
0
200
400
600
800
1000
1200
0 1 2 3 4 5 6 7 8
Degradation time [year]
Ela
sti
cit
y [
Mp
a]
Fig. 6. Unit Module of Elasticity (d = 1 mm) over degradation
time [X-axis: 0 = 2009,1c
aasadpjtcttataobi
mtwpscw8tis
F0ta
0
20
40
60
80
100
120
140
160
0 1 5 7degradation years
mo
istu
re c
on
ten
t [%
]
F1r
4
taP
F
wdcs
(t
Gesi
z
ww
= 2008, 3 = 2006, 5 = 2004 and 7 = 2002]. Degradation rate =
4.8%; annulment (inter-ept of X-axis) after 20.9 years R2 =
0.8105.
nd MoE (Fig. 5) decrease with increasing root degradation
time,nd consequently the examined samples of old plants have
lowertrength and stiffness properties than of younger plants. A
drynd decayed root should be very fragile and so undergo
scarceeformation until it reaches a quick break, with extremely
lowlastic field. In our case, as the roots were kept in closed
bagsust after sampling and before testing, they should have
conservedheir original moisture content. It has to be noticed that
MoE isonsidered not dependent on the root state of decay accordingo
other studies, but they were carried out after short degrada-ion
times and for other species (OLoughlin and Watson, 1979) ort
different altitudes and at very low, not constant and far fromhe
actual soil moisture conditions (Ammann et al., 2009). Actu-lly
there might be a combination effect of the moisture contentf the
root itself: degradation time being equal, the MoE shoulde
inversely proportional to moisture, at least within
determinedntervals.
A root from a dead plant for a few years has a slightly
loweroisture content (Fig. 8) and so there would be various
fac-
ors at work: root degradation lowers the MoE (stiffness),
butater loss could increase it. If there was a combination of
com-ensating factors (moisture + degradation) it might be possible
toee with a multiple regression the weight of each componentompared
to the MoE, but the moisture content of our samplesas almost always
higher than 50% (average values higher than0%): due to these high
values, an influence of the moisture on
he Tr and MoE was not expected. Possibly effects could ver-fy
only at lower values (under 3040%), if achievable into theoil.
y = -5,7428 x + 189 ,54R = 0,251 5
y = -16,125x + 288,21
R = 0,8029
0
50
100
150
200
250
300
350
0 1 2 3 4 5 6 7 8Degradation time [year]
Ela
sti
cit
y [
Mp
a]
diameter 4 mm
ig. 7. Average Module of Elasticity [MPa] in Y-axis over
degradation time [X-axis: = 2009, 1 = 2008, 3 = 2006, 5 = 2004 and
7 = 2002] in samples grouped by a diameterhreshold (< and >4
mm): degradation rate = 6 and 3%; annulment (intercept of X-xis)
after 18 and 33 years, respectively.
btenfdc
h
wt
2gT
weoH
ig. 8. Box-plot of moisture content (Y-axis) over degradation
time [X-axis: 0 = 2009, = 2008, 5 = 2004 and 7 = 2002] for all
samples (2006 dry weight not available);egression line for average
values gives y = 2.6872x + 105.71 R2 = 0.8868.
.2. Slope stability analysis
The failure condition is examined using the MohrCoulomb
cri-erion within an infinite slope model accounting for the
apparentdditional cohesion provided by the root system (e.g. Sidle,
1992;reti and Giadrossich, 2009):
s = cv + z cos2 S tan
satz sin S cos S(14)
here S is the slope angle; z is the breaking or failure
surfaceepth (m); is the soil friction angle; is the submerged
spe-ific weight (kN/m3) = sat w (specific water weight); sat is
thepecific saturated weight (kN/m3).
Eq. (14) is written in the hypothesis that the soil cohesion is
nullas in the study case) and the vegetation weight is not relevant
forhe slope stability (e.g. Preti, 2006).
In the hypothesis of vertical infiltration according to
areenAmpt approximation with a piston like front (which is thexact
solution for soil with diffusivity equal to a Diract function
ataturation) and no infiltration excess, the depth of saturation
frontncreases according to the relation
(t) = h(t)
S i
(15)
here i and S are the initial and saturated soil water
content,hile h(t) is the rainfall height after time t.Other
solutions have been proposed for assessing the slope sta-
ility factor during a storm event and particularly interesting
arehose solutions based on a linear approximation of the
Richardsquation (Iverson, 2000; DOdorico et al., 2005), even if
consideredot necessary for the purposes of the present paper.
Rainfall heightor a given duration t can be usefully expressed by
rectangularesign hyetograph according to local intensity-duration
frequencyurve:
(t) = a(rT )tn (16a)
here n is a constant parameter and a(rT) is a scaling factor,
func-ion of the return period rT.
In the present study we considered the equation h =1.985r0.167T
t
0.537 corresponding to rainfall data recorded at theauge closest
to the experimental plots (Preti et al., 1996; Regioneoscana,
2007).To analyse the effect on the stability of a representative
slope,e calculated the return time rT associated with a single
rainfallvent, which in saturated conditions would lead to the
attainmentf the limit value of the safety factor Fs and the
associated hazardna for different rainfall durations during a fixed
period of time.
-
640 F. Preti / Ecological Engineering 61P (2013) 633 645
0,00
0,20
0,40
0,60
0,80
1,00
1,20
1,40
1,60
1,80
0 1 2 3 4 5 6 7 8 9 10
Degradation years
Lan
dslid
e d
ep
th [
m]
Cv(0) = 28,08 kPa
Cv(0) = 0 kPa
Cv(0) = 178,8 kPa
Cv(0) = 54 kPa G = 10 mq/ha
December 2010 landslide measured scarp depth (rainfall 316
mm)
Fig. 9. Landslide failure depth associated to root decay over
root degradation time in agreement with the values registered in
Preti et al. (2001), Casagli et al. (2006), Schwarzet al. (2010a,b)
and recently measured in the landslides occurred in the area under
examination. Slope characteristics: soil cohesion c = 0 kPa, soil
angle of internal friction = 30 , saturated specific weight = 2. 2
kN/m3, submerged specific weight = 1.2 kN/m3 (porosity = 0.35),
slope angle = 32; root cohesion Cv up to 0 m of depth,r ges or d to
c
c
H
Hl
d
ccrCfti3mFddi
csvFod
cF((
(dooCa(
F(
sat
espectively = 0 kPa; 28.08 kPa and 178.8 kPa. In the case of
felling, the effect of chaned point represents the measured
landslide scarp depth at Crespole slope subjecte
The corresponding hazard in a fixed period (range time)
wasalculated as:
na = 1 (1 1
rT
)na(16b)
na is the probability of occurrence of an event with a return
periodarger than or equal to rT in a period of na years.
Information about the increasing risk of collapse with the
degra-ation of root system was obtained.Figs. 9 and 10 show the
value of the critical z depth and
orresponding rainfall for several years of root
degradationonsidering three representative conditions of root
reinforcement,espectively: Cv(0) = 0 kPa that means complete
absence of roots;v(0) = 28.08 kPa that would give at 1 m depth
Cv(1) = 1 kPa (felledorest and/or newly sprouting spruces and
pioneer flora case withree basal area G = 5.2 m2/ha) and Cv(0) =
178.8 kPa correspond-ng to the tree basal area of the slope under
examination G of3 m2/ha and to a value of Cv(1) = 6.38 kPa (in
agreement with oureasurements and the back analysis in Schwarz et
al., 2010a,b).
igs. 11 and 12 show the behaviour of the return time over
rootegradation time and the behaviour of the hazard over root
degra-ation time for different durations (Eqs. (14)(16)). For
examplen Fig. 11a the value of the return time of a 48-hours
rainfall is
r2
s
0
100
200
300
400
500
600
0 1 2 3 4 Degradat
Rain
fall lan
dlisd
e t
rig
geri
ng
[m
m]
Cv(0) = 28 ,08 kPaCv(0) = 0 kPa
Cv(0) = 178 ,8 kPa
ig. 10. Critical rainfall triggering landslides over root
degradation time in agreement w2010a,b) and in recent field
surveys.s
f above-ground biomass overload was conservatively considered
not relevant. Thelear cut just 9 years before.
ritical for slope instability (triggering threshold for shallow
land-lides) and decreases in value from 100 years (in the case of
healthyegetation) to 10 years in about a decade of degradation,
while inig. 11b a 48-h rainfall has a return time 100 years (in the
casef healthy vegetation) and
-
F. Preti / Ecological Engineering 61P (2013) 633 645 641
11
00
10
00
01
00
00
00 0 1 2 3 4 5 6 7 8 9 10
Degradation years
Re
turn
tim
e
1 hour 3 hours 6 hours12 hours 24 hours 2 days
11
00
10
00
01
00
00
00 0 1 2 3 4 5 6 7 8 9 10
Degradation years
Degradation years
Re
turn
tim
e
1 hour 3 hours 6 hours12 hours 24 hours 2 days
1
10
100
1000
10000
0 1 2 3 4 5 6 7 8
retu
rn t
ime
Qua rtil e I
Qua rtil e II
Qua rtil e II I
Qua rtil e I
Qua rtil e II
Qua rtil e II I
(a)
(b)
(c)
Fig. 11. (a) Return time (years) of landslide triggering
rainfall threshold associated to root decay over root degradation
time. Slope characteristics in Fig. 9 caption with rootcohesion
Cv(0) = 28.08 kPa and rainfall-duration equation h =
21.985rT0.167t0.537 with h = rainfall amount (mm), rT = return time
(years), t = rainfall duration (h). (b) Return time(years) of
landslide triggering rainfall threshold associated to root decay
over root degradation time. Slope characteristics in Fig. 9 caption
with root cohesion Cv(0) = 178.8 kPaand rainfall-duration equation
h = 21.985r0.167
Tt0.537 with h = rainfall amount (mm), rT = return time (years),
t = rainfall duration (h). (c) Return time (years) associated to FS
= 1,
concerning the slope characteristics in Fig. 9 caption for the
values of the median root cohesion and of the first and third
quartiles of the experimental T data (tensile strenghtl es)
rait
aiJtlRt
daf
ike in Fig. 2) per each degradation year with 24 h (blue curves)
and 48 h (red curvhe reader is referred to the web version of the
article.)
severe number of landslides per unit area were triggered offn
abandoned forest areas (Johnson and Wilcock, 2002). On 19thune 1996
the Versilia-Garfagnana area of the Apuanian Alps (in
he same Province of Lucca, Fig. 1) has been affected after
out-ier rainstorms (a 24-h total of 400 mm of rain was observed
atetignano gauge). 15 people died and great damages were
regis-ered. The majority were soil slips, but a fraction developed
into
tem2
rr
nfall-duration. (For interpretation of the references to color
in this figure caption,
ebris flows. The area is characterized by shallow soils (0.351.1
ms in Fig. 9) that develop on acid bedrock. The texture rangesrom
sandyclayloam to loamy-sand. The slopes were originally
erraced for chestnut orchards that have been abandoned at thend
of the 19th century and that have subsequently turned intoixed
broad leaves woods with little undergrowth (Preti et al.,001).
Another severe rainstorm of high intensity occurred on
-
642 F. Preti / Ecological Engineering 61P (2013) 633 645
0,00
0,01
0,10
1,00
0 1 2 3 4 5 6 7 8 9 10Degradation years
Ha
za
rd
1 hour 3 hours 6 hours12 hou rs 24 hou rs 2 da ys
0,00
0,01
0,10
1,00
0 1 2 3 4 5 6 7 8 9 10Degradation years
Ha
za
rd
1 hour 3 hours 6 hours12 hou rs 24 hou rs 2 da ys
(a)
(b)
Fig. 12. (a) Hazard of landslide associated to root decay over
time for different rainfall durations in a time range na = 100
years. Slope characteristics in Fig. 11 captionw 0.167t0.537
H s in a tC ll am
1assdSe2foe1wf1Ioofwooa
eotTSa
sosstriasa
ith root cohesion Cv(0) = 28.08 kPa and rainfall-duration
equation h = 21.985rT
azard of landslide associated to root decay over time for
different rainfall durationv(0) = 178.8 kPa and rainfall-duration
equation h = 21.985rT0.167t0.537 with h = rainfa
9th21st November 2000, in the same Province of Lucca (Fig. 1)nd
in the close-by Province of Pistoia which triggered tens of
land-lides. These landslides can be broadly defined as complex
earthlidesearth flows, originating as rotational slides that
developownslope into a flow (Casagli et al., 2006; Schwarz et al.,
2010a,b).torms are fairly frequent in this area, but cyclonic
storms ofxtreme intensity such as the one in November 20th and
21st,000, are rare. This storm precipitated around 200220 mm of
rain-all within a period of 3840 h. Rainfall intensity hit a
maximumf 17 mm/h, near Montecatini Terme and Pescia, and it has
beenstimated that the event has a return time period of more than00
years. Intensive antecedent rainfall was also recorded over 3eeks
prior to the event giving a total of 545 mm of precipitation
or November 2000, which is the highest quantity recorded
since970, exceeding the monthly average by 328% (Casagli et al.,
2006)n a little catchment near the village of Vinchiana in the
Provincef Lucca (Fig. 1), a number of shallow landslides have
occurred, onef them resulting in 2 human casualties during a
moderate rain-all event that followed on three week period of
prolonged rainfall,
ith a cumulative rainfall of 360 mm having a return time periodf
more than 100 years. This landslide affected a small portionf the
slope (area 1000 m2) between 175 and 260 m a.s.l. withn angle of 38
and a soil thickness of less than 1 m (Schwarz
ca
c
with h = rainfall amount (mm), rT = return time (years), t =
rainfall duration (h). (b)ime range na = 100 years. Slope
characteristics in Fig. 11 caption with root cohesionount (mm), rT
= return time (years), t = rainfall duration (h).
t al., 2010a), as in Fig. 9. The vegetation cover consisted
mainlyf chestnut trees (Castanea sativa Mill.), with presence of
locustrees (Robinia pseudoacacia L.) and cluster-pines (Pinus
Pinaster A.).he mobilized sediments reached the main stream as a
debris flow.imilar scenarios were observed for other shallow
landslides in thisrea.Important factors for evaluating the risk of
landslides or ero-
ion after tree death are of course not only the time spanf root
decomposition, but also the slope steepness and theoil material
susceptibility to landslides. Slopes with a gradientteeper than the
angle of internal friction of the soil in ques-ion are more
susceptible to landslides as a consequence of theeduced protective
function of the vegetation. Actually we foundn the case under
examination the slope angle s = 32 > the soilngle of internal
friction = 30 and in Schwarz et al. (2010a)s = 3538 > = 33.4 (in
Casagli et al., 2006, = 29.935), in oururveys in Versilia (Preti et
al., 2001) and in the recently affectedreas s = 3136 > =
24.530).The proposed methodology here presented has been tested
byalculating the landslide depth for all the 29 investigated
slopesnd obtaining the results shown in Fig. 13 (Montgomery,
1994).Considering Crespole landslide, we could assess the total
root
ohesion equal to 3.52 kPa (C.V. = 1.159) up to 1 m of depth. At
the
-
F. Preti / Ecological Engineerin
y = 0,96x
R2 = 0,63
0
0,5
1
1,5
2
2,5
3
3,5
4
0 1 2 3 4measured landslide scarp depth [m]
sim
ula
ted
la
nd
sli
de
de
pth
[m
]
Fig. 13. Comparison between the measured and calculated scarp
depth obtainedwith the proposed method and considering the effect
of vegetation. The stabilityanalyses of these slopes have been
carried out with the following soil mechan-ical parameters observed
in the field: soil cohesion c = 0 kPa (C.V. = 0), frictionangle =
29.96 (C.V. = 0.192), specific weight of saturated soil sat = 2.2
kN/m3,sdr
i1soocetistb
i
Ftstcsidru
mtii
ic(ssttisate
rmclnteilRep
gaistlis
oil porosity = 0.35, slope angle 36.80 (C.V. = 0.218), root
cohesion Cv at 1 mepth = 3.52 kPa (C.V. = 1.159), average rooting
depth b = 0.83 m (C.V. = 0.958), unitoot tensile strength Tr(1) =
41 MPa).
nitial condition (live plants), we would obtain by the model
ca..6 m failure depth, while the observed failure depth of the
Cre-pole landslide scarp is 0.9 m (Fig. 9). This depth can be
predictednly if we assume that root strength is reduced precisely
to 10%f the original value, due to 9 years of root degradation
after treeut. We find similar results even if we account for the
root lat-ral cohesion (Schwarz et al., 2010a,b). It is interesting
to observehat the root lateral cohesion is not significant for the
slope stabil-ty in this case as far as the root strength is above
90% of the roottrength when plants were alive. Significant
differences between
he FS values computed with and without the lateral cohesion cane
observed only for a root strength reduction factor of 10%.Finally
Fig. 14 shows the effect of all experimental data variabil-
ty on the return time estimation: in the Box-plot the minimum
and
0,00 1
0,01
0,1
1
10
100
1000
10000
0 1 2 3 4 5 6 7 8 9
Degradation years
Re
turn
tim
e
Serie1
Serie2
Serie3
Serie4
Serie5
'
ig. 14. Box-plot (plotted values are: minimum and maximum, first
quartile,hird quartile, median) concerning the 29 investigated
slopes with the cohe-ion degradation as in Fig. 3 (the degradation
years here are 10, as we usedhe experimental degradation llnear raw
applied to a decade), soil cohesion = 0 kPa (C.V. = 0), angle of
internal friction = 29.96 (C.V. = 0.192), saturatedpecific weight
sat = 2.2 kN/m3, submerged specific weight = 1.2 kN/m3 (poros-ty =
0.35), slope angle s = 36.80 (C.V. = 0.218), root cohesion Cv up to
1 mepth = 3.52 kPa (C.V. = 1.159), average rooting depth b = 0.83 m
(C.V. = 0.958), unitoot tensile strenght Tr1 = Tr(1) = 41 MPa); for
all the slopes we considered the sat-ration and FS = 1 conditions.
Logarithmic scale gives distortion in box widths.
wontqestbta(
5
gp
iami
i
g 61P (2013) 633 645 643
aximum, first quartile, third quartile, median values of the
returnime are represented concerning the 29 investigated slopes by
vary-ng all the slope parameters and with the cohesion degradation
asn Fig. 3.
That is, such slopes may remain stable if they are covered
withntact protective vegetation, but they will become unstable if
theonditions of the forest deteriorate or after a wooded area dies
offSelby, 1993; Ammann et al., 2009): in our case from root
cohe-ion Cv(1) = 6.38 kPa to soil cohesion null. The results of
this studyhow that within a decade of tree death the root system of
protec-ion forests loses most of its soil-stabilising function. The
length ofime span depends on the decomposition rate of the roots,
whichn turn is a function of site parameters such as altitude
(climate) oroil water regime and it can be assumed that,
particularly at highltitudes, this period of time is not long
enough for new genera-ions of trees to have grown enough to have
the same stabilisingffect on the soil (Ammann et al., 2009).The
model here used considers, at present, the hydrological
esponse of the soil which is not dependent on the topological
andorphological connections (the slope being the only morphologi-al
parameter employed). Since this problem represents a crucialimit to
the procedure applied, a control is required in order to allowot
using a distributed hydrological approach and consequentlyo
overcome this shortcoming (e.g. Borga et al., 2004; Erminit al.,
2005). The upslope contributing area represents obviously
anmportant hydrological factor selected for this analysis. In
particu-ar, as widely recognized (e.g. Kirkby, 1971; Tarboton et
al., 1992;inaldo et al., 1995), it controls the water flow at a
point, influ-ncing how soil saturates, with particular reference to
landslidingrocesses.Moreover, slope stability analysis in areas
dominated by active
eomorphologic processes (such as soil erosion and landsliding)nd
covered by vegetation, is often impeded by the lack of
reliablendirect methods for the spatial estimation of soil depth.
Actuallyoil thickness (or the potential failure depth) z can vary
as a func-ion of many different and interplaying factors, such as
underlyingithology, climate, gradient, hillslope curvature, upslope
contribut-ng area, and vegetation cover, making the distributed
estimation ofoil depth challenging and often unreliable. While the
relationshipith gradient and curvature should reflect the kinematic
stabilityf the regolith cover, allowing greater soil thicknesses
over pla-ar and concave areas, the distance from the hill crest (or
fromhe valley bottom) accounts for the position within the soil
topose-uence. This last parameter is fundamental: according to
Catanit al. (2010) points having equal gradient and curvature can
haveignificantly different soil thickness due to their dissimilar
posi-ion along the hillslope profile. Finally, slope stability
appears toe especially important particularly where installations
and infras-ructures (such as road networks, railway lines,
electrical lines, etc.)re present and it has to be carefully
considered regarding the latterBorga et al., 2004).
. Conclusions
Steep slopes are often covered by tree and shrub vegetation
thatives an increased stability, technically and legaly recognized
asrotection forests.The potential use of protection forests to
combat shallow slope
nstabilities here investigated is becoming increasingly
importantnd relevant, especially in the light of the recent
landslides and
ud flows in various regions triggered by rainfalls with
increased
ntensity.Thus in the analysis of slope stability it is necessary
to take
nto account that the characteristics of vegetation (above
and
-
6 neerin
boTwpsdirtmJ2fmo
faottftotbaffaasyi
utleCeuPdt
A
AdTNp
R
A
A
A
A
B
B
B
B
C
C
C
C
C
D
D
E
E
F
F
F
F
F
F
F
G
G
G
G
G
I
J
44 F. Preti / Ecological Engi
elow-ground biomass, root density, etc.) do not remain
constantver time, but evolve along with their effects on slope
stability.ree vegetation is also subject to dying (beetle outbreak,
storms,ild-fires, etc.) and cutting operations (forestry
production,lant felling, fire control strips, replacement of
species, etc.). Theoil root reinforcement (soil-bolstering
function) of dead treeseclines as their strength and stiffness
decreases. If the vegetations not replaced soon enough, erosion
processes may be aggravated,esulting in increased weathering and
more water penetrating intohe soil through spaces created by the
decayed roots: this, in turn,ay trigger off shallow landslides
(OLoughlin, 1974; Sidle, 1992;
ohnson and Wilcock, 2002; Reubens et al., 2007; Ammann et
al.,009). Consequently, after a large-scale tree die-off in
mountainorests, a triggering event such as heavy rain or massive
snowelting can be a serious threat to slope stability as in
Tuscanyccurred in the last decades.Based on field investigation and
numerical modelling, a model
or estimating the decrease in cut/senescent root strength over
timend its effect on slope stability when combined with rainfall
eventsf different scale is presented. The goal was the evaluation
of varia-ions over time in the shallow landslide hazard associated
with theriggering rainfall threshold as a result of root
degradation afterorest cutting (or death). The study case is
related to experimen-al investigations to determine the tensile
strength and elasticityf root samples of plants dead within a
decade, which correspondo decreasing values of root cohesion. Tr
and MoE were tested onoth living beech roots and degraded ones and
they decreasedlmost linearly over the number of years after death.
The rain-all return time corresponding in saturated conditions to
safetyactor FS = 1 and the associated hazard for different
durations in
fixed period of time have been calculated, obtaining results in
good agreement with recent inventories at regional scale: soillope
stabilization by plants would enter a critical phase someears after
a plant die-off, caused either naturally or by
humanntervention.
Further study will be performed to assess the role of
parameterncertainty in the overall model prediction as well as to
explorehe possibility of distibuted mapping the relative hazard of
shal-ow landslide over large spatial scales (Wu and Sidle, 1995;
Roeringt al., 2003; Pollen and Simon, 2005; Schwarz et al., 2010a,
2010b;atani et al., 2010), also by using remote sensing data
(Forzierit al., 2009, 2011a,b, 2012), to design soil-bioengineering
meas-res (Petrone and Preti, 2008, 2009; Stokes et al., 2009; Preti
andetrone, 2012), and to investigate the damages to forest cover
itselfue to slope erosion and instability or other related
processess ashe transport ones (e.g. Rossi Pisa et al., 1999;
Ziemer et al., 1991).
cknowledgments
Special thanks to Ergys Alliu (DEISTAF Thesis student),ndrea
Dani (DEISTAF Fellowship funded by Fondazione Cassai Risparmio di
Trento e Rovereto), Marco Togni (DEISTAF Woodechnnology Lab),
Giovan Battista Chirico (DIAAT, Universit diapoli Federico II), and
to DREAM Italia (funding the researchroject Dissesto idro-geologico
e vegetazione).
eferences
bernethy, B., Rutherfurd, A.I., 2001. The distribution and
strength of riparian treeroots in relation to riverbank
reinforcement. Hydrol. Process. 15, 6379.
lila, Y., Kuras, P.K., Schnorbus, M., Hudson, R., 2009. Forests
and floods: a new
paradigm sheds light on age-old controversies. Water Resour.
Res. 45, W08416,http://dx.doi.org/10.1029/2008WR007207.
mmann, M., Bll, A., Rickli, C., Speck, T., Holdenrieder, O.,
2009. Significance of treeroot decomposition for shallow
landslides. For. Snow Landsc. Res. 82 (1), 7994,79.
J
g 61P (2013) 633 645
ndriola, P., Chirico, G.B., De Falco, M., De Crescenzo, P.,
Santo, A., 2009. A compari-son between physically-based models and
a semiquantitative methodology forassessing suceptibility to
flowslides triggering in pyroclastic deposits of south-ern Italy.
Geografia Fisica e Dinamica Quaternaria 32, 213226.
ischetti, G.B., Chiaradia, E.A., Simonato, T., Speziali, B.,
Vitali, B., Vullo, P., Zocco, A.,2005. Root strength and root area
ratio of forest species in Lombardy (NorthernItaly). Plant Soil
278, 1122.
orga, M., Dalla Fontana, G., Da Ros, D., Marchi, L., 1998.
Shallow landslide hazardassessment using a physically based model
and digital elevation data. EnvironGeol 35 (23), 8188.
orga, M., Tonelli, F., Selleroni, J., 2004. A physically based
model of the effects offorest roads on slope stability. Water
Resources Research 40, 12202.
urroughs, E.R., Thomas, B.R., 1977. Declining root strength in
Douglas-Fir afterfelling as a factor in slope stability. Research
Paper INT-190. USDA Forest Service.Ogden. 27 S. DIN 52188, 1979:
Bestimmung der Zugfestigkeit parallel zur Faser.In: DIN Taschenbuch
31. 2000. Berlin, Beuth. 109110.
atani, F., Segoni, S., Falorni, G., 2010. An empirical
geomorphology-based approachto the spatial prediction of soil
thickness at catchment scale. Water Resour. Res.46, W05508,
http://dx.doi.org/10.1029/2008WR007450.
asagli, Dapporto, N.S., Ibsen, M.L., Tofani, V., Vannocci, P.,
2006. Anal-ysis of the landslide triggering mechanism during the
storm of20th21st November 2000 in Northern Tuscany. Landslides 3,
1321,http://dx.doi.org/10.1007/s10346-005-0007-y.
hirico, G.B., Grayson, R.B., Western, A.W., 2003a. A downward
approach to iden-tifying the structure and parameters of a
process-based model for a smallexperimental catchment. Hydrol.
Process. 17 (11), 22392258.
hirico, G.B., Grayson, R.B., Western, A.W., 2003b. On the
computation of the quasi-dynamic wetness index with
multiple-flow-direction algorithms. Water Resour.Res. 39 (5),
1115.
ommandeur, P.R., Pyles, M.R., 1991. Modulus of elasticity and
tensile strength ofDouglas-fir roots. Can. J. For. Res. 21,
4852.
e Beats, S., Poesen, J., Reubens, B., Wemans, K., De
Baerdemaeker, J., Muys, B.,2008. Root tensile strength and root
distribution of typical Mediterranean plantspecies and their
contribution to soil shear strength. Plant Soil 305, 207226.
Odorico, P., Fagherazzi, S., Rigon, R., 2005. Potential for
landsliding:dependence on hyetograph characteristics. J. Geophys.
Res. 110, F01007,http://dx.doi.org/10.1029/2004JF000127.
kanayake, J.C., Marden, M., Watson, A.J., Rowan, D., 1997. Tree
roots and slopestability: a comparison between Pinus radiata and
Kanuka. N. Z. J. For. Sci. 27 (2),216233.
rmini, L., Catani, F., Casagli, N., 2005. Artificial Neural
Network applied to landslidehazard assessment. Geomorphology 66
(14), 327343.
an, C.C., Su, C.F., 2008. Role of roots in the shear strength of
root-reinforced soilswith high moisture content. Ecol Eng 33,
157166.
orzieri, G., Guarnieri, L., Vivoni, E.R., Castelli, F., Preti,
F., 2009. Multiple attributedecision-making for individual tree
detection using high-resolution laser scan-ning. For. Ecol. Manage.
258 (11), 25012510.
orzieri, G., Guarnieri, L., Vivoni, E.R., Castelli, F., Preti,
F., 2011a. Spectral-ALSdata fusion for different roughness
parameterizations of forested floodplains,15351459. River Res.
Appl.
orzieri, G., Degetto, M., Righetti, M., Castelli, F., Preti, F.,
2011b. Satellite multispec-tral data for improved floodplain
roughness modelling. J. Hydrol.
orzieri, G., Castelli, F., Preti, F., 2012. Advances in remote
sensing of hydraulicroughness. Int. J. Remote Sens.
ournier, M., Stokes, A., Coutand, C., Fourcaud, T., Moulia, B.,
2006. Tree biomechanicsand growth strategies in the context of
forest functional ecology. In: Herrel,A., Speck, T. (Eds.), Ecology
and Biomechanics: A Mechanical Approach to theEcology of Animals
and Plants. CRC Taylor & Francis, Boca Raton.
rei, M., Bll, A., Graf, F., Heinimann, H.R., Springmann, S.,
2003. Quantification ofthe influence of vegetation on soil
stability. In: Proceedings of the InternationalConference on Slope
Engineering, Hong Kong, Department of Civil Engineering,The
University of Hong Kong, pp. 872877.
enet, M., Stokes, A., Salin, F., et al., 2005. The influence of
cellu-lose content on tensile strength in tree roots. Plant Soil
278, 19,http://dx.doi.org/10.1007/s11104-005-8768-6.
enet, M., Stokes, A., Fourcaud, T., Li, M., Luo, T., 2006.
Effect of altitude on rootmechanical and chemical properties of
Abies georgei in Tibet. In: Salmen,L. (Ed.), Proceedings 5 Plant
Biomechanics Conference. Sweden, 28 August1September, 2006, pp.
305310.
enet M., Stokes, A., Fourcaud, T., Li, M., Luo, T., 2008.
Differences in roottensile strength and sugar content along an
altitudinal gradient in Tibet,.
iadrossich, F., Schwarz, M., Cohen, D., Preti, F., Or, D., 2012.
Mechanicalinteractions between neighbouring roots during pullout
tests. Plant Soil,http://dx.doi.org/10.1007/s11104-012-1475-1.
ray, D.H., Barker, D., 2004. Rootsoil mechanics and
interactions. Riparian vegeta-tion and fluvial geomorphology. Water
Sci. Appl. 8, 113123.
verson, R.M., 2000. Landslide triggering by rain infiltration.
Water Resour. Res. 36(7), 18971910.
i, J., Kokutse, N.K., Genet, M., Fourcaud, T., Zhang, Z.Q.,
2012. Effect of spatial varia-
tion of tree root characteristics on slope stability. A case
study on Black Locust(Robinia pseudoacacia) and Arborvitae
(Platycladus orientalis) stands on the LoessPlateau, China. Catena
92, 139154.
ohnson, A.C., Wilcock, P., 2002. Association between cedar
decline and hillslope sta-bility in mountainous regions of
southeast Alaska. Geomorphology 46, 129142.
dx.doi.org/10.1029/2008WR007207dx.doi.org/10.1029/2008WR007450dx.doi.org/10.1007/s10346-005-0007-ydx.doi.org/10.1029/2004JF000127dx.doi.org/10.1007/s11104-005-8768-6http://www.meeting.edu.cn/webmedia/jingpin/icgbe2/pic/abstract_book.pdfdx.doi.org/10.1007/s11104-012-1475-1
-
neerin
K
K
L
M
M
M
N
O
O
O
P
P
P
P
P
P
P
P
P
P
R
R
R
R
R
S
S
S
S
S
S
S
S
S
S
T
T
W
W
W
W
W
W
W
W
F. Preti / Ecological Engi
irkby, M.J., 1971. Hillslope process-response models based on
the continuity equa-tion. In: Slopes, Forms and Processes, Inst. of
British Geographers Spec. Publ. No.3 (1971).
utschera, L., Lichtenegger, E., 2002. Wurzelatlas
mitteleuropischer Waldbumeund Strucher. Graz, Stocker. 604 S.
aio, F., DOdorico, P., Ridolfi, L., 2006. An analytical model to
relate the vertical rootdistribution to climate and soil
properties. Geophys. Res. Lett. 33 (18).
ickovski, S.B., Bengough, A.G., Bransby, M.F., Davies, M.C.R.,
Hallet, P.D., Sonnen-berg, R., 2007. Material stiffness, branching
pattern and soil matric potentialaffect the pullout resistance of
model root systems. Eur. J. Soil Sci. 58, 14711481.
ickovski, S.B., Hallett, P.D., Bransby, M.F., Davies, M.C.R.,
Sonnenberg, R., Ben-gough, A.G., 2008. Mechanical reinforcement of
soil by willow roots: impactsof root properties and root failure
mechanism. Soil Phys. Soil Sci. Soc. Am. J. 73,12761285,
http://dx.doi.org/10.2136/sssaj2008.0172.
ontgomery, D., 1994. Road surface drainage, channel initiation,
and slope insta-bility. Water Resour. Res. 30 (6), 19251932.
ilaweera, N.S., Nutalaya, P., 1999. Role of tree roots in slope
stabilisation. Bull. Eng.Geol. Env. 57, 337342.
Loughlin, C.L., 1974. A study of tree root strength
deterioration following clear-felling. Can. J. For. Res. 4,
107113.
Loughlin, C., Watson, A.J., 1979. Root-wood strength
deterioration in Radiata Pineafter clearfelling. N. Z. J. For. Sci.
9 (3), 284293.
perstein, V., Frydman, S., 2000. The influence of vegetation on
soil strength. GroundImprov. 4, 8189.
ollen, N., Simon, A., 2005. Estimating the mechanical effects of
riparian vegetationon stream bank stability using a fiber bundle
model. Water Resour. Res. 41,w07025.
etrone, A., Preti, F., 2008. Suitability of soil bioengineering
techniques in Cen-tral America: a case study in Nicaragua. Hydrol.
Earth Syst. Sci. 12,
12411248www.hydrol-earth-syst-sci.net/12/1241/2008/
etrone, A., Preti, F., 2009. Soil bio-engineering for risk
mitigation and environ-mental restoration in a humid tropical area.
Hydrol. Earth Syst. Sci. 6,
138www.hydrol-earth-syst-sci-discuss.net/6/1/2009/
reti, F., 2006. Stabilit dei versanti vegetati. Cap. 10, pp.
137168 in: Sauli, G.,Cornelini, P., Preti, F (Eds.). Manuale 3
Ingegneria Naturalistica Sistemazionedei versanti. Regione Lazio,
http://www.regione.lazio.it/web2/contents/ingegneria
naturalistica/manuale versanti/Cap 10 10.pdf (in Italian).
reti, F., Bresci, E., Ravenna, V., 2001. Field measurements for
hydrologic mappingin old terraced land. In: International
Conference on Sustainable Soil Manage-ment for Environmental
ProtectionSoil Physical Aspects, Florence, Italy, 27July 2001 (in
Italian).
reti, F., Dani, A., Laio, F., 2010. Root profile assessment by
means of hydrological,pedological and above-ground vegetation
information for bio-engineering pur-poses. Ecol. Eng. 36, 305316,
http://dx.doi.org/10.1016/j.ecoleng.2009.07.010.
reti, F., Forzieri, G., Chirico, G.B., 2011. Forest cover
influence on regionalflood frequency assessment in Mediterranean
catchments. Hydrol. EarthSyst. Sci. 15, 30773090,
http://dx.doi.org/10.5194/hess-15-3077-2011www.hydrol-earth-syst-sci.net/15/3077/2011
reti, F., Giadrossich, F., 2009. Root reinforcement and slope
bioengineering sta-bilization by Spanish Broom (Spartium junceum
L.). Hydrol. Earth Syst. Sci. 13,17131726.
reti, F., Petrone, A., 2012. Soil bio-engineering for watershed
management anddisaster mitigation in South America: local species
survivability and growthrates in Ecuador, iForest,
http://dx.doi.org/10.3832/ifor0636-005
reti, F., Settesoldi, D., Mazzanti, B., Paris, E., 1996. Criteri
e procedure per la val-utazione delle piene nel territorio toscano.
Atti del XXV Convegno di Idraulicae Costruzioni Idrauliche. Torino,
Settembre 1996. CNR-GNDCI, pp. 317328, inItalian.
egione Toscana, 2007. Aggiornamento e sviluppo del del sistema
di region-
alizzazione delle portate di piena in ToscanaAlTo,
http://www.rete.toscana.it/sett/pta/suolo/difesa
suolo/alto/index.5htm, 2007. in Italian.
eubens, B., Poesen, J., Danjon, F., Geudens, G., Muys, B., 2007.
The role of fine andcoarse roots in shallow slope stability and
soil erosion control with a focus onroot system architecture: a
review. Trees 21, 385402.
Z
Z
g 61P (2013) 633 645 645
inaldo, A., Vogel, G.K., Rigon, R., Rodriguez-Iturbe, I., 1995.
Can one gauge the shapeof a basin? Water Resour. Res. 4,
11191127.
oering, J.J., Schmidt, K.M., Stock, J.D., Dietrich, W.E.,
Montgomery, D.R., 2003. Shal-low landsliding, root reinforcement,
and the spatial distribution of trees in theOregon Coast Range.
Can. Geotech. J. 40, 237253.
ossi Pisa, P., Preti, F., Rossi, M., Ventura, F., Mazzanti, B.,
1999. Water, soil andchemical losses: field experiments and model
analysis, Water Sci. Technol., 39,1999.
aklas, M.E., Sidle, R.C., 2004. A spatial and temporal model of
root cohesion in forestsoils. Can. J. For. Res. 34 (4), 950958.
chwarz, M., Preti, F., Giadrossich, F., Lehmann, P., Or, D.,
2010a. Quantifying therole of vegetation in slope stability: a case
study in Tuscany (Italy). Ecol. Eng. 36(March (3)), 285291,
http://dx.doi.org/10.1016/j.ecoleng.2009.06.014.
chwarz, M., Lehmann, P., Or, D., 2010b. Quantifying lateral root
reinforcement insteep slopesfrom a bundle of roots to tree stands.
Earth Surf. Process. Landf.35 (354367),
http://dx.doi.org/10.1002/esp.1927.
chmidt, K.M., Roering, J.J., Stock, J.D., Dietrich, W.E.,
Montgomery, D.R., Schaub,T., 2001. The variability of root cohesion
as an influence on shallow landslidesusceptibility in the Oregon
Coast Range. Can. Geotech. J. 38, 9951024.
elby, M.J., 1993. Hillslope Materials and Processes, 2nd ed.
Oxford University Press,Oxford, p. 451 S.
idle, R.C., 1992. A theoretical model of the effects of timber
har-vesting on slope stability. Water Resour. Res. 28 (7),
18971910,http://dx.doi.org/10.1029/92WR00804.
idle, R.C., Ziegler, A.D., Negishi, J.N., Nik, A.R., Siew, R.,
Turkelboom, F., 2005. Erosionprocesses in steep terraintruths,
myths, and uncertainties related to forestmanagement in Southeast
Asia. For. Ecol. Manage. 224 (March (12) 2006),199225.
idle, R.C., Ochiai, H., 2006. Landslides: processes, prediction,
and land use. WaterResour. Monogr. 18, American Geophysical
Union.
imon, A., Collison, A.J.C., 2002. Quantifying the mechanical and
hydrologic effectsof riparian vegetation on streambank stability.
Earth Surf. Process. Landf. 27,527546.
tokes, A., Atger, C., Bengough, A.G., Fourcaud, T., Sidle, R.C.,
2009. Desirable plantroot traits for protecting natural and
engineered slopes against landslides. PlantSoil 324, 130,
http://dx.doi.org/10.1007/s11104-009-0159-y.
arboton, D.G., Bras, R.L., Rodriguez-Iturbe, I., 1992. A
physical basis for drainagedensity, Geomorphology, 5 (12) Fractals
in Geomorphology, May 1992. 5976,ISSN 0169-555X, DOI:
10.1016/0169-555X(92)90058-V.
osi, M., 2007. Root tensile strength relationships and their
slope stability implica-tions of three shrubs species in the
Northern Apennines (Italy). Geomorphology87, 268283.
atson, A., OLoughlin, C., 1985. Morphology, strength, and
biomass of manuka rootsand their influence on slope stability. N.
Z. J. For. Sci. 15 (3), 337348.
atson, A., Marden, M., Rowan, D., 1997. Root-wood strength
deterioration inKanuka after clearfelling. N. Z. J. For. Sci. 27
(2), 205215.
atson, A., Pillips, C., Marden, M., 1999. Root strength, growth,
and rates of decay:root reinforcement changes of two tree species
and their contribution to slopestability. Plant Soil 217, 3947.
aldron, L.J., 1977. The shear stress resistance of
root-permeated homogeneousand stratified soil. Soil Sci. Soc. Am.
Proc. 41, 843849.
aldron, L.J., Dakessian, S., 1981. Soil reinforcement by roots:
calculationof increased soil shear resistance from root properties.
Soil Sci. 132,427435.
ang, W.L., Yen, B.C., 1974. Soil arching in slopes. J. Geotech.
Eng. Div., ASCE 100(GT1), 6178.
u, T.H., 1976. Investigation on landslides on Prince of Wales
Island. Alaska Geotech.Rpt. No 5, Dpt. Of Civil Eng., Ohio State
Univ., Columbus, USA.
u, W., Sidle, R.C.A., 1995. A distributed slope stability model
for steep forested
basins. Water Resour. Res. 31 (8), 20972110.
iemer, R.R., 1981. Roots and the stability of forested slopes.
Erosion and SedimentTransport in Pacific Rim Steeplands, 343361
(IAHS Publ. No. 132).
iemer, R.R., Lewis, J., Rice, R.M., Lisle, T.E., 1991. Modeling
the cumulative watershedeffects of forest management strategies. J.
Environ. Qual. 20, 3642.
dx.doi.org/10.2136/sssaj2008.0172http://www.hydrol-earth-syst-sci.net/12/1241/2008/http://www.hydrol-earth-syst-sci-discuss.net/6/1/2009/http://www.regione.lazio.it/web2/contents/ingegneria_naturalistica/manuale_versanti/Cap_10_10.pdfhttp://www.regione.lazio.it/web2/contents/ingegneria_naturalistica/manuale_versanti/Cap_10_10.pdfdx.doi.org/10.1016/j.ecoleng.2009.07.010dx.doi.org/10.5194/hess-15-3077-2011http://www.hydrol-earth-syst-sci.net/15/3077/2011http://dx.doi.org/10.3832/ifor0636-005http://www.rete.toscana.it/sett/pta/suolo/difesa
suolo/alto/index.5htmhttp://www.rete.toscana.it/sett/pta/suolo/difesa
suolo/alto/index.5htmdx.doi.org/10.1016/j.ecoleng.2009.06.014dx.doi.org/10.1002/esp.1927dx.doi.org/10.1029/92WR00804dx.doi.org/10.1007/s11104-009-0159-y
Forest protection and protection forest: Tree root degradation
over hydrological shallow landslides triggering1 Introduction2
Study sites3 Materials and methods3.1 Data collection3.2 Root
reinforcement
4 Results and discussion4.1 Root reinforcement decay after tree
death4.2 Slope stability analysis
5 ConclusionsAcknowledgmentsReferences