COMPARING ERODIBILITY MEASUREMENTS FROM JET METHOD TO

THOSE FROM EXPERIMENTAL LEVEE BREACH TESTS OF COHESIVE

SOILS

Ali Asghari Tabrizi, Senior Staff Professional (PhD)1

M. Hanif Chaudhry, Associate Dean, Professor (PhD)2

ABSTRACT

Riverine floods and consequent levee failures have increased in recent years. Estimating
the embankment erodibility is a crucial step in predicting levee failure and breach
evolution. JET Erosion Test (JET) is a widely accepted method for characterizing the soil
erodibility. The erodibility coefficients obtained from the JET method can be used as an
input into breach models such as WinDAM to estimate erosion of an earthen
embankment. In this study, four homogenous, cohesive, compacted soil mixtures were
made by varying the silt and clay content. Erodibility coefficients of these samples were
first estimated using an in-house fabricated JET apparatus. Four laboratory overtopping
levee breach tests were then performed using the same soil compositions to study the
erodibility characteristics and failure process by recording the breach shape. The breach
evolution is monitored quantitatively by using a sliding-rods technique and image
processing. The water surface elevation over the breach invert was recorded using an
ultrasonic sensor device. The rate of widening, deepening, and erosion volume were used
along with the estimated applied shear stress to determine the erodibility coefficients. The
erodibility coefficients from the JET method and levee breach tests were then compared.
The erodibility coefficients from the JET method were found to be smaller than those
from the breach tests. Moreover, the experimental levee breach test results presented in
this study can be used by other researchers to validate numerical models and evaluate the
effects of cohesion on levee failure.

INTRODUCTION

Flood disasters are common worldwide, due to levee breach and occasionally dam failure
resulting in fatalities and considerable economic losses. Failure of earthen embankments
may be due to various reasons, such as overtopping, seepage, internal erosion and piping,
and slope instability. An accurate prediction of the embankment failure by overtopping
(i.e., breach shape, breach outflow and flow field) is necessary for emergency planning,
proper risk assessment and management, and protection measures. Estimating the soil
erodibility by flowing water is an essential step in studying, modeling, and predicting
earthen embankment failures.

An embankment breach is influenced by both hydraulic load and geotechnical properties

of the embankment material (Schmocker and Hager, 2012). The most important

1
Senior Staff Professional, Schnabel Engineering, Greensboro, NC 27407. Email: atabrizi@schnabel-eng.com.
2
Mr. and Mrs. Irwin B. Kahn Professor and Associate Dean (International Program and Continuing Education),
Department of Civil and Environmental Engineering, University of South Carolina, Columbia, SC 29208. Email:
chaudhry@sc.edu.

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 1

geotechnical parameters affecting the embankment erodibility are: grain size distribution
which implies cohesion and plasticity, compaction energy, and water content. Estimating
erodibility of cohesive soil is more complex because of the large number of parameters
controlling the erosion behavior and the difficulty of estimating these parameters.

Hanson and Cook (2004) assumed a sediment detachment process for the erosion rates of
cohesive soils rather than the sediment transport processes and defined erodibility by two
parameters in a widely accepted detachment model: the erodibility coefficient and the
critical shear stress. The detachment model is a fundamental and widely accepted
equation used in characterizing erodibility for embankment overtopping (Temple et al.,
2005) and has been used by several investigators (Hutchinson, 1972; Foster et al., 1977;
Dillaha and Beasley. 1983; Temple, 1985; Hanson, 1989; Stein and Nett, 1997; Wan and
Fell, 2004).

𝜀 = 𝑘𝑑 (𝜏𝑒 − 𝜏𝑐 ) (1)

where 𝜀 is the erosion rate (m/s), 𝑘𝑑 is the erodibility or detachment coefficient (m3/N-s),
𝜏𝑒 is the hydraulically applied shear stress (Pa), and 𝜏𝑐 is the critical shear stress (Pa).
Historically, many investigators tried to develop simple relationships between these
parameters and soil properties (Smerdon and Beasley, 1959; Kamphius and Hall, 1983;
Briaud et al., 2001). However, no widely accepted and reliable relation was obtained due
to the complexity of soil properties. The aforementioned previous investigations show
that the best approach to determine these parameters is to measure them by using both
field and laboratory methods developed for assessing earthen material erodibility,
including large and small flumes, submerged JET test, hole erosion test, flume with
lifting mechanism test, slot test. Clark and Wynn (2007) compared different methods for
determining 𝑘𝑑 and 𝜏𝑐 and field validation of these methods over a wide range of soil
types was recommended. Hanson and Hunt (2007) conducted a submerged JET test to
investigate the effects of compaction energy and water content on erodibility. They also
compared the 𝑘𝑑 values from the JET test to those from three large-scale, outdoor-
laboratory breach widening tests with the flow direction perpendicular to the
embankment as reported by Hunt et al. (2005). They found a general agreement between
the 𝑘𝑑 values from the JET test and those from the field, breach widening tests.

The objectives of the present study is to conduct the JET test on four homogenous,
cohesive, compacted soil mixtures by varying the silt and clay content and determine
their erodibility. Also, four levee embankment overtopping experiments were conducted
using the same soil mixtures in which the levee was aligned parallel to the dominant flow
direction, similar to field conditions. The effects of cohesion on the failure process of
overtopped homogenous levees (i.e., breach evolution using a sliding-rod technique)
were quantitatively determined. Erodibility coefficients in both vertical and horizontal
directions were estimated from the experimental results and were compared to those from
the JET test.

PROJECT DESCRIPTION

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 2

JET Test

To measure the cohesive soil erodibility based on the detachment model (𝑘𝑑 and 𝜏𝑐 in Eq.
1) a submerged JET apparatus, made by the author, was used (Figure 1). The test is
simple, quick, and relatively inexpensive to perform and it can be used for laboratory and
in-situ measurements of soil erodibility. The JET apparatus, procedures, and analyses are
similar to the device described by Hanson and Cook (2004). Four cohesive soil samples
were prepared and compacted using a rammer and the standard proctor compaction mold
and their erodibility characteristics were measured using the JET test. Table 1 represents
the soil sample properties used in this study.

The laboratory JET test apparatus consists of a submergence tank, jet tube with an orifice
at the bottom, head tank to provide a head of water, and point gage to measure scour
depth. The soil samples are placed in the submergence tank and directly beneath the jet
orifice. A deflection plate is attached to the bottom of the tube to deflect the jet and
protect the soil sample during initial filling of submergence tank and also at given time
intervals for reading the scour depth. The soil scouring starts when the deflection plate is
rotated out of the way of the jet, allowing direct impingement on the soil surface. The
scour depth is measured at given time intervals using the point gage. A spreadsheet is
used to enter and analyze the JET test data and to estimate the erodibility coefficients of
each soil sample.

Figure 1. Submerged JET test apparatus and the standard proctor compaction mold and
rammer

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 3

Table 1. Soil sample properties
Grain Size Dry Density Moisture
Soil % Sand % Silt % Clay (kg/m3) Content (%)
1 74 20 6 2008.7 10
2 64 30 6 1965.5 10
3 54 40 6 1973.8 10
4 68 20 12 2011.8 10

Levee Breach Test (LBT)

Figure 2 shows the schematic diagram of the plan view of the physical model test facility
which consists of a flume, a floodplain adjacent to one side of the flume, and a calibrated
sharp-crested weir (height = 0.075 m) at the downstream side of the flume. The flume is
constructed with vertical plywood walls on one side and levee-shaped wall on the other
side, except for a 0.70 m long centrally located section where the earthen embankment
test section is emplaced. The earthen levee cross section has a trapezoidal shape with 0.2
m height, 1:2 (v:h) side slopes, and 0.10 m crest width. A flow straightener honeycomb
and a wave suppressor were used at the intake section of the flume to reduce the inflow
turbulence. The inflow discharge was kept constant at 0.05 m3s-1.

Four soil mixtures similar to those in the JET test were prepared to build compacted
earthen levees. The soil material was placed in four loose layers, each layer about 10 cm
thick, and it was then compacted and trimmed into a 5 cm layer. An attempt was made to
attain the same level of compaction as in the JET soil mixtures. To initiate overtopping, a
pilot channel perpendicular to the levee crest, 0.05 m deep and 0.10 m wide, is carved at
one third of the length from the upstream edge of the earthen levee. The upstream face of
the embankment is covered with a thin layer of clay to control the seepage through the
embankment.

Breach evolution was measured using rods technique. This technique consists of four
rows of thin rods with 0.10 m distance passing vertically through two horizontal parallel
Plexiglas sheets with a few centimeter distance in-between to ensure that the rods are
dropping vertically and each row has seven rods (Figure 3). Each rod is of a specific
color representing a specific location along the levee. The rods are placed on the
downstream face of the levee, with the first row of the rods along the centerline of the
levee crest. The bottom of the rods rest on the soil and drops when the soil is eroded. To
make sure that the rods do not float in water, small weights are attached to the top of the
rods. The top of the rods are tracked by a high definition video camera. The video is then
split into single frames, each is then digitized and the top of the rods is tracked by
applying simple threshold technique to the image. Thus, the breach shape with time is
determined using these measurements. The widening of the breach is monitored using an
overhead video camera to delineate the boundary between the water and the soil mixture.

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 4

 5.18

Floodplain

2.30
0.70

Wooden Earthen Wooden Calibrated

Levee Levee Levee sharp-
crested
weir
2:1 0.40 2:1

0.20

4.80 0.36 2.24 2.24 0.36 1.20

0.61 Qin Channel Qout

Vertical wall

Figure 2. Plan view of experimental setup (dimensions are in meter)

Figure 3. Front view of the sliding rods technique for measuring the breach shape with
time.

RESULTS
JET Test Results

Figure 4 shows JET test results of the measured values of 𝑘𝑑 and 𝜏𝑐 for four levels of
cohesion. Increasing the clay content was found to be more effective on soil erosion
resistance than the silt content. Time series of scour depth are shown in Figure 5 for four
levels of cohesion. Again, the soil mixtures with different clay contents were showing

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 5

similar souring trend, while increasing the clay content decreased the depth of scour
significantly.

10
kd (cm3/N-s)

1
JET 20% Silt, 6% Clay
JET 30% Silt, 6% Clay
JET 40% Silt, 6% Clay
JET 20% Silt, 12% Clay

0.1
0.1 1 10 100
τc (Pa)

Figure 4. Erodibility measurements for different level of cohesion from JET test

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 6

 0.1
20% Silt, 6% Clay
0.09 30% Silt, 6% Clay
40% Silt, 6% Clay
0.08 20% Silt, 12% Clay

0.07
Depth of scour (m)

0.06

0.05

0.04

0.03

0.02

0.01

0
0 500 1000 1500 2000 2500
Time (s)

Figure 5. Time changes of scour depth

Levee Breach Test (LBT) Results

Three-dimensional breach evolutions are shown in Figures 6 to 9 for Tests 1 to 4,

respectively. The erosion process starts on the downstream face with head-cuts
progression from the downstream toe of the levee towards the crest, while the breach
width remains constant during this stage. After the head-cut reaches the crest, the breach
along the centerline of the crest first starts to deepen until it almost reaches the bottom
and then the widening stage starts. Comparing to non-cohesive levees, the transition from
the breach deepening to breach widening is more distinct in the tests with cohesive
material. The time changes of the breach width and breach depth along the crest
centerline are shown in Figures 10 and 11 for all the tests, respectively. Figure 12 shows
the time changes of the breach eroded volume. Similar to the results from the JET test,
increasing the clay content was found to be more effective on the erosion resistivity of
the levees than increasing the silt content.

The erodibility coefficient 𝑘𝑑 was determined from three different breach parameters:
breach widening rate (∆𝑊⁄∆𝑡), breach deepening rate (∆𝐷⁄∆𝑡), and rate of breach
erosion load (𝑞𝑏 ). These breach rates and the estimated applied shear stress were used to
determine the erodibility coefficient 𝑘𝑑 , similar to the methodology in Hunt et al. (2005)
for breach widening rate. Applied shear stress on the bed (𝜏𝑒𝑑 ) is estimated using the
measured water depth at the crest centerline and the following equation

𝜏𝑒𝑑 = 𝛾𝑤 𝑑𝑆𝑓 (2)

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 7

where 𝛾𝑤 is unit weight of water, 𝑑 is water depth, and 𝑆𝑓 is the energy slope which is
estimated using the Manning’s equation as follow:
1
𝑆𝑓 = (𝑔𝑁 2 )⁄𝑑 3 (3)

where 𝑔 is the gravitational acceleration, and 𝑁 is the Manning’s roughness associated

with soil grain sizes. The applied shear stress on sidewalls of the breach (𝜏𝑒𝑤 ) can be
estimated as follow (Chow, 1959)

𝜏𝑒𝑤 = 0.7𝜏𝑒𝑑 (4)

Breach erosion load can be estimated as

∆𝑉 1
𝑞𝑏 = (1 − 𝑛) ∆𝑡 𝐿 (5)

where 𝑛 is soil porosity, ∆𝑉⁄∆𝑡 is the rate of eroded volume, and 𝐿 is the characteristic
length which is the length of the earthen levee section herein.

Erodibility coefficient 𝑘𝑑 may then be estimated from the breach widening and
deepening are, and breach erosion load using the following equations

∆𝐷⁄∆𝑡 = 𝑘𝑑 (𝜏𝑒𝑑 − 𝜏𝑐 ) (6)

∆𝑊 ⁄∆𝑡 = 2𝑘𝑑 (𝜏𝑒𝑤 − 𝜏𝑐 ) (7)

𝑞𝑏 /𝐵 = 𝑘𝑑 (𝜏𝑒𝑑 − 𝜏𝑐 ) (8)

where B is a characteristics width.

Table 2 summarizes the erodibility measurements from the levee breach test using
different breach parameters. Results from all three LBT measurement approaches show
that in overall the erodibility decreases as the silt content increases. However, similar to
the results from the JET test, clay content was introducing more erosion resistivity than
the silt content.

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 8

 Elevation (m)
Elevation (m)

t = 5s t = 20s

x (m) x (m)

Elevation (m)
Elevation (m)

t = 40s t = 60s

x (m) x (m)
Elevation (m)

Elevation (m)

t = 80s t = 110s

x (m) x (m)
Figure 6. Breach evolution for Test 1 at t = 5, 20, 40, 60, 80, and 110 s

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 9

 Elevation (m)
Elevation (m)

t = 10s t = 40s

x (m) x (m)
Elevation (m)

Elevation (m)
t = 80s t = 120s

x (m) x (m)
Elevation (m)

Elevation (m)

t = 150s t = 170s

x (m) x (m)
Figure 7. Breach evolution for Test 2 at t = 10, 40, 80, 120, 150, and 170 s

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 10

 Elevation (m)
Elevation (m)

t = 10s t = 40s

x (m) x (m)

Elevation (m)
Elevation (m)

t = 80s t = 120s

x (m) x (m)
Elevation (m)

Elevation (m)

t = 150s t = 170s

x (m) x (m)
Figure 8. Breach evolution for Test 3 at t = 10, 40, 80, 120, 150, and 170 s

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 11

 Elevation (m)
Elevation (m)

t = 1min t = 5min

x (m) x (m)
Elevation (m)

Elevation (m)
t = 10min t = 15min

x (m) x (m)
Elevation (m)

Elevation (m)

t = 19min t = 21min

x (m) x (m)
Figure 9. Breach evolution for Test 4 at t = 1, 5, 10, 15, 19, and 21 min

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 12

 0.6
Test 15 20% Silt
0.5
Test 26 30% Silt
Top width of breach, W (m)

0.4 Test 37 40% Silt

Test 48 12% Clay

0.3

0.2

0.1

0
1 10 100 1000 10000
Time (s)

Figure 10. Time changes of breach top width for levees with different cohesion
0.25
Test 15 20% Silt

0.2 Test 26 30% Silt

Breach depth, D (m)

Test 37 40% Silt

0.15
Test 48 12% Clay

0.1

0.05

0
1 10 100 1000 10000
Time (s)

Figure 11. Time changes of breach depth for levees with different cohesion

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 13

 0.025
Test 15 20% Silt

0.02 Test 26 30% Silt

Erosion volume, V (m3)

Test 73 40% Silt

0.015
Test 84 12% Clay

0.01

0.005

0
1 10 100 1000 10000
Time (s)

Figure 12. Time changes of breach eroded volume for levees with different cohesion

Comparing Erodibility from JET Test and Levee Breach Test

The erodibility coefficients 𝑘𝑑 from the JET test were compared to those from the levee
breach test. Generally, the erodibility coefficients from the JET method were found to be
smaller than those from the breach tests (Table 2). However, the estimated erodibility
coefficients from the widening rate and deepening rate were in better agreement with the
JET results as compared to those from the eroded volume.

Table 2. Erodibility coefficient measurements from the JET and LBT test for different
soil mixtures
LBT LBT LBT
JET
widening deepening eroded volume
𝒌𝒅 𝒌𝒅 𝒌𝒅 𝒌𝒅
Test (m3/N-h) (m3/N-h) (m3/N-h) (m3/N-h)
1 0.100 0.096 0.111 0.019
2 0.081 0.012 0.089 0.013
3 0.034 0.063 0.091 0.013
4 0.006 0.023 0.020 0.005

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 14

 REFRENCES

Briaud, J. L., F. C. K. Ting, H. C. Chen, Y. Cao, S. W. Han, and K. W. Kwak. 2001.

Erosion function apparatus for scour rate predictions. J. of Geotechnical and
Geoenvironmental Engineering, ASCE 127(2): 105-113.

Chow, V. T. (1959). Open-Channel Hydraulics. New York, N.Y.: McGraw-Hill.

Clark, L.A., and T. M. Wynn. 2007. Methods for determining streambank critical shear
stress and soil erodibility: implications for erosion rate predictions. Trans ASABE 50(1):
95–106.

Dillaha, T. A., and D. B. Beasley. 1983. Distributed parameter modeling of sediment

movement and particle size distribution. Transactions of the ASAE 26(6): 1716-1722.

Foster, G. R., L. D. Meyer, and C. A. Onstad. 1977. An erosion equation derived from
basic erosion principles. Transactions of the ASAE 20(4): 678-682.

Hanson, G. J. 1989. Channel erosion study of two compacted soils. Transactions of the
ASAE 32(2): 485-490.

Hanson, G. J., and K. R. Cook. 2004. Apparatus, test procedures and analytical methods
to measure soil erodibility in situ. Applied Engineering in Agriculture 20(4): 455-462.

Hanson, G. J., and S. L. Hunt. 2007. Lessons learned using laboratory JET method to
measure soil erodibility of compacted soils. Applied Eng. in Agric. 23(3):305-312.

Hunt, S. L., Hanson, G. J., Cook, K. R., and Kadavy, K. C. (2005). “Breach widening
observations from earthen embankment tests.” Trans. ASAE, 48(3), 1115–1120.

Hutchinson, D. L. 1972. Physics of erosion of cohesive soils. Ph.D. thesis. New Zealand:
University of Aukland.

Kamphius, W. J., and K. R. Hall. 1983. Cohesive material erosion by unidirectional

current. J. of Hydr. Eng., ASCE 109(1): 1076-1081.

Schmocker, L., & Hager, W.H. (2012). Plane dike breaching due to overtopping: Effects
of sediment, dike height and discharge. J. Hydraulic Res. 50(6), 576–586.

Schmocker, L., & Hager, W.H. (2012). Plane dike breaching due to overtopping: Effects
of sediment, dike height and discharge. J. Hydraulic Res. 50(6), 576–586.

Smerdon, E. T., and R. P. Beasley. 1959. The tractive force theory applied to stability of
open channels in cohesive soil. Research Bull. 715. Univ. of Missouri. Ag. Exp. Station.
Columbia, Mo.

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 15

Stein, O. R., and D. D. Nett. 1997. Impinging jet calibration of excess shear sediment
detachment parameters. Transactions of the ASAE 40(6): 1573-1580.

Temple, D. M. 1985. Stability of grass-lined channels following mowing. Transactions of

the ASAE 28(3): 750-754.

Temple, D.M., Hanson, G.J., Nielsen, M.L., and Cook, K., 2005. Simplified breach
analysis model for homogeneous embankments: Part 1, Background and model
components. USSD Technologies to Enhance Dam Safety and the Environment, 25th
Annual USSD Conference, Salt Lake City, Utah.

Wan, C. F., and R. Fell. 2004. Investigation of rate of erosion of soils in embankment
dams. J. of Geotechnical and Geoenvironmental Engineering, ASCE 130(4): 373-380.

Copyright © 2018 U.S. Society on Dams. All Rights Reserved. 16