Proceedings of the ASME 2013 International Pipeline Geotechnical Conference

IPG2013
July 24-26, 2013, Bogota, Colombia

IPG2013-1931

DETERMINATION OF THE SEPARATION OF DIVERTER BERMS FOR EROSION

CONTROL
Carlos Eduardo Rodríguez Pineda1 Paula Andrea Rodríguez
Technical Director Assistant Engineer

Juan Sebastián Gutiérrez Gómez

Assistant Engineer

Civil and Tech Ingenieros S.A.S.

Bogotá, Colombia

ABSTRACT 1. INTRODUCTION

Diverter berms are structures that are intended to control Diversion berms are structures that are designed to capture
erosion by means of the reduction of the dragging forces due to the runoff and minimize the erosive effects runoff. This
runoff. In a hillside, as the slope and length increases, the requires an appropriate design of the diverter berms to prevent
towing capacity of the runoff increases; diverter berms aim to water concentrations, erosion of roadways and instabilities of
reduce the flow length and with this the water potential erosion. filled slopes. The correct design criteria for diversion berms are
In Colombia is common to determine the separation of based on the basic principles of erosion caused by rainfall. The
diverter berms based in the graphics presented in the NIO 0802 factors that control these types of erosion are related to the
standard (Ecopetrol, 2001), in which separation is function of amount and characteristics of the precipitation, the type of soil,
the slope of the hillside, the type of soil and the rainfall the topography, and the type and extension of any kind of
intensity. vegetal cover. One of the most important aspects when this
In this paper the equations proposed by Morgan (1995) kind of structures is designed is the berms distance so that the
and Mirtskhoulava (2001) are applied to determine the superficial flow does not cause soil erosion.
separation of diverter berms based on the maximum non- There are several methods by which the distance between
eroding velocity, the discharge flow, the surface roughness and diverters can be defined. Through engineering experience it has
the slope steepness. been possible to create tables or graphs in which with the slope
In order to compare the results of the application of the steepness and the observed erosion of the soil, typical spacing
Morgan (1995) and Mirtskhoulava (2001) models, with those have been obtained. In some countries is common the
curves of the NIO 0802 standard (Ecopetrol, 2001), an construction of structures that have the same function as the
application was made to the right-of-way for the Medellín- diverter berms; for this kind of structures not only horizontal
Cartago pipeline, along the called Sinifana variant. The distances are defined but also the vertical spacing. In order to
information regarding topography, soils and coverage was define the diverter berms dimensions various alternatives have
provided by Civil and Tech SAS. been proposed as typical dimensions that are a function of the
kind and materials of those diverters (i.e. sacks of soil-cement
mixture or concrete berms).

1
Associate Professor at Pontificia Universidad Javeriana – Assistant Professor at Universidad Nacional de Colombia

1 Copyright © 2013 by ASME

 In this paper two analytical methods to determine the Results obtained from the application of analytical
diversion berms spacing are presented. Other empirical methods were compared with the NIO 0802 standard
methods or simple specifications to define diverter berms (Ecopetrol, 2001), which proposes the use of the abacus
spacing are allowable but they usually are not supported by any showed in the Figure 1:
type of theoretical analysis, so these methods are usually of
very specific applicability, while analytical methods, as those
presented here, are applicable in any possible scenario.

Figure 1. Diverter berms spacing as function of the type of soil, the slope of the ground and the precipitation. (Ecopetrol, 2001)

For the adequate use of this graph is necessary to establish (𝜗 5
2 𝑛3
2 )
if the diversion berms are going to be in erodible soils (silts and 𝑋 (𝑚) = Eq. 1
sands) or in resistant soils (clays, compact soils, rock, etc.) 𝑃 − 𝐼 𝑆𝑒𝑛𝜃 3
4 𝐶𝑜𝑠𝜃
Also, the longitudinal slope (in percentage) of the ground
should be known and the precipitation needs to be classified Where:
between A (when is less than 10 mm/hr) and B (when is greater L = diverter berms spacing
than 15 mm/hr). Unfortunately, the procedure for obtaining this  = maximum non-eroding velocity
graph is unknown, which makes its application subjective. n = Manning´s roughness coefficient
P = precipitation rate, mm h-1
2. METHOD I = infiltration rate, mm h-1
θ = slope angle, deg
To determine the separation of the diverter berms it should
be taken into account the runoff velocity and the size of the soil The second method corresponds to the analytical equation
particles; these two parameters determine soil behavior and proposed by Mirtskhoulava (2001), based on Chézy’s
define if any problem with soil erosion, sedimentation or expression to determine flow velocity:
transport is expected.
Two methods were taking into account to determine the
diversion berms spacing. The first one is that proposed by 0.000034 𝑣𝑐 3.32
𝑋 𝑚 = Eq. 2
Morgan (1995) and defined by the following equation: 𝑄 𝑚1 2.32 𝑠1.16 𝑛

Where:
Q = discharge flow, m3 s-1

2 Copyright © 2013 by ASME

m1 = coefficient to describe ground surface roughness difference between the rainfall intensity and the infiltration of
s = slope, m m-1 water into the soil for the design storm, which should reflect
n = Manning’s roughness coefficient the intensity and the duration of the characteristic rainfall in the
vc = maximum non-eroding velocity area. Morgan et al. (2003) recommend working with a design
rainfall correspondent to a return period of 10 years. The
The maximum non-eroding velocity is determined as a roughness is considered based on the Manning’s roughness
function of the soil particles size, which can be obtained by coefficient n, which is also determined as a function of soil size
using the graph proposed by Hjulström (1935), presented in and the kind of soil cover (Morgan, 2005). The slope steepness
Figure 2, also those values presented in Table 1, proposed by is taken directly from the topography.
Morgan (2005), can be used; the discharge flow was
determined based on the effective rainfall, being this the

Material Bare Medium grass cover Very good grass cover

Very lightly silty sand 0.3 0.75 1.5
Light loose sand 0.5 0.9 1.5
Coarse sand 0.75 1.25 1.7
Sandy soil 0.75 1.5 2.0
Firm clay loam 1.0 1.7 2.3
Stiff clay or stiff gravelly soil 1.5 1.8 2.5
Geotextil mat 1.5 2.5 3.5
Coarse gravels 1.5 1.8 n/a
Shale, hardpan, soft rock 1.8 2.1 n/a
Hand-pitched stone 2.0 n/a n/a
Hard cemented conglomerates 2.5 n/a n/a
Rip-rap (D50: 250-400 mm) 3.0 n/a n/a
Box gabions, grouted stone 5.0 n/a n/a
Concrete block system 6.0 n/a n/a
n/a, not applicable, since a medium or very good grass cover is unlikely to be obtained
Intermediate values may be selected.
Table 1. Maximum safe velocities (m/s) in channels based on covers expected after two seasons. Morgan (2005)

 𝜓𝑓
𝑓 = 𝐾
1 + Eq. 4
𝐹

Where:
K = hydraulic conductivity, cm h-1
Δθ = effective porosity
Ψf = suction head at wetting front, cm
F = infiltration, cm
f = infiltration rate, cm h-1

The effective porosity and the suction head should be

established from laboratory tests. The procedure consists in the
calculation of the infiltration rate and the total volume
Figure 2. Curves for calculating the flow rate according to the size of particles infiltrated, taking into account the total precipitation during the
(Dingman, 2009) storm duration, assuring that the infiltrated volume is not
greater than precipitation. Values for the infiltration parameters
For the present analysis the model of Green-Ampt (1911) used with the Green-Ampt equation for the present analysis
was used to determine the infiltration. This model is based on were determined from Rawls et al. (1983). The infiltration rate
the following equations: used was that one calculated for a lapse of 60 minutes after the
storm started, according to the value suggested by Morgan
𝐹 (1995).
𝐾. 𝑡 = 𝐹 −  𝜓𝑓 𝐿𝑛
1 + Eq. 3
 𝜓𝑓
3. APPLICATION AND RESULTS

3 Copyright © 2013 by ASME

 It was carried out an application example for the Sandy Loam
calculation of diversion berms for the called Sinifaná variant in Manning’s roughness coefficient 0,024
the Medellín-Cartago pipeline owned by Ecopetrol. For this Maximum non erosive velocity (m/s) 0,5
particular case there were available hydrological data registered
Roughness coefficient of the hillside’s
by the Fredonia station, which belongs to the IDEAM hydro- 1,1
surface
meteorological network. The diverter berms spacing was Table 2: Properties adopted for soil in the application area which
determined by three different ways, the first, is using Morgan’s correspond to a sandy loam
equation; the second, is using Mirtskhoulava’s equation; and
the third, is using the graphs established by the NIO 0802 Rainfall infiltration was calculated by using the Green-
standard (Ecopetrol, 2001). Ampt (1911) method; hydrological data from Fredonia’s station
The diverter berms spacng is found for slopes between were used to determine the Intensity-Duration-Frequency (IDF)
10% and 100%, varying each 5%. The type of soil in the curves shown in Figure 3.
surface considered for this example is a Sandy loam, for which
the parameters adopted are presented in the Table 2.

INTENSITY - DURATION - FREQUENCY (IDF) CURVES

VARGAS & DÍAZ-GRANADOS METHOD
350

300
Tr = 2.33 Years
Tr = 5 Years
250 Tr = 10 Years
Tr = 20 Years
Intensity [mm/h]

Tr = 50 Years
200 Tr = 100 Years

150

100

50

0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180
Duration [min]

Figure 3. IDF curves for Fredonia’s Station.

4 Copyright © 2013 by ASME

 SANDY LOAM (Morgan)
Vc = 0,5 m/s --- n = 0,024

100,000
Separation (m)

Tr = 2 years
Tr = 10 years
Tr = 15 years
Tr = 25 years
Tr = 50 years

10,000
0 10 20 30 40 50 60 70 80 90 100
Slope (%)

Figure 4. Diverter berms spacing for a Sandy Loam based on Morgan (2005) method.

SANDY LOAM (Mirtskhoulava)

m1 = 1,1 --- Vc = 0,5 m/s --- n = 0,024

100,0
Separation (m)

Tr = 2 years
Tr = 10 years
Tr = 15 years
10,0
Tr = 25 years
Tr = 50 years

1,0
0 10 20 30 40 50 60 70 80 90 100
Slope(%)

Figure 5. Diverter berms spacing for a Sandy Loam based on Mirtskhoulava (2001) method.

Once the intensity was calculated, the Morgan (1995) and

Mirtskhoulava (2001) equations were applied to calculate
diverter berms spacing for the different return periods, which
are shown in Figure 4 and Figure 5, respectively; values
obtained are shown in Table 3, in which values obtained using
the NIO 0802 standard (Ecopetrol, 2001) are also included.

5 Copyright © 2013 by ASME

 Tr = 2 years Tr = 10 years Tr = 15 years Tr = 25 years Tr = 50 years
Separation (m)
Slope NIO
Mirts. Morgan Mirts. Morgan Mirts. Morgan Mirts. Morgan Mirts. Morgan
(%) Norm
10 52.16 118.31 38.83 88-07 35.82 81.26 32.39 73.48 28.29 64.18 27
20 23.34 72.17 17.38 53.73 16.03 49.57 14.50 44.82 12.66 39.15 18
30 14.58 55.48 10.86 41.30 10.02 38.11 9.06 34.46 7.91 30.10 13
40 10.45 47.22 7.78 35.15 7.17 32.43 6.49 29.32 5.67 25.61 11
50 8.06 42.64 6.00 31.74 5.54 29.29 5.01 26.48 4.37 23.13 10
60 6.53 40.04 4.86 29.81 4.48 27.50 4.05 24.87 3.54 21.72 9
70 5.46 38.63 4.06 28.76 3.75 26.53 3.39 23.99 2.96 20.96 8
80 4.67 38.01 3.48 28.30 3.21 26.11 2.90 23.61 2.54 20.62 8
90 4.08 37.93 37.93 28.24 2.80 26.05 2.53 23.56 2.21 20.58 8
100 3.61 38.25 38.25 28.48 2.48 26.27 2.24 23.76 1.96 20.75 7.5
Table 3. Diverter berms for different return periods according to Morgan (1995), Mirtskhoulava (2001) and NIO 0802 standard.

4. CONCLUSIONS of Mirtskhoulava (2001). Based on the present analysis is not

possible to assure which method is more reliable. In order to
Results show that curves of the NIO 0802 standard establish which model is optimal is necessary to study the
(Ecopetrol, 2001) tend to give smaller separations than those conditions of historic cases where erosion was evidenced,
obtained with the other two applied methods. The advantage in similar to the analysis present by Morgan et al. (2003), in
applying these analytical methods is that a return period could which they conclude that generally analytical methods tend to
be associated to the rainfall, and thought this, to a rainfall give greater separations than empirical methods, but they do
probability of occurrence and erosion probability, which allows not concluded over which of the methods is more reliable for
making designs by on reliability. Values obtained with the NIO design.
0802 standard (Ecopetrol, 2001) shown that these curves could The NIO 0802 standard (Ecopetrol, 2001) does not
be associated with recurrence periods much greater than 10 establishes clearly what is the basis of the curve presented in
years which is the value suggested by Morgan (2005); this Figure 1, so is not possible to make a direct comparison
indicates that its application could be overestimating the between all methods, however from the analysis carried out for
number of diverter berms needed for erosion control, with the this particular case of study, the values of precipitation that the
consequent cost increase, as suggested by Morgan et al. (2003). NIO 0802 standard uses corresponds to rainfalls of long
Besides the diverter berms spacing it is necessary for an duration, up to 3 hours, which will not always correspond to
adequate design, to define other factors that must be taken into the characteristic rainfall of the zone.
account such as the diverter cross section, the longitudinal
slope, and the estimated time needed to let the protective 5. RECOMMENDATIONS AND FUTURE WORKS
vegetation to be established (Morgan et al., 2003).
The cross section must guarantee that the intercepted In order to obtain an appropriate design is necessary to
runoff could be evacuated effectively, for which its design must have a good hydrologic record within the zone of study since
be based on using open channel theory (Morgan, 2005). analytical methods are based in precipitation data and the
According to the NIO 0802 standard (Ecopetrol, 2001), calculation of the infiltration rate. When there is no rainfall
the diverter berms should be constructed with a longitudinal data available on the basin is not possible to obtain reliable
slope between 3% and 5%; this, in order to avoid aggressive results. Several empirical methods determine the diverter
erosive processes caused by a high flow velocity in a steep berms spacing without taking into account rainfall (Morgan,
slope or slow drainage in the case of a gentle slope. 2005), however is not possible to determine its reliability with
Additionally, the bottom should be protected with resistant the available data for the present study, hence it should be
materials against erosion and the crest and the outer face should useful in the future study this point.
be cover by grass to avoid erosion around the berms. Conclusion of Morgan et al. (2003), related to the
According to Morgan et al. (2003), diverter berms are a differences obtained between the empirical and analytical
temporary measure for erosion control while the protective models, contradicts the one obtained in the present analysis,
vegetation is established, for which is considered that its which indicates that the application of these methods are very
efficiency should be guaranteed at least between 2 and 3 years. sensitive to the parameters adopted for the modeling, among
which Manning’s roughness coefficient, the characteristics of
According to the results shown in Table 3 it could be the design rainfall and the Green-Ampt soil parameters were
evidenced that the Morgan (1995) method tend to give much identified as the most important; these are the ones that cause
greater separations compared to those obtained by the method

6 Copyright © 2013 by ASME

more variation in the results, therefore it is necessary that its
determination obeys to an accurate characterization.
For future works is suggested taking into account different
methodologies for the infiltration calculation, which lets to
determine the concentration time based on certain
characteristics of the watershed and the event duration; in order
to define adequately this concentration time for design.

6. ACKNOWLEDGMENTS

We thank the engineer Manuel Felipe Ríos Gaona for its

valuable contribution for the calculation of the IDF curves and
in general for his support regarding the topics related to
hydraulic and hydrology during the development of this work.
We also thank Civil & Tech Ingenieros SAS for its
endorsement in the use of the data with which the presented
analyses were carried out.

7. REFERENCES

[1] R.P.C. Morgan, M.J. Han., 2005, “Design of diverter

berms for soil erosion control and biorestoration along
pipeline rights-of-way”, Soil Use and Management, 306-
311.

[2] Ecopetrol, 2001, Normas de Ingeniería de Oleoductos

“Sistemas de Drenaje”, NIO 0802, 1-11.

[3] R.P.C. Morgan, 2005, “Soil erosion and conservation”, 3rd

Edition. Blackwell Publishing.

[4] R.P.C. Morgan, 1995, “Soil Erosion and Conservation”.

2nd Edition. Longman, Harlow.

[5] W.H. Green., G. A. Ampt (1911), “Studies on soil

physics”, J. Agric. Sci., 4(1), 1-24.

[6] W.J. Rawls, M. ASCE, D. L. Brakensiek, N. Miller.,

(1983), “Green-Ampt Infiltration Parameters from Soils
Data”. ASCE Journal of Hydraulic Engineering, 109 (1),
62-70.

[7] R.P.C. Morgan, Ts.E. Mirtskhoulava, V. Nadirashvili, M.J.

Hann, A.H. Gasca., 2003, “Spacing of Berms for Erosion
Control along Pipelines Right-of-way”. Biosystems
Engineering 85 (2), 249-259.

[8] F. Hjulström, 1935, Studies of the morphological activity

of rivers as illustrated by the River Fyris. Bulletin of the
Geologial Intitute, University of Uppsala, 25, 221-527.

7 Copyright © 2013 by ASME