Slope Stability Then and Now: Professor Emeritus, Virginia Tech, 1600 Carlson DR., Blacksburg, VA, 24060

Geo-Congress 2013 ASCE 2013
2191
SLOPE STABILITY THEN AND NOW

J. Michael Duncan, Ph.D, PE, Distinguished M. ASCE
Downloaded from ascelibrary.org by dml u on 03/08/14. Copyright ASCE. For personal use only; all rights reserved.
Professor Emeritus, Virginia Tech, 1600 Carlson Dr., Blacksburg, VA, 24060
Abstract: Many changes in geotechnical engineering for slopes and embankments

have taken place since the 1966 and 1992 conferences on this subject (ASCE 1966.
1992). Most of the changes are due to the revolutionary influences of computers in
nearly every aspect of the field. Highly significant changes include possibilities for
very thorough and detailed evaluations of slope stability and performance, use of 3D
stability analyses, finite element and finite difference analyses of slope movements,
examination of probability of failure, in situ and laboratory measurement of shear
strength, improved tools for slope monitoring, and many new and effective methods
of slope stabilization. Although the tools we use are much more sophisticated than in
1966, experience, judgment and thorough quality control remain as important as ever.
FIRST, A LOOK BACK
Geotechnical engineering in general, and slope design in particular, has been
fortunate to have been guided by many very talented and insightful engineers and
engineering geologists who laid the groundwork on which we are still building.
Many of those who made important contributions to the 1966 and 1992 conferences
(ASCE 1966, ASCE 1992) are no longer with us. Notable among those in 1966 were:
David Henkel stressed the importance of the geological setting in stability of
natural slopes.
Juul Hvorslev emphasized fundamental aspects of clay strength, and stresses
in embankments and embankment foundations.
John Lowe described the state of the art with respect to analysis of the
stability of embankments.
Ralph Peck discussed the state of the art with respect to the stability of natural
slopes.
Harry Seed described analyses of earthquake-induced landslides, and the
stability of sloping core earth dams.
Bob Whitman discussed the fundamental mechanics of slope stability and the
use of computers for stability analysis.
Stan Wilson described instruments for measuring pore pressures and
movements in dams, and their use to investigate embankment behavior during
and following construction.
Notable among the leaders at the 1992 conference were:
Roger Foott wrote about severe cracking that had been observed in the
Sherman Island (California) levees.
Lyman Reese described a method for designing drilled shafts to stabilize
potentially unstable slopes.
George Sowers described ten case histories of landslides in natural slopes and
developed a grouping according to their predominant motion.
Geo-Congress 2013
2192
Others who showed the way in 1966 and 1992 are still active, continuing to guide
the development of the field. We are fortunate that geotechnical engineering has
benefitted from the work of so many exceptional engineers and engineering
geologists. We should strive to continue in their footsteps.
Judging by the proceedings of the 1966 conference on Stability and Performance
of Slopes and Embankments, the principal concerns of the profession at that time
were slope stability analysis, shear strength of soil, and monitoring of movements in
slopes. Those topics continued to be the subjects of many contributions to the 1992
conference, with growing interest in methods of slope reinforcement and ground
improvement. Judging by the titles of abstracts received for this conference, these
topics are still at the top of the list. Each of these will be examined here. First,
however, it is important to review the changes in computer technology which
influence every aspect of the field.
TRANSFORMATION OF THE WORK ENVIRONMENT BY COMPUTERS
The environment we work in today is very different from the way things were in
1966. In 1966, most slope stability calculations were done by hand, using a slide
rule, paper and pencil, a process unfamiliar to engineers who began their careers in
1980 or later. The alternative to hand calculations, not available to most practicing
geotechnical engineers in 1966, was use of a main frame computer. This was
revolutionary, but a far cry from the way we are able to use computers today.
Engineers had to develop their own computer programs there were no commercially
available programs. The computer program and input data, both on punched cards,
were submitted to the operators of a computer center. Alphanumeric output (no
graphics) was available hours later. It was expensive as well as inconvenient - $300
per hour for mainframe use which would be over $2,000 per hour today.
In a paper at the first conference on stability and performance of slopes and
embankments (ASCE 1966), Whitman and Bailey (1966) described what seemed like
an unimaginably convenient way to use these mainframe computers. They wrote:
Let us begin by imagining how we might wish to perform slope stability
analyses using a computerAn engineer, with experience in the design and
analysis of slopes, seats himself in front of a keyboard
Whitman and Bailey suggested that besides the keyboard, the engineer would
have a monitor, a light pen, and a pointing device so that graphical as well as
numerical input would be possible. Despite how visionary this was, and how
incredibly convenient it seemed at the time, reality has now exceeded the fondest
hopes of 1966. Neither Whitman and Bailey nor anyone else could anticipate that
computers could be so small, so powerful, and so convenient as they are now.
Consider the contrast between then and now: The IBM 7094 mainframe
computer at Berkeley in 1966 was capable of performing 3.6 x 108 floating point
calculations in an hour, which, as noted above, cost the equivalent of $2,000 today. A
laptop computer, available today for about $500, can perform the same number of
floating point calculations in 0.11 seconds, at negligible cost.
Geo-Congress 2013
2193
Wide availability of this computer power has transformed the environment in

which we work. Computations can be done so rapidly that we become impatient if a
process requires as long as a minute. Data can be transmitted at the speed of light
over the internet, making it possible to deliver a document with intricately formatted
text and color graphics anywhere in the world almost instantaneously. In addition,
storage of data has become incredibly efficient. One 4 GB flash drive, weighing 30
grams, holds as much information as 27,000 boxes of punch cards weighing a total of
196 tons.
EVOLUTION OF COMPUTER PROGRAMS FOR SLOPE STABILITY
When mainframe computers became available for use in civil engineering, efforts
began to develop computer programs to make use of their power, among them
programs for slope stability analysis. Before commercial programs were available,
the programs were written by the engineers who used them graduate students and
practicing engineers who had a strong interest in being able to use computers as a
means of learning about slope stability analysis, or who wanted to use the computer
to enhance productivity. Most of the programs were written in FORTRAN, and
FORTRAN was taught to engineering students (including civil engineering students)
at many universities. When PCs with FORTRAN compilers became available in the
early 1980s, some of these programs were adapted for use on personal computers.
Kai Wong, working at Berkeley, adapted several computer programs that had been
developed by Steve Wright and Guy Lefebvre during their doctoral and post-doctoral
studies. These batch-process programs were used by a number of firms and
geotechnical engineers for a few years. The programs were crude by todays
standards, with typed data files for input and only alphanumeric output, but they were
used because there were no better alternatives. The developments at Berkeley were
undoubtedly paralleled at many universities, and a large number of slope stability
programs were in use at that time, each with its own small group of users.
In the three decades since the advent of personal computers, geotechnical
engineers have teamed with professional programmers (or have themselves become
sophisticated programmers) to develop commercial programs that take advantage of
the speed and capacity of the powerful desktop and laptop computers now available.
These present-generation programs provide great analytical power, easy-to-use
graphical interfaces, computer-aided drafting and word processing features within the
stability analysis programs. Used in conjunction with color printing and essentially
instantaneous document transmission over the internet, they provide a range of
capabilities not even dreamed about in the mid-1980s, let alone in 1966. Whitman
and Bailey were indeed visionary when they described in 1966 how we might wish
to perform slope stability analyses using a computer, but reality now exceeds that
vision by a wide margin. Today we take for granted what no one could imagine in
1966. It behooves us to use these powerful tools carefully, with a thorough
understanding of soil mechanics principles, careful judgment, and a high level of
quality control to ensure that results are correct and meaningful.
Many commercially available computer programs for soil slope stability are now
available. A few are listed in Table 1, but many more can be found by searching
Geo-Congress 2013
2194
online. These programs have a wide variety of the features needed to make them
useful and efficient for practical application. My personal experience is limited to a
few Clara-W for 3D analyses, GSTABL7 with STEDwin, SLIDE, SLOPE/W, and
UTEXAS. An appreciable amount of time is needed to become familiar with any of
these programs and efficient in their use, because each has its own style and features.
Selecting from among available programs the one or two best suited to a particular
engineers needs can be difficult. It would be very helpful if a committee of
professionals, or a graduate student supervised by one or more experienced engineers,
would undertake a thorough review of some of the available programs, and present
the results in a Consumer Reports-like format, to assist engineers faced with selection
of programs for analysis of slope stability.
TABLE 1. A Sampling of commercially available

computer programs for soil slope stability analysis
CHASM
9 SLIDE
Clara-W (2D and 3D)
10 Slope 2000
GALENA
11 SLOPE/W
Geo-Tec B
12 Stable for Windows
GGU-Stability
13 STABLPRO
GSLOPE
14 SVSLOPE (2D and 3D)
GSTABL7 with STEDwin
15 TSLOPE (2D and 3D)
8 LimitState:GEO
16 UTEXAS3
Links to these programs and others can be found at geotechnicaldirectory.com
Each program has a web page that describes capabilities and availability.
Validity of the results of an analysis is of course the responsibility of the user.
No matter which program is used, it is possible that the results will be incorrect if the
program is applied incorrectly. With complex programs that have a great number of
options, this is always possible. Steve Wright recommends that analyses should be
done using two or more computer programs, because the computations are so
complex that it is virtually impossible to check results thoroughly in any other way.
During the investigation of flood wall failures in New Orleans after Hurricane
Katrina, Tom Brandon, Steve Wright, Ron Wahl and Noah Vroman and I used both
SLIDE and UTEXAS to perform analyses and confirm results. In the process, we
found that neither program had the features necessary to perform analyses for the
condition where a crack was formed down the back of a floodwall by hydraulic
fracturing. Later investigation showed that SLOPE/W also had to be modified for
this previously unforeseen condition. All three programs have now been modified to
deal with this type of loading correctly. Had the analyses been performed using a
single program, the results might have been erroneously judged correct, with serious
consequences for the results of the investigation.
Geo-Congress 2013
2195
One of my most interesting discoveries while writing this paper has been the
potential of spreadsheets for complex analyses, including searching for critical
circular and non-circular slip surface, and probabilistic analyses using the Hasofer
Lind method and Monte Carlo simulations (Low 2003; Low and Tang 2004, 2007).
Such analyses require that the user be adept at Visual Basic programming to describe
complex conditions, and are thus not suitable for general use. However, it is clear
that spreadsheets have great power, and may provide computational engines for future
generations of user-friendly programs.
Fast computers with large memory and storage have made possible types of
analyses that were not possible using hand calculation. Among these are finite
element analyses, three-dimensional analyses, and probabilistic analyses, discussed in
the following paragraphs.
FINITE ELEMENT ANALYSES
Strength Reduction Method of Stability Analysis. Slope stability analyses can be
performed using the finite element method or the finite difference method, by means
of the strength reduction technique described by Griffiths and Lane (1999). The
soil is modeled as an elastic-perfectly plastic material. Gravity stresses are
established within the finite element model of the slope by the gravity turn-on
method, where the full force of gravity is applied in one increment. The factor of
safety against slope instability is determined by gradually reducing the strength of the
soil. As this is done, deformations occur, and at some value of the strength reduction
factor, the deformations become very large, and the numerical process will no longer
converge. This numerical instability is interpreted as physical instability, and the
value of the strength reduction factor when this happens is the factor of safety the
factor by which the strength must be reduced to bring the slope to a state of barely
stable equilibrium. This is the same definition of factor of safety used in limit
equilibrium analyses of slope stability.
An example using the strength reduction method for a hypothetical slope is
shown in Figure 1, after Griffiths and Lane (1999). The example shows a rapid
increase in displacement as the strength reduction factor approached 1.38, the
minimum factor of safety determined by limit equilibrium analysis. The two sets of
strength reduction data show the results with and without a compressible foundation
layer. The nodal point displacement vectors at the right side of Figure 1 show a
failure mechanism that is shaped much like a circular rupture surface.
While it may seem like overkill to use finite element analyses to calculate the
factor of safety against instability of a slope, the strength reduction method does have
one significant advantage as compared to limit equilibrium analysis: It is not
necessary to specify the position or shape of the slip surface. As shown subsequently,
this is particularly advantageous for 3D analyses.
Incremental construction of the slope can be modeled. However, with assumed
linear elastic soil stress-strain properties, the stresses computed by gravity turn-on or
incremental construction are the same. The calculated displacements are not the
same, however. If meaningful displacements as well as factor of safety are desired,
Geo-Congress 2013
2196
Dimensionless Displacement,
E'max/H2
the analysis should model the actual sequence of fill placement or excavation
involved in construction of the slope.
Rigid Foundation
Nodal point displacement vectors

SRF = 1.4
Compressible
Foundation
Limit Equilibrium
F = 1.38
4
0.8
1.2
1.4
Strength Reduction Factor
FIG. 1. Variation of displacement magnitude with strength reduction

factor and nodal point displacement vectors for a hypothetical slope.
(After Griffiths and Lane, 1999)
A number of commercially available finite element programs provide options for
slope stability analysis using the strength reduction method, as well as for analysis of
stresses and movements in earth masses and soil-structure interaction problems.
Among these are PLAXIS, SIGMA/W, Phase2 and a number of academic programs.
FLAC 7, a finite difference program, also has this capability. These and many other
finite element programs can be found through geotechnicaldirectory.com.
Finite Element Analysis of Progressive Failure. Finite element analysis can also
be used to study the possibility of progressive failure. As shown in Figure 2, from
Filz, et al. (2001), shearing resistance of a geomembrane liner beneath a landfill
increases with displacement up to a peak, then decreases with further displacement.
FIG. 2. Progressive failure of waste impoundment on a geomembrane liner.

(Filz, et al. 2001)
With this type of shearing resistance, it is not possible to mobilize the peak
strength simultaneously at all points along the liner because displacements vary from
point to point and with time during placement of the waste fill. By means of finite
element analyses that modeled the nonlinear shear stress-displacement curves from
Geo-Congress 2013
2197
laboratory interface shear tests, Filz et al. were able to track progressive development
of failure as shear occurred between the waste fill and the liner, and at interfaces
within the composite liner. Applying this method to the Kettleman Hills landfill, it
was concluded that the average shearing resistance at the time when the waste fill
became unstable would have been considerably less than the peak shearing resistance,
and only very slightly larger than the residual shearing resistance.
Finite Element Analysis of Reinforced Slopes. If the stress-strain properties used
in finite element analyses are realistic, and the sequence of events involved in
construction and evolution of the slope are modeled realistically, the calculated
movements can be expected to be meaningful approximations of the actual
movements in the field. Such analyses can be useful in a number of ways, including
modeling the performance of reinforced slopes. Limit equilibrium analyses can be
used to determine the increase in factor of safety of a slope achievable with a
reinforcement force of a given magnitude, but they do not show what force will
actually be developed in the reinforcement. Unless the reinforcement is prestressed,
the force in reinforcement depends entirely on the movements of the slope, and the
consequent strains that occur in the reinforcement. Determining the magnitude of the
force that will develop under working conditions requires a soil-structure interaction
analysis finite element or finite difference.
A soil-structure interaction analysis was performed as part of the design studies
for Mohicanville Dike No. 2, shown in Figure 3. The dike, which was constructed on
a very weak foundation, was reinforced with a steel reinforcing mat at the bottom of
the embankment (Franks, et al., 1988). As the embankment was constructed on top of
the steel mat, and settled on the weak foundation, the embankment and foundation
tended to spread laterally, inducing tension in the reinforcement. The resulting
reinforcement force stabilized the embankment. Figure 4 shows the calculated and
measured tensile forces in the steel reinforcing mat.
FIG 3. Mohicanville Dike No. 2, reinforced with a steel mat at the base of the
embankment. (Franks, et al., 1988)
Geo-Congress 2013
2198
FIG. 4. Mohicanville Dike No. 2 reinforcement forces.

(Franks, et al., 1988)
3D ANALYSES OF STABILITY
3D analyses are needed where a three-dimensional failure mechanism may have
a lower factor of safety than the usual 2D failure mechanism. An example is the
Kettleman Hills waste landfill, shown in Figure 5. A slope failure in 1988 (Seed, et
al. 1990) occurred by sliding along weak interfaces in the composite liner that
separated the waste from the ground beneath. Figure 6 shows factors of safety
FIG. 5. Plan view of Kettleman Hills waste landfill

(Seed, et al. 1990)
Geo-Congress 2013
2199
computed for several cross sections through the waste fill. Normally, for design, only
the maximum section would be analyzed, because in most cases the maximum section
has the lowest factor of safety. However, this is not true for the Kettleman Hills
landfill geometry. It can be seen that the factors of safety for sections K, J, and H are
lower than for the maximum section, B or G. The failure mechanism was threedimensional and the position of the failure surface was known, with the mass moving
as shown by the arrows in Figure 5, by sliding within the relatively weak lining. The
actual value of factor of safety, F=1.0 at failure, is not represented by any single 2D
section, and computing it requires consideration of the equilibrium of the mass as a
whole. Seed, et al. (1990) performed what they described as a multiblock analysis,
considering the equilibrium of various blocks within the mass, and their interaction.
This approach resulted in a factor of safety equal to 0.96, close enough to unity to
confirm that the three-dimensional interaction between the various section of the
mass were an important factor in this case.
FIG. 6. Factors of safety for six 2D sections through the

Kettleman Hills waste landfill (Seed, et al. 1990)
Figures 7 and 8 show an example of a 3D finite element analysis used to
determine the factor of safety of a curved section of the Wolf Creek Dam
embankment slope by the strength reduction method. The Wolf Creek Dam
embankment wraps around the end of the concrete section of the dam, as shown in
Figure 7. The slope at the end of the embankment is curved, and a 2D stability
analysis does not represent the conditions there. To evaluate the effect of the
curvature of the slope on the factor of safety, Jeremic (2010) used the strength
reduction method for a 3D finite element analysis of the wraparound section. The 3D
finite element mesh shown in Figure 8 was used to model the curved section of the
embankment, and a plane strain mesh (shown grey) was used to model the straight
section.
Geo-Congress 2013
2200
FIG. 7. Wolf Creek Dam embankment.
FIG. 8. 3D finite element mesh for straight and wraparound

sections of the Wolf Creek Dam embankment.
Comparison of the critical strength reduction factors for the straight and the
wraparound sections showed that the factor of safety for the wraparound section was
only 6 percent lower than for the straight section, indicating that the curvature of the
embankment in the wrap-around section did not have a significant effect on stability.
The advantage of the strength reduction method was that it was unnecessary to
describe and analyze a 3D slip surface, which is problematical unless the shape of the
slip surface is known a priori, as at Kettleman Hills.
PROBABILISTIC ANALYSES
Probabilistic analyses provide a useful supplement to factors of safety against
slope instability. Because these analyses reflect the effects of uncertainties in the
information used in stability analyses, they provide a viewpoint on safety that can be
helpful to geotechnical engineers and their clients. Computing both factor of safety
and probability of failure is better than computing either one alone. Much experience
Geo-Congress 2013
2201
is available regarding acceptable factors of safety for slopes, and this experience is
useful for judging safety. Estimating probability of failure (or its counterpart,
reliability) provides a measure of the uncertainty involved in factor of safety
calculations, which enhances our ability to judge their significance.
Vanmarcke (1977) described the fundamental basis for these analyses, and
Baecher and Christian published their landmark book on reliability and statistics in
geotechnical engineering in 2003. As the value of probabilistic analyses for use in
practice has become evident, probabilistic methods have been applied with increasing
frequency to practical slope stability problems (Christian, et al. 1994; Duncan 2000;
Low 2003; Low et al. 2007; Low 2008).
A variety of probabilistic methods have been rendered more practical and
transparent, including the Taylor series technique (Wolff 1994; U.S. Army Corps of
Engineers 1997, 1998), the first-order reliability method (FORM), the Hasofer-Lind
method (Low, 2008), and Monte Carlo simulation, which can be applied to any
analysis done in a spreadsheet using the computer program @Risk (Palisades
Corporation, 2012). As noted earlier, these analyses have the added benefit of
showing the greatest sources of uncertainty in analyses. It is very important that all
significant sources of uncertainty be represented in the analyses.
SHEAR SRENGTH
It has often been said that the three most important things in slope stability
analysis are shear strength, shear strength and shear strength. Other things are also
important of course, but there is no possibility of correctly assessing the stability of a
slope if the shear strength is not evaluated correctly, no matter what type of analysis
is performed.
As described by Tom Brandon in his keynote paper to this conference (Brandon
2013), computers have changed testing methods in many ways. Laboratory tests are
almost always computer controlled, and both laboratory and in situ test data is almost
always collected by automatic data acquisition systems. The result is a reduction in
personnel costs, more complete data, and greater efficiency because the automated
equipment can operate around the clock. Although testing efficiency and accuracy
have improved, the usefulness of the data still depends on sample quality are the
samples representative of field conditions, and as nearly undisturbed as possible?
Along with the changes in testing methods that have occurred since the 1960s,
there is growing appreciation and improved understanding of the importance of fully
softened strength. As described by Terzaghi (1936), and by Skempton (1964, 1970,
1977) and his colleagues, strength loss occurs over years or decades in fissured clays
that is not reflected in relatively short term laboratory tests on undisturbed samples.
Kayyal and Wright (1991) found that softening also occurs in highly plastic clay
embankments where desiccation cracks develop due to cyclic wetting and drying.
The gradual softening that occurs is caused by influx of water through the fissures or
cracks. Through this process, the strength of the clay is eventually reduced to, or
close to, the normally consolidated strength. In cases where softening occurs, tests on
undisturbed samples overestimate the long-term strength in the field, and the strength
for the fully softened condition is best evaluated by performing tests on remolded
Geo-Congress 2013
2202
normally consolidated samples. The applicability, measurement and use of fully

softened shear strength are discussed in a paper to this conference (Duncan et al.
2013).
There is also a growing realization that the variation of soil shear strength with
effective normal stress is often nonlinear, and a number of slope stability analysis
computer programs provide capabilities for using curved strength envelopes in slope
stability analyses. The result is more realistic representation of soil shear strength
characteristics, especially at low confining pressures, and improved agreement
between the positions and shapes of calculated critical slip surfaces and those
observed in the field.
INSTRUMENTATION AND SLOPE MONITORING
Stan Wilson, in his paper Investigation of Embankment Performance (Wilson,
1966) listed the purposes for slope monitoring and the instruments in use at that time.
The reasons for monitoring slope performance (check behavior during construction,
monitor the performance of the completed structure, and obtain basic information for
future projects) remain the same today. What has changed significantly as a result of
new technology are the instruments being used to monitor performance, and the
methods used for collecting, storing, processing, and displaying the data they collect.
The inclinometer developed by Wilson (Wilson, 1962) is still the most frequently
used means of monitoring subsurface movements. The operation of inclinometers has
been speeded up through the use of data loggers and improvements in the hardware,
but obtaining readings is still quite labor intensive. In-place inclinometers (IPIs),
available from Durham Geo Slope Indicator (dgsi.info) and others, can be read
automatically, reducing labor costs. Shaped Accelerometer Arrays (SSAs), available
from Measurand, Inc. (measurand.com), can also be read automatically. SSAs consist
of a string of accelerometers spaced at 30 cm or 50 cm along a cable. Both IPIs and
SSAs are relatively expensive because more hardware is required, and because the
hardware must be dedicated to a single hole (Mikkelsen 2012). The advantages are
that readings can be obtained rapidly, and that labor costs are lower.
Vibrating wire piezometers, now used widely, have a number of advantages over
earlier types of piezometers. They monitor water pressures with extremely small
movement of the pressure diaphragm. As a result, they can be grouted in place and
they equilibrate almost instantaneously (Contreras, et al. 2012). The data can be
collected automatically on a data logger, or can be transmitted by cellular telephone
or other telemetry, and accessed over the internet. Engineers can thus monitor pore
water pressures in real time from their office (Mikkelsen 2012).
Surface movements can be monitored using GPS (global positioning system)
technology with horizontal accuracy of about 1 mm and vertical accuracy of about 1.5
mm (Bond and Nyren 2012; Shimizu et al. 2012); accuracies of 10 mm vertical and
15 mm horizontal are routinely available. Computerized total stations are able to
provide about the same accuracy for points as far as 500 meters apart. Surface
movements can be measured with millimeter or sub-millimeter accuracy using inSAR
(interferometric synthetic aperture radar). Figure 9 shows this equipment being used
to measure the movements within an area of the wall of an open pit mine in South
Geo-Congress 2013
2203
Africa. No targets are required for the observations, and movements smaller than a
millimeter can be detected (Little 2011).
FIG. 9. Sandsloot open pit mine in South Africa, and mobile radar
scanning antenna for inSAR measurements of slope movements.
(M. J. Little, 2011)
SLOPE STABILIZATION
In 1966 consultants and owners were reluctant to accept new technologies for
ground improvement or slope stabilization, but the climate is more open to innovation
now.
MSE walls, soil nailing, light-weight foam fill, column-supported
embankments, and reinforced embankments have been used on many projects.
Consultants and owners are often receptive to these new methods when they are
proposed by contractors they view as reliable, and when construction quality will be
thoroughly verified by a comprehensive quality control program. Computers and
instrumentation provide data in real time to verify procedures and make
comprehensive quality control possible. The climate for acceptance of new methods
is also enhanced by design-build contracts, which put the responsibility for success
of innovation in the hands of the people who will do the work (Sehn 2012).
In times past, knowledge about and adoption of innovative techniques was
usually driven by an individual or a group personally familiar with the techniques,
who could judge their suitability for a particular project. A recent SHRP2 (Strategic
Highway Research Program 2) project has made information about a wide range of
techniques more widely available by organizing a great deal of information about
ground improvement techniques (including slope stabilization) in a very convenient
computer-based form (Schaefer et al. 2012). This interactive system which contains
information for 46 ground improvement techniques, has been called "a
comprehensive web-based information and guidance system for ground
improvement." Engineers can use the information system: (1) by reviewing a
catalogue of the 46 technologies, or (2) by using an interactive system, which
identifies the technologies that are most suitable based on the users description of
project characteristics. For each technology, the system provides these types of
information:
Technology Fact Sheets

Photographs
Case Histories
Geo-Congress 2013
2204
Design Procedures
Quality Control/Quality Assurance Procedures
Cost Estimating
Specifications
Bibliography
Having these resources available more or less instantly, at the click of a mouse,
saves a great deal of research time. Final selection of a method and final design can
proceed quickly with this information readily available. Nonetheless, proper use of
this new resource still requires careful evaluation by experienced geotechnical
engineers with full appreciation of the actual conditions unique to each specific
project and problem.
The potential for efficient use of engineering resources is clear. It seems likely
that this innovative system provides a model that could be useful for geotechnical
applications beyond the transportation projects for which it was developed.
IMPROVEMENTS IN GEOTECHNICAL ENGINEERING PRACTICE
Improvements in geotechnical engineering practice do not arise automatically
from the availability and application of the advances in technology discussed in the
preceding pages. Achieving high quality in geotechnical engineering has always
required experience, judgment, and comprehensive quality control. It is my belief
that when these basic requirements are met, appropriate use of new technologies does
lead to improvements in practice.
However, if these basic requirements are not met, problems can arise in even
simple projects. The experience at Silver Lake Dam serves as an example.
The spillway capacity of the Silver Lake Dam, in the Upper Peninsula of
Michigan, was increased in 2003 to accommodate the probable maximum flood
(FERC 2003). To achieve the required discharge capacity, the fuse plug embankment
shown in Figure 10 was constructed on the rim of the reservoir. The fuse plug was
designed to erode away within about an hour when it was overtopped, providing an
additional emergency spillway 250 feet wide and five feet high. The embankment
FIG.10. Silver Lake fuse plug embankment at

end of construction, October 8, 2002
Geo-Congress 2013
2205
was designed and constructed using a standard cross section developed and tested by
the U. S. Bureau of Reclamation.
The embankment shown in Figure 10 was only five feet high. It had been
constructed using imported materials that were carefully graded and compacted. The
embankment cross section had been tested in the Bureau of Reclamation laboratory
and used successfully on other projects requiring fuse plugs. What could go wrong?
Through an error, the elevation of the fuse plug pilot channel (the elevation at
which erosion would begin) was designed too low, and the fuse plug embankment
was washed away less than year after it was constructed, by a flood far smaller than
the probable maximum flood. Furthermore, the fuse plug embankment was
constructed on an erodible sand foundation rather than non-erodible rock or concrete.
As a result, erosion continued for 25 feet below the base of the fuse plug
embankment, releasing virtually the entire reservoir. Consequences included
temporary evacuation of about 1800 people from the city of Marquette and about
$100 million in damages (ASDSO 2003). Fortunately, no lives were lost.
This case shows that even on a simple project, success can only be assured
through judgment, experience and diligent quality control. Omission of any of these
can lead to problems.
SUMMARY
Methods of geotechnical engineering for slopes have been transformed since the
first ASCE conference on stability and performance of slopes and embankments in
1966, largely because computers have changed so many aspects of geotechnical
engineering and geo-construction. While the new tools available for shear strength
evaluation, computation, communication, construction, and monitoring have
revolutionized the way we work, the need for judgment and the value of experience
have not diminished.
Computer programs for slope stability analysis have been developed that can
perform analyses and provide results in figures of professional report-quality, in very
little time. However, as pointed out many times by Steve Wright, results should not
be accepted at face value. They should be checked thoroughly. Because the
computer programs now available are so complex, it is virtually impossible to check
the results using hand calculations. The only feasible way of checking the results is
by using a second computer program to analyze the problem. Wright has shown the
value of thorough checking through many examples.
Finite element and finite difference analyses are finding use for several types of
analyses of slopes. They can be used to evaluate factors of safety against instability
through the strength reduction method, which defines the factor of safety in terms
of shear strength the same definition of safety factor used in limit equilibrium
analyses. The advantage of the strength reduction procedure is that it is not necessary
to prescribe the position or shape of the rupture surface it develops naturally as the
shear strength is reduced.
Finite element and finite difference analyses can also be used to estimate
displacements in slopes and embankments. If the stress-strain properties assigned to
the soil are realistic, and if the sequence of events involved in construction and
Geo-Congress 2013
2206
evolution of the slope are modeled realistically, the calculated movements are
meaningful approximations of the actual movements in the field. Such analyses can
be useful in a number of ways, including modeling the performance of reinforced
slopes.
Finite element and finite difference analyses are also capable of tracking changes
of stress and strain within a slope that can lead to progressive failure in strainsoftening materials. Analyses of this type showed that the shearing resistance that
could be mobilized within the liner at the Kettlemen Hills Landfill was only
marginally greater than the residual strength because of the nature of the
displacements that developed as the waste was placed in the landfill.
Finite element and finite difference analyses provide an effective means for
three-dimensional analyses of slope stability because it is unnecessary to locate the
critical rupture surface by trial and error. Analyses of this type showed that the factor
of safety for the wraparound section of the Wolf Creek Dam embankment was only
slightly lower than that for the adjacent straight section of the embankment.
The ability to perform many analyses very quickly has made probabilistic
analyses feasible for use in practice. Probabilistic analyses provide a supplement to
deterministic analyses of stability, and indicate the most important sources of
uncertainty in stability calculations. A variety of analytical methods have been
developed, including the Taylor series technique, the Hasofer-Lind method, and
Monte Carlo simulation.
Methods of measuring and representing the shear strength of soils have
advanced, notably through greater appreciation of the importance of fully softened
strength, and the use of curved shear strength envelopes to characterize soil strengths
more accurately. Laboratory tests to evaluate shear strength can be performed more
quickly, and with less manual labor, using new computer-controlled equipment and
data acquisition systems. Although testing efficiency and accuracy have improved,
the usefulness of the data still depends on selecting or preparing samples that are
representative of field conditions and that are as little affected by disturbance as
possible. In situ tests have been made more accurate and efficient through use of
electronic data acquisition systems.
Field monitoring has been made more efficient and more accurate through use of
computer-controlled instruments. Computerized total stations, GPS, and inSAR
provide position data with accuracy as refined as 1 mm horizontally and 1.5 mm
vertically, with very little labor required for the measurements. Vibrating wire
piezometers provide fast response, accurate pressure readings, and capabilities for
automatic recording and transmission of data.
Innovative techniques for slope stabilization and ground improvement are now
used widely, aided by more receptive consultants and owners, more thorough quality
control procedures, and more frequent use of design-build contracts.
A
comprehensive web-based information and guidance system for ground
improvement developed under a SHRP2 research project provides practically
immediate access to information on 46 ground improvement procedures (Schaefer, et
al. 2012). This online system is a model for efficient use of geotechnical engineering
knowledge that has potential for use in other areas of geotechnical engineering.
Geo-Congress 2013
2207
Computers have revolutionized geotechnical engineering for slopes in many

ways, freeing engineers to focus on the issues that control stability and performance.
When results are achievable so quickly, and when they are presented in such
polished-looking forms, it is more important than ever to remember that the most
important thing to know about any piece of information is that it may be wrong.
Geotechnical engineers need to polish their quality control skills to ensure that the
power of computers improves the quality of their work.
ACKNOWLEDGEMENTS
The writer is indebted to many people who have provided valuable assistance
with preparation of this paper, and wise counsel over many years. Matthew Sleep,
who was a post-doctoral student at Virginia Tech when this paper was written, was
very helpful in the literature review. Al Sehn helped the writer to understand the state
of practice in geo-construction, and the reasons that the profession is more open to
innovation than it was in 1966. Erik Mikkelsen helped the writer understand how the
instruments for measuring movements and pore pressures have been changed by
computer technology, and how these new tools have led to more efficient methods for
monitoring slope performance. Robert Werner was very helpful in explaining the use
of shaped accelerometer arrays in in-place inclinometers, and by providing a very
instructive example of the use of this new technology. George Filz, Vern Schaefer,
and Jim Mitchell gave generously of their time to provide access to, and explain the
function of the web-based information and guidance system for ground improvement.
They were also very helpful by providing their views on the state of the art and state
of practice in slope stabilization and ground improvement, and on the state of practice
for slope stability. Jim Mitchell also made other suggestions that extended the scope
and improved the accuracy of the paper. Tom Brandon very generously provided
photographs of modern equipment for laboratory and in situ testing for the San Diego
lecture on this subject, and explained how the state of practice in strength evaluation
is being changed by this new equipment. Bak Kong Low made me aware of the
potential of spreadsheets for very complex deterministic and probabilistic analyses.
Steve Wright, Kai Wong, Tom Brandon, Dan VandenBerge and Garry Gregory have
given generously of their time to provide their thoughts and recollections of the
evolution of slope stability analysis, and the capabilities of some of the computer
programs that are currently available. Pat Lucia helped me to consider the
relationship between improvements in technology and improvements in practice, and
encouraged me to address this topic directly. All of these talented, capable and
generous colleagues have contributed greatly to this paper. This assistance is much
appreciated and is gratefully acknowledged.
Geo-Congress 2013
2208
REFERENCES
ASCE (1966). Stability and Performance of Slopes and Embankments, Proceedings
of an ASCE Specialty Conference, Berkeley, California.
ASCE (1992) Stability and Performance of Slopes and Embankments-II, Proceedings
of an ASCE Specialty Conference, Berkeley, California, Geotechnical Special
Publication No. 31.
ASDSO (2003) Association of State Dam Safety Officials Water Webster News
Summary 11/10/2003, Community Partners, 1000 North Alameda Street, Suite
240, Los Angeles, CA 90012.
Baecher, G.B. and Christian, J.T. (2003). Reliability and Statistics in Geotechnical
Engineering. Wiley: Chichester, UK; Hoboken, NJ, 618 p.
Baecher, G.B. and Christian, J.T. (2008). Spatial variability and geotechnical
reliability, Chapter 2 (pages 76-133) of Reliability-Based Design in Geotechnical
Engineering- Computations and Applications, Taylor and Francis, ed. K.K.Phoon.
Bond, J. and Nyren, R. (2012). Remote monitoring of deformations using
Differential Global Positioning System (D-GPS, Geotechnical News, Vol. 30, No.
2, p. 28, June 2012.
Christian, J.T., Ladd, C.C. and Baecher, G.B. (1994). Reliability applied to slope
stability analysis. J. Geotech. Eng. ASCE, 120(12): 2180-2207.
Contreras, I. A., Grosser, A. T., and Ver Strate, R. H. (2012). Update of the fullygrouted method for piezometers installation, Geotechnical News, Vol. 30, No. 2,
p. 20-25, June 2012.
Duncan, J. M. and Buchignani, A. L., (1973), "Failure of an Underwater Slope in San
Francisco Bay," Journal of the Soil Mechanics and Foundations Division, ASCE, 99,
No. SM9, pp. 687-703.
Duncan, J. M., (2013) Impacts of Time on the Performance of Reinforced Slopes,

Proceeding of the ASCE Geo-Congress, San Diego.
Duncan, J.M. (2000). Factors of safety and reliability in geotechnical engineering.
J. Geotechnical and Geoenvironmental Engrg. 126(4): 307-316.
Duncan, J. M., VandenBerge, D. R., and Brandon, T. L. (2013) Fully Softened Strength of
Natural and Compacted Clays for Slope Stability, Proceeding of the ASCE Geo-
Congress, San Diego.

Fell, R., Hungr, O., Leroueil, S. and Riemer, W. (2000). Keynote Lecture - Geotechnical
engineering of the stability of natural slopes, and cuts and fills in soil, GeoEng 2000,
Vol.1, Invited Papers, Technomic Publishing, Lancaster, pp.21-120, ISBN: 1-58716067-6, November 2000.
FERC Independent Consultant Review Panel (2003) Technical Reasons for the
Release of Silver Lake Reservoir on May 14-15, 2003, Report to FERC Division
of Dam Safety and Inspections, Washington, DC.
Filz, G. M., Esterhuizen, J. J. B., and Duncan, J. M., (2001). Progressive Failure of
Lined Waste Impoundments, Journal of Geotechnical and Geoenvironmental
Engineering, Vol. 127, No. 10, October, 2001.
Franks, L. W., Duncan, J. M., Collins S. A., Fowler, J., Peters, J. F., and Schaefer V.
R. (1988). "Use of reinforcement at Mohicanville Dike No. 2." Proceedings of
the Second International Conference on Case Histories in Geotechnical
Engineering, St. Louis, Missouri.
Geo-Congress 2013
2209
Griffiths, D. V., and Lane, P. A., (1999), Slope Stability Analysis by Finite Elements,
Geotechnique 49, No. 3, pp. 387-403.
Jeremic, B. (2010) Comparison of 2D vs 3D Slope stability for the Wolf Creek Dam,
Report to the Nashville District, USACE.
Kayyal, M. K., and Wright, S. G. (1991). Investigation of long-term properties of
Paris and Beaumont Clays in earth embankments, Center for Transportation
Research, University of Texas at Austin, Austin, 134
Little, M. J. (2011). Slope Monitoring Strategy at PPRust Open Pit Operation,
International Symposium on Stability of Rock Slopes in Open Pit Mining and Civil
Engineering, Vancouver, pp. 211-230.
Low, B.K. (2003). Practical probabilistic slope stability analysis. Proceedings, Soil
and Rock America 2003, Cambridge, Massachusetts, USA, June 22-26, 2003,
Verlag Glckauf GmbH Essen, Vol. 2, 2777-2784.
Low, B.K. (2008). Practical reliability approach using spreadsheet, Chapter 3 (pages
134-168) of Reliability-Based Design in Geotechnical Engineering-Computations
and Applications, Taylor and Francis, ed. K.K.Phoon.
Low, B. K., and Duncan, J. M., (2013) Testing bias and parametric uncertainty in
analyses of a slope failure in San Francisco Bay mud, Proceeding of the ASCE
Geo-Congress, San Diego.
Low, B.K., Lacasse, S. and Nadim, F. (2007). Slope reliability analysis accounting
for spatial variation. Georisk: Assessment and Management of Risk for
Engineered Systems and Geohazards, Taylor & Francis, London, 1(4): 177-189.
Low, B.K., and Wilson H. Tang (2004). Reliability analysis using object-oriented
constrained optimization. Structural Safety, Elsevier, Amsterdam, 26(1), 69-89.
Low, B.K. and Wilson H. Tang (2007). Efficient spreadsheet algorithm for firstorder reliability method. J. Engineering Mechanics, ASCE, 133(12): 1378-1387.
Mikkelsen, P. E. (2012). Personal communication.
Palisades Corporation (2012) @Risk computer program for Monte Carlo Simulation,
Ithaca, New York, online at @Risk.
Schaefer, V. R., Mitchell, J. K., Berg, R. R., Filz, G. M., and Douglas, S. C., (2012),
Ground improvement in the 21st century: a comprehensive web-based information
system, Geotechnical Special Publication No. 226, ASCE Geo-Institute.
Seed, R. B., Mitchell, J. K., and Seed, H. B. (1990), Kettleman Hills Waste Landfill
Slope Failure, II: Stability analyses, Journal of Geotechnical Engineering, Vol.
116, No. 4, April, 1990.
Sehn, A. L. (2012). Personal communication.
Shimizu, N., Masunari, T., and Iwasaki, T. (2012). GPS displacement monitoring
system for the precise measuring of rock movements, Proceedings of the 12th
ISRM International Congress on Rock Mechanics, pp. 1117-1120.
Skempton, A. W. (1964). Long-term stability of clay slopes, Gotechnique, 14(2),
77-102.
Skempton, A. W. (1970).
First-time slides in over-consolidated clays,
Geotechnique, 20(3), 320324.
Skempton, A. W. (1977). Slope stability of cuttings in brown London clay,
Proceedings of the 9th International Conference on Soil Mechanics and
Foundation Engineering, 3, 261-270.
Geo-Congress 2013
2210
Terzaghi, K. (1936). Stability of slopes of natural clay, Proceedings of the 1st

International Conference of Soil Mechanics and Foundations, 161-165.
U.S. Army Corps of Engineers. (1998). Risk-based analysis in geotechnical
engineering for support of planning studies. Engrg. CircularNo. 1110-2-554,
Department of the Army, Washington, D.C., ^www.usace.army.mil/usace-docs&
(27 Feb. 1998).
Vanmarcke, E.H. (1977). Reliability of earth slopes. J. Geotech. Eng. ASCE,
103(11): 1247-1266.
Whitman, R. V., and Bailey, W. A. (1967). "Use of computers for slope stability
analyses." ASCE, Journal of the Soil Mechanics and Foundations Division,
Vol. 93, No. SM4, July, 475-498. [Also in Stability and Performance of
Slopes and Embankments, Proceedings of an ASCE Specialty Conference,
Berkeley, California, August 22-26, 1966, pp. 519-542.]
Wilson, S. D. (1962). The Use of Slope Measuring Devices to Determine
Movements im Earth Masses, Field Testing of Soils, ASTM STP 322. pp. 187197.
Wilson, S. D. (1966). Investigation of Embankment Performance, Proceedings of
the ASCE Conference on Stability and Performance of Slopes and Embankments,
pp. 153-177.
Wolff, T. F. (1994). Evaluating the reliability of existing levees. Report prepared
for U.S. Army Engineer Waterways Experiment Station Geotechnical Laboratory,
Vicksburg, Miss.
Wright, S. G. (2013). Application of Computer Software for Slope Stability
Computations, ASCE Geo-Congress, San Diego.
Geo-Congress 2013

Slope Stability Then and Now: Professor Emeritus, Virginia Tech, 1600 Carlson DR., Blacksburg, VA, 24060

Uploaded by

Copyright:

Available Formats

Slope Stability Then and Now: Professor Emeritus, Virginia Tech, 1600 Carlson DR., Blacksburg, VA, 24060

Uploaded by

Document Information

Original Title

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

Slope Stability Then and Now: Professor Emeritus, Virginia Tech, 1600 Carlson DR., Blacksburg, VA, 24060

Uploaded by

Copyright:

Available Formats

Geo-Congress 2013 ASCE 2013