Underwater Slope Stabilization

Thaismara Garcia
Southern Illinois University Edwardsville Geotechnical Engineering Master of Science student
1042 Enclave Blvd, Edwardsville, IL; thgarci@siue.edu

ABSTRACT
The underwater analysis came into importance because more and more companies are
trying to exploit gas and fuel undersea. For the companies it is vital to know the risks
that the zone where they decide to make this kind of explorations have. The stability
of the zone where they are going to work can save money according to the risk. The
present paper includes information about the principal failure mechanism that affect
underwater slopes, how the slope can be built, the required test for the analysis, the
data recollection equipment necessary and the stability analysis under static and
pseudo-static condition. Additional, two case histories are presented. One of them
consists of a natural slope, and one consists of a slope created by dredging.
INTRODUCTION
The study of underwater slopes has increased with interest in offshore facilities and the
interest of the protection communities and facilities near the sea. Common offshore
facilities included the ones that are used for the extraction of resources such as oil and
natural gas, underwater roads, and electrical lines.
The stability analysis for underwater slopes is an adaptation of the analysis that is made
to above water slopes, adding elements to make them suitable for the underwater slopes
like the buoyant or submerged unit weight of the soil.
Various studies demonstrate that the odds of failure for underwater slopes due only by
gravity alone is almost unlikely. They tend to fail due to the reduction of the strength
of the material that is caused by the reduction of the effective stress and excess on pore
water pressures. These conditions can be the following results of an external event that
make the slope to fail. For example, earthquakes increase the stresses on the soil by the
seismic acceleration and reduce the shear strength of the soil, rapid deposit of
sediments increases the total stresses at a rate too fast for the excess of pore water
pressure to dissipate, consequently causing the slope to fail, the gas hydrates (ice-like
substance consisting in natural gas that are stable under specifics pressure and
temperature conditions (Locat and Lee, 2002)) under change of pressure or temperature
disassociate into natural gas that can trigger the slope to fail, by creating voids that

Page 1
induce reduce the effective stresses and shear strength of the soil, create excess pore
air pressures and excess pore water pressures. Telling this, considering that underwater
slopes are most likely to fail under a trigger mechanism. FIG. 1 shows the trigger
mechanism more common for submarine slopes. The database created by Hauce in
2003, shows that the most likely trigger mechanism is Earthquake and faulting,
following by rapid sedimentation and gas hydrates disassociation.

FIG. 1. Triggering mechanisms for 534 submarine slides (Hauce, 2003).

This paper includes hints about the construction, study, and analysis of underwater
slopes. For the stability analysis, the assumption of an infinite slope is made. The
stability should be made under static and pseudo-static conditions, to take into
consideration different possible trigger mechanism. Additionally, the present of two
cases of study is presented, involving a dredge made slope and a natural slope.

CONSTRUCTION OF AN UNDERWATER SLOPE

Direct hydraulic placement is the most used method for the construction of underwater
slopes. It can be done using a hopper dredge or bottom dump barge (FIG.2), or when
the water is shallow, alternative methods are used, such as pipeline discharge (FIG.4)
or subaerial rainbow discharge (FIG.3) (Hamidi et al., 2015). Each hydraulic placement
method has different grades of precision and achieves different density levels. Hicks et
al. (1985) mention that the pipeline discharge was used in the Nerlerk Berm case study
to produce slopes as steep as 1:4, and it helps to improve the material density and
gradation separation fines from coarser materials.

Page 2
 FIG. 2. Subaqueous discharge by hopper or bottom dump barge (Lee et al.,
1999)

FIG. 3. Subaerial rainbow discharge (Lee et al., 1999)

FIG. 4. Pipeline discharge redrawn (Lee et al., 1999).

PROPERTIES RECOLLECTION
The data recollection for an underwater slope came with uncertainties due to
difficulties and cost. The geometry of the slope is collected using devices such as echo
or sonar sounding and multibeam bathymetric data. Samples are recollected using
sediment sample collection methods. For instance, the Kullenberg corer is a single-
drive, wire-deployed piston corer that is dropped into the sediment from a short

Page 3
distance, propelled by the momentum of the heavy lead weights on the core head. This
device can recollect samples between 2 and 350m water depth and recollect 12m of
sediment. FIG.5 shows the process of sediment collection and a photo of the device
itself. (Kullenberg, 1947) (Kelts et al., 1986)

FIG. 5. Kullenberg Corer Sediment sample collection process

(http://c2fn.dt.insu.cnrs.fr/spip/spip.php?rubrique49, Access Nov 1, 2019)

STABILITY ANALYSIS OF UNDERWATER SLOPES

Underwater slope stability analysis is a straightforward analysis addressing the use of
buoyant or submerged unit weight and total stress. Two different approaches should be
made during the analysis that includes static and pseudo-static conditions. For the static
condition, it should be considered the normally and underconsolidated state. These
approaches are made according to the failure mechanism that could be a trigger for the
slope to fail. For example, during rapid sedimentation or earthquake, the influence on
the slope is fast and does not leave time for the soil to drain. During these events, it
should be considered the analysis under normally consolidated undrained conditions.
Faulting mechanism tends to increase the inclination of a slope progressively, and the
slope tends to be loaded at a rate slow enough to dissipate the excess pore water
pressures. In this case, the stability analysis should address the drained shear strength.
The equations described in the following sections applied to infinite slopes with fully
underwater conditions. These equations are the results of Morgenstern's (1967)
investigation of submarine slides.

Stability Analysis Under Static Condition

Normally Consolidated Soils
Many underwater soils tend to be normally consolidated. FIG.6 shows the forces that
are imposed on submerge infinite slopes.

Page 4
FIG. 6. Driving and resisting forces on a submerged infinite slope (Hance, 2003)

For undrained conditions, the shear strength is express in terms of total stresses
(equation 1).

𝑠 = 𝑐 + 𝜎 𝑡𝑎𝑛 𝜑 (1)

Underwater soils are saturated, so the friction angle is zero. Subsequently, the
undrained shear strength (𝑠𝑢 ) is equal to the cohesion (𝑐). The factor of safety under
undrained condition can be expressed as:

𝑠 𝑠
𝐹 = 𝜏 = 𝛾′ 𝑧 cos𝑢𝛽 sin 𝛽 (2)

Where 𝛽 is the slope angle, z is the depth below the seafloor, and γ’ is the submerged
or buoyant unit weight.
The undrained shear strength is often normalized with the in situ effective overburden
stress (𝜎𝑣′ ) so 𝑠𝑢 /𝜎𝑣′ , called c/p ratio. The in situ effective vertical overburden stress
(𝜎𝑣′ ) can be approximated by the product of submerged unit weight (γ’) and the vertical
depth (z) the equation before can be written as:

𝑐/𝑝
𝐹 = cos 𝛽 sin 𝛽 (3)

For drained conditions, the shear strength is expressed in terms of effective stresses
(equation 4)

𝑠 = 𝑐′ + 𝜎′ 𝑡𝑎𝑛 𝜙 ′ (4)

Page 5
Morgenstern (1967) and Skempton (1970) noted that many marine sediments are fine-
grained, normally consolidated to slightly overconsolidated, and the cohesion (𝑐′) for
this type of soils is usually zero. For an infinite slope with no water flow and cohesion
of zero, the factor of safety under drained condition can be expressed as:

𝑡𝑎𝑛 𝜙′
𝐹= (5)
𝑡𝑎𝑛 𝛽

Based on this equation is expected that failures occur when the slope angle (β) is higher
than the effective friction angle (𝜙 ′ ).

Under-consolidated conditions
A soil is under-consolidated when the soil is still in the process of consolidation or
has not yet reached an equilibrium state under a previously applied load when new
load is applied. This condition usually is related to the rapid sedimentation deposit
and tends to generate excess pore water pressures (ū) in soils. The effective cohesion
(𝑐′) is assumed zero, and the shear strength along the slip surface(s) was evaluated in
terms of effective stresses. In this case, the factor of safety (F) can be expressed as:

𝑠 (𝛾′ 𝑧 𝑐𝑜𝑠2 𝛽−𝑢

̅) 𝑡𝑎𝑛 𝜙′
𝐹= = (6)
𝜏 𝛾′ 𝑧 𝑐𝑜𝑠 𝛽 𝑠𝑖𝑛 𝛽

Pseudo-Static Condition Analysis

Earthquakes are the most common trigger mechanism for slope failure. A pseudo-static
approach should be made to include the acceleration caused by earthquake movements.
FIG.7 shows the forces considered for this analysis. It adds a submerged weight of the
sliding mass (W’) and a horizontal static force (kW) due to earthquake loading.

FIG. 7. Schematic of driving and resisting forces present in a pseudo-static slope

stability analysis on a submerged infinite slope (Hance, 2003)

Page 6
The undrained shear strength should be used for this analysis because most of the
sediments are not drain immediately during seismic load. The factor of safety can be
expressed as:

𝑠 𝑠𝑢
𝐹 = 𝜏 = 𝛾′ 𝑧 𝑐𝑜𝑠 𝛽 𝑠𝑖𝑛 𝛽+𝑘𝛾′𝑧 (7)
𝑐𝑜𝑠2 𝛽

In terms of c/p ratio and the seismic coefficient (k), the equation can be rewritten as:

𝑐/𝑝
𝐹= 𝛾 (8)
𝑐𝑜𝑠2 𝛽[𝑡𝑎𝑛𝛽+𝑘( ⁄ ′ )]
𝛾

CASE STUDY
Case Study I- Nerlerk Berm Failure (Hicks et al., 1998)
The Nerlerk Berm is located on the Beaufort Sea, Canada. The berm was created to be
used as a platform for hydrocarbon exploration. The construction of the berm starts in
July of 1982, at it supposed to finish after two years of construction. FIG. 8 shows the
preliminary design of the berm. Under the berm, a clay layer of 1m thick, over a
granular floor compose by sand and silty sand was encounter. The clay layer could not
be removed due to dredging limitations.

FIG. 8 – Geometry Design of Nerlerk Berm (Hicks et al., 1998)

The construction began in mid-august of 1982. By October, 3 million cubic meters of

sand had been placed. During the construction, the slopes were lower than expected
(1:8 to 1:10). Later investigations show that the Nerlerk sand contains between 5 to 8%
more fines content than the Ukalerk sand use for the core of the berm. It was expected
that the excavation and the placement would improve material quality.

Page 7
The construction restarts in July of 1983. Before the end of the construction, five slides
on different locations and times where register, causing millions of dollars lost, which
cause the abandon of the construction. The slides caused a displacement of 410 000
m³. The failure occurred very quickly and involved entirely the sand placed in 1983.
The changes were estimated using remote sensing, and there were changes of 5 to 12
m of height, and there were changes in slope inclination to 1:30 to 1:50 (FIG 9).

FIG. 9. Section Through the Berm (Hicks et al., 1998)

The berm was analyzed using the program MONICA (MONot incremental Computer
Algorithm), which is the finite element code which encompasses Monot. For the monot
model, the parameters of the soil needed to be calibrated. Table 1 shows a summary of
the monot material parameter used for the analysis.

Table 1 – Summary of Monot material parameters (Hicks et al., 1998)

Model Monot
Parameter Description Nerlerk sand
Component parameters
Poisson’s ratio V 0.1
Non-Linear
A 0.00042
Elastic Stiffness scalar and curvature
AP 0·3
Plastic B 0·00081
Stiffness scalar and curvature
Compressive BP 0.55
E 0.08
Stiffness scalar and curvature EP 2.6
Plastic
LB 0·0
Deviatoric
Yield surface curvature DP 1·0
Constant volume friction angle FICV 31·1°
0·59
Failure Peak friction angle scalar and C
(φ’=31·1°)
surface curvature CP
1·0

The berm was modeled according to the dimension at slide 4 location:

Page 8
 • Depth of clay layer 1.5m
• height of 26m
• upper berm slope 1:4
• lower berm slope 1:9.
The layer of sand was not modeled. Two hundred ninety-nine elements elongated
quadrilaterals to save computer storage and triangular shape where put on the
boundaries with 180 degrees as the internal angle to have three nodes instead of 5
(FIG.10). The following boundary conditions were made:
• Fixed mesh base
• Rollers allowing only vertical movement along the left and right mesh edges
• Free drainage in all exposed soils
The initial stresses have been generated simulating the actual construction process,
adding 2m thick row of elements, and assuming drained conditions and a buoyant unit
weight of 9.4 KN/m³. For each analysis, the berm was loaded by increasing the gravity
in an extra row of elements to simulate the application of a further 2m of surcharge.

FIG. 10. Plane Strain idealization of Nerlerk slope: (a) finite element mesh
details; (b) distribution of fill and foundation materials (Hicks et al., 1998)

The analyses were made under different undrained shear strength:

• Cu=0.1 σ′v and Cu=0.2σ′v (the effective vertical stress in the clay layer due to
the 1982 fill)
• A constant cohesive strength of 10 kPa
The influence of the clay strength was analyzed under undrained and drained
conditions. An additional analysis was made to study the possible construction-induced
pore water pressures failure.
Results
• For the drained conditions, the analysis shows that at the absence of the
generation pore pressure, the berm is stable, discharging this failure possibility.
Page 9
 • The possible construction-induced pore water pressures analysis shows that
under the undrained condition, the construction rate should be 2m/3min and the
drained condition is 2m/5h. The actual construction rate was 2m/day. Thus, the
possibility that the berm had failed for the rapid dumping of the material was
discarded.
• Under undrained condition for the lowest value of cohesive strength FIG. 11
shows the failed zone occur in the clay layer; the movement was originating in
the clay layer near the toe of the Ukalerk fill. FIG. 12 shows the failed region
of the Nerlerk Berm.

FIG. 11. Vector and contour of shear strain invariant (0.02) for Cu=0.1 σ’v
(Hicks et al., 1998)

FIG. 12. Vector and contour of mobilized shear stress ratio (0.1) for Cu=0.1 σ’v
(Hicks et al., 1998)

It was concluded that the failure happens due to an earthquake that induces the
movement in the clay layer triggering that way the possible liquefaction and
subsequently fails of the upper Nerlerk fill.

Case Study II – Upper Continental Vesterålen Slides

The Lofoten-Vesterålen passive continental margin is located west of Andoya,
Norway. In 2010, the Norwegian Petroleum Directorate estimated that 202 million
standard m³ of oil equivalent could be found in the area. The area is characterized by
a deep canyon and numerous slides and so the risk associated with the slope stability
in this zone needed to be studied to determine the possibility of failures the oil
exploration could cause. FIG.13 shows the study area characterized by high seismic
activities (earthquakes with magnitudes close to 6 on the Richter scale) and small
isolated slides.

Page 10
FIG. 13. The Lofoten-Vesterålen continental margin (study area in a rectangle)

The profile consists of two main soil types: a sandy clay with Ip of 15-32% and silty
clay with Ip around 25%. Table 2 summarizes the soil parameters. The undrained
shear strength was obtained from the fall cone, torvane, anisotropically consolidated
triaxial compression (CAUC), and direct shear (DSS) tests. The oedometer test on
three silty clay samples indicated that the sediments were normally-consolidated. The
CAUC and DSS shear test showed moderate strain-softening in the silty clay with
reduction of about 22% in the peak strength for a shear strain of 20%. The sandy clay
showed little to no strain-softening.

Table 2 - Soil Parameters for stability analyses (based on L’Heureux et al.

2012b).
Layer γ(kN/m³) Clay w St
c/p c/p ratio α Cv c’ φ
Content ratio (CAUC)
(dss)
Sandy clay 17.7 <40% 45% 1.3 - 0.52 0.75
Laminated 18.4 50-65% 40- 3.4 0.27 0.46 0.75 3 to 5 28°
silty clay 47% 7x10- kPa
8
m²/s

FIG.14 shows the model used to analyses the slope. The surface was assumed to be
parallel to the adjacent seafloor. Two layers were defined below the seafloor: one unit

Page 11
of sandy clay and a second thinner unit of weaker laminated silty clay from a depth of
8m.

FIG. 14. – Model for slope stability analyses; vertical scale is multiplied by 10
(L’Heureux et al. 2012b)

Results
A pseudo-static analysis, dynamic analysis, and post-earthquake pore pressure
modeling were made. Table 3 summarizes the results of the slope stability analyses.
The analysis shows that the shear strength, undrained or drained, was not sufficiently
low to explain the observed failures. An external trigger was the cause, an earthquake,
for example, or high excess pore pressures. The slope stability analysis revealed that
the slope is currently stable. A 10,000 yr earthquake (PGA of 0.22g) would be needed
to cause a failure. Earthquake response analysis showed that deformations and excess
pore pressures after a strong earthquake would be accumulated mostly in the top 10 m
of the soil profile. Earthquake activity combined with pore pressure generation might
have played an important role in destabilizing the underwater slope

Table 3 – Summary of stability analyses of Vesterålen slopes

Analysis The factor of Remarks

Safety, FS
Pseudo-static analysis, Su/σ’vc FS>2 For 475 yr earthquake
=0.27 FS=1 For 10000-yr earthquake
Su/σ’vc increased to 0.35

Page 12
 1D dynamic earthquake FS=1 Accumulation of
analyses Fails in layer 11.5m deformations in top 10m,
10,000 yr earthquake
Post-earthquake slope stability FS=1 For 475-year earthquake
analysis

CONCLUSION
The probability of failure and failure mechanism of an underwater slope are studied to
analyze the risks of projects developed in the area of study.
For underwater slope stability analysis is vital to take into consideration all failure
mechanism and do an analysis both for the static and pseudo-static conditions. Most
failures could happen by a different trigger mechanism acting at the same time. For the
two cases study, earthquake was mostly the failure trigger mechanism, that agree with
the statement Hence (2003) did that the most important mechanism for a slope to fail
are earthquakes.
Like the Nerlerk Berm failure results show, the seafloor material is an important
paramenter of a slope construction Also, Soils that have strong strain-softening
characteristics and high sensitivity are most susceptible to failure during earthquake
shaking. The material the seafloor is important; it’s stability can make more stable the
slopes berm created. The material used to the construction of underwater slopes needs
to be good sand with low permeability and low content of fines. The inclination angle
seems to be an important determinant of th slope stability. Hence (2003), mention that
most of slope failures occurred in slopes with inclination greater than 10 degrees (flatter
than 10). Thus, the slope inclination should be less than 10 degrees to decrease the
possibilities of failure.
Relatively small final slope angle maximum 10 degree are reached to guarantee
stability, this results came from a study of 534 submarine slides.
Submarine slopes with overcritical slope inclination can be stable for a while when
prepared with suction dredging, Less impact on marine life
for the construction and protection mostly are made assurance the angle of the slope
and the correct dredge method used. For example, is mention that a pipeline dredge
that corresponds to high vacuum has better accuracy at the time of creating the design
slope angle.
REFERENCES
Delo, E.A., Burt, T.N., (1987). “stability of cohesive dredged slopes.”
Hamidi, B. Varaksin, S. (2015).” Dynamic Compaction and Dynamic Surcharging at
Dubai’s Palm Jumeira Sewage Treatment Plants.” Ground Improvement Case
Histories: 309-335.

Page 13
Hance, J.J. (2003). “Development of a Database and Assessment of Seafloor slope
stability based on Published Literature.” MS Thesis, Dept. Engineering, University of
Texas at Austin, Austin, TX.
Hicks, M.A., Boughrarou, R. (1998). “Finite element analysis of the Nerlerk
underwater berm failures.” J. Geótechnique 48 (2): 169-185.
Lacasse, S., Nadim, F., Vanneste, M., L’Heureux, J.S., Forsberg, C.F., Kvalstad, T.J.
(2013). “Case Studies of Offshore Slope Stability.” Geo-Congress 2013.

Page 14