PUBLICATIONS

Geophysical Research Letters

RESEARCH LETTER Vegetation as a driver of temporal variations in slope stability:
10.1002/2017GL073174
The impact of hydrological processes
Key Points: John H. Kim1,2 , Thierry Fourcaud1, Christophe Jourdan3 , Jean-Luc Maeght4,5 , Zhun Mao1,
• Land uses with trees provided greater
slope stability that is also less variable James Metayer1 , Louise Meylan3,6, Alain Pierret4,7 , Bruno Rapidel6,8 , Olivier Roupsard3 ,
compared to those with shrub or Anneke de Rouw9, Mario Villatoro Sanchez10, Yan Wang1, and Alexia Stokes1
herbaceous communities
1
• Analyzing the literature, slope stability AMAP, INRA, IRD, CIRAD, CNRS, and University of Montpellier, Montpellier, France, 2Now at Max Planck Institute for
with woody vegetation was less Biogeochemistry, Jena, Germany, 3CIRAD, UMR Eco&Sols, Montpellier, France, 4IRD, Soils and Fertilizers Research Institute,
sensitive to soil and climate factors
Hanoï, Vietnam, 5IRD Institute of Ecology and Environmental Sciences of Paris (iEES-Paris UMR 242), Paris, France, 6CATIE,
than with herbaceous plants
• To reduce risks from earthquake- and Programa de Agroforestería y Agricultura Sostenible, Turrialba, Costa Rica, 7Department of Agricultural Land Management,
rainfall-induced landslides, the IRD, Vientiane, Laos, 8CIRAD, UMR System, Montpellier, France, 9IRD, Université Pierre et Marie Curie, Paris, France, 10Centro
additional stabilization by trees may de Investigaciones Agronómicas, Universidad de Costa Rica, San José, Costa Rica
need to be quite high

Supporting Information:
Abstract Although vegetation is increasingly used to mitigate landslide risks, how vegetation affects the
• Supporting Information S1 temporal variability of slope stability is poorly understood, especially in earthquake-prone regions. We
• Table S1 combined 3-year long soil moisture monitoring, measurements of soil physical properties and plant
functional traits, and numerical modeling to compare slope stability under paired land uses with and without
Correspondence to:
J. H. Kim,
trees in tropical, subtropical, and temperate landslide- and earthquake-prone regions. Trees improved
[email protected] stability for 5–12 months per year from drawdown of soil moisture and resulted in less interannual variability
in the duration of high-stability periods compared to slopes without trees. Our meta-analysis of published
Citation:
data also showed that slopes with woody vegetation were more stable and less sensitive to climate and
Kim, J. H., et al. (2017), Vegetation as a soil factors than slopes with herbaceous vegetation. However, estimates of earthquake magnitude necessary
driver of temporal variations in slope to destabilize slopes at our sites suggest that large additional stabilization from trees is necessary for
stability: The impact of hydrological
processes, Geophys. Res. Lett., 44,
meaningful protection against external triggers.
4897–4907, doi:10.1002/2017GL073174.

Received 21 FEB 2017 1. Introduction

Accepted 18 APR 2017
Accepted article online 24 APR 2017 Almost a fifth of global land surface is classified as highly susceptible to shallow, rainfall-induced landslides
Published online 29 MAY 2017 [Hong et al., 2007]—common phenomena in hilly and mountainous regions, and on road and river embank-
ments. Worldwide, landslides are responsible for over 4500 fatalities and up to U.S. $3.2 billion in damages
per year, creating both onsite and far-reaching downstream impacts on energy, agricultural, forest, riverine,
and marine resources [Glade, 2003; Petley, 2012; Schuster and Highland, 2001]. Landslides can also prompt
large-scale policy responses [Guiang, 2001; Panagos et al., 2012; Phien-Wej et al., 1993; Xu et al., 2002], under-
lying the need for sound scientific evidence on how land use affects slope stability. One major controversy is
whether planting trees significantly improve, have little effect, or even decrease slope stability compared to
herbaceous and shrubby vegetation [Fattet et al., 2011; Genet et al., 2010]. This dilemma becomes particularly
relevant in the current contexts of global food and fuel needs and rapid anthropogenic and socioeconomic
changes that will also affect other environmental benefits and costs [Ferrara et al., 2015; Kim et al., 2016].

Using vegetation for enhancing stability and resilience of hillslopes is increasingly seen as a cost-effective and
sustainable alternative to man-made structures [Basu et al., 2014; National Research Council, 2006; Stokes et al.,
2014]. Plant roots in particular stabilize hillslopes by anchoring into deeper soil layers and forming a fiber-
reinforced soil composite, resulting in additional shear resistance (“mechanical cohesion”) that help counter
the destabilizing gravitational forces [Glade, 2003; McGuire et al., 2016; Sidle et al., 2006]. Vegetation also
removes soil water through evapotranspiration, increasing soil’s cohesive forces against mechanical failure
[Collins and Znidarcic, 2004; van Asch et al., 1999]. These forces between air, water, and soil particles (suction
or “hydric cohesion”) are significantly diminished when soil water content is high, leading to a loss of shear
resistance and ultimately slope failure [Lu et al., 2010; Vanapalli et al., 1996].

Even though vegetation is the main conduit for water fluxes out of the soil and has strong and predictable
influences on subsurface hydrology [Kim and Jackson, 2012], studies examining the effect of vegetation on
©2017. American Geophysical Union.
All Rights Reserved. slope stability have largely focused on the effects of mechanical cohesion [Schwarz et al., 2010; Wu, 2013].

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4897

 Geophysical Research Letters 10.1002/2017GL073174

Despite some progress addressing this discrepancy [e.g., Arnone et al., 2016; Chirico et al., 2013; Gonzalez-
Ollauri and Mickovski, 2017; Hales and Miniat, 2016; Hwang et al., 2015], the effect of vegetation on temporal
variability of slope stability is still poorly understood [Band et al., 2012]—we found only three field-based stu-
dies comparing hydrological effects of contrasting vegetation types on slope stability (Table S1 and Text S3 in
the supporting information) [Rahardjo et al., 2014; Simon and Collison, 2002; Simon et al., 2006]. The lack
of published data is especially apparent for interannual and deep-soil comparisons; the studies lasted only
5–14 months, and only one study examined soil beyond a 1.2-m depth, although potential shear surfaces
are often located below 1.0 m [van Asch et al., 1999]. Long-term observations are crucial to assessing changes
in slope stability, as landslides tend to initiate after long-term accumulations of water [Sidle et al., 2006]. Thus,
quantifying temporal uncertainties remains an outstanding priority in landslide risk management [Dai et al.,
2002]. Our study is the first attempt to our knowledge to compare interannual variation in slope stability
under different vegetation types in the field.
Along with rainfall, earthquakes are common triggers for landslides, with the proximity of many mountainous
regions to seismically active areas adding to the inherent vulnerability from steep slopes. Unlike rainfall,
earthquakes are less seasonal and occur less predictably. Although many earthquake-induced landslides
are deep-seated, recent studies have shown that vegetation can be an important spatial predictor of
earthquake-induced landslide occurrence [e.g., Peduzzi, 2010]. Understanding the temporal variability of
slope stability under contrasting vegetation types thus could be especially useful regarding seismic land-
slides, but these links are rarely explored due to the expense and difficulty of full-scale experimentation
[Liang and Knappett, 2015]. Therefore, we also tested whether the positive influence of vegetation on slope
stability would extend to protecting slopes against earthquake-induced shallow landslides.
To test the temporal effect of land-use type on slope stability, we monitored soil moisture over 3 years under
paired land-use plots with trees and without trees (but with shrubby or herbaceous vegetation) in tropical,
subtropical, and temperate landslide- and earthquake-prone regions. We hypothesized that land-use type
would drive temporal variations in slope stability through hydrological processes. We measured soil physical
properties and plant functional traits and modeled slope stability to quantify (1) the effect of trees and the
potential protection that trees could confer to slopes during earthquakes and (2) how these factors varied
interannually. In addition, we explored whether general relationships existed between vegetation, climate,
soil, and slope stability through a meta-analysis of data from the literature. Identifying such interactions is
a key requisite for better predictive models of landslides [Sidle and Bogaard, 2016].

2. Study Area and Methods

Our sites in mountainous regions (Laos, Costa Rica, and France) were part of a study on ecosystem services
under paired locally common vegetation types [Prieto et al., 2015]. Geohazard risks are recurrent issues at
all three sites that were contrasted in terms of climate (subtropical, tropical, and temperate), soil (Ultisols-clay,
Alfisols-clay, and Cambisols-sandy loam), and vegetation (Table S2). The differences between sites allowed us
to test the generic nature of our results. At each site, we chose a pair of tree-dominated (“with tree”) and her-
baceous and/or shrub (“without tree”) vegetation types that correspond to a contrast in management history.
In Laos, we compared a 22-year-old teak plantation to a naturally regenerating herbaceous 1-year-old slash-
and-burn that was left fallow (typically prior to upland rice cultivation). In Costa Rica, we compared coffee
plantations grown under the shade of Erythrina poeppigiana trees and under full sun. In France, we compared
natural mixed spruce-beech-fir forest to open canopy gaps dominated by herbaceous species (Table S2).
Landslides occur when external factors disrupt the equilibrium between resisting (frictional and cohesive)
and driving forces (gravitational) [Greenway, 1987]. To examine vegetation effects on slope stability, we used
a geotechnical model to estimate the factor of safety (FoS, dimensionless) under the different vegetation
types, calculated as the ratio of resisting and driving forces acting on a slope as follows:

ci þ cri þ chi þ tan ɸi W i cos β

FoS ¼ (1)
W i sin β

where ci is effective cohesion at saturation of ith soil layer (kPa); cri is mechanical cohesion from roots (kPa;
equation (S2)); chi is the additional hydric cohesion produced by suction on the unsaturated soil (kPa;

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4898

 Geophysical Research Letters 10.1002/2017GL073174

Figure 1. (a) Reduction of shear resistance with respect to increasing soil volumetric water content. Data shown for soil
from 1.0 to 1.2 m depth at the three field sites. Error bars indicate standard error. Regression lines represent fitted
exponential functions following Matsushi and Matsukura [2006] except for the French site where a power law function was
fitted (Text S1.2 and equation (S4)). (b–f) Interaction effects of vegetation on the Factor of Safety (FoS) from multiple
regression of data from our literature review (no borders) and our fieldwork (black borders). Residuals for minimum FoS
(Figures 1b and 1c) and maximum FoS (Figures 1d–1f) were calculated without the vegetation and its interaction terms to
visualize the differential impacts vegetation type had on relationships between FoS and soil and climate variables, namely,
soil hydraulic conductivity estimated based on soil texture [Rawls et al., 1982] (Figure 1b), depth of soil layer with the
lowest FoS as a proxy for failure depth (Figures 1c and 1e), potential water excess (PWE = Precipitation  Potential
Evapotranspiration; Text S3) [Allen et al., 1998] as an index of climatic humidity (Figure 1d), and effective soil cohesion
(Figure 1f). Slopes of the residuals against all but one interaction term for woody vegetation regression lines are signifi-
cantly more gradual (P < 0.05; Table S4) than those for herbaceous vegetation, indicating a buffering of FoS against climate
and soil variability with woody vegetation.

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4899

 Geophysical Research Letters 10.1002/2017GL073174

equation (S4)); Wi is cumulative weight of soil, water, and biomass of the ith and above layers per area (kPa;
equation (S5)); ɸi is effective angle of internal friction (degrees); and β is slope angle (degrees from horizon-
tal). chi has typically been described with soil suction [Bishop, 1960; Fredlund et al., 1978; Vanapalli et al., 1996].
We empirically estimated chi based on direct shear tests at different soil moisture levels and normal loads
following methods developed by Matsushi and Matsukura [2006] (Figure 1a, Texts S1.2 and S2.2, and
equation (S4)) and verified by others [Bai and Liu, 2012; Zydron and Mietus, 2016; Zydron et al., 2016].
Estimated matric potential using the highest observed moisture values at the most vulnerable depths for slope
stability (342 to 6 kPa) suggest positive pore pressures were unlikely at our sites (Text S1.2) [Cosby et al., 1984].
FoS < 1 indicates an unstable slope, whereas FoS > 1.3 denotes a stable slope, and monitoring is generally
recommended for FoS < 1.3 [U.S. Army Corps of Engineers, 2003]. Because 0.3 is the difference in FoS between
stable and unstable slopes by convention, we used this difference of 0.3 as one of the thresholds for compar-
ing the effect of vegetation on the FoS between years. We analyzed our FoS data by examining the number of
days per year when (1) land-use types with trees had a FoS higher than those without trees, (2) the difference
in FoS between land-use types was >0.3, (3) the hydric-cohesion component contributed >0.3 additional FoS
to the plot, (4) hydric cohesion was greater than mechanical cohesion, and (5) FoS was >0.3 greater than the
minimum calculated for the site (see Text S4.2 for more details).
For input to the slope stability model (equation (1)), we collected the following hydrological and mechanical
data. Soil moisture sensors (TDR or electrical resistance) under the paired vegetation types (with and without
trees) at various depths logged data every 30 min from January 2012 to December 2014 in Costa Rica and
France, and monthly in Laos from January 2012 to February 2014 to track temporal changes in hydric cohesion
(chi; Table S2). Missing data were estimated using the Station Relative Difference method (Text S2.1) [Defossez
et al., 2015; Dumedah and Coulibaly, 2011]. Additionally, 2–3 kg of soil were collected at each plot at a depth of
1.0–1.2 m from a 3.0 × 2.0 × 1.6 m3 trench for the direct shear tests (Text S2.2). We measured shear resistance on
reconstituted soil samples without accounting for hysteresis (Text S1.2) [Anagnostopoulos et al., 2015; Han
et al., 1995; Schuppener et al., 1999]. We selected the locations of trenches to be representative of the plant
communities while minimizing variations in geomorphology between the paired vegetation plots (Text S2.1).

We estimated mechanical cohesion (cri) by applying Wu and Waldron’s model [Waldron, 1977; Wu et al., 1979]
using root tensile strength data from the literature [Eqs. S2, S3, Text S1.1; Genet et al., 2005; Kuriakose and
van Beek, 2011; Mao et al., 2012] and root impacts counted from open pits, cores, or rhizotrons (Text S2.1).
We chose a conservative root orientation factor of 0.5 [Ji et al., 2012], because Wu and Waldron’s suggested
value (1.2) overestimates cri by 33–68% [Fan and Su, 2008; Hubble et al., 2010; Thomas and Pollen-Bankhead,
2010]. To calculate the load (Wi) on the slopes, aboveground biomass was estimated from harvests or
from species-specific allometric equations based on measurements of woody stem diameter and height
(Text S2.3) [Mao et al., 2015; Pache et al., 2003; Ruiz-Peinado et al., 2011; Zianis et al., 2005].
To estimate the importance of water on slope stability at each site, we calculated the proportion of differ-
ences in FoS resulting from the difference in soil water content between the vegetation types relative to
FoS differences resulting from the difference in all factors affected by vegetation, i.e., water (chi, Wi), root
(cri), and biomass (Wi; Text S4.1 and equation (S9)). To compare the effects of climate on slope stability, we
separated the wet and dry seasons defined by precipitation in Laos and Costa Rica and temperature in
France (Text S2.4) [National Oceanic and Atmospheric Administration, 2011; Peel et al., 2007].
To characterize earthquakes necessary to destabilize slopes at our sites, we estimated ground acceleration
using slope geometry and FoS of each vegetation type (equation (S10)) [Newmark, 1965]. We used the annual
median FoS to estimate acceleration required to overcome stability for half of the year. We converted accel-
eration to earthquake magnitude based on distance to epicenter and earthquake depth following Mickey
(Text S4.3 and equation (S11)) [Mickey, 1971]. We used a fixed distance of 10 km to earthquake hypocenter
to standardize the comparison between the sites and used focus depths from a 92-year record of earth-
quakes in our study areas [United States Geological Survey, 2016].

While our paired-plot approach allowed a robust test of land-use changes independent of covarying factors,
we expanded our analysis to the literature to quantify relationships between vegetation, climate, soil, and
slope stability across sites using stepwise multiple regression. The predictor terms tested in the meta-analysis
were selected based on our conceptual models of slope stability and soil hydrology (equations (1), (S6), and

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4900

 Geophysical Research Letters 10.1002/2017GL073174

Table 1. Seasonal and Interannual Patterns of Factor of Safety (FoS) and Differences in FoS Between Plots With and Without Trees
a a a,d f
ch/Fd > 0.3 ch > cr High FoS ΔFoS M
FoST > ΔFoS > ΔFoS ∵
a a,b c
Site Years FoSNT 0.3 Water Tree NT Tree NT Tree NT Dry Wet Tree NT

Laos 2012 278 278, 66 0.82 366 366 366 366 366 366 0.84 0.76 6.92 6.77
2013 365 365, 0 0.87 365 196 365 232 365 208 1.61 1.01 6.76 6.31
2014 — — 0.88 — (0) — (183) — — — (0) — (66) 1.17 — 6.65 6.19
Costa Rica 2012 149 101, 0 0.71 151 165 268 366 156 160 0.41 0.09 5.99 6.06
2013 149 103, 0 0.67 150 166 234 365 153 158 0.21 0.07 5.99 6.05
2014 151 102, 0 0.68 150 (1) 174 (5) 303 365 154 (1) 165 (3) 0.5 -0.01 6.04 6.08
France 2012 357 87, 0 0.46 57 33 121 59 101 41 0.24 0.04 6.46 6.42
2013 364 78, 0 0.48 44 1 127 59 107 5 0.27 0.04 6.48 6.42
2014 365 116, 0 0.54 59 (8) 14 (16) 149 43 117 (7) 21 (15) 0.36 0.05 6.51 6.42
a
FoST: factor of safety (FoS) on plots with trees, FoSNT: FoS of plots without trees, ΔFoS: FoST  FoSNT, ΔFoS ∵ water: proportion of FoS differences between
plots with trees and without trees attributable to differences in soil water content between the vegetation types, ch/Fd > 0.3: ch increases FoS by >0.3, ch: hydric
cohesion, Fd: driving force (equation (1)), cr: mechanical cohesion, NT: without tree plots, Dry: dry season for Laos and Costa Rica and summer for France, and Wet:
wet season for Laos and Costa Rica and winter for France.
b
Duration in days per year. Values in parentheses indicate the standard deviation of values of the three monitoring years.
c
Values for FoST  FoSNT and FoSNT  FoST shown, respectively.
d
Proportion of the differences in FoS between plots with and without trees attributable to differences in hydric cohesion (equation (S9)).
e
FoS greater than the minimum calculated for the site by 0.3.
f
M is the annual median magnitude of earthquake necessary to destabilize slopes based on our FoS results and hypothetical distances to hypocenters of 10 km
(equations (S10) and (S11)). Likely distance between our sites and hypocenters are 177, 58, and 34 km for Laos, Costa Rica, and France, respectively [United States
Geological Survey, 2016].

(S8) and Text S3) [Dunne et al., 1991; Gale and Grigal, 1987; Gray and Leiser, 1982; Thornthwaite and Mather,
1957]. We compiled information from studies that estimated FoS from soil water monitoring at multiple
time points [Bittelli et al., 2012; Blatz et al., 2004; Damiano et al., 2012; Godt et al., 2009; Harris et al., 2012; Li
et al., 2003; Matsushi and Matsukura, 2007; Rahardjo et al., 2014; Rinaldi and Casagli, 1999; Simon and
Collison, 2002; Simon et al., 2006, 2000; Springman et al., 2013; Terlien, 1997].

3. Results
Slopes with trees had higher FoS than those without trees for 321, 149, and 362 d yr1 in Laos, Costa Rica,
and France, respectively, demonstrating greater landslide susceptibility of slopes with shrubby/herbaceous
vegetation for longer periods of time in Laos and France (Table 1 and Figure 2). The differences in FoS
between plots with and without trees were driven mainly by the seasonal effect of vegetation on water,
as opposed to mechanical cohesion or surcharge from weight of the biomass. Proportions of the difference
in FoS between the vegetation types due to differences in soil water (hydric cohesion and water weight)
averaged 86, 68, and 50% for Laos, Costa Rica, and France, respectively (Table 1 and Figure 2). Hydric cohe-
sion exceeded mechanical cohesion for 121–365 and 43–365 d yr1 in plots with and without trees, respec-
tively (Table 1). Hydric cohesion under trees also contributed >0.3 of additional FoS for 190 d yr1 on
average, compared to 155 days on plots without trees, indicating that vegetation can significantly impact
slope stability through its influence on the soil water balance (Table 1). The effect of vegetation was much
more apparent during the dry/summer season when hydric cohesion was clearly differentiated between
vegetation types, with a mean difference in FoS of 1.21, 0.37, and 0.28 in Laos, Costa Rica, and France,
respectively (Table 1 and Figure 2).
Our multiyear assessment also revealed how these periods of high FoS fluctuated interannually (Figure 2).
The days per year whereby FoS on plots with trees was >0.3 greater than the FoS of plots without trees varied
by >80 days, highlighting interannual climate variability as an important determinant of success and reliabil-
ity of ecological engineering approaches to slope stabilization (278–365, 101–103, and 78–116 days for Laos,
Costa Rica, and France, respectively; Table 1). Standard deviations for annual periods of high FoS (i.e., FoS
exceeding the site minimum by 0.3) were also always lower on plots with trees compared to those without
trees, indicating a buffering against climate variability and greater certainty in stability with trees (Table 1).
In the rainy (Laos and Costa Rica) and winter (France) seasons, the differences in hydric cohesion between the
plots with and without trees varied from persistent differences to a complete reversal of the effects of

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4901

 Geophysical Research Letters 10.1002/2017GL073174

Figure 2. Factor of safety (FoS) at the most vulnerable depths under contrasting land uses for Laos (1.8 m), Costa Rica
(1.5 m), and France (1.2 m) during the monitoring period (2012–2015). Contributions to FoS from vegetation, soil, and
water are shown for the following additive components for plots with trees: soil mechanical and hydric cohesions
c cr ch
correspond to W sin β, W sin β, and W sin β from equations (1) and (S2)–(S4) and soil, water, and biomass loads correspond

to tanWɸisin
zρb cos β tan ɸi θzA cos β
β , W sin β , and tanWɸsin
i B cos β
β from equations (1) and (S5). The FoS of plots without trees are shown as red
lines. Boxed areas indicate dry seasons (for Laos and Costa Rica) and the summer (for France), based on monthly
climate data (see section 2). Arrows indicate pruning dates of shade-inducing Erythrina trees during coffee flowering
and fruiting periods. Note that vertical axis scales differ between panels for easier viewing.

vegetation. FoS of the recently slash-and-burned fallow slope in Laos was initially similar to that of the teak
plantation but decreased massively in the years following vegetation removal. In contrast, the low antece-
dent soil moisture under teak prior to the 2013 wet season buffered the infiltration of rainwater and loss
of shear resistance (Figure 2). Ranges of soil moisture under teak at the most vulnerable depth for slope
stability (1.8 m) were narrower and contributed to less variable FoS compared to the slash-and-burned plot
(Figure S1). Hydric cohesion accounted for almost all the differences in FoS between the land-use types,
indicating that the degradation of slope stability was driven primarily by hydrology, rather than root decom-
position after vegetation shifts (Figures 2 and S2) [O’Loughlin and Ziemer, 1982]. Surprisingly, although
Erythrina shade trees increased FoS in coffee plantations during the dry season, a complete reversal occurred
during the wet season, when the contribution of hydric cohesion from Erythrina was lower than that from cof-
fee plants alone—low enough to offset the greater mechanical cohesion of the trees compared to that of the
coffee (Figures 2 and S3). In France, FoS and hydric cohesion were at their lowest during the winter, with little
difference between dense forest and canopy gaps (Figure 2). Instead, higher mechanical cohesion of trees
maintained greater FoS relative to gaps during this period (Figure S3).
The most vulnerable depths to slope failure at our sites (i.e., lowest FoS) were the deepest monitored depths
(1.2–1.8 m; Table S3). At these depths, mechanical cohesion was minimal as root density decreased rapidly
with increasing depth, and hydric cohesion instead contributed more to slope stability. At plots with trees,

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4902

 Geophysical Research Letters 10.1002/2017GL073174

average hydric cohesion was 22.6 (range: 6.4–33.9), 8.2 (0.0–7.3), and 1.2 (0.0–29.6) kPa, compared to 3.6, 0.9,
and 0.9 kPa from mechanical cohesion, in Laos, Costa Rica, and France, respectively (Figures S2 and S3). Shear
resistance at these depths can be sensitive to variations in soil moisture (Figure 1a), indicating that land uses
affecting deep soil moisture may have the most destabilizing potential on slopes.
Earthquake magnitude necessary to destabilize slopes with trees was higher than slopes without trees in
Laos and France in all years (Table 1). The pattern was reversed in Costa Rica, although by a small margin,
due to the seasonal reversal of slope stability between vegetation types. An appreciable protection from
earthquake-triggered landslides with trees was possible only in Laos, where we calculated large differences
in FoS between slopes with and without trees (Table 1).
Finally, in our meta-analysis of the literature, woody vegetation also had higher minimum and maximum FoS
by 0.11 and 0.52, respectively, compared to herbaceous vegetation (Table S4). Slope stability under woody
vegetation was also less sensitive to climate and soil factors such as climatic humidity, soil hydraulic conduc-
tivity, and depth to the potential shear surface than herbaceous vegetation, indicating the positive role of
woody plants in increasing FoS while also reducing its variability (Figures 1b–1f and Table S4).

4. Discussion
Our analyses of both field and literature data reveal an important role for vegetation type regarding slope
stability and its variability. The longer and less variable periods of higher FoS with trees compared to herbac-
eous vegetation from our long-term monitoring (Table 1) and the higher FoS from woody plants that was less
sensitive to interacting soil and climate factors in our meta-analysis (Figures 1b–1f) indicate that trees may be
an attractive ecological engineering solution for landowners, public planners, and policy makers in landslide-
prone communities. By lowering soil moisture and maintaining greater shear resistance, trees may increase
thresholds for landslide triggers such as rainstorms or earthquakes [Peduzzi, 2010]. Antecedent soil moisture
is one of the main determinants of slope failure [Hawke and McConchie, 2011; Kirschbaum et al., 2015], and
maintaining low soil moisture could guard against the loss of shear resistance from infiltrating water, for
example, in the Laotian teak plantation. Similarly, vigorous evapotranspiration may maintain soil moisture
below the threshold for soil liquefaction, a process that weakens shear resistance and triggers landslides
during earthquakes [Bray and Sancio, 2006; Seed and Idriss, 1982]. Nevertheless, our results suggest that
differences in FoS with and without trees need to be large to achieve appreciable protection against such
landslide triggers.
Interactions between soil, climate, and vegetation identified in our meta-analysis may help explain how slope
stability responds to land-use changes at our sites. For example, differences in FoS between woody and her-
baceous vegetation in the meta-analysis were greater in clayey soils with low permeability than in sandy soils
(Figures 1b–1f and Table S4). At our Laotian site, evapotranspiration by deep-rooted teak in a clayey soil pro-
pagated large increases in slope stability with minimal seasonal effects, despite the highly seasonal rainfall
(Figure S4). At the French site, permeable sandy loam possibly maintained high moisture under both forest
and gaps, obscuring the FoS differences between vegetation types (Figure 2). The consistent seasonal rever-
sals of slope stability between coffee plantations and coffee-Erythrina agroforestry in Costa Rica may arise
from frequent pruning that reduces differences in leaf area between the land uses (Figure 2). At the onset
of the rainy season, rainwater infiltration to the most unstable depth took 0–12 more days under Erythrina
trees than under coffee plantations. Therefore, our soil moisture data did not support alternative hypotheses
of faster water movement by preferential flow along tree roots or hydraulic redistribution as mechanisms for
higher soil moisture under Erythrina trees [Ghestem et al., 2011].
We propose three broad recommendations for stakeholders to guide land-use decisions that do not compro-
mise slope stability. First, land-use changes on slopes with a dry soil profile maintained by the current vege-
tation or with moisture-sensitive shear resistance should be considered judiciously, as they may be most
vulnerable to increased landslide risk (Figure 1a) [Kim and Jackson, 2012]. Second, because climate variability
can negatively influence the reliability of ecological engineering approaches to slope stability (Table 1), care-
ful site and species selection should precede using vegetation to protect against shallow landslides. Lastly,
the capacity to maintain consistently low soil moisture by preventing deep infiltration should be a top criteria
when designing vegetation communities for slope rehabilitation [Stokes et al., 2009]. Traits such as evergreen

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4903

 Geophysical Research Letters 10.1002/2017GL073174

leaf cover and deep roots may buffer against the destabilizing effects of climate variability, though soil and
climate also likely modify the efficacy of such traits on soil water removal (Figures 1b–1f).
Based on the scarcity of data in our literature review, we call on the scientific community for greater
investments in examining slope stability under different land uses that integrate deeper and longer-term
monitoring of soil hydrology and root dynamics. For example, although our meta-analysis highlighted poten-
tial interactions between vegetation, soil, climate, and geotechnical factors on slope stability, we could not
experimentally test for these interactions with our field data. Such experiments will help understand the tem-
poral variability and interactive effects between vegetation and abiotic factors on hillslope stability [Sidle and
Bogaard, 2016]. Geophysical imaging methods, which are less invasive and less costly, may be well suited for
these efforts [Jayawickreme et al., 2014].
We know of no other studies that compared multiyear variations in FoS under different vegetation types
using field observations of soil moisture. Nevertheless, several authors using a modeling framework or
shorter field monitoring found similar trends: (1) FoS improved under woody plants compared to herbaceous
plants or bare soil [Arnone et al., 2016; Rahardjo et al., 2014; Simon and Collison, 2002], (2) the seasonality of
FoS is driven by precipitation and/or evapotranspiration [Arnone et al., 2016; Bittelli et al., 2012; Damiano
et al., 2012; Harris et al., 2012], (3) mechanical cohesion decreases drastically with depth, becoming negligible
[Arnone et al., 2016; Hwang et al., 2015; Ji et al., 2012] or still-providing critical cohesion during low-FoS per-
iods [Chirico et al., 2013; Simon and Collison, 2002], and (4) fluctuations in hydric cohesion are smaller in sandy
soils relative to silty/clayey soils [Arnone et al., 2016; Chiu et al., 2012; Lepore et al., 2013; Matsushi and
Matsukura, 2007]. Regarding (2), highly seasonal rainfall (>90% of precipitation during the 8-month-long
wet seasons) produced large intra-annual variability in the stability of our Costa Rican slopes with and without
trees but not on the Laotian slope with trees (Figure 2). Moreover, despite the less pronounced rainfall
seasonality (Figure S4), the stability of French slopes with and without trees exhibited distinct seasonality,
indicating that evaporative demand (France) or rainfall patterns (Costa Rica) may enhance seasonality of
FoS, while vegetation (Laos-teak) may dampen it.
Current socioeconomic drivers of land-use changes in our study areas such as the booming Vietnamese fur-
niture industry [Rerkasem et al., 2009], premiums for coffee with environmental certifications [Le Coq et al.,
2011], and lack of viable alternatives for cropping on steep landscapes continue increasing pressures on hill-
slopes. We showed that these drivers can make slopes more susceptible to landslides, which negatively affect
hydropower generation (Laos and Costa Rica) [Melendez Marin, 2010], recreational hiking or skiing (France)
[de Jong et al., 2014], or flow of commerce and people (Laos) [Sidle and Ziegler, 2012]. Landslide risks from
land-use changes are compounded by increases in both susceptibility (likelihood of failure) and exposure
(presence of people or infrastructure), as documented for coffee production and ski slope development
Acknowledgments [de Jong et al., 2014; García-Rodríguez et al., 2008]. Alternative land uses that result in a more permanent
This work was supported by Agence vegetation cover may help abate landslide risks. These may include longer-rotating rubber or mixed native
Nationale de la Recherche, France
fruit tree plantations compared to the relatively short rotation of teak in Laos, where harvests of the first-
(Ecosfix ANR-10-STRA-003-001) and a
European Commission’s Marie Curie generation large-scale teak plantations are starting, or higher shade tree densities with less pruning in
fellowship (FP7-PEOPLE-2013-IEF- Costa Rica. Policies incentivizing such land-use changes should also consider potential consequences on pro-
626334). We thank the Ecosfix team for
vision and delivery of other services and products [Band et al., 2012; Kim et al., 2016].
their field assistance and also land-
owners and staff who provided access
to sites or logistical support: UMR AMAP,
Montpellier (France) the Scientific
Partnership Platform on Agroforestry
References
Systems with Perennial Crop led by Allen, R. G., L. S. Pereira, D. Raes, and M. Smith (1998), Crop Evapotranspiration-Guidelines for Computing Crop Water Requirements-Irrigation
CIRAD and CATIE in Turrialba (Costa and Drainage Paper 56, pp. 1–6541, Food and Agriculture Organization of the United Nations, Rome.
Rica) and the SOERE F-ORE-T supported Anagnostopoulos, G. G., S. Fatichi, and P. Burlando (2015), An advanced process-based distributed model for the investigation of
by Ecofor, Allenvi, and the French rainfall-induced landslides: The effect of process representation and boundary conditions, Water Resour. Res., 51, 7501–7523, doi:10.1002/
national research infrastructure (http:// 2015WR016909.
www.anaee-france.fr/fr/); and Houay Arnone, E., D. Caracciolo, L. Noto, F. Preti, and R. Bras (2016), Modeling the hydrological and mechanical effect of roots on shallow landslides,
Pano Watershed research station in Ban Water Resour. Res., 52, 8590–8612, doi:10.1002/2015WR018227.
Lak Sip, Luang Prabang (Laos) managed Bai, F., and S. Liu (2012), Measurement of the shear strength of an expansive soil by combining a filter paper method and direct shear tests,
by International Water Management Geotech. Test. J., 35(3), 451–459.
Institute and LTER-Europe. Data sup- Band, L. E., T. Hwang, T. Hales, J. Vose, and C. Ford (2012), Ecosystem processes at the watershed scale: Mapping and modeling
porting our analyses can be found at ecohydrological controls of landslides, Geomorphology, 137(1), 159–167.
https://doi.pangaea.de/10.1594/ Basu, D., A. Misra, and A. J. Puppala (2014), Sustainability and geotechnical engineering: Perspectives and review, Can. Geotech. J., 52(1),
PANGAEA.868491. The authors declare 96–113.
no competing financial interests. Bishop, A. W. (1960), The Principles of Effective Stress, pp. 1–15, Norges Geotekniske Institutt, Oslo.

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4904

 Geophysical Research Letters 10.1002/2017GL073174

Bittelli, M., R. Valentino, F. Salvatorelli, and P. R. Pisa (2012), Monitoring soil-water and displacement conditions leading to landslide
occurrence in partially saturated clays, Geomorphology, 173, 161–173.
Blatz, J. A., N. J. Ferreira, and J. Graham (2004), Effects of near-surface environmental conditions on instability of an unsaturated soil slope,
Can. Geotech. J., 41(6), 1111–1126.
Bray, J. D., and R. B. Sancio (2006), Assessment of the liquefaction susceptibility of fine-grained soils, J. Geotech. Geoenviron. Eng., 132(9),
1165–1177.
Chirico, G. B., M. Borga, P. Tarolli, R. Rigon, and F. Preti (2013), Role of vegetation on slope stability under transient unsaturated conditions,
Prog. Environ. Sci., 19, 932–941.
Chiu, C., W. Yan, and K.-V. Yuen (2012), Reliability analysis of soil–water characteristics curve and its application to slope stability analysis,
Eng. Geol., 135, 83–91.
Collins, B. D., and D. Znidarcic (2004), Stability analyses of rainfall induced landslides, J. Geotech. Geoenviron. Eng., 130(4), 362–372.
Cosby, B., G. Hornberger, R. Clapp, and T. Ginn (1984), A statistical exploration of the relationships of soil moisture characteristics to the
physical properties of soils, Water Resour. Res., 20(6), 682–690.
Dai, F., C. Lee, and Y. Y. Ngai (2002), Landslide risk assessment and management: An overview, Eng. Geol., 64(1), 65–87.
Damiano, E., L. Olivares, and L. Picarelli (2012), Steep-slope monitoring in unsaturated pyroclastic soils, Eng. Geol., 137, 1–12.
de Jong, C., G. Carletti, and F. Previtali (2014), Assessing impacts of climate change, ski slope, snow and hydraulic engineering on slope
stability in ski resorts (French and Italian Alps), in Engineering Geology for Society and Territory, edited by G. Lollino et al., pp. 51–55,
Springer, Switzerland.
Defossez, E., B. Courbaud, O. Lasbouygues, K. Schiffers, and G. Kunstler (2015), Are variations of direct and indirect plant interactions along a
climatic gradient dependent on species’ strategies? An experiment on tree seedlings, Oikos, 125(5), 708–717.
Dumedah, G., and P. Coulibaly (2011), Evaluation of statistical methods for infilling missing values in high-resolution soil moisture data,
J. Hydrol., 400(1), 95–102.
Dunne, T., W. Zhang, and B. F. Aubry (1991), Effects of rainfall, vegetation, and microtopography on infiltration and runoff, Water Resour. Res.,
27(9), 2271–2285.
Fan, C.-C., and C.-F. Su (2008), Role of roots in the shear strength of root-reinforced soils with high moisture content, Ecol. Eng., 33(2),
157–166.
Fattet, M., Y. Fu, M. Ghestem, W. Ma, M. Foulonneau, J. Nespoulous, Y. Le Bissonnais, and A. Stokes (2011), Effects of vegetation type on soil
resistance to erosion: Relationship between aggregate stability and shear strength, Catena, 87(1), 60–69.
Ferrara, C., P. M. Barone, and L. Salvati (2015), Unravelling landslide risk: Soil susceptibility, agro-forest systems and the socio-economic
profile of rural communities in Italy, Soil Use Manag., 31(2), 290–298.
Fredlund, D. G., N. R. Morgenstern, and R. Widger (1978), The shear strength of unsaturated soils, Can. Geotech. J., 15(3), 313–321.
Gale, M., and D. Grigal (1987), Vertical root distributions of northern tree species in relation to successional status, Can. J. For. Res., 17(8),
829–834.
García-Rodríguez, M. J., J. Malpica, B. Benito, and M. Díaz (2008), Susceptibility assessment of earthquake-triggered landslides in El Salvador
using logistic regression, Geomorphology, 95(3), 172–191.
Genet, M., A. Stokes, F. Salin, S. B. Mickovski, T. Fourcaud, J.-F. Dumail, and R. van Beek (2005), The influence of cellulose content on tensile
strength in tree roots, Plant and Soil, 278(1–2), 1–9.
Genet, M., A. Stokes, T. Fourcaud, and J. E. Norris (2010), The influence of plant diversity on slope stability in a moist evergreen deciduous
forest, Ecol. Eng., 36(3), 265–275.
Ghestem, M., R. C. Sidle, and A. Stokes (2011), The influence of plant root systems on subsurface flow: Implications for slope stability,
BioScience, 61(11), 869–879.
Glade, T. (2003), Landslide occurrence as a response to land use change: A review of evidence from New Zealand, Catena, 51(3), 297–314.
Godt, J. W., R. L. Baum, and N. Lu (2009), Landsliding in partially saturated materials, Geophys. Res. Lett., 36, L02403, doi:10.1029/
2008GL035996.
Gonzalez-Ollauri, A., and S. B. Mickovski (2017), Plant-soil reinforcement response under different soil hydrological regimes, Geoderma, 285,
141–150.
Gray, D. H., and A. T. Leiser (1982), Biotechnical Slope Protection and Erosion Control, pp. 1–271, Van Nostrand Reinhold, New York.
Greenway, D. (1987), Vegetation and slope stability, in Slope Stability: Geotechnical Engineering and Geomorphology, edited by M. Anderson
and K. Richards, pp. 187–230, John Wiley, Chichester, U. K.
Guiang, E. S. (2001), Impacts and effectiveness of logging bans in natural forests: Philippines, in Forest Out of Bounds: Impacts and
Effectiveness of Logging Bans in Natural Forests in Asia-Pacific, edited by P.B. Durst et al., pp. 103–135, Agricultural Organization, Bangkok,
Thailand.
Hales, T. C., and C. F. Miniat (2016), Soil moisture causes dynamic adjustments to root reinforcement that reduce slope stability, Earth Surf.
Process. Landf., 42, 803–813.
Han, K., H. Rahardjo, and B. Broms (1995), Effect of hysteresis on the shear strength of a residual soil, paper presented at The Proceedings of
the 1st International Conference on Unsaturated Soils (UNSAT 95), Paris, France, 6–8 Sept.
Harris, S., R. Orense, and K. Itoh (2012), Back analyses of rainfall-induced slope failure in Northland Allochthon formation, Landslides, 9(3),
349–356.
Hawke, R., and J. McConchie (2011), In situ measurement of soil moisture and pore-water pressures in an “incipient” landslide: Lake Tutira,
New Zealand, J. Environ. Manage., 92(2), 266–274.
Hong, Y., R. Adler, and G. Huffman (2007), Use of satellite remote sensing data in the mapping of global landslide susceptibility, Nat. Hazards,
43(2), 245–256.
Hubble, T., B. Docker, and I. Rutherfurd (2010), The role of riparian trees in maintaining riverbank stability: A review of Australian experience
and practice, Ecol. Eng., 36(3), 292–304.
Hwang, T., L. E. Band, T. Hales, C. F. Miniat, J. M. Vose, P. V. Bolstad, B. Miles, and K. Price (2015), Simulating vegetation controls on
hurricane-induced shallow landslides with a distributed ecohydrological model, J. Geophys. Res. Biogeo., 120, 361–378, doi:10.1002/
2014JG002824.
Jayawickreme, D. H., E. G. Jobbágy, and R. B. Jackson (2014), Geophysical subsurface imaging for ecological applications, New Phytol., 201(4),
1170–1175.
Ji, J., N. Kokutse, M. Genet, T. Fourcaud, and Z. Zhang (2012), Effect of spatial variation of tree root characteristics on slope stability. A case
study on Black Locust (Robinia pseudoacacia) and Arborvitae (Platycladus orientalis) stands on the Loess Plateau, China, Catena, 92,
139–154.

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4905

 Geophysical Research Letters 10.1002/2017GL073174

Kim, J. H., and R. B. Jackson (2012), A global analysis of groundwater recharge for vegetation, climate, and soils, Vadose Zone J., 11(1).
Kim, J. H., E. G. Jobbágy, and R. B. Jackson (2016), Trade-offs in water and carbon ecosystem services with land-use changes in grasslands,
Ecol. Appl., 26(8), 1633–1644.
Kirschbaum, D., T. Stanley, and J. Simmons (2015), A dynamic landslide hazard assessment system for Central America and Hispaniola,
Nat. Hazards Earth Syst. Sci., 15(10), 2257–2272.
Kuriakose, S. L., and L. van Beek (2011), Plant root strength and slope stability, in Encyclopedia of Agrophysics, edited by J. Gliński, J. Horabik,
and J. Lipiec, pp. 622–627, Springer, Dordrecht, Netherlands.
Le Coq, J., G. Soto, G. Hernández, C. Rapidel, F. DeCleerck, J. Beer, A. Varsa, S. Suazo, and P. Pacheco (2011), PES and Eco-Label: A comparative
analysis of their limits and opportunities to foster environmental services provision. Ecosystem services from agriculture and agroforestry:
Measurement and payment, in Ecosystem Services From Agriculture and Agroforestry: Measurement and Payment, edited by B. Rapidel et al.,
pp. 237–264, Earthscan Publications, London, U. K.
Lepore, C., E. Arnone, L. Noto, G. Sivandran, and R. Bras (2013), Physically based modeling of rainfall-triggered landslides: A case study in the
Luquillo forest, Puerto Rico, Hydrol. Earth Syst. Sci., 17, 3371–3387.
Li, A., Z. Yue, L. Tham, and C. Lee (2003), Soil moisture and suction measurement and its effect on slope stability, Chin. J. Geotech. Eng., 25(3),
278–282.
Liang, T., and J. Knappett (2015), Centrifuge modelling of vegetated slopes under earthquake loading, paper presented at 6th International
Conference on Earthquake Geotechnical Engineering, Christchurch, New Zealand, 1-4 November 2015.
Lu, N., J. W. Godt, and D. T. Wu (2010), A closed-form equation for effective stress in unsaturated soil, Water Resour. Res., 46, W05515,
doi:10.1029/2009WR008646.
Mao, Z., L. Saint-André, M. Genet, F.-X. Mine, C. Jourdan, H. Rey, B. Courbaud, and A. Stokes (2012), Engineering ecological protection against
landslides in diverse mountain forests: Choosing cohesion models, Ecol. Eng., 45, 55–69.
Mao, Z., Y. Wang, C. Jourdan, L. Cécillon, J. Nespoulous, H. Rey, L. Saint-André, and A. Stokes (2015), Characterizing above-and
belowground carbon partitioning in forest trees along an altitudinal gradient using area-based indicators, Arct. Antarct. Alp. Res.,
47(1), 59–69.
Matsushi, Y., and Y. Matsukura (2006), Cohesion of unsaturated residual soils as a function of volumetric water content, Bull. Eng. Geol.
Environ., 65(4), 449–455.
Matsushi, Y., and Y. Matsukura (2007), Rainfall thresholds for shallow landsliding derived from pressure-head monitoring: Cases with
permeable and impermeable bedrocks in Boso Peninsula, Japan, Earth Surf. Process. Landf., 32(9), 1308–1322.
McGuire, L. A., F. K. Rengers, J. W. Kean, J. A. Coe, B. B. Mirus, R. L. Baum, and J. W. Godt (2016), Elucidating the role of vegetation in the
initiation of rainfall-induced shallow landslides: Insights from an extreme rainfall event in the Colorado Front Range, Geophys. Res. Lett., 43,
9084–9092, doi:10.1002/2016GL070741.
Melendez Marin, L. (2010), Plan de Manejo de Cuenca del Río Pirrís, Instituto Costarricense de Electricidad, San José, CR.
Mickey, W. (1971), Strong motion response spectra, Seismol. Res. Lett., 42(1), 5–8.
National Oceanic Atmospheric Administration (2011), National Climatic Data Center, National Environmental Satellite, Data and information
Service. [Available at http://www.ncdc.noaa.gov/oa/ncdc.html.]
National Research Council (2006), Geological and Geotechnical Engineering in the New Millennium: Opportunities for Research and
Technological Innovation. Rep. 128034928X, pp. 1–222, The National Academies Press, Washington, D. C.
Newmark, N. M. (1965), Effects of earthquakes on dams and embankments, Geotechnique, 15, 139–160.
O’Loughlin, C., and R. R. Ziemer (1982), The importance of root strength and deterioration rates upon edaphic stability in steepland forests,
paper presented at Proceedings of International Union of Forest Research Organizations Workshop P.1.07-00 Ecology of Subalpine
Ecosystems as a Key to Management, Oregon State University, Corvallis, Oregon, Corvallis, Oregon, 2–3 Aug.
Pache, G., D. Joud, N. Magnan, and R. Michalet (2003), Direct and indirect effects of shade on four forest tree seedlings in the French Alps,
Ecology, 84(10), 2741–2750.
Panagos, P., M. Van Liedekerke, A. Jones, and L. Montanarella (2012), European Soil Data Centre: Response to European policy support and
public data requirements, Land Use Policy, 29(2), 329–338.
Peduzzi, P. (2010), Landslides and vegetation cover in the 2005 North Pakistan earthquake: A GIS and statistical quantitative approach,
Nat. Hazards Earth Syst. Sci., 10, 623–640.
Peel, M. C., B. L. Finlayson, and T. A. McMahon (2007), Updated world map of the Köppen-Geiger climate classification, Hydrol. Earth Syst. Sci.
Discuss., 4(2), 439–473.
Petley, D. (2012), Global patterns of loss of life from landslides, Geology, 40(10), 927–930.
Phien-Wej, N., P. Nutalaya, Z. Aung, and T. Zhibin (1993), Catastrophic landslides and debris flows in Thailand, Bull. Int. Assoc. Eng. Geol., 48(1),
93–100.
Prieto, I., C. Roumet, R. Cardinael, C. Dupraz, C. Jourdan, J. H. Kim, J. L. Maeght, Z. Mao, A. Pierret, and N. Portillo (2015), Root functional
parameters along a land-use gradient: Evidence of a community-level economics spectrum, J. Ecol., 103(2), 361–373.
Rahardjo, H., A. Satyanaga, E. Leong, V. Santoso, and Y. Ng (2014), Performance of an instrumented slope covered with shrubs and
deep-rooted grass, Soils Found., 54(3), 417–425.
Rawls, W. J., D. L. Brakensiek, and K. E. Saxton (1982), Estimation of Soil-water Properties, Trans. ASABE, 25(5), 1316–1320.
Rerkasem, K., N. Yimyam, and B. Rerkasem (2009), Land use transformation in the mountainous mainland Southeast Asia region and the role
of indigenous knowledge and skills in forest management, For. Ecol. Manage., 257(10), 2035–2043.
Rinaldi, M., and N. Casagli (1999), Stability of streambanks formed in partially saturated soils and effects of negative pore water pressures:
The Sieve River (Italy), Geomorphology, 26(4), 253–277.
Ruiz-Peinado, R., M. del Rio, and G. Montero (2011), New models for estimating the carbon sink capacity of Spanish softwood species,
Forest Syst., 20(1), 176–188.
Schuppener, B., J. Bohac, and M. Dysli (Eds.) (1999), Laboratory Methods for Direct Shear Tests, pp. 87–92, Beuth Verlag, Berlin.
Schuster, R. L., and L. Highland (2001), Socioeconomic and Environmental Impacts of Landslides in the Western Hemisphere, pp. 1–886, United
States Geological Survey Denver, Colo.
Schwarz, M., P. Lehmann, and D. Or (2010), Quantifying lateral root reinforcement in steep slopes–from a bundle of roots to tree stands,
Earth Surf. Process. Landf., 35(3), 354–367.
Seed, H. B., and I. M. Idriss (1982), Ground Motions and Soil Liquefaction During Earthquakes, vol. 5, pp. 1–134, Earthquake Engineering
Research Institute, Berkeley, Calif.
Sidle, R. C., and T. A. Bogaard (2016), Dynamic earth system and ecological controls of rainfall-initiated landslides, Earth Sci. Rev., 159,
275–291.

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4906

 Geophysical Research Letters 10.1002/2017GL073174

Sidle, R. C., and A. D. Ziegler (2012), The dilemma of mountain roads, Nat. Geosci., 5(7), 437–438.
Sidle, R. C., A. D. Ziegler, J. N. Negishi, A. R. Nik, R. Siew, and F. Turkelboom (2006), Erosion processes in steep terrain—truths, myths, and
uncertainties related to forest management in Southeast Asia, For. Ecol. Manage., 224(1), 199–225.
Simon, A., and A. J. Collison (2002), Quantifying the mechanical and hydrologic effects of riparian vegetation on streambank stability,
Earth Surf. Process. Landf., 27(5), 527–546.
Simon, A., N. Pollen, and E. Langendoen (2006), Influence of two woody riparian species on critical conditions for streambank stability: Upper
Truckee River, California, J. Am. Water Resour. Assoc., 42(1), 99–113.
Simon, A., A. Curini, S. E. Darby, and E. J. Langendoen (2000), Bank and near-bank processes in an incised channel, Geomorphology, 35(3),
193–217.
Springman, S., A. Thielen, P. Kienzler, and S. Friedel (2013), A long-term field study for the investigation of rainfall-induced landslides,
Geotechnique, 63(14), 1177.
Stokes, A., C. Atger, A. G. Bengough, T. Fourcaud, and R. C. Sidle (2009), Desirable plant root traits for protecting natural and engineered
slopes against landslides, Plant Soil, 324(1–2), 1–30.
Stokes, A., G. B. Douglas, T. Fourcaud, F. Giadrossich, C. Gillies, T. Hubble, J. H. Kim, K. W. Loades, Z. Mao, and I. R. McIvor (2014), Ecological
mitigation of hillslope instability: Ten key issues facing researchers and practitioners, Plant Soil, 377(1–2), 1–23.
Terlien, M. T. (1997), Hydrological landslide triggering in ash-covered slopes of Manizales (Columbia), Geomorphology, 20(1), 165–175.
Thomas, R. E., and N. Pollen-Bankhead (2010), Modeling root-reinforcement with a fiber-bundle model and Monte Carlo simulation,
Ecol. Eng., 36(1), 47–61.
Thornthwaite, C., and J. Mather (1957), Instructions and Tables for Computing Potential Evapotranspiration and the Water Balance, 5th Printing,
pp. 1–311, CW Thornthwaite Associates, Laboratory of Climatology, Elmer, NJ, USA, 10(3).
U.S. Army Corps of Engineers (2003), Engneering and Design: Slope Stability, EM 1110-2-1902, pp. 1–205, Department of the Army,
Washington, D. C.
United States Geological Survey (2016), Earthquake archives. [Available at http://earthquake.usgs.gov/earthquakes/map/.]
van Asch, T. W., J. Buma, and L. P. Van Beek (1999), A view on some hydrological triggering systems in landslides, Geomorphology, 30(1),
25–32.
Vanapalli, S., D. Fredlund, D. Pufahl, and A. Clifton (1996), Model for the prediction of shear strength with respect to soil suction, Can. Geotech.
J., 33(3), 379–392.
Waldron, L. (1977), The shear resistance of root-permeated homogeneous and stratified soil, Soil Sci. Soc. Am. J., 41(5), 843–849.
Wu, T. H. (2013), Root reinforcement of soil: Review of analytical models, test results, and applications to design, Can. Geotech. J., 50(3),
259–274.
Wu, T. H., W. P. McKinnell III, and D. N. Swanston (1979), Strength of tree roots and landslides on Prince of Wales Island, Alaska, Can. Geotech.
J., 16(1), 19–33.
Xu, J., E. Katsigris, and T. A. White (2002), Implementing the Natural Forest Protection Program and the Sloping Land Conversion Program:
Lessons and Policy Recommendations, pp. 1–340, China Forestry Publishing House, Beijing.
Zianis, D., P. Muukkonen, R. Mäkipää, and M. Mencuccini (2005), Biomass and stem volume equations for tree species in Europe, Silva Fennica
Monographs, 4, 1–63.
Zydron, T., and P. Mietus (2016), Shear strength of unsaturated soils from vicinity of Gorlice, Acta Sci. Pol. Form. Circum., 15(1), 151.
Zydron, T., T. Zaleski, and D. Janik (2016), Influence of moisture content and shearing rate on shear strength of silty soils from the
neighbourhood of Kotlina Sądecka, Acta Sci. Pol. Form. Circum., 15(1), 165.

KIM ET AL. VARIABLE STABILITY OF VEGETATED SLOPES 4907