geotechnics of soil erosion gsp167

8/2/2019 Geotechnics of Soil Erosion GSP167

1/65

G E O T E C H N I C A L S P E C I A L P U B L I C A T I O N N O . 1 6 7

GEOTECHNICS OFSOIL EROSIONPROCEEDINGS OF SESSIONS OF GEO-DENVER 2007

February 1821, 2007

Denver, Colorado

SPONSORED BY

Geotechnics of Soil Erosion Committee of

The Geo-Institute of the American Society of Civil Engineers

EDITED BY

G. Biscontin, Ph.D.

Published by the American Society of Civil Engineers


2/65

Notices

Any statements expressed in these materials are those of the individual authors anddo not necessarily represent the views of ASCE, which takes no responsibility for any

statement made herein. No reference made in this publication to any specific method,

product, process or service constitutes or implies an endorsement, recommendation,

or warranty thereof by ASCE. The materials are for general information only and do

not represent a standard of ASCE, nor are they intended as a reference in purchase

specifications, contracts, regulations, statutes, or any other legal document.

ASCE makes no representation or warranty of any kind, whether express or implied,

concerning the accuracy, completeness, suitability, or utility of any information,

apparatus, product, or process discussed in this publication, and assumes no liability

therefore. This information should not be used without first securing competentadvice with respect to its suitability for any general or specific application. Anyone

utilizing this information assumes all liability arising from such use, including but not

limited to infringement of any patent or patents.

Copyright 2007 by the American Society of Civil Engineers.

All Rights Reserved.

Manufactured in the United States of America.

American Society of Civil Engineers

ASCE International Headquarters

1801 Alexander Bell DriveReston, VA 20191-4400 USA

Call Toll-Free in the U.S.: 1-800-548-2723 (ASCE)

Call from anywhere in the world: 1-703-295-6300

Internet: http://www.pubs.asce.org


3/65


4/65

Contents

Geotechnics of Soil ErosionOrganizer: G. Biscontin, Texas A&M University

How does Water-Soil Interaction Lead to Erosion?George W. Annandale

Current State-of-the-Art of Rock Scour TechnologyGeorge W. Annandale

Bureau of Reclamation Erosion Testing for Evaluation of Piping and Internal Erosion ofDams

Jeffrey A. Farrar, Roger L. Torres, and Zeynep Erdogan

The Effect of Abutment Length for Abutment Scour in Cohesive Soil: Initial ResultsSeung Jae Oh, Xingnian Chen, Jean-Louis Briaud, Kuang-An Chang,and Hamn-Ching Chen

Evaluation of Soil Erosion Using the Rainsplash TechniqueJ. L. Smith, M. R. Davieau, and S. K. Bhatia

No-Fines Concrete as Ecologic Stream Bank Erosion ControlYuewen Huang and Xiong (Bill) Yu
http://40897-1483.pdf/http://40897-1485.pdf/http://40897-3646.pdf/http://40897-3646.pdf/http://40897-2165.pdf/http://40897-1986.pdf/http://40897-1482.pdf/http://40897-1483.pdf/http://40897-1485.pdf/http://40897-3646.pdf/http://40897-3646.pdf/http://40897-2165.pdf/http://40897-1986.pdf/http://40897-1482.pdf/


5/65

Author Index-GSP 167 Geotechnics of Soil Erosion

Annandale, George W. How does Water-Soil Interaction Lead to Erosion?

Annandale, George W. Current State-of-the-Art of Rock Scour Technology

Bhatia, S. K. Evaluation of Soil Erosion Using the Rainsplash Technique

Briaud, Jean-Louis The Effect of Abutment Length for Abutment Scour in Cohesive Soil: Initial Results

Chang, Kuang-An The Effect of Abutment Length for Abutment Scour in Cohesive Soil: Initial Results

Chen, Hamn-Ching The Effect of Abutment Length for Abutment Scour in Cohesive Soil: Initial Results

Chen, Xingnian The Effect of Abutment Length for Abutment Scour in Cohesive Soil: Initial Results

Davieau, M. R. Evaluation of Soil Erosion Using the Rainsplash Technique

Erdogan, Zeynep Bureau of Reclamation Erosion Testing for Evaluation of Piping and Internal Erosion o

Farrar, Jeffrey A. Bureau of Reclamation Erosion Testing for Evaluation of Piping and Internal Erosion o

Huang, Yuewen No-Fines Concrete as Ecologic Stream Bank Erosion Control

Oh, Seung Jae The Effect of Abutment Length for Abutment Scour in Cohesive Soil: Initial Results

Smith, J. L. Evaluation of Soil Erosion Using the Rainsplash Technique

Torres, Roger L. Bureau of Reclamation Erosion Testing for Evaluation of Piping and Internal Erosion o

Yu, Xiong (Bill) No-Fines Concrete as Ecologic Stream Bank Erosion Control
http://40897-1483.pdf/http://40897-1485.pdf/http://40897-1986.pdf/http://40897-2165.pdf/http://40897-2165.pdf/http://40897-2165.pdf/http://40897-2165.pdf/http://40897-1986.pdf/http://40897-3646.pdf/http://40897-3646.pdf/http://40897-1482.pdf/http://40897-2165.pdf/http://40897-1986.pdf/http://40897-3646.pdf/http://40897-1482.pdf/http://40897-1482.pdf/http://40897-3646.pdf/http://40897-1986.pdf/http://40897-2165.pdf/http://40897-1482.pdf/http://40897-3646.pdf/http://40897-3646.pdf/http://40897-1986.pdf/http://40897-2165.pdf/http://40897-2165.pdf/http://40897-2165.pdf/http://40897-2165.pdf/http://40897-1986.pdf/http://40897-1485.pdf/http://40897-1483.pdf/


6/65

No-Fines Concrete as Ecologic Stream Bank Erosion Control

Yuewen Huang1

and Xiong (Bill) Yu2

ABSTRACT

No-fines concrete is a pervious concrete obtained by eliminating the sand from normal

concrete mix. Compared with conventional concrete, no-fines concrete has uniqueproperties desirable for various applications. Because of the presence of large voids,no-fines concrete has lower density, cost and thermal conductivity, smaller dying

shrinkage, no segregation, larger contaminant retaining capability, and reduced capillary

movement of water. No-fines concrete is used for construction of pavement, stormwater control utilities and green houses. This paper discusses the application of no-

fines concrete as an ecology preservative method for stream bank erosion control. Soil

erosion is an important factor that can trigger the instability of an embankment. A

sustainable revetment should provide soil erosion protection without significantlychanging the existing ecologic environment. The strength of no-fines concrete provides

sufficient protection against scour of embankments. Although grasses are found to be

hard to survive with ordinary types of bank revetment, especially when subjected toperiodic inundation from river water level fluctuations, the large pore spaces of no-fines

concrete protect grass seeds and provide an environment for grasses to grow. By

installing artificial access holes in the no-fines concrete revetment, the ecologicconditions can be preserved to the maximal extent. The design criteria and durability of

no-fines concrete revetments are discussed in this paper. This paper also provides an

example application of a no-fines concrete revetment that achieved the desirableecologic effects.

INTRODUCTION

Erosion control is important from a variety of environmental considerations.

Excessive river bank erosion not only poses hazards to engineered structures adjacent to

1Senior Engineer, Guangzhou Investigation Design and Research Institute of Water

Conservancy & Hydropower, Guangzhou, Guanggong, China, 510640,

[email protected] Professor, Department of Civil Engineering, Case Western Reserve

University, Cleveland, OH 44106-7201, 216-368-6247, [email protected],

GSP 167 Geotechnics of Soil Erosion


7/65

the bank, but also contributes significant amounts of sediments from the tributary river.

This ultimately can lead to high solids concentration, excessive nutrition and pollutionin the river. Proper river erosion control is needed to prevent these potential

environmental hazards and to maintain the existing ecosystem. Erosion control

involves the use of engineering measures to maintain the desired ecosystem.

Concrete has been widely used to construct engineering structures. Its advantages

include high strength, low cost, and wide availability. Special types of concrete arecurrently being developed, such as self-consolidating concrete, fiber reinforced polymer

concrete and roller compacting concrete. Research on innovative concrete technology

has triggered applications of new materials in construction practice. The concept ofgreen concrete is a relatively new development from the concrete industry, which refers

to concrete designs that not only achieve the required structural performance, but also

bring the least impact to the surrounding environment. Examples of green concrete

include porous concrete, which can be used to increase the thermal efficiency ofresidential buildings.

No-fines concrete is another type of green concrete that can help achieve a sustainableenvironment. No-fines concrete refers to concrete mix a mixture of coarse aggregates,

cement and water without fine aggregates. The cement acts to hold the mixture together.

Coarse aggregate can be gravel, pebble or other artificial aggregates. Because of theabsence of fine aggregates, large void volume exists inside the concrete. The sizes of

the voids are typically just slightly smaller than the sizes of coarse aggregates. These

make the non-fine concrete to be highly permeable. It has been used to construct

permeable concrete pavements. The large pore sizes help to reduce the noise generatedat the tire/pavement interface. The high permeability makes it a suitable material for

storm water management. The large pore sizes can also accommodate the growth of

vegetations (Dong et al. 2002). The pores in the no-fine concrete can retain significantamount of sludge, mature and seeds. This creates a favorable environment for seeds to

grow up. On the other hand, because of the strong adhesion by the cement, no-fine

concrete provides relatively high scour resistance. To accommodate the development ofaquatic creatures, proper large holes can be created artificially. The size of holes can be

determined by surveying the size of typical local aquatics. These artificial holes can

provide shelters for a variety of aquatic creatures. The growth of vegetation is not onlyaesthetically amiable; the root of vegetation can also contribute to the improved scour

protection.

This paper discusses the application of no-fines concrete for river bank erosion control.The design utilizes the high strength of no-fine concrete together with its ecological

friendly factors. A project on river bank revitalization is used to illustrate the

application of this type of concrete.

USE OF NO-FINES CONCRETE FOR RIVER BANK EROSION CONTROL

Erosion of soil by run-off takes place when the dragging force acting on soils particles

is larger than the resisting force from the adjacent soil layers (or the strength of soil).



8/65

The dragging force is influenced by factors such as river water velocity and topology of

the river channel. Typically, these are determined by the geometry of the river bankand local hydrologic conditions (i.e., precipitation). Since these factors are not easily

changed, the resistance force has to be increased to prevent or minimize soil erosion.

this can be done by protecting river bank by methods such as stone toeing, geotexile and

riprap. However, these measures typically do not work without disrupting the existingecologic system, i.e., they typically do not support the growth of vegetation particularly

due to water level fluctuations.

An ecologic revetment is a hydraulic concept that embraces hydraulic, environmental,

biologic, ecologic and aesthetic factors. It is a new concept for revetment design thatemphasizes the creation of biologic amiable and aesthetic environment. It helps to

create a clean river and sustainable ecologic system. The ecologic revetment considers

the strength, safety and durability of the revetment structure as well as its ecologic

effects. It departs from the use of traditional stone-cement mixes in that it prompts thecreation of an ecologic amiable river environment. The following of this paper uses an

example to illustrate the use no-fine concrete for revetment design.

MECHANISM OF NO-FINE CONCRETE TO ACHIEVE ECOLOGIC

EROSION PROTECTION

It is known that pore fluid in commonly used concrete are strongly basic due to

dissolution of ions during hydration reactions hindering vegetation growth. Because of

the large pore size in no-fines concrete, the basic pore solution is drained or washed out

during water level fluctuations. As a result, the drainable pore water inside no-finesconcrete is typically neutral. This provides a favorable environment for vegetation

growth without the need of additional chemical treatment of the pore solutions.

Another factor contributing to the control of the pore environment is the carbonizationprocess. Exposure of concrete to water and carbon dioxide leads to the production of

inert chemicals. As no-fines concrete is typically applied at the location where river

water fluctuates, the alternating exposure to water and carbon dioxide speeds up thecarbonization process. As a result, an impermeable layer forms at the exposed surface

of concrete, preventing the diffusion of basic chemicals generated inside the concrete

into the pore water, which also help to maintain a neutral PH pore water environment.Both factors result in an environmentally favorable condition for vegetation and aquatic

creatures.

CASE HISTORY

The project is located in the Southern province of Guangdong, China. The region was

reclaimed from lakes by creating a variety of land patches (referred as polders by localresidents). It is within the river network of the Zhujiang Delta, with the Tiger-Gate

River to the east, the South China Sea to the south and the Hong Qili River, the Shang

River and the Gongbao River to the north. The Jiaomen River intercepts the district inthe middle. Most of this district is used for farms, banana production and fish ponds.

To construct the polders, gravel was placed in the outside and the inner part was filled



9/65

with borrow soils. Sludge from the Zhujiang River was also used in some instances for

soil supply was scare. Because of the high water content (above 60%) of the thicksludge layer and the absence of foundation treatments, the bearing capacity of the

foundation soils was very low after the natural consolidation processes. The reclamation

zones were typically characterized by large settlements. Record settlements larger than

1m within one year were observed. Construction of the polders typically requiredseveral years because of the slow speed of construction. After one polder was

constructed, construction of another polder would begin. An artificial channel typicallyexisted between the adjacent polders, called Chong by local residents.

Because of extensive scour and erosion, the banks of the artificial rivers werecharacterized by the scarcity of vegetation but large presence of aquatic life. These

conditions made the channel banks susceptible to scour triggered by water level

fluctuations and heavy rains. Significant numbers of failures occurred along the

channel banks, reducing run-off capacity and also creating undesirable aesthetic effects(Figure 1). Dredging every two years was required to sustain a given flow capability.

Figure 1. Example of a collapsed river bank

DESIGN OF REVETMENT SYSTEM

The goal of this project was to alleviate the extent of channel bank erosion. Ratherthan frequently dredging the channel, it was regarded as more economically viable to

design and construct a sustainable revetment system. The revetment design included

protection for both the toe and slope. As erosion of channel bed was likely, it wasdeemed necessary for the toe revetment to accommodate certain amount of deformation.

The design selected different types of toe revetments based on a comprehensive



10/65

consideration of various factors. Large stones were used for protecting the toe at

sections of the river that were relatively wide (Figure 2A). Pipe piles were installed fortoe protection at sections of the river that were narrow (Figure 2B).

Figure 2. A) Use of stones for toe protection; B) Use of pipe piles for toe protection

The design schematics of both types of revetments are shown in Figure 3. Compared

with the use of stones, the use of pipe piles required a lesser amount of embankment cutand backfills. However, the cost of installing the pipe piles was typically higher than

casting stones.

Figure 3. Schematic of design the revetments (length unit: mm; elevation unit: m)

Different designs were also used for the slope revetments. A no-fines concrete layer

was used at locations with toe protections. The protected toes provided support for theno-fines concrete layer. The thickness of the no-fines concrete layer was around 20cm.Holes were also neated in the no-fine concrete layers for the growth of vegetation and

aquatic habitat. Although not discussed here, a three-dimensional (3D) geogrid was

used at elevations higher than river level fluctuations. 3D geogrid is also an emergingtype of ecologic revetment. It has the advantageous of easy construction, low cost and

good erosion protection. It can prevent precipitation-induced river bank erosion at

locations where grasses are yet to grow. Later the grass, geogrid and near surface soils

road

3D geogrid

3D geogrid

road

Non-fine concrete

a) stone as toe revetment b) pipe pile toe revetment

Non-fine concrete

Original slope Original slope



11/65

became a unit with excellent resistance to erosion. This method is an effective

protection for banks with relatively steep slopes prone to erosion. However, geogridsare ineffective for protecting grass seeds from washing out in the zone with water level

fluctuations. In contrast, it was found that seeds could grow in the pore spaces of the no-

fines concrete even under frequent water level fluctuations. It was partly attributed to

the fact that the tortuosity of pores in the no-fine concrete helped retain the seeds andthe necessary nutrition.

The design of erosion control revetments using no-fines concrete embraced

comprehensive considerations of various factors, both mechanical and ecologic. From

the mechanical aspect, the factors included the ability of erosion prevention (shearstrength or compression strength of no-fines concrete), and accommodation of

settlement (flexure strength and thickness). From the ecologic aspect, factors considered

included grass seeds growth and nutrition retention (size of aggregates or air voids), and

pre-installed holes (survey of the typical size of aquatic communities). Another designstep from the ecologic consideration was the screening of the vegetation to determine

plant compatibility with no-fines concrete, aesthetic value and contribution to erosionresistance.

An experimental program was conducted to determine the design parameters for the

no-fines concrete embracing the structural, environmental and ecologic factors. The no-fines concrete design for this project was based on the experimental program and prior

experience. Granite aggregates with nominal diameters of 31.5 to 63mm were used as

coarse aggregate. The on-fines concrete consisted of 1620kg of aggregate, 240kg of

cement and 84kg of water respectively. The 28-day design strength was required to beno smaller than 5.6MPa. The strength for flexural failure was required to be no less than

1.2MPa. The thickness of the protection layer was at least 20cm. 7cm diameters holes

were required to be preinstalled every 25cm to accommodate existing aquatic life.Structural joints were set every 1.5m.

The selection of grasses was also based on a pilot experimental program. The criteriawas that the grasses must be able to survive the frequent inundation during river level

fluctuations. Thirty-four different types of grasses were tested in the screening program.

The growth of grasses under simulated conditions was measured in a two-month period.The record included their survival rate, area of cover, height, and aesthetic condition. A

few grasses were found to perform satisfactory and were recommended for use for this

project.

CONSTRUCTION AND PERFORMANCE OF THE NEW REVETMENT

SYSTEM

The revetment system included both toe and bank protection. Stone or pipe piles were

used at locations where toe protection was deemed necessary. Both the stone and pipe

pile toe revetments were found to be able to accommodate the deformation of the softsoils. They were both found to achieve the desired erosion improvements within

reasonable costs. The pipe pile toe protection had a relatively high construction costs.



12/65

However, compared with the gravel toe revetment, it also had a few advantages: 1) pipe

piles provided increased stability. It was found that there was no significant deformationof the pipe pile revetment after one year in service. Collapse, however, did happen at

certain locations using tone toe revetment; 2) improved construction quality control can

be achieved for pipe pile construction. The construction process is relatively

straightforward. The construction quality can be easily evaluated by observing the pipepile alignment. In comparison, strict construction control was not easy to achieve with

the stone toe revetment. Overexcavation or overfilling were found during theconstruction process; 3) pipe pile toe revetment typically resulted in better visual effects,

especially when exposed during low water levels. Construction of pipe piles was

relatively simple. A photo taken during construction is shown in Figure 4.

Figure 4. Construction of pipe pile toe protection

No-fines concrete and pre-casted bricks were used for bank protection in the zonewithin water level fluctuations. An example of a no-fines concrete revetment is

illustrated in Figure 5. Cured conventional concrete frames provided confinement

during the curing period of no-fines concrete. Holes were created in the no-fines

concrete. Seeds and nutrition were also sprayed in the no-fines concrete. Pre-castedbricks were used at a few locations for comparison.



13/65

Figure 5. Appearance of no-fines concrete revetment

Both the no-fines concrete and pre-casted bricks were found to retain good ecologicconditions. The bricks were found to have better strength and ability to accommodate

larger settlements than no-fines concrete. The cost of the pre-casted bricks was also

slightly higher. However, for locations with no vegetation, the no-fines concrete wasfound to provide better retention of grasses than bricks, possibly due to the turtuosity of

the pores which prevented seeds from washing out. As its pores accommodated the

sludge, the no-fines concrete also visually appeared to be closer to the natural slope than

the precasted bricks (Figure 6).

Figure 6 Appearance of precasted bricks (left) and no-fines concrete (right)

The designed revetment systems performed satisfactory. Vegetation grew very well

within three months after the construction was accomplished. A variety of aquatic life

was observed in the grasses. River crabs and other aquatics were spotted in thepreinstalled holes in the no-fines concrete. After the construction of the revetment

No-fines concrete

Conventional

concrete frame



14/65

system, the channel became much more aesthetically pleasant. A variety of colonies

were found around the channels. Figure 7 shows views of the river after the ecologicrevetment systems were constructed.

Figure 7. View of the river after the construction of the revetment system

CONCLUSIONS

The use of no-fines concrete was introduced in conjunction with a project involvingriver revetment design. The natural embankment was subjected to extensive erosion and

collapse due to inadequate scour protection. The erosion protection revetment involving

the use of no-fines concrete achieved the desired ecological function. As a result, the

local river ecologic system was retained and aquatics achieved prosperities. No-finesconcrete was found to be an effective ecologic method for river bank erosion control.

Factors such as strength and ecologic compatibility need to be taken into account in thedesign. Upon the application of proper design, the use of no-fines concrete can achieveadequate protection and retain the existing river ecosystem.

ACKNOWLEDGEMENTS

The authors would like to thank Mr. Lihua Wang, research engineer at the Hydraulic

and Hydropower Institute of Guangdong for his assistance in testing properties of non-

fines concrete and Mr. Shourong Wu, senior engineer at Guangzhou InvestigationDesign and Research Institute of Water Conservancy & Hydropower, for initializing the

use of no-fines concrete in the revetment design.

REFERENCE

Dong, J.W., Pei, Y.B., and Wang, L.Q. (2002). Study and practice of environmentprotection concrete. Chinese Journal of Jilin Hydraulics, 2, 1-4.



15/65

How does Water-Soil Interaction lead to Erosion?

George W. Annandale, D.Eng., P.E., D.WRE, F.ASCE1

1President, Engineering and Hydrosystems, Inc., 8122 SouthPark Ln., Suite 205,

Littleton, Colorado. PH 303-683-5191 e-mail: [email protected]

ABSTRACT

By making use of the Shields (1936) diagram the paper illustrates that the

commonly held belief that erosion of earth material results from the interaction

between shear stress caused by flowing water acting on it is incorrect. It is shownthat erosion results from shear stress under laminar flow conditions only. Fluctuating

pressure is responsible for erosion when the flow is turbulent. This observation isparticularly important when investigating erosion and scour of earth materials otherthan non-cohesive soils. For example, scour of rock formations and cohesive soils

subject to turbulent flow can be satisfactorily explained only in terms of the effects of

fluctuating turbulent pressures (Annandale 2006; 2007).

INTRODUCTION

A common misconception is that erosion of soils is caused by the shear stressof water flowing over a water-soil interface. This paper explains that such an

interpretation is only valid in the case of laminar flow. It continues to show that

fluctuating pressure is the principle cause of erosion associated with both smooth andrough turbulent flow. This is a very important observation because it assists

practitioners in conceptualizing and understanding erosion scenarios for solving

practical problems, particularly erosion and scour in earth materials like cohesivesoils and rock formations (Annandale 2006). The difference in the interaction

between flowing water and earth material for laminar and turbulent flow is illustrated

by making use of the Shields (1936) diagram.Determination of the threshold of incipient motion of non-cohesive sediment

impacted by flowing water is commonly accomplished by making use of the Shields

diagram (Figure 1), which relates dimensionless shear stress to the particle

Reynolds number*

Re . The dimensionless shear stress is also known as the Shields

parameter and is expressed as,

( )s d

=

(1.1)

where = boundary shear stress;s

= unit weight of sediment; = unit

weight of water; d= particle diameter.



16/65


17/65

forces results in the flowing water attempting to drag the assembly of particles at the

surface along with it.One of the particles, with forces acting on it, is shown in the lower portion of

Figure 2. The force F represents the action of the laminar flow, and the forceR

F the

resistance offered by the soil particle. The forcegW is the submerged weight of the

particle, and the angle represents the angle of friction between this particle and the

ones adjacent to and below it. The shear stress can be found by dividing the forces

F and RF by the projected horizontal surface area A of a particle, i.e.

0

F

A = and RR

F

A = . When the bed shear stress is just large enough to initiate

movement of the particles it is known as the critical shear stress, represented by the

symbolc

. Incipient motion is imminent whenc R

= = .

The resisting force FR is a function of the submerged weight of the soil

particle and the angle of friction, which is expressed as:

( )

3

tan tan6

R g sdF W g = = (1.3)

Wheres = mass density of the soil; g = acceleration due to gravity.

Figure 2 Forces active during scour of earth material underlaminar flow conditions.

The critical shear stressc

that should be exceeded for incipient motion to

occur can be expressed as,

( )24 2

tan3

Rc s

Fgd

d

= = (1.4)

F

gW

RF

Friction Angle =



18/65

The critical Shields parameter is therefore expressed as,

( )

2tan

3

c

sgd

= =

(1.5)

If the angle of friction of the soil 30o = then the value of the Shields

parameter becomes,

0.4 = (1.6)

Shields himself never measured this exact value of the minimum critical

dimensionless shear stress. He extrapolated his data and obtained a value of 0.1.

Subsequent research by Mantz (1973) and by White (1970) indicates a tendency to a

maximum value of 0.4, which confirms Equation (1.6).

TURBULENT FLOW

The critical Shields parameter characterizing incipient conditions of sediment

movement under turbulent flow conditions is about ten times lower than that required

by laminar flow. The reason for this is that the pressure fluctuations characteristic ofturbulent flow interact with individual elements in a non-cohesive soil, as opposed to

an assembly of elements when the flow is laminar. Less power is required to move

an individual elements than what is required to move an assembly of elements.

Croad (1981) illustrates the role of fluctuating pressures during incipientmotion by calculating the conditions for turbulent pressure fluctuations and

comparing it with the Shields diagram. Hinze (1975) found that the average

fluctuating pressures

'

p on a streambed under turbulent flow conditions can bequantified as

' 3p = (1.7)

This means that the root mean square of the fluctuating pressures is

approximately equal to three times the boundary shear stress . It is very important

to note that Equation (1.7) represents an empirical correlation and is not a

fundamental physical relationship.Emmerling (1973) further found that the positive and negative pressure peaks

can be as high as '6p , which means that the maximum instantaneous turbulent

pressure peaks maxp can be up to eighteen times the boundary shear stress, i.e.

max18p = (1.8)

If one now considers a sand grain with diameter d that is acted upon by an

upwards pressuremas

p (Figure 3), the total uplift force active on the particle can be

calculated as,



19/65


20/65

of sediment particles. As the flow conditions over the particles change from smoothto rough turbulent flow, the characteristic size of sediment particles become larger

relative to the characteristic size of the pressure fluctuation footprints (see e.g.

Annandale 2006). What this means is that incipient motion of sediment particles

under rough turbulent flow is the result of the combined effect of uplift due to

turbulent pressure fluctuations and rolling due to the effect of drag on the particles.This results in the increased value of critical dimensionless shear stress.

QUANTIFICATION OF FLUCTUATING PRESSURE

If the potential for incipient motion of non-cohesive granular material isinvestigated the Shields diagram can be used for both laminar and turbulent flow by

quantifying the magnitude of the shear stress imposed by the flowing water on the

streambed. The Shields diagram automatically allows for the effects of fluctuating

pressures resulting from turbulent flow, justifying the use of shear stress as anindicator of the relative magnitude of the erosive capacity of water.

However, when considering the possibility for erosion and scour to occur inearth materials other than non-cohesive granular earth material, such as rockformations or cohesive soils, it is necessary to quantify the relative magnitude of the

fluctuating turbulent pressures (Annandale 2006). This can be accomplished by

making use of Equations (1.7) and (1.8) or by making use of techniques developed by

Bollaert (2002) for turbulent plunging jets. Another approach is to use the rate ofenergy dissipation of flowing water (also known as stream power) as an indicator of

the relative magnitude of the erosive capacity of water.

Figure 4. Relationship between rate of energy dissipation and fluctuatingturbulent pressure (Annandale 1995).

160

180

200

220

240

260

280

300

320

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05

Rate of Energy Dissipation (W/m2)

Std.

DeviationofPressureFluctuations(Pa)



21/65

Turbulence is the principal cause of energy loss in flowing water, resulting ina correlation between rate of energy dissipation and turbulence intensity (Figure 4).

A simple formulation of the total rate of energy dissipation in flowing water can be

found by multiplying the mass rate of flow ( Q ) with the energy loss E (expressed

in units of length), i.e.

totalP Q E= (1.13)

Annandale (2006) showed that all the stream power is not applied to astreambed. The stream power applied to a streambed, which is less than the total

stream power, can be calculated as,

3

*7.853bedP u= (1.14)

The Erodibility Index Method (EIM) (Annandale 1995;2006) provides aconvenient way for determining the erosion potential of earth materials when using

stream power to quantify the relative magnitude of the erosive capacity of water. TheEIM is based on an erosion threshold relating the relative ability of earth materials to

resist erosion to the erosive capacity of water. The relative ability of earth materialsto resist erosion is quantified by means of a geo-mechanical index, known as the

Erodibility Index. The relative magnitude of the erosive capacity of flowing water is

equal to its stream power. Annandale (1995; 2006) developed a thresholdrelationship between stream power and the Erodibility Index by making use of field

and laboratory data (Error! Reference source not found.).

Figure 6. Erosion threshold relating stream power and the Erodibility Index

(Annandale 1995).

0.10

1.00

10.00

100.00

1000.00

10000.00

1.00E-02 1.00E-01 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04

Erodibility Index

StreamPowerKW/m2

Scour

No Scour

Scour-CSU

Threshold



22/65

The relative ability of rock to resist scour is quantified by means of theErodibility Index. This index is identical to Kirstens Excavatability Index (Kirsten

1981), and is expressed as,

(1.15)

where K = Erodibility Index; sM = mass strength number; bK = block size number;

dK = shear strength number; sJ = orientation and shape number.

The values of the four variable comprising the Erodibility Index can beaccomplished by making use of field identification techniques and tables presented in

Annandale (1995;2006).

SUMMARY

The paper illustrates that the commonly held belief that erosion of soils is

caused by shear stress is limited to laminar flow only. By making use of the Shields

diagram it is shown that incipient motion of sediment particles under turbulent flowconditions is caused by fluctuating turbulent pressures. The interaction between

fluctuating turbulent pressures and non-cohesive soil particles explains why lesseffort is required to cause erosion when flow is turbulent than when it is laminar.

The significance of this observation is that it assists in understanding of

erosion and scour of earth materials. For example, scour of rock formations cannot

possibly result from the action of shear stress, but is due to the effects of fluctuatingturbulent pressures (Annandale 2006; Bollaert 2002). Similarly, the erosion of clays,

which is characterized by a process that is commonly known as plucking is due to

the presence of fluctuating turbulent pressures and not due to the effects of shearstress (see e.g. Annandale 2006; Annandale 2007).

Four approaches for quantifying the relative magnitude of the erosive capacityof flowing water due to the presence of fluctuating pressures in turbulent flow are

presented. When considering the erosion potential of non-cohesive granular materialshear stress can be used to quantify the relative magnitude of the erosive capacity of

water provided the Shields diagram is used as the threshold criterion.

Direct quantification of the magnitude of fluctuating pressure can be

accomplished by using an empirical correlation between average and peak fluctuatingpressure and shear stress. Additionally, methods by Bollaert (2002) can be used to

quantify the magnitude of fluctuating pressure resulting from turbulent plunging jets.

When directly quantifying the magnitude of fluctuating pressures it is necessary touse physically-based methods for assessing the potential for scour. Such methods are

presented by Bollaert (2002) and Annandale (2006).The fourth option is to quantify the stream power of turbulent flowing water

and use Annandales (1996;2006) Erodibility Index Method as the threshold criterion.

REFERENCES

Annandale, G.W. 1995. Erodibility, J. of Hydraulic Research, vol. 33, No. 4, pp.471 494.



23/65

Annandale, G.W. (2006). Scour Technology, McGraw-Hill, New York.Annandale, G.W. (2007). Erosion of Clay: What do we know? Proc. GeoDenver

2007, ASCE, Denver, Colorado.

Bollaert, E. (2002). Transient Water Pressures in Joints and Formation of Rock

Scour due to High-Velocity Jet Impact, Communication No. 13. Laboratory of

Hydraulic Constructions, Ecole Polytechnique Federale de Lausanne, Switzerland.Croad, R.N. (1981). Physics of Erosion of Cohesive Soils, PhD Thesis, Department of

Civil Engineering, University of Auckland, Auckland, New Zealand.Gessler, J. (1970). Self-stabilizing Tendencies of Alluvial Channels, J. Waterways

and Harbors Division, ASCE, Vol. 96, No. WW2, pp. 235 249.

Emmerling, R. (1973). The instantaneous of the wall pressure under a turbulent

boundary layer flow, Max-Planck-Institut fur Stromungsforschung, Rep. No. 9.

Hinze, J.D. (1975). Turbulence, 2nd. Ed. McGraw-Hill, New York.

Kirsten, H.A.D. 1982. A classification system for excavation in natural materials,

The Civil Engineer in South Africa, pp. 292-308, July, (discussion in Vol. 25, No.5,

May, 1983).

Mantz, P.A. (1973). Cohesionless, Fine Graded, Flaked Sediment Transport byWater, Nature, Physical Science, Vol. 246, pp. 14-16.Shields, A. (1936). Application of Similarity Principles, and Turbulence Research to

Bed-Load Movement, California Institute of Technology, Pasadena, California

(translated from German).White, S.J. (1970). Plane Bed Threshold of Fine Grained Sediments, Nature, Vol.

228(5267), pp. 152-153.



24/65

Current State-of-the-Art of Rock Scour Technology

George W. Annandale, D.Eng., P.E., D.WRE, F.ASCE1

1President, Engineering and Hydrosystems, Inc., 8122 SouthPark Ln., Suite 205,

Littleton, Colorado.

ABSTRACT

A number of research efforts since 1991 resulted in the development of

technology that can be used to predict the erodibility and extent of scour in rock. This

research indicated four potential failure modes; i.e. abrasion, block removal, brittlefracture and fatigue failure. Scour can be predicted by making use of semi-empirical

and physically based methods. An empirical approach that was found to work well isknown as the Erodibility Index Method (Annandale 1995; 2006). Physically basedmethods, known as the Comprehensive Fracture Mechanics Approach, which is used

to predict scour by brittle fracture and fatigue failure, and the Dynamic Impulsion

Method, which is used to predict scour by block removal, have been developed byBollaert (2002). Methods to predict scour by abrasion are not developed as yet. This

paper summarizes current understanding and methods.

INTRODUCTION

Scour of rock, which occurs when the erosive capacity of water exceeds the

ability of rock formations to resist it, is of concern in infrastructures safety. Rock canscour in unlined water bearing tunnels, around bridge piers and downstream of dams

(Annandale 2006). Such events compromise the safety of infrastructure and pose a

threat to public safety and property. Methods for identifying rock scour potential andthe extent and rate of scour of rock are therefore required by practicing engineers

responsible for infrastructure safety.

Research on scour of rock that resulted in pragmatic methods for quantifyingits occurrence, extent and rate has been executed by Annandale (1995) and Bollaert

(2002). Annandale (1995; 2006) developed a threshold relationship defining incipient

conditions for rock scour by making use of a geo-mechanical index, known as the

Erodibility Index, to quantify the relative ability of rock to resist the erosive capacityof water. He quantifies stream power to estimate the relative magnitude of the erosive

capacity of water. The threshold defining incipient conditions for rock scour relates

the Erodibility Index and stream power. It is possible to estimate the maximum extentof rock scour by making use of the Erodibility Index Method (EIM) (see e.g.

Annandale 2006).

Bollaert (2002) investigated rock scour processes in detail by making use of aphysics-based approach. By making use of fracture mechanics and basic Newtonian

force relationships he developed techniques that can be used to directly quantify the



25/65

erosive capacity of water and use this information to determine whether rock will

scour by brittle fracture, in fatigue or by means of block removal. Using the methodsdeveloped by Bollaert (2002) it is possible to quantify both the extent and rate of rock

scour. Practice in rock scour assessment currently entails jointly using the methods by

Annandale (1995; 2006) and Bollaert (2002) to cross-check findings for developing

defensible engineering decisions about the safety of infrastructure as it relates to rockscour. In what follows rock scour mechanisms are presented as are methods for

quantifying the erosive capacity.

ROCK SCOUR MECHANISMS

The common belief that shear stress resulting from flowing water causes scourof earth materials is incorrect when considering turbulent flow. Annandale (2006)

demonstrates that the erosive capacity of turbulent flowing water is the result of

fluctuating pressures and advises against using shear stress concepts to explain erosivecapacity of water.

Rock can scour through four mechanisms, i.e. block removal, brittle fracture,sub-critical failure and abrasion. Fluctuating turbulent pressure lead to scour of rock

by brittle fracture, fatigue or by block removal (Bollaert 2002). Scour of rock byabrasion is currently considered of secondary importance and principally results from

impact by rock suspended in the turbulent flow or by rock elements sliding along the

surface of the intact rock. The four scour mechanisms are briefly described in whatfollows.

Figure 1 Rock scour by block removal (also known as dynamic

impulsion)

Fluctuating

Pressure

Transient

Pressure

1sF2sF

gW

upF

1sideF2sideF

downF



26/65

Block Removal

The magnitude of fluctuating pressure resulting from turbulent flow varies as a

function of space and time. Pressures introduced into rock joints (downF in Figure 1)

can result in increased pressure (up

F ) directly underneath the rock due to transients

and resonance that occur due to the introduction of fluctuating pressures into thediscontinuity surrounding the rock block. When upward pressure underneath the rock

exceeds the submerged weight of a rock block (g

W ) and the friction forces along its

sides (1sF and 2sF ), the rock will be removed from the rock formation. This failure

type is immediate and occurs as soon as the upward forces exceed the downward

forces.

Bollaert (2002) developed an impulse force balance equation for theconfiguration in Figure 1 for the case where a turbulent jet plunges into a pool of

water. This equation was solved by Annandale (2006), who found that the height

h through which a block of rock will be lifted when subjected to a turbulent plunging

jet falling into a pool can be expressed as,

( )

22

2 2 2

1 2 2

2

j

I s b sh

s b

v L L Lh C A Az F

g A z gc c c

=

(2.1)

if ( )2

2 2j I s b sh

v L L LC A Az F

gc c c + and h 0= otherwise.

where L = length of the open-ended joint around the rock block; c = pressure

wave celerity of the water; A = horizontal surface area below the rock block onto

which the water pressure acts;sh

F = total shear force on the sides of the rock block;

bz = vertical dimension of the rock block; jv = jet velocity; g = acceleration due togravity; = unit weight of water; s = unit weight of the rock; s = mass density of

rock;I

C = impulsion coefficient. The latter can be expressed as (Bollaert 2002),

2

0.0035 0.119 1.2Ij j

Y YC

D D

= +

(2.2)

where Y= plunge pool depth; jD = jet thickness at point of entry into the

plunge pool.

The appearance of the pressure wave celerity of water in Equation (2.1)reflects the role of resonance of fluctuating pressures in scour of rock by block

removal. Because the pressure wave celerity in water is a function of the free air

content in the water, the air content in a plunge pool can significantly affect rockscour. A relationship between the pressure wave celerity of water and air content is

shown in Figure 2.



27/65

0 0.2 0.4 0.6 0.8 1 1.20

200

400

600

800

1000

Air Content (%)

MixturePressureWaveCelerity(m/s)

Figure 2 Relationship between pressure wave celerity of water andwater air content

Figure 3 Rock scour by block removal for varying pressure wave

celerity of the water

The impact of air content on the potential for rock block removal is shown inFigure 3. The figure relates dimensionless vertical displacement of a rock block(ordinate) and the aspect ratio of the block. The aspect ratio is the quotient between

the blocks vertical dimensionb

z and its plan dimensionb

x (assuming the rock block is

square in plan). The dimensionless displacement of the rock block is the ratio between

the vertical distance h through which the rock block is displaced divided by its vertical

dimensionb

z . When the dimensionless displacement exceeds 1 it is concluded that

0 2 4 6 80

0.5

1

1.5

2

c = 1000 m/s

c = 600 m/s

c = 100 m/s

Displacement by Dynamic Impulsion

Aspect Ratio (zb / xb)

DimensionlessDisplacement(h/zb)



28/65

the rock block is definitely removed from its matrix. Figure 3 therefore demonstrates

that if the air content of the water in a plunge pool increases to only about 1% (i.e. c =

100 m/s), the chances for removing the block from its matrix increases significantly as

compared to a rock block subjected to the same turbulent pressure fluctuation, but

with zero air content in the water ( c = 1,000 m/s).

Brittle Fracture

Brittle fracture of rock occurs when the stress intensity at the edges of close-

ended fissures resulting from the introduction of fluctuating pressures into the fissuresis greater than the fracture toughness of the rock. When this occurs the rock fails in an

explosive manner (Figure 4). Such failure can result in the rock breaking up into

smaller pieces. An example of rock scour by brittle fracture has been found at SantaLuzia Dam, Portugal (Annandale 2006). This failure type is immediate, i.e. it occurs

instantaneously as soon as the stress intensity introduced by the fluctuating pressures

into the rock fissure exceeds the fracture toughness of the rock.

Figure 4 Scour of rock by brittle fracture

Bollaert (2002) and Annandale (2006) provide detailed procedures forcalculating the potential for brittle fracture of rock due to the erosive capacity of

water. The stress intensity at the edge of a close-ended rock fissure is calculated with

the equation,

RockFissure

Fluctuating pressures

Fracture toughness

of rockStress Intensity


Introduced into fissure


acting on a rock with

fissures

Brittle fracture occurs if stress

intensity

exceeds fracture toughness (sudden

explosive failure).



29/65

waterK a f = (2.3)

where K= stress intensity in the rock due to pressure p ; f = stress intensity

factor, depending on the shape of the fissure;f

L = length of the close-ended fissure.

The stress intensity in the fissure thus calculated is compared to the fracturetoughness of the rock. An equation developed by Bollaert (2002) using data published

in Atkinson (1987) is currently used to quantify the fracture toughness of rock,

( ) ( ), , 0.008 0.010 0.054 0.42 I insitu UCS iK to UCS = + + (2.4)

whereinsitu

K = fracture toughness; UCS = unconfined compressive strength of

the rock;i = confining pressure of the rock.

Brittle fracture will occur when,

K > KI,insitu,UCS (2.5)

Such failure is immediate and explosive.

Sub-Critical Failure (Fatigue)

Scour of rock by sub-critical failure occurs when the fluctuating stressintensities at the edges of close-ended fissures do not exceed the fracture toughness of

the rock. Continued application of the fluctuating pressures in the close-ended rock

fissures eventually results in break-up of the rock due to fatigue (Figure 5). Thisfailure type is time-dependent.The time it takes for rock to fail in fatigue can be estimated with the equation

by Paris et al. (1961),

( )mdL

C KdN

= (2.6)

where N = number of cycles of the fluctuating pressure that will lead to fatigue

failure; ,C m = material properties; K = range of stress intensities introduced to the

material by the fluctuating pressures; L = distance of crack growth required for thematerial to fail.

This equation can be integrated to directly calculate the number of pressure

cycles required for fatigue failure, i.e.



30/65

( )

( ) 1 12 2

11

12

m m

fm

water

N a LmC f

=

(2.7)

where waterDs = range of fluctuating water pressure; a = length of the close-ended

fissure; fL = overall material thickness.

Once the number of cycles for the rock to fail in fatigue has been calculated,

the actual time to failure can be estimated by dividing by the frequency of the pressure

fluctuations. A reasonable estimate for such a frequency is about 25Hz. Relationshipsbetween discharge of a plunging jet and time to fail in fatigue for a basalt rock in

Iceland are shown in Figure 6 (Annandale 2006). Such relationships provide

practicing engineers with the ability to make decisions as to development ofappropriate designs for protection against scour of rock.

Figure 5 Rock scour by sub-critical failure, also known as fatigue it

time-dependent (Annandale 2006). Failure proceeds according to theprogression from 1 to 4.

1 2

3 4

Fluctuating

Pressure

Close-ended

Fissures

ProgressiveFissure

Growth

Eventual

Fatigue Failure



31/65

AbrasionResearch into scour of rock by abrasion is in its infancy. However, current

interpretation is that scour of rock by abrasion can occur if the fluid interacting with

the rock is abrasive enough relative to the resistance offered by the rock to cause it toscour by breakage. Two potential mechanisms exist: a) the rock formation is impacted

by other blocks of rock suspended by the turbulence in the water impacting onto it, b)the other mechanism consists of such blocks of rock sliding over the rock formation.

In the former case, the rock is considered to fail by pieces breaking off. In the lattercase it is deemed to scour due to removal of rock on a layer by layer basis.

Figure 6 Example result of the relationship between scour depth and

time to scour for rock failing in fatigue as a function of varying

discharge into a plunge pool (Annandale 2006).

10 20 30 400

1000

2000

3000

4000

5000

6000

950 m3/s

800 m3/s400 m3/s

200 m3/s

100 m3/s

50 m3/s

Depth (m)

Time(days)



32/65

ERODIBILITY INDEX METHOD

The EIM is based on an erosion threshold relating the relative ability of rock to resist

scour and the erosive capacity of water. The relative ability of rock to resist scour is

quantified by means of a geo-mechanical index, known as the Erodibility Index. Therelative magnitude of the erosive capacity of flowing water is equal to its stream

power. Annandale (1995; 2006) developed a threshold relationship between streampower and the Erodibility Index by making use of field and laboratory data (Figure

7Error! Reference source not found.).

Ability of Rock to Resist ScourThe relative ability of rock to resist scour is quantified by means of the

Erodibility Index, which is expressed as,

sdbs JKKMK = (2.8)

where K= Erodibility Index; sM = mass strength number; bK = block size

number;d

K = shear strength number;s

J = orientation and shape number.

The mass strength number is directly related to the Unconfined Compressive

Strength (UCS) of the rock and represents the potential for brittle fracture or fatigue

failure of the rock. The block size number represents the relative size of rock blocksand is related to the potential for scour by block removal (dynamic impulsion). The

other two numbers, i.e. the shear strength and the orientation and shape numbers,

relate the difficulty to remove rock from the matrix due to the introduction of friction

and orientation of forces and is therefore related to block removal. The number values

can be quantified by referring to tables published in Annandale (1995; 2006).

Erosive Capacity of WaterThe relative magnitude of the erosive capacity of flowing water, for use in the

EIM, is equal to its stream power. The simplest way to explain the concept is to

compare it to the power that is available for hydroelectric generation. For example,the available power at a hydroelectric facility is quantified as,

Power = Q H (2.9)

where Q = total discharge; H= total head.

So, for example, if one wishes to calculate the stream power per unit area of a

jet plunging over a dam (for use in Figure 7) the result of Equation (2.9) is divided by

the footprint area of the jet at the point of impingement.In order to facilitate quantification of stream power for other flow types

Annandale (2006) derived a number of equations that can be used to quantify the

stream power for hydraulic jumps, head-cuts, flow over knick-points and in channels.Annandale also illustrates how the stream power for flow in rock-lined water bearing



33/65

tunnels, in flow around bends in open channels and around bridge piers can becalculated. In cases where it is not possible to theoretically quantify the stream power,

e.g. very complex flow conditions, one can use physical hydraulic model studies.

Annandale (2006) provides techniques for calculating the stream power by making useof measured flow velocities and flow depths obtained from such models.

Scour ExtentThe extent of scour in rock using the EIM can be quantified by comparing

available and required stream power. Available stream power is the stream power

contained in the flowing water, i.e. its erosive capacity. Required stream power is

another name for threshold stream power, i.e. the stream power that is required toscour rock. When the available stream power is greater than the required stream

power, the rock will scour. As soon as the available stream power is less than the

required stream power, scour will cease. Therefore, in calculating the extent of scour,it is necessary to quantify the change in stream power as a scour hole develops.

Calculation of such change depends on the flow situation. Annandale (2006) outlines

methods that can be used to quantify the decay of stream power in common flowsituations.

Figure 7 Erosion threshold relating stream power and the Erodibility

Index (Annandale 1995; 2006)

The change in threshold stream power, i.e. the stream power required to scour

rock, is determined as a function of elevation below the ground surface by indexing

0.10

1.00

10.00

100.00

1000.00

10000.00

1.00E-02 1 .0 0E- 01 1. 00 E+0 0 1.00E+01 1.0 0E+02 1.00 E+0 3 1.00E+04

Erodibility Index

StreamPowe

rKW/m2

Scour

No Scour

Threshold



34/65

the rock using the Erodibility Index and core logs. Once the Erodibility Index of therock at a particular elevation is known, Figure 7 is used to determine the stream power

required to scour the rock. This threshold stream power is then plotted as a function of

elevation below the ground surface.Comparison of the available and threshold stream power as a function of

elevation below the ground surface (Figure 8) provides an indication of the maximumscour depth. This occurs when the erosive capacity of the water, i.e. the available

stream power, is less than the threshold stream power.

Figure 8 Calculation of scour depth by comparing available stream

power of flowing water and threshold (required) stream power of the

rock.

SUMMARYThe mechanisms leading to scour of rock have been presented. Rock scours due to the

effects of fluctuating pressures resulting from turbulent flowing water by following

four failure mechanisms, i.e. brittle fracture, fatigue failure, dynamic impulsion andabrasion. Methods for calculating scour resulting form the first three mechanisms are

presented. Scour by abrasion is considered of secondary importance and research into

quantifying scour due to its effect is required. Technology for estimating scourpotential, extent and rate is presented. Application of Bollaerts (2002)

Eleva

tion

Eleva

tion

E

levation

Erosive Capacity Erosion Threshold

Erosive Capacity /

Erosion Threshold

Maximum Scour

Depth

Erosion Threshold of Earth MaterialErosive Capacity of Water



35/65

Comprehensive Fracture Mechanics and Dynamic Impulsion approaches concurrentlywith Annandales (1995; 2006) Erodibility Index Method provides a means of cross-

checking results and is the favored approach to rock scour analysis.

REFERENCES

Annandale, G.W. (1995). Erodibility, J. of Hydraulic Research, vol. 33, No. 4, pp.471 494.

Annandale, G.W. (2006). Scour Technology, McGraw-Hill, New York.

Atkinson, B. K. (ed.) (1987). Fracture Mechanics of Rock, Academic Press.

Bollaert, E. (2002). Transient Water Pressures in Joints and Formation of Rock Scour

due to High-Velocity Jet Impact, Communication No. 13. Laboratory of Hydraulic

Constructions, Ecole Polytechnique Federale de Lausanne, Switzerland.

Paris, P.C., Gomez, M.P. and Anderson, W.E. (1961). Trend Engineering, Vol. 13, pp.9-14.



36/65

1

Evaluation of Soil Erosion Using the Rainsplash Technique

J.L. Smith1, M.R. Davieau

2, and S.K. Bhatia

3

1Research Fellow, Civil and Environmental Engineering (CIE) Department, Syracuse

University (SU), 151 Link Hall, Syracuse, NY 13244; PH (315) 443-2313; FAX

(315) 443-1243; email: [email protected] Research Assistant, CIE Department, SU, 151 Link Hall, Syracuse,

NY 13244; PH (315) 443-2313; FAX (315) 443-1243; email: [email protected] J. and L. Douglas Meredith Professor, CIE Department, SU, 151 Link Hall,

Syracuse, NY 13244; PH (315) 443-3352; FAX (315) 443-1243; email:

[email protected]

ABSTRACT

Soil erosion by raindrop impact is an important source of soil erosion from slopes

and other non-concentrated flow areas. Vegetation is one of the most effective ways

for minimizing soil erosion by raindrop impact. Ground covers, such as mulches,

rock, and rolled erosion control products (RECPs) are also effective in reducing the

effects of raindrop impact by limiting the amount of soil exposed.This paper presents the results of a laboratory study that compares the rainsplash

performance of 4 different RECPs (wood excelsior, coconut, jute, and polypropylene

(PP)), with and without vegetation, using a rainsplash simulator. The results are used

to evaluate the contribution of vegetation and the role of RECPs in minimizing soil

erosion.

Based on the results of 44 tests, it was found that the use of RECPs on bare soil

slopes reduces soil losses by 74% to 91%. The use of vegetation in combination with

RECPs further reduces soil losses by 95% to 99%. Rainsplash tests provide a useful

venue for comparing the performance of RECPs and vegetation and the effects of

different RECP index properties on soil loss results. It is recommended that

rainsplash results be validated with large-scale ASTM testing and field results.

INTRODUCTION

Soil erosion is the detachment and transport of soil particles from the ground surface

by raindrops, water, or wind. It can lead to decreased soil productivity, increased

flooding, and increased turbidity in surface waters. Soil erosion by raindrop impact is



37/65

2

one of the most important sources of soil erosion from slopes and other non-

concentrated flow areas, particularly on unprotected construction sites.

In the raindrop erosion process, soil particles are detached from the ground surface

by raindrops; entrained in the sediment load; transported by thin films of water; and

deposited (Toy et al. 2002). The extent to which soil is eroded by rainfall is

dependent on the characteristics of the storm event, such as raindrop velocity andsize, and storm intensity, the erodibility of the soil, slope of the surface, and cover

type (Ellison 1944). The steeper the gradient, the faster dislodged soil particles travel

down a slope. Raindrops can also dislodge soil particles from underneath films of

water that are less than three times the raindrop diameter (Mutchler and Young 1975).

The importance of cover in reducing the erosivity of raindrops is well known (e.g.,

Gray and Sotir 1996; and Toy et al. 2002). Vegetation is one of the most important

ways for minimizing soil erosion by raindrop impact. Vegetation reduces the energy

of raindrops by intercepting them; provides soil reinforcement with its roots and

stems; decreases runoff velocities with its roughness; and maintains the infiltration

properties of the soil (Gray and Sotir 1996). Ground covers, such as mulches, rock,

and rolled erosion control products (RECPs), are also effective in reducing the effectsof raindrop impact by limiting the amount of soil exposed (Toy et al. 2002).

The effectiveness of vegetation and ground covers in protecting soil slopes against

raindrop impact is dependent on the percentage of ground covered by the vegetation

or ground cover, its density, and its closeness to the ground (Toy et al. 2002). Evans

(1980) noted how erosion and runoff rapidly increase on soils with less than 70%

vegetative cover. Rickson and Morgan (1988) found that soil particle detachment

rates by raindrops increase exponentially with decreasing ground cover.

This emphasizes the need for providing suitable cover protection on exposed soil

surfaces, whether it is by providing dense vegetative cover or through the selection of

proper ground cover. This is particularly important on construction sites, where bare

soils are exposed until vegetation can be established. This is often difficult on steepslopes and in high flow areas (Gray and Sotir 1996).

Rolled Erosion Control Products (RECPs) were developed to fill this gap. RECPs

are temporary degradable or long-term non-degradable materials manufactured or

fabricated into rolls that can be easily installed on sites (ECTC 1998). They provide

immediate protection for bare ground and assist with the growth, establishment,

protection, and reinforcement of vegetation.

Many researchers have conducted field studies to compare the effectiveness of

RECPs in establishing vegetation and minimizing soil erosion (e.g., Fifield 1992;

Krenitsky and Carroll 1994; and Bhatia et al. 2002). However, few studies have been

conducted to evaluate the performance of vegetation and vegetation in combination

with RECPs under controlled large-scale (e.g., Sutherland 1998) or laboratoryconditions.

This paper presents the results of a laboratory study that compares the rainsplash

erosion performance of four different RECPs (wood excelsior, coconut, jute, and

polypropylene (PP)), with and without vegetation, using a rainsplash simulator

constructed at Syracuse University. The results are used to evaluate the contribution

of vegetation and the role of RECPs in minimizing soil erosion.



38/65

3

MATERIALS AND METHODS

Soil. The soil selected for the evaluation is a topsoil that is commercially available in

New York. The soil was selected because of its similarity to the ASTM sand

specified in the ECTC Draft Test Method #2 for testing RECPs under simulated

rainfall conditions (ECTC 2003) and its similarity to planting mediums used at localconstruction sites. The topsoil has an optimum moisture content of 19.6% at 1.68

g/cm3

dry density. Its grain-size distribution in comparison to the ASTM testing sand

is shown on Figure 1.

Figure 1. Grain-size distribution of the topsoil in comparison to ASTM sand

RECPs. Four different RECPs were selected for this study. The RECPs were

selected based on their differences in structure type, matrix type, and index

properties. Two of the RECPs (wood excelsior and coconut) are erosion control

blankets (ECBs), temporary degradable RECPs composed of processed natural orpolymer fibers mechanically, structurally, or chemically bound together to form a

continuous matrix (ECTC 2001). The third RECP (jute) is an open weave textile, a

temporary degradable RECP composed of processed natural or polymer yarns woven

into a matrix (ECTC 2001). The fourth RECP (PP) is a turf reinforcement mat

(TRM), a long-term, non-degradable RECP composed of UV-stabilized, non-

degradable, synthetic fibers, nettings, and/or filaments processed into 3-D

reinforcement matrices (ECTC 2001). The index properties of the RECPs, as

determined by the authors in accordance with ECTC (2001), are given in Table 1.

The results given in Table 1 are average values of five specimens tested.

Table 1. RECPs used in this study and their average measured index propertiesRECP Matrix/Structure

Type

Mass/ Area

(g/cm2)

Thickness

(cm)

Light

Penetration (%)

Flexural

Rigidity (g-cm)

W2 Wood ECB 0.0721 0.959 12 130.4

C2 Coconut ECB 0.0246 0.350 20 18.8

J1 Jute OWT 0.0451 0.344 50 4.44

T2 PP TRM 0.0569 1.109 25 589

0%

20%

40%

60%

80%

100%

0.0010.010.1110100

Grain Size (mm)

PercentFiner

ASTM Sand

Topsoil



39/65

4

Rainsplash erosion tests. Rainsplash erosion tests were performed with a rainsplash

simulator constructed at Syracuse University. The simulator was constructed in

accordance with ECTC (2003), in consultation with TRI/Environmental, Inc. (TRI).

The simulator is approximately 2-m tall and produces raindrops that fall onto an

adjustable slope table (see Figure 2.) The simulator produces raindrops with a

medium diameter of 2.2 mm from a drop height of 2 m above the lowest point of theslope table. Raindrops are produced using four, 30.5-cm long, 2-cm diameter

Schedule 40 PVC pipes, with 0.08-cm diameter holes at 2.54-cm spacing. The lateral

pipes are attached perpendicularly to a gear motor that rotates at a frequency of 6

revolutions per minute (rpm.) The holes are directed upwards to encourage the

freefall of water from the lateral pipes. Water is fed to the system using a 1 to 4 liter

per minute (lpm) flow meter and 0 to 6.9 kPa pressure regulator to control the flow of

water to the pipes. A 15.24 cm per hour rainfall intensity was simulated for the tests.

Figure 2. Photographs of the rainsplash simulator at Syracuse University

The adjustable slope table (0 to 3 horizontal:1 vertical) contains three 89-cm long by

25-cm wide channels that are separated by 25-mm metal dividers. The base of each

channel contains a recessed hole where prepared soil cores are placed and tested. The

soil cores are watertight containers approximately 20 cm in diameter and 10 cm deep.

Soil was compacted in the cores at 90 + 3% of standard dry density at optimum

moisture content + 2%, in accordance with ASTM D698.Soil was placed at a moist unit weight of 1.36 g/cm

3+ 0.08 g/cm

3and 35 to 40%

moisture content for the vegetated specimens. The soil was then sown with Kentucky

31 tall fescue grass seeds and watered. Seeds were placed at a rate of 0.75 g per soil

core, which is approximately equal to 350 seeds. This amount of grass seed is twice

the rate specified in the ECTC Draft Test Method #4 for encouraging seed

germination and plant growth under bench-scale conditions (ECTC 2004) to ensure

that an adequate stand of grass would be established after 21 days. The soil cores



40/65

5

were placed in an environmental chamber capable of sustaining conditions at 27 +

2oC, 45 + 5% relative humidity, and 9687 lumens/m

2+ 1076 lumens/m

2. The soil

cores were watered twice per week with 100 ml of water. The vegetated samples

were tested for rainsplash performance after 21 days of growth.

Tests were performed for durations of 1 hour. Soil was collected, runoff was

measured, and turbidity readings were taken at increments of 5 minutes through 30minutes and at 45 and 60 minutes during the test. A minimum of five tests was

performed for each condition tested. Note that testing was only conducted in the

outside channels of the simulator because the rainfall conditions were more similar

there than in the middle of the channel, which exhibited higher flow rates (nearly

double) due to the nature of the simulator.

RESULTS

Bare soil. Bare soil samples were tested in the rainsplash simulator to serve as a

baseline for evaluating and comparing the effectiveness of the RECPs, vegetated

covers, and RECP/vegetated systems. As shown on Figure 3, there was goodagreement between the seven tests performed, demonstrating the consistency of the

simulator within and between different tests. The average rate of soil erosion for the

conditions tested was nearly linear (R2

= 0.9994) over the 1-hour period tested,

indicating that the rate of soil erosion was fairly constant throughout the test. The

soil concentration and turbidity of the runoff were also fairly constant throughout the

test, averaging 6.92x10-3 g/ml (=0.000196) and 820 NTUs (=14.1), respectively.

The average cumulative soil loss was 125 g (2.5%) after a 1-hour testing period.

Figure 3. Bare soil rainsplash results

Vegetated samples. Five soil cores were planted with Kentucky 31 grass seed and

allowed to grow for 21 days. After 21 days, the vegetated soil cores were tested in

the rainsplash simulator. The results are shown on Figure 4. As shown on the figure,

the rates of soil erosion were fairly constant throughout the test, similar to the bare

soil conditions (see Figure 3), although the rates of erosion were less, 0.925 average

slope for the vegetated specimens versus 2.015 for the bare soil specimens. The total

amount of soil loss after the 1-hour period was also 57% less, averaging 54 g.

y = 2.0152x + 4.7474

R2

= 0.9994

0

20

40

60

80

100

120

140

0 20 40 60

Time (min)

CumulativeSoilLo

ss(g)

Test #1

Test #3

Test #4

Test #5

Test #6

Test #7

Test #8

Average



41/65

6

Figure 4. Rainsplash results for the vegetated soil specimens

There was slightly more variability among the vegetated tests than among the baresoil tests, because of the nature of the vegetated specimens. A comparison of

biomass versus total soil loss, with the exception of specimen #5 (biomass 0.457g),

indicates an increasing trend of soil loss with decreasing biomass (see Figure 5). The

variation in specimen #5 could be due to differences in the diameter and the quality

of the shoots or their spacing. A similar correlation was made between the total

number of shoots and total soil loss. Overall, the vegetated specimens had an average

biomass of 0.421g ( = 0.035) and averaged approximately 162 shoots (=8.3.)

Figure 5. Vegetation biomass versus total soil loss

RECPs. Four different RECPs (wood excelsior, coconut, jute, and PP) were included

in the study. Average rainsplash results for the RECPs are given in Figure 6. Each

line represents the average of five different specimens. As shown in Figure 6, the

average total soil loss for the RECPs tested was between 74% and 91% less than for

the bare soil samples and between 39% and 78% less than for the vegetated soil

samples. Overall, the wood excelsior (W2) and coconut (C2) ECBs were the most

effective, with 91% and 89% reduction in soil erosion, respectively, in comparison to

0

20

40

60

80

100

120

140

0.30 0.35 0.40 0.45 0.50

Biomass (g)

TotalSoilLoss(g)

Test #1Test #2

Test #3

Test #4

Test #5

y = 0.9253x - 0.9178

R2

= 0.999

0

20

40

60

80

100

120

140

0 10 20 30 40 50 60

Time (min)

CumulativeSoilLoss(g)

Veg #1 (0.376g, 151)

Veg #2 (0.399g, 159)

Veg #3 (0.417g, 167)Veg #4 (0.454g, 173)

Veg #5 (0.457g, 162)

Ave-Veg (0.421g, 162)

biomass/# shoots



42/65

7

the average bare soil condition. This was followed by the jute OWT (J1) and PP

TRM (T2), with overall reductions of 84% and 73%, respectively. It is believed that

the ECBs (W2 and C2) were more effective than the OWT (J1) and TRM (T2)

because of their higher ground cover percentages (between 80% and 88% for the

ECBs and 50% and 75% for the OWT and TRM, based on light penetration results).

Figure 6. RECP rainsplash results

RECP/vegetated systems. Twelve soil cores were planted with Kentucky 31 grass

seed, covered with the four different RECPs (wood excelsior, coconut, jute, and PP)

(three specimens each), and allowed to grow for 21 days. After 21 days, the

RECP/vegetated systems were tested in the rainsplash simulator. The results are

shown on Figure 7. As shown on Figure 7, the average total soil loss for the

RECP/vegetated systems tested was between 88% and 99% less than for the

vegetated soil samples and between 54% and 95% less than for the RECP alone

samples. Overall, the wood excelsior ECB was the most effective, with an average99% reduction in soil erosion from the average vegetated condition. This was

followed by the PP TRM (T2), jute OWT (J1), and coconut ECB (C2), with overall

reductions of 92%, 90%, and 88%, respectively.

Figure 7. RECP/vegetated system rainsplash results

0

2

4

6

8

10

0 10 20 30 40 50 60

Time (min)

CumulativeSoilLoss(g)

W2-Veg (0.399g, 113)

C2-Veg (0.468g, 145)

J1-Veg (0.611g,125)T2-Veg (0.688g, 130)

biomass/# shoots

0

10

20

30

40

0 10 20 30 40 50 60

Time (min)

CumulativeSoilLoss(g

W2 (wood)

C2 (coconut)

J1 (jute)

T2 (PP)



43/65

8

DISCUSSION

The results of this comparative laboratory rainsplash study confirm that it is

important to protect bare soils from erosion caused by raindrop impact. This is

particularly important on construction sites prior to the establishment of vegetation.

The use of RECPs and RECPs in combination with vegetation can significantlyreduce soil erosion from these sites.

Figure 8 presents a comparison of the performance of the four different RECPs

(wood excelsior, coconut, jute, and PP) tested in this study versus bare soil

conditions. As shown in the figure, the use of RECPs reduces soil losses on bare soil

slopes, although to varying extents depending on the type of RECP used. Figure 9

presents a comparison of how the use of vegetation in combination with RECPs

further reduces soil losses on slopes. Although there were differences in the

performance of the RECP/vegetated samples, soil losses were in a relatively narrow

range. In both cases, the wood excelsior (W2) RECP was the most effective,

although all of the RECPs significantly reduced soil losses. These results also show

how RECPs can significantly improve the performance of vegetation for erosioncontrol. This is attributed to the fact that RECPs provide a suitable environment for

vegetation to grow, reinforce vegetative shoots, and provide protection in areas where

vegetation is sparse.

Figure 8. Comparison of RECP versus bare soil performance

Figure 9. Comparison of performance of vegetation versus RECP/vegetation systems

0

2040

60

80

100

120

140

Cumula

tiveSoilLoss(g)

Bare Soil

W2 (wood)

C2 (coconut)

J1 (jute)

T2 (PP)

0

20

40

6080

100

120

140

CumulativeS

oilLoss(g)

Bare Soil

Vegetated Soil

W2 (wood) w/vegC2 (coconut) w/veg

J1 (jute) w/veg

T2 (PP) w/vegW2 C 2 J1 T 2

W2 C2 J1 T2



44/65

9

Although the four RECPs were effective in reducing soil losses, the type of RECP

plays an important role in the overall performance. With more than 100 different

RECPs available in the US alone, rainsplash studies such as this are important to

understand how different RECPs, which have very different structure types and index

properties (light penetration, absorptive capacity, flexural rigidity, etc.) perform. For

example, there appears to be a relationship between light penetration and averagecumulative soil loss for the natural fiber (C2-coconut, J1-jute, and W2-wood

excelsior) RECPs tested (see Figure 10.) However, it is likely that other properties,

such as water absorption and flexural rigidity, play interrelated roles. Rainsplash

tests provide a useful venue for studying the effects of these properties on their

performance, evaluating quality control, and developing new products. It is

important to note, however, that rainsplash tests are conducted under controlled,

small-scale conditions. It is recommended that results be compared with large-scale

ASTM testing and field results to evaluate their applicability to field conditions.

Figure 10. Trends in light penetration results versus average cumulative soil loss

CONCLUSION

Based on the results of 44 simulated rainsplash tests (bare soil, vegetated soil,

RECPs, RECP/vegetated systems) using a 15.24 cm/hour rainfall intensity, the

following conclusions can be drawn:

1. The use of RECPs on bare soil slopes reduces soil losses by 74% to 91%.

2. The use of vegetation in combination with RECPs further reduces soil losses

by 95% to 99%. RECPs provide a suitable environment for vegetation to

grow. They reinforce vegetative shoots and provide protection in areas where

vegetation is sparse.3. There appears to be a relationship between light penetration and soil loss for

the natural fiber RECPs tested. However, it is likely that other properties,

such as water absorption and flexural rigidity, also play interrelated roles.

4. Rainsplash tests provide a useful venue for comparing the performance of

RECPs and vegetation and the effects of different RECP index properties on

reducing soil loss. It is recommended that rainsplash results be validated with

large-scale ASTM testing and field results.

0

10

20

30

40

0% 20% 40% 60% 80% 100%

Ave Light Penetration (%)

AveCumulativeSoilLo

ss(g)

C2 (coconut)

J1 (jute)

W2 (wood)

T2 (PP)



45/65

10

ACKNOWLEDGEMENTS

The authors would like to acknowledge support received from the National

Science Foundation (Award #3535848), the Department of Education (Foreign

Language and Area Studies Fellowship in South Asia), and the Syracuse

University Graduate School (University Fellowship Award.) The authors wouldalso like to thank Sam Allen and Jarrett Nelson (TRI) for their technical assistance

and Tony Johnson (American Excelsior Company), Bob Moran (Belton Industries),

and Roy Nelsen (North American Green) for providing materials used in this study.

REFERENCES

Bhatia, S.K., Smith, J.L., Lake, D., and Walowsky, D. (2002). A technical and

economic evaluation of geosynthetic rolled erosion control products in highway

drainage channels. Geosynthetics International, 9(2), 125-148.

Ellison, W.D. (1944). Studies o

geotechnics of soil erosion gsp167

Documents