Unsaturated Soil

PanAm Unsaturated Soils 2017 GSP 302 153
Rainfall Induced Shallow Slope Failure over Yazoo Clay in Mississippi

M. S. Khan1; M. Nobahar2; J. Ivoke3; and F. Amini4
1
Assistant Professor, Dept. of Civil and Environmental Engineering, Jackson State Univ., 1400
J.R. Lynch St., JSU Box 17068, Jackson, MS 39217-0168. E-mail: j00797693@jsums.edu
2
Graduate Student, Dept. of Civil and Environmental Engineering, Jackson State Univ., 1400
J.R. Lynch St., JSU Box 17068, Jackson, MS 39217-0168. E-mail:
j00816771@students.jsums.edu
Downloaded from ascelibrary.org by Tufts University on 07/01/18. Copyright ASCE. For personal use only; all rights reserved.
3
Graduate Student, Dept. of Civil and Environmental Engineering, Jackson State Univ., 1400
J.R. Lynch St., JSU Box 17068, Jackson, MS 39217-0168. E-mail: johnivokx@gmail.com
4
Professor, Dept. of Civil and Environmental Engineering, Jackson State Univ., 1400 J.R. Lynch
St., JSU Box 17068, Jackson, MS 39217-0168. E-mail: famini@jsums.edu
Abstract
Yazoo clay soil is expansive in nature, and extended over central Mississippi, Alabama and
Southern Louisiana. Shallow slope failures are frequent in embankments constructed on
expansive Yazoo clay in Mississippi due to the rainfall volume and climatic variation. The
current study investigates the effect of rainfall over the factor of safety of fill slopes constructed
with Yazoo clay soil using the finite element method (FEM). The historical rainfall pattern of
Mississippi obtained from the National Oceanic and Atmospheric Administration (NOAA), is
evaluated at different intensities (2 to 50 mm/h), and durations of rainfall (30 min to 7 days) with
the data being, based on a return periods of 50, 100, and 500 years. Three fill slopes were
constructed using Yazoo clay soil, with a slope ratios of 2H: 1V, 3H: 1V, and 4H: 1V. The
rainfall pattern is applied over the top boundary of the slope in Plaxis, using flow analysis. The
preliminary investigation results indicated that rainfall water infiltrated the top part of the slope,
near the crest, resulting in the reduction of both the matric suction and the factor of safety of the
slope. Moreover, the factor of safety of the slope is directly influenced by the total rainfall
volume, regardless of the intensity and duration.
Introduction
Rainfall-induced slope failures are reported to occur during or immediately following

periods of intense or prolonged heavy rainfall in fine grained soil (Day, R. W. 1989, Tohari et al,
2007, Harp et al. 1996; Baum and Chleborad 1999, Khan et al., 2017, Barik et al., 2017). The
unsaturated state of the soil plays a major role in protecting natural slopes from failure. With
even a very small amount of suction, the shallow soil deposit often has high factor of safety,
depending on slope morphology and the triggering of a landslide is not a common phenomenon
(Damiano and Mercogliano, 2013). Moreover, the soil covers often offer a low hydraulic
conductivity at unsaturated state that usually prevents the soil from approaching saturation. This
happens even during a high intensity rainfall period where the soil would be normally
oversaturated. Furthermore, during the dry periods between rainfall events, the soil loses part of
its water content, mainly by evapotranspiration, favored in many cases by the presence of
vegetation: consequently, only intense rainstorms occurring after prolonged rainy periods may
lead the soil to such wet conditions as to trigger slope instability (Damiano and Mercogliano,
2013).
© ASCE
PanAm Unsaturated Soils 2017

Highly plastic soil is expansive in nature and cover more than 25% of the total area of the United
States. Usually, the shear strength of expansive clay is high immediately after compaction.
However, with seasonal moisture variation, expansive soil experience cyclic shrink-swell
behavior, which eventually reduces its strength to fully soften shear strength that reduces the
factor of safety of the slope (Wright S.G., 2005). Moreover, during an intense and prolonged
rainfall period, rain water will infiltrate into the slope and create a temporary perched water
condition. This, reduces the effective stress on the soil and increases the excess pore water
pressure. The combination of the fully softened shear strength and perched water zone due to
rainfall mostly cause the shallow slope failure (Khan et al. 2015, Hossain et al., 2017).
Generally, the failure depth is usually within 3-6 ft. and the failure plane remains parallel to the
slip surface. This type of slope failure is typically referred to as shallow/surficial slope failure
(Titi and Helwany, (2007), Loehr and Bowders, (2007), Khan et al., (2016)). Surficial failures of
slopes are common throughout the United States. Shallow slope failure refers to surficial slope
instabilities along cuts, fill slopes, embankments, earth dam, and levees constructed using
marginal, to highly plastic clay soil.
This current study investigates the effect rainfall has, regarding the factor of safety, of fill slopes
constructed with Yazoo Clay soil by using the Finite Element Method (FEM). The historical
rainfall pattern of Mississippi from NOAA is evaluated at different intensities, 2 mm/ hour to 50
mm/hour, and duration of rainfall, 30 mins to 7 days, with the data based on a return period of
50, 100 and 500 years. Three fill slopes were constructed using Yazoo Clay soil, with slope
ratios of 2H: 1V, 3H: 1V and 4H: 1V. Based on the unsaturated finite element analysis results,
the variation of suction under rainfall is investigated.
Precipitation Pattern
NOAA collects the rainfall data all over the US and develops the precipitation pattern of
any locality based on the historical data (NOAA Atlas 2014). The Partial Duration Series (PDS)
based Intensity Duration (PDS) and frequency (IDF) curve of precipitation, based on the NOAA
Atlas 2014 of Jackson, Mississippi, was collected. The IDF curve of the Jackson area is
presented in Figure 1, which indicated that high intensity rainfall ranges between 9.25 inch/hour
(0.234 m/hr) (for 15 mins. duration) and 0.86 inch/hour (0.021 m/hr) (for 12 hours duration) with
a 500-year return period.
Intensity-Duration-Frequency Curve for Jackson,
MS
50 Years 100 years 500 years
Precipitation Intensity
1000
100
(mm/hr)
10
1
0.1
0.001 0.01 0.1 1 10 100
Duration (days)
Figure 1 PDS based intensity duration frequency (IDF) curve for Jackson MS.
© ASCE

Different intensities (3.09 m/day to 0.05 m/day) and duration of rainfall (30 mins to 7 days) was
selected, based on a return period of 50, 100 and 500 years, from the PDS based IDF curve of
Jackson, MS. Three fill slopes were constructed using Yazoo Clay soil with slope ratios of 2H:
1V, 3H: 1V and 4H: 1V. The FEM matrix for the selected slope inclination and rainfall
intensities are presented in Table 1.
Table 1 Selected Precipitation pattern for FEM analysis
Rainfall Intensity (m3/day/m)

Slope
Duration
Inclination 50 Years 100 Years 500 Years
30-min 3.09 3.41 4.12
60-min 2.13 2.36 2.93
2-hr 1.36 1.51 1.89
2H:1V,
6-hr 0.635 0.71 0.92
3H:1V and
12-hr 0.369 0.41 0.53
4H:1V
1 day 0.209 0.23 0.29
3 day 0.0823 0.09 0.12
7 day 0.0406 0.044 0.053
Finite Element Model

The FEM program PLAXIS 2D was used to run the flow analyses, using the flow
modeling suit. A 15-node triangular element was used, which provides a fourth order
interpolation for displacements and the numerical integration involved twelve Gauss points. This
is automatically taken to be compatible with the soil element type as selected in PLAXIS.
Undrained or short-term material behavior is selected for soil types. The Van Genuchten model
is considered as the hydraulic model. Rigid interface was used in order not have a reduced
strength with respect to the strength in the surrounding soil. The soil parameters, as shown in
Table 2, were used in the numerical analysis using PLAXIS 2D. They were established from
field investigation results. Moreover, the numerical values are considered as Max step as 1000,
tolerated error as 0.01, over-relaxation factor as 1.2, desired Min number of iterations as 6,
desired Max number of iterations as 15, Max number of iterations as 60, first, Min and Max time
step as 1.00E-9 Day which are suggested values by PLAXIS.
© ASCE

Table 2 Soil parameters for FEM analysis

Parameter Name Unit Soil Type Soil Type Soil Type Soil Type
1 (Yazoo 2 (Silty 3 (Sandy 4 (Clay)
Clay) Clay) Clay)
Bulk unit ϒunsat kN/m3 19.6 19.6 20.5 22
weight
Saturated unit ϒsat kN/m3 21 21 22 22
weight
Horizontal kx m/day 5.50E-07 5.50E-07 5.50E-07 5.50E-07
permeability
Vertical ky m/day 1.23E-05 5.50E-07 5.50E-07 5.50E-07
permeability
Residual water Sres - 0.018 0.018 0.018 0.018
content
Saturated water Ssat - 0.41 0.41 0.41 0.41
content
Van Genuchten n=gn - 2.7 2.7 2.7 2.7
fitting
parameter
Van Genuchten 𝛼=ga 1/m 2.90E-03 2.90E-03 2.90E-03 2.90E-03
fitting
parameter
Van Genuchten m=gl - 0.095 0.095 0.095 0.095
fitting
parameter
Cohesion C kN/m2 5.221 4.788 11.97 143.640
Friction angle Ф - 18.7 23 15 35
Precipitation of different intensities was applied to the soil model to assess the flow
behavior during rainfall. The analysis was carried out at three rainfall intensities. The rainfall
intensities were selected based on 50, 100, and 500-year periods of Mississippi rainfall data.
These return periods are selected due to climate change and extreme weather condition in recent
years collected from National Oceanic and Atmospheric Administration (NOAA, 2014). The
flow through the top soil was determined for each of the intensities assuming rainfall durations
lasting from: 30 min, 60 min, 2 hours, 6 hours, 12 hours, 1 day, 3 days and 7 days. The
representative soil model is presented in Figure 2. It should be noted that during the dry period,
the highly plastic clay soil developed desiccation cracks which might have significantly
increased the permeability along the vertical direction of the top soil at the active zone. Albrecht
and Benson (2001) conducted a study on the effect of desiccation on compacted clay. The study
reported that the hydraulic conductivity of clay soil increased as much as 500 times during
drying, where the largest increase in hydraulic conductivity occurred after first drying cycle.
However, due to the desiccation crack, the permeability along the horizontal direction might
have had no effect and could have remained unchanged. The permeability along all remaining
directions was uniform and considered as uniform clay. Rainfall with uniform intensity was
applied during the flow analysis, and the Van Genuchten model was used for the flow
parameters. During this study, the depth of the active zone was considered as 2.13 m (7 ft), the
area in which a desiccation crack could occur during summer period. Therefore, a high vertical
permeability value of ky = 1.063 m/day (1.23 ∗ 10−5 m/s) was used for the top 2.13 m (7 ft) of
© ASCE

the slope to simulate the effect of the desiccation crack. In other clay layers, the permeability for
both horizontal and vertical directions were selected as 0.0475 m/day (5.5 ∗ 10−7 m/s). Hence,
the Van Genuchten fitting parameters were utilized for the flow analysis are tabulated in Table 2.
45 Ft (13.7 m) 77 Ft (23.4 m) 48 Ft (14.6 m)

Point A
Soil Type 1 Infiltration
55
Ft
(1 Point B
Soil Type 2
6.
7 Point C 28.
Water Table 9 Ft
m)
Soil Type 3 (8.8
m)
Soil Type 4 Seepage
170 Ft (51.8 m)
Closed Boundary
Figure 2 The soil model and boundary condition for FEM analysis.
Finite Element Modeling Results

In the flow analysis, the left and right boundaries were selected as the closed boundaries,
and the top of the slope was selected as rainfall infiltration (Figure 2). The water table was
assumed to be positioned at 6 m (19.6 ft.), to define the initial unsaturated condition. The
variations of suction at the crest of the 2H: 1V slope are presented in Figure 3 (a) to (l) with
rainfall intensities 2 hrs, 1 day and 7 days. As indicated in Figure 3, that the suction at the top of
the slope immediately dropped not only after rainfall, but also for the three slopes and continued
to drop during rainfall, representing the accumulation of water at the corresponding depth. The
initial suction starts at 130 kPa, and continue decreasing to 120 kPa and 100 kPa after one day
and 30 days respectively. These initial suction values are validated based on water table which is
similar to the field condition. It was also observed that the suction decrease continued for a few
hours, after the rainfall had, which has taken place due to the water ponding at the top for all
three slopes (Figure 3(c), (g) and (k)). It is also noticed that after few days of rainfall, the suction
increased and almost regained its original profile of steeper slope. However, (Figure 3(l)),
virtually no suction variation occurred throughout the slope except for a small amount of
increase at the very top of the crest of the slope.
© ASCE

High suction at top before Immediate drop of suction at Drop of suction continues
rainfall top during rainfall up to 7 days
2H:1V 2H:1V 2H:1V

(a) (b) (c)
Suction profile close to initial High suction at top before Immediate drop of suction at
condition after 30 days rainfall top during rainfall
2H:1V 2H:1V 2H:1V

(d) (e) (f)
Drop of suction not interrupted Suction profile almost close High suction at top before
and it continues to initial condition rainfall
up to 7 days after 30 days
2H:1V 2H:1V 2H:1V

(g) (h) (i)
Immediate drop of suction at Drop of suction uninterrupted No suction occurs after 30
top during rainfall and it continues days
up to 7 days
2H:1V 2H:1V 2H:1V

(j) (k) (l)
Figure 3 Suction profile for 2 hrs rainfall intensity (a) prior to rainfall (b) during rainfall (1 day)
(c) 7 days after rainfall (d) 30 days after rainfall, Suction profile for 1 Day rainfall intensity (e)
prior to rainfall (f) during rainfall (1 day) (g) 7 days after rainfall (h) 30 days after rainfall and
Suction profile for 7 Days rainfall intensity (i) prior to rainfall (j) during rainfall (1 day) (k) 7
days after rainfall (l) 30 days after rainfall.
The variations for the change in the suction versus time graph is a result of the 8 rainfall
durations at 100-year return periods for each of the 3 slope ratios at point A (a) to (c) are
presented in Figure 4. It should be noted, however, that Point A is located at the crest. Moreover,
points A, B, and C are located at the surface whereas, point D and E are located at 1.5 m (5 ft.)
depth, below the top surface, and it is worth mentioning that the variation of change in suction
due to the 8 rainfall durations at 100-year return period for the three slopes at points B and C
have the same trend. The change of suction refers to the change from the initial suction value
prior to rainfall.
© ASCE

Slope 2H:1V @ Point A (17.00 m) - 100 Slope 3H:1V @ Point A (16.00 m) - 100
yrs Return Period yrs Return Period
RI-30 M RI-60 M RI-30 M RI-60 M

RI-2 HR RI-6 HR RI-2 HR RI-6 HR
RI-12HR RI-1 D RI-12HR RI-1 D
RI-3 D RI-7 D RI-3 D RI-7 D
Ѱ (kN/m2)
Ѱ (kN/m2)
130 130
80 80
30 30
-20 0 10 20 30 -20 0 10 20 30
Time (day) Time (day)
(a) (b)
Slope 4H:1V @ Point A (13.00 m) - 100 yrs Return
Period
RI-30 M RI-60 M RI-2 HR

RI-6 HR RI-12HR RI-1 D
RI-3 D RI-7 D
130
Ѱ (kN/m2)
110
90
70
50
30
10
-10 0 10 20 30
Time (day)
(c)
Figure 4 Variations of change in suction versus time for 8 rainfall durations at 100 years return
period for 3 slope ratios at point A, (a) 2H:1V, (b) 3H:1V and (C) 4H:1V.
Based on Figure 4, the change in suction was more significant at the initial period,
whereas it starts decreasing with the milder slope. The largest change in suction was observed at
the steeper slope (2H: 1V) rather than that the flatter slope (4H: 1V). For example, the matric
suction value is about 130 kPa for 2H: 1V slope, which is higher than the matric suction values
are 110 kPa and 90 kPa for 3H: 1V and 4H: 1V slope ratios respectively. As shown in the above
© ASCE

figure, the matric suction value of the crest dropped from 130 kN/m2 to 90 kN/m2 in the first day
of rainfall. The figure shows various rainfall durations with a 100-year return period. The suction
was observed that it dropped remarkably with rainfall at a higher intensity and longer duration,
but then continued with only slight changes. Moreover, the change in suction was more
significant at the crest of the slope, when compared to the middle and toe of the slope. The drop
of suction was instantaneous at deeper layers as compared to the soil Type 1 surface for different
rainfall intensities and durations, as depicted in Figure 4. In contrast, the change in suction
continued from a period of several days to several weeks post-rainfall to reach a steady value at
point A in depth. The constant values of suction at the crest of the slope (point A), indicated that
the water that is moved and filtered in porous voids (percolation theory) could not drain out from
the slope due to the low permeability of the highly plastic clay soil.
Figure 5 presents the suction profile for low (60 min), medium (1 day) and high (7 days)
rainfall intensities for the three slope ratios 2H:1V, 3H:1V and 4H:1V. The FEM analysis results
indicate that for the each of the three slope ratios, at low rainfall intensity suction dropped
immediately at the surficial soil, and then dropped uninterrupted for a rainfall intensity of 1-day.
It then continued until there was no significant suction occurrence at high rainfall intensity such
as 7 days. It should be noted that the change in suction was more prominent at a lower depth
(point A) than that of the longer period of high rainfall intensity for almost all three slope ratios.
The change in suction was not substantial at point B or C.
60min Rainfall Duration 60min Rainfall Duration 60min Rainfall Duration
2H:1V slope 3H:1V slope 4H:1V slope

(a) (b) (c)
1 day Rainfall Duration 1 day Rainfall Duration 1 day Rainfall Duration

(d) (e) (f)
7 days Rainfall Duration 7 days Rainfall Duration 7 days Rainfall Duration

(g) (h) (i)
Figure 5 Low rainfall intensity suction profile (a) 2H:1V (b) 3H:1V (c) 4H:1V, Suction profile
for medium rainfall intensity (d) 2H:1V (e) 3H:1V (f) 4H:1V and Suction profile for high rainfall
intensity (g) 2H:1V (h) 3H:1V (i) 4H:1V.
© ASCE

Conclusion
Shallow slope failure is typical and recurring in most of the highway slopes in Mississippi due to
the existence of the high plasticity index (PI) clay soil. A flow analysis for three types of slope
conditions with 24 rainfall intensity was conducted to investigate the effect of rainfall on the
suction matrix of fill slopes constructed with Yazoo Clay soil. Based on the flow analysis results,
it was observed that there is high suction at the initial moment before rainfall. With a short
rainfall duration, a decrease in the suction takes place. This decrease in suction is immediate at
the shallow surface and subsequently much faster than the deeper layers. It was also noticed that
an increase in rainfall intensity does not affect infiltration to deeper layers, due to the low
permeability of the highly plastic clay soil. The total volume of rainfall plays a major role in the
infiltration behavior of highly plastic clay soil. As a result of the rainfall data analysis, it has
been shown that successive rainfall can have a significant influence on slope failure in the
shallow depth of Yazoo Clay soil.
Reference:
Albrecht, B.A., Benson, C.H., 2001. Effect of desiccation on compacted natural clays. J.
Geotech. Geoenviron. Eng. 127 (1), 67–75.
Barik, M. G., Adam, J. C., Barber, M. E., & Muhunthan, B. (2017). Improved landslide
susceptibility prediction for sustainable forest management in an altered climate. Engineering
Geology. Online Publication.
Baum, R.L., Chleborad, A.F., and Schuster, R.L., 1998, Landslides triggered by the December
1996 and January 1997 storms in the Puget Sound area, Washington: U.S. Geological Survey
Open-file Report 98-239, 16 p.
Damiano and Mercogliano, 2013.A simulation chain for early prediction of rainfall-induced
landslides.
Day, R.W., Axten, G.W., 1989. Surficial stability of compacted clay slopes. J. Geophys. Eng. 15
(4), 577–580.
Griffiths, D.V., Lane, P.A., 1999. Slope stability analysis by finite elements. Geotechnique 49
(3), 387–403.
Harp JM, Timm DE, Bunick GJ. Macromolecular crystal annealing: overcoming increased
mosaicity associated with cryocrystallography. Acta Crystallogr D Biol Crystallogr. 1998 Jul 1;
54(Pt 4):622-8
Loehr, J. E., and Bowders, J. J. (2007). ―Slope stabilization using recycled plastic pins—Phase
III.‖ Final Rep. RI98-007D, Missouri Dept. of Transportation, Jefferson City, MO.
Khan, M. S., Hossain, S., Ahmed, A., & Faysal, M. (2017). Investigation of a shallow slope
failure on expansive clay in Texas. Engineering Geology, 219, 118-129.
Hossain, S., Khan, S., & Kibria, G. (2017). Sustainable Slope Stabilisation using Recycled
Plastic Pins. CRC Press, Boca Raton.
Khan, M. S., Hossain, S., & Kibria, G. (2015). Slope Stabilization Using Recycled Plastic Pins.
Journal of Performance of Constructed Facilities, 30(3), 04015054.
National Oceanic and Atmospheric Administration Daily Climate Report, 2014a.
http://www.nws.noaa.gov/climate/index.php?wfo=fwdN July 29, 2014.
Titi, H., Helwany, S., 2007. Investigation of vertical members to resist surficial slope
instabilities, (WHRP 07–03). Wisconsin Department of Transportation, Madison, WI.
© ASCE

Tohari, A., Nishigaki, M., & Komatsu, M. (2007). Laboratory rainfall-induced slope failure with
moisture content measurement. Journal of Geotechnical and Geoenvironmental
Engineering, 133(5), 575-587.
Wright, S.G., 2005. Evaluation of soil shear strengths for slope and retaining wall stability
analyses with emphasis on high plasticity clays (FHWA/TX-06/5–1874–01-1). Federal Highway
Administration, Washington, D.C.
© ASCE

Unsaturated Soil

Uploaded by

Copyright:

Available Formats

You might also like

Unsaturated Soil

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Unsaturated Soil

Uploaded by

Copyright:

Available Formats

PanAm Unsaturated Soils 2017 GSP 302 153

Rainfall Induced Shallow Slope Failure over Yazoo Clay in Mississippi

Rainfall-induced slope failures are reported to occur during or immediately following

PanAm Unsaturated Soils 2017

PanAm Unsaturated Soils 2017

Table 1 Selected Precipitation pattern for FEM analysis

Rainfall Intensity (m3/day/m)

Finite Element Model

PanAm Unsaturated Soils 2017

Table 2 Soil parameters for FEM analysis

PanAm Unsaturated Soils 2017

45 Ft (13.7 m) 77 Ft (23.4 m) 48 Ft (14.6 m)

Finite Element Modeling Results

PanAm Unsaturated Soils 2017

2H:1V 2H:1V 2H:1V

2H:1V 2H:1V 2H:1V

2H:1V 2H:1V 2H:1V

2H:1V 2H:1V 2H:1V

PanAm Unsaturated Soils 2017

RI-30 M RI-60 M RI-30 M RI-60 M

RI-30 M RI-60 M RI-2 HR

PanAm Unsaturated Soils 2017

60min Rainfall Duration 60min Rainfall Duration 60min Rainfall Duration

2H:1V slope 3H:1V slope 4H:1V slope

2H:1V slope 3H:1V slope 4H:1V slope

2H:1V slope 3H:1V slope 4H:1V slope

PanAm Unsaturated Soils 2017

PanAm Unsaturated Soils 2017

PanAm Unsaturated Soils 2017

You might also like