TYPiMG GUIDE SHEE1
EPA 600/7-81-117
B;tGi?-: I
USE OF A VEGETATIVE FILTER
Z0TJE TO CONTROL FHSE-GRAINED SEDIMENTS
FROM SURFACE MIKES
by
Steve C. Albrecht Billy J. Barfield
Hittman Associates, Inc. ' University.of Kentucky
Lexington, Kentucky 40504 : Lexington, Kentucky 40506
Cooperative Agreement No. CS-805632
Project Officer
• Edward R. Bates
Energy Pollution Control Division
Industrial Environmental Research Laboratory
'. Cincinnati, Ohio 45268
This study was conducted
in cooperation with
Kentucky Department 'for Natural Resources
and Environmental Protection
INDUSTRIAL ESVIROMEENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND,DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
W- TEXT'*;
PAG" IV.JV:':\r
A-2C? {Ci.i.;
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory-Ciraciaraatij Of. S. Environmental Protection Agency, and approved
for publication. Approval does mot signify that the contents necessarily
reflect the views and policies of the D. S. Environmental Protection Agency,
nor does mention of trade names of commercial products constitute endorse-
ment of recommendation for use.
-------
FOREWORD
I When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
; health often require that new and increasingly more efficient pollutional
1 control methods be used. The Industrial Environmental Research Laboratory-
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
Federal and state regulatory agencies are involved in the reclamation
of mined lands and the elimination or reduction of sediment-laden drainage
.which pollutes rivers and streams. This report evaluates the effectiveness
of using vegetative filters to trap fine-grained sediments through the
results of a field demonstration. This report is intended to aid private
industry and government agencies in the reclamation of mined lands. For
further information, contact the Energy Pollution Control Division.
David G. Stephan
Director
Industrial Environmental Research Laboratory
111
-------
ABSTRACT
The objective of this study was "to demonstrate the
effectiveness of a vegetative filter zone in trapping fine-
grained sediments from surface mining operations". The area
selected for study was located in Whitley County, Kentucky,
directly below an active surface mining operation. The
outslope above the filter was the primary drainage area
monitored during the study.
This project was initiated with the specific purpose of
conducting a field test on vegetation as a viable sediment
trapping medium. From the onset, the project was wholly
designed for a field evaluation under typical mining conditions
The filter area was constructed directly below an abandoned
surface mine bench, on typical soil types found in mined
areas of Eastern Kentucky. The outslope located above the
filter was the primary area, from which sediment-laden
drainage was to be diverted to the inlet monitoring_station.
Sediment-laden water samples were collected at the inlet
'flume for comparison with samples collected at the outlet
flume to permit evaluation of the sediment removal capability
of the vegetative filter.
Results of the monitoring efforts revealed that a
dramatic reduction in sediment load was achieved by vegetative
filtration for particle sizes larger than clay. Based on
results of this study, it is concluded .that vegetative
filters are an effective control for reducing the quantity
of sediment transported into surface streams and rivers from
disturbed mined lands.
This report was submitted in fulfillment of Cooperative
Agreement CS-805632 by Hittman Associates, Inc., Natural
Resource Consultants, and the University of Kentucky, all of
Lexington, Kentucky. This report covers the period of June
31, 1979 to March 31, 1981.
IV
-------
TABLE OF CONTENTS
FOREWORD ii:L
ABSTRACT 1V
FIGURES V11
TABLES x:*-
ACKNOWLEDGEMENTS xii
SECTION 1. Introduction !
SECTION 2. Summary and Conclusions 3
SECTION 3. Background 5
Overview of Erosion and Sediment Control
Technology 5
Legislation, Funding and Site Selection 21
Description of Project Site 25
Project Objectives and Applications 27
SECTION 4. Preliminary Site Conditions and Vegetative Filter
Design 28
Geology and Mining of the Project Site 28
Soil Conditions 28
Hydrologic Design Considerations 31
Vegetative Design 54
Monitoring System Design 62
SECTION 5. Installation of Vegetative Filter 68
Site Alterations and Soil Handling 68
Vegetative Establishment and Performance VI
Water Monitoring System Installation 76
Time and Material Requirements of Filter
Installation. 89
SECTION 6. Sediment and Water Quality Monitoring 91
Sampling Procedures 91
Laboratory Analyses 93
SECTION 7. Description of Results 94
Background and Calculations 94
Storm Event Descriptions. 95
Steady State Tests 11]-
Summary
SECTION 8. Computer Model Validation
Introduction
Results 12°
v
-------
REFERENCES 127
APPENDICES 132
A. Computation Procedures for Runoff Hydrographs 132
B. Plots of Sensitivity Analysis of Kentucky Grass
"" Filter Model : 145
C. Analytical Procedures Utilized for Project
laboratory Testing 152
.GLOSSARY 157
vx
-------
FIGURES
Number Page
1 Schematic of the process of soil erosion by water
2 Schematic of the sediment deposition process in
erect media
3 Fraction trapped versus time for field test 1 ................. 18
4 Outflow concentration versus time for field test 1 ............ 19
5 Site location map ............... . ............................. 26
6 Previously mined areas above the project site ................. 29
7 Map and cross section ........... . ............................. 32
8 Runoff hydrograph for 2 , 5 and 10 year storm .................. 33
9 Slope segments used in sediment predictions ................... 35
10 Flowchart for erosion determination ........ ................... 37
11 Predicted size distribution of eroded sediment
aggregates .................................................. 38
12 Predicted runoff pollutographs ....... ............. . ........... 43
13 Predicted hydrograph and pollutographs for the design
filter with a 10 year storm. . ............................... 46
14 Predicted inflow and outflow sediment ......................... 47
15 Predicted hydrograph and pollutographs for the design
filter with a 5 year storm .................................. 48
16 Predicted inflow and outflow sediment size distributions
for the design filter using a 5 year storm .................. 49
17 Predicted hydrograph and pollutographs for the design
filter with a 2 year storm.
50
Vll
-------
FIGURES
Number
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Predicted inflow and outflow sediment size distribution
for the design filter using a 2 year storm
Predicted instantaneous trapping percentages for the
design filter with a 2 , 5 and 10 year storm
Predicted outflow concentrations for the design
filter for a 2 , 5 and 10 year storm
Tall fescue producing areas of the U.S. (37)
Ryegrass producing areas of the U.S. (37)
Calibration curves for the H-Flumes
Design drawings for site plan, monitoring system, H-Flume,
and flow divider
Design drawings for inlet and outlet flume implacements
Removal of topsoil for stockpiling
Rough grading of the subsoil
Stockpiled topsoil replacement
Qn-site erosion control using a straw bale dike
Fescue and Ryegrass after maturity
Diversions to direct flow into the filter area
Leveling concrete for flume supports
Positioning of H-Flumes „
Placement of metal flow divider.
Attachment of 4 mil. plastic to level spreader
Soil filling between 2x4 flow divider plates ,
Page
51
52
53
59
59
64
65
66
69
70
72
73
75
78
80
81
82
83
... 84
Vlll
-------
FIGURES
Number Page
37 Completed flow divider network 85
38 Plastic diversion leading to lower flume 86
39 Completed inlet and outlet monitoring stations 88
40 Point gauge mounted in H-Flume for use in hand sampling. 92
41 Inflow and outflow concentrations for storm on January 11, 1980. 96
42 Inflow and outflow concentrations for storm on April 3, 1980.... 97
43 Inflow and outflow hydrograph for storm on .April 3, 1980 98
44 Aggregate size distribution of sediment for storm on April 3,
1980 99
45 Inflow and outflow sediment concentrations for storm on
April 11, 1980 1°]-
46 Inflow and outflow hydrograph for storm on April 11, 1980 102
47 Inflow and outflow sediment concentrations for storm on
May 17, 1980 103
48 Inflow and outflow hydrograph for storm on May 17, 1980 104
49 Inflow and outflow sediment concentrations for storm on
May 24, 1980 106
50 Inflow and outflow sediment concentrations for storm on
July 10, 1980 107
51 Inflow and outflow hydrograph for storm on July 10, 1980 108
52 Inflow and outflow sediment concentrations for storm on
August 29, 1980 109
53 Aggregate size distribution of sediment for storm on
August 29, 1980 110
IX
-------
FIGURES
Number
54 Inflow and outflow sediment concentrations for storm on
September 3, 1980 ............. .
55 Aggregate size distribution of sediment for storm on
September 3, 1980 ............. .
57 Inflow and outflow sediment concentrations for
steady-state test 2.
59 Inflow and outflow hydrograph for steady-state tests
56 Inflow and outflow sediment concentrations for
steady-state test 1 ....... ...................................
115
58 Inflow and outflow sediment concentrations for
steady-state test 3 .......................................... '
60 Sediment trapping efficiency vs time for storm on
April 3, 1980 ................................................ 12]-
61 Sediment trapping efficiency vs time for test 1 ................ 123
62 Sediment trapping efficiency vs time for test 2 ................ 124
63 Sediment trapping efficiency vs time for test 3 ................ 125
^•fSS^SS^^^f^^y^few^^^^S^^^^^
Filter Model 122
-------
TABLES
Number Page
1 Summary of the Flow and Transport Equations for
Steady-State, Homogeneous Sediment 14
2 Parameters for Field Tests.
Required for Filter Installation.
20
3 Percent of 24 Hour Rainfall With an Intensity
Greater than 0.6 Inches (1.52 Centimeters) Per
Hour Using an SCS Type II Curve ............. ................... 39
4 Sediment Yield From the Source Area to the Filter
Area From a Design Storm. . ..................................... 40
5 Calculated Average and Peak Concentrations of
Storm Runoff ....................... ............................ 4-1-
6 Kentucky Vegetative Filter Model Inputs and
Outputs [[[ 42
7 Values Used for Sensitivity Analysis of Kentucky
Grass Filter Design Model ......... ... ...........................
8 Design Vegetative Filter Characteristics ......................... 45
9 Soil Test Results ............... ................................. 61
10 Monitoring Instrumentation ....... . ............................... 63
11 Approximate Number of Manual and Equipment Hours
89
12 Type and Amount of Materials Required for Filter
Installation. 90
13 Summary of Storms Sampled 119
14 Inputs Necessary for Running the Kentucky Grass
-------
ACKNOWLEDGEMENTS
The Vegetative Filter Zone Demonstration Project was a
cooperative effort between the U.S. Environmental Protection
Agency (EPA) and the Kentucky Department for Natural Re-
sources and Environmental Protection (KY. DNREP). The work
was conducted under contract by Hittman Associates, Inc.
(HAI), Natural Resource Consultants/ (NRC) and the^Uni-^
versity of Kentucky Department of Agricultural Engineering.
During the course of this project, many individuals
contributed to the study. Without the contributions of
these individuals, the project's success would never have
been realized.
Direct management of the project was provided by Steve
C. Albrecht, (HAI), The project consultant was Dr. Billy J.
Barfield, (NRC). The project manager for the KY. DNREP was
Richard Rohlf. The EPA project officer was Edward Bates of
the Industrial Environmental Research Laboratory (EPA),
Cincinnati.
Other individuals who provided valuable contributions
during the course of the project include: Dr. R. I. Barn-
hisel, Mike Ruetten, David Kerr, Bill Harris, Loyd Dunn, Jim
Penman, and Julia Peace.
Savoy Coal's cooperation in providing the project study
site is also gratefully acknowledged.
Xll
-------
SECTION 1
INTRODUCTION
The coal regions of the United States are richly endowed
with abundant mineral resources. The surface mining of
these valuable resources results in a major disturbance of
the land. This disturbance interrupts a natural balance
between soils and geologic formations established over many
years. When highly erodible surface material is broken up,
a natural pollutant, sediment, is introduced into rivers and
streams. Surface mining operations often serve to accelerate
the amount of sediment introduced into the environment. The
resulting sedimentation can adversely affect local and
surrounding ecosystems.
Various structures such as ponds, rock dams, silt
fences, as well as other devices can be used to control
sediment on surface mining operations. This report addresses
the usefulness of one particular type of sediment control
structure, the vegetative filter zone.
The objective of this study was to demonstrate under
actual field conditions the technical, economic, and environmental
feasibility of using a cultivated vegetative filter zone to
assist in controlling fine-grained sediments originating
from surface mining operations. Project evaluation criteria
focused on the sediment trapping efficiency of the filter,
the suitability of selected types of vegetation -for use
within the filter zone, the filter's usefulness in improving
water quality, and its cost feasibility. In the past,
vegetation has been used extensively as a method of controlling-
erosion and to reduce the amount of sediment generated on
surface mining operations. Unfortunately, little information
was available concerning design criteria for the use of
vegetation in trapping sediment, thus it was very difficult
to incorporate this trapping method into a complete mining
plan. Through the course of this demonstration project,
strong emphasis was placed on developing practical vegetative
field applications which could effectively be used by surface
mine operators to control sediment. Specific design criteria
were utilized for constructing the vegetative filter. The
use of such criteria was necessary to provide for adequate
control of the demonstration monitoring program. Mine
operators wishing to utilize vegetative filters, will not
-------
have to utilize such design criteria to the extent required
in this project.
This report provides detailed background information
concerning sediment control technology, legislation and
funding, site selection criteria, and project objectives
(Section 3). In addition, the study provides baseline
information on site characteristics, mining, geology, soils,
vegetative design, hydrologic design, and water monitoring
system design (Section 4). Installation of the filter,
including site alterations, vegetative establishment and
performance, monitoring system installation, time and materials
are discussed in detail (Section 5).
Discussions of lab analyses, storm events, and computer
model applications are also provided (Sections 6-8). The
appendices contain storm event data, design storm calcula-
tions and descriptions of all analytical procedures utilized
during the project.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
A field demonstration project was conducted near the
city of Corbin in Southeastern Kentucky to evaluate the
effectiveness of a vegetative filter zone in trapping fine-
grained sediments from surface mining operations. Activ-
ities conducted during the project were: construction of a
filter below an area generating sediment-laden runoff from
surface mining; evaluation of vegetation as a trapping
medium and the adaptability of the vegetation to_the_pro-
longed stress conditions; water and sediment monitoring of
inlet and outlet flows through the filter area; water sample
analyses for total solids; and analysis of particle size
distribution of the trapped sediment.
Conclusions drawn from this study are as follows:
1) By vegetative filtration, a range of from
70-99 percent of all fine-grained sediments
can be removed. This percentage does not
include clays held in suspension. This range
was determined from all data recorded during
storm events.
2) vegetative filter areas can normally be con-
structed using equipment found on most
surface mining operations.
3) Vegetation used in the filter are subjected
to high stress conditions. Response to this
stress was excellent despite acid inflow into
the area. Based upon the results of this
study, the vegetative trapping medium can
adapt to severe environmental conditions with
very little difficulty.
4) The Kentucky Grass Filter Model was found to
be excellent in predicting trapping efficiencies,
given parameters of the individual storm
events.
-------
5) The trapping efficiency in storms early in the
monitoring period were very high (95-100
percent), but toward the end of the period,
trapping efficiencies had decreased approx-
imately 75 percent.
-------
SECTION 3
BACKGROUND
OVERVIEW OF EROSION AND SEDIMENT CONTROL TECHNOLOGY
The Erosion Process
Soil erosion occurs as a result of the exposure of
detachable soil particles to the erosive forces of falling
raindrops and flowing water. Soil is least protected from
these forces when the surface is bare of vegetation or
mulches. These conditions are typically associated with
active surface mining and construction activities. It is
impossible to perform such earth moving activities without
leaving the surface exposed to these erosive forces for a
period of time. It is possible, however, to control excessive
stream pollution by trapping a significant amount of sediment
before it leaves the permitted area.
Soil erosion occurs in rill and interrill areas on
exposed hillsides or slopes. Rills are small fingerlike
channels that are formed by the localized concentration of
runoff water. Erosion in interrill areas is primarily due
to the impact force of raindrops on the soil surface, while
rill erosion is primarily due to the shearing forces of
runoff. Gully erosion occurs when large volumes of runoff
are concentrated onto an unstabilized area. Areas where the
potential for gully erosion is high include uncontrolled
drainage ditch outlets, drainage culvert outlets, and tile
outlets.
Soil erosion involves detachment, transport, and the
ultimate redeposition of sediment. Soil is detached by
raindrop impact and the shearing forces of runoff, and is
then transported by.flowing water. Runoff and resulting
transport of sediment does not occur until the rainfall rate
exceeds the infiltration rate. Once runoff starts, the
quantity and size of material transported increases with the
velocity of runoff.
The rate of erosion from an exposed slope will be con-
trolled by gravity and the soil available for transport, or
the transport capacity of the flow, whichever is smaller
(Figure 1). On long steep slopes, the transport capacity of
-------
CO
Ul
o
o
ac
a.
o
hi
O
a.
-IS
o
«l
o
M Z
M
0
1
o
o
1
I-
1
-1
i
O
|!
ri
o
W «•"
o
to
M
0)
0)
a
0)
-p
«
c
o
•H
O
M
-H
o
n
0
O
o
•H
4J
ITS
I
'o
w
0)
-------
the flow will typically be less than the sediment available
for transport, resulting in deposition. The use of properly
designed vegetative filters optimizes deposition in a natural
filtration process.
Erosion and Sediment Control Technology
Sediment control can be influenced by either limiting
the detachment rate or transport capacity (Figure 1). De-
tachment due to raindrop splash can be primarily reduced by
absorbing the energy of raindrops on some type of cover,
which may be permanent vegetation, temporary mulches, or
chemical binders used to stabilize soil and to make it more
resistant to erosion. The rate of detachment due to runoff
can also be reduced by making the soil more resistant to
erosion, as with chemical stabilizers, the addition of
organic matter to the soil, or by reducing the runoff velocity.
Any practice that roughens the surface, such as mulches or
contour tillage, decreases the runoff velocity and hence the
shearing forces of runoff. The sediment transport capacity
of runoff also monotonically increases with velocity, and
can be reduced by the application of practices used to
reduce erosion by runoff.
Erosion and sediment control techniques can be grouped
into on-site and off-site practices. On-site practices such
as mulching, establishment of a vegetal cover, and runoff
control are used to reduce both detachment and transport
capacity. These practices provide a "first line of defense"
against sediment pollution, but can only be used practically
after grading has been completed.
Offsite techniques such as the use of sediment detention
reservoirs, sediment traps, check dams, and vegetative
filters can help reduce the transport capacity of runoff.
Sediment traps, check dams, and sediment detention reservoirs
typically trap sediment a considerable distance from the
point of detachment. Vegetative filters can be located at
points downslope or downstream of a sediment source, but are
most effective when located as close as possible to the
point of detachment. In the demonstration project described
in this report, a vegetative filter was located near the
base of an abandoned surface mine spoil outslope in order to
monitor its effectiveness in trapping sediment near the
point of soil detachment.
Description of Previous Work
Flow Retardation by Vegetation —
Early research by Ree (1949) considered the hydraulic
characteristics of vegetation used as vegetative waterways.
-------
He noted that resistance of a vegetation to flowing water is
its most important hydraulic characteristic. Determinations
of the retardance coefficient, using Manning's equation,
were made for different vegetations under a variety of
conditions. Manning's equation is given by
Vm = 1 R V3 c 1/2
± R
n
c
where Vm is the mean velocity, R is the hydraulic radius, Sc
is the channel slope, and n is the retardance coefficient
which is usually referred to as Manning's roughness. The
values of Manning's n ranged widely, even for the same veg-
etation in the same channel. A gradual increase in n was
noted by Ree as the discharge increased. As discharge
further increased, a depth was reached where the grass bent
over and became submerged. The n values then rapidly decreased
with flow depth until the grass was completely bent over, at
which time the Manning's n begin to approach a constant
value. Permissible velocities for a vegetation to protect
against erosion were found to depend on the type and quality
of vegetation, the soil texture in the channel bed, and
channel slope. Ree cautioned that complete sod vegetations
allow the highest permissible velocities, and bunch-type
permit the lowest. A total sod cover gives complete channel
protection, while the bunch-type grasses leave unprotected
bare areas between the clumps.
Fenzl and Davis (1964) discussed the difficulties in
adequately characterizing the hydraulic resistance for open-
channel flow through vegetation in irrigation borders. Two
essential design criteria, which must be adhered to in order
to satisfactorily design vegetative channels, are: 1), the
channel must be able to carry the flow, and 2), the channel
bed must remain stable. Dimensional analysis of relevant
variables was used in the development of a proposed pro-
cedure. An analytical model defining flow resistance for
vegetated channels was presented by Thompson and Roberson
(1976). This relationship considered the total flow resistance
to be composed of flow resistance caused by the vegetative
elements in the flow and by the roughness of the channel
surface unoccupied by vegetation. The resistances were
separated by source component, and basic laws of fluid
dynamics were used to describe the flow. Kouwen et al.
(1969) presented a quasitheoretical analysis for flow retardance
in vegetative channels by anticipating that most of the
parameters found to affect flow resistance in channels with
rigid roughness will also be present with flexible roughness,
such as represented by grass and other vegetation. This
approach based the flow equation for vegetated channels on
the vegetative characteristics, including vegetative-blocked
cross-sectional area and roughness coefficient.
8
-------
In response to Kouwen et al. (1969), Gourlay (1970)
observed that a complete solution for resistance in a vegetated
channel should strive toward expressing the flow velocity as
a function of the independent variables of flow rate, slope,
and physical characteristics of the grass. Typical grass
physical characteristics would be grass height, the number
of stems per unit area, a factor accounting for grass resilience,
and the shape of the stems or blades. Since several flow
regimes are possible, a prediction of which flow regime
occurs for a given condition is also essential.
Li and Shen (1973) believed that vegetation serves as
an effective method of reducing sediment yield by protecting
the soil from rainfall impact energy, retarding the flow,
reducing soil resistance and aiding infiltration. After
testing a number of different patterns of vegetation, Li and
Shen concluded that vegetation has a significant effect on
retardation of flow rates and sediment yields, and that
vegetation planted in staggered patterns is much more effective
in reducing flow rate than any other pattern for the same
number of plants.
Vegetative Buffer Strips --
Both natural and seeded vegetative strips have been
recommended for filtering runoff from a variety of disturbed
areas. EPA (1976) recommends use of both natural and in-
stalled buffer strips downslope from areas disturbed by
mining. Brush barriers are recommended by Dallaire (1976)
as a means of controlling erosion losses from construction
sites.
Ohlander (1976) discussed sediment trapping by a vegetative
strip such as might exist below a road drainage outlet. A
regression model was developed which defined the filter
effectiveness in terms of slope, soil stability, and a soil
cover complex index.
The use of grass filter areas to improve the quality of
runoff polluted by manure" has received widespread attention
during the last few years. Bingham (1978), Norman (1978),
Khallel (1979), Doyle and Stanton (1977), Young (1978), and
Vanderholm and Dickey (1978) are some of the individuals
responsible for grass filter research. However, the recommenda-
tions presented by these studies are only gross estimates
and do not include the physical basis necessary for developing
prediction equations. These empirical studies were based
primarily on monitoring results from a limited variety of
conditions, were restricted to mild slopes, and did not
consider variations in particle diameter.
-------
Wilson (1967) studied the effectiveness of grass to
filter sediment from flood waters. Results indicated that
an inverse relationship exists between the filtration length
to produce a maximum concentration of a given particle size
in the deposited sediment and the diameter of that particle
size.. Thus, smaller particles require longer filters for
maximum deposition than larger particles. Visual observation
at the beginning of the next growing season of the Bermuda
grass used in the tests seemed to indicate no growth inhibition
due to deposited sediment. Wilson concluded that filter
length, initial sediment concentration, flow rate, slope,
grass height, grass density, and degree of submergence are
interrelated factors affecting sediment removal. Grasses
selected for filtration should have deep root systems to
resist scouring if swift currents develop. In addition,
these grasses should possess dense top growth, resistance to
flooding and drought, and the ability to recover growth
subsequent to inundation with sediment. The ability to
yield economic returns either through the production of seed
or hay would also be advantageous. However, Wilson does not
provide a method for estimating the relationship between
these parameters.
A rainfall simulator study conducted by Neibling and
Alberts (1979) indicated the potential effectiveness of
grass filtration. Sod strips as short as 0.61 meters reduced
sediment discharge rates by more than ninety (90) percent,
and sediment discharge of particles less than 0.002mm through
the 0.61 meter strips was reduced by thirty-seven (37)
percent. Neither empirical nor theoretical relationships
were presented which could be used to estimate the influence
of the various parameters on outflow conditions.
A new filter arrangement was proposed by Kao, et al.
(1975) to solve the problems of sediment inundation and
killing of the vegetation. The filter alternated grass
strips with bare ground strips. Test results indicated
that, when appropriate width-ratio of the grass to bare
ground strips is chosen, filter efficiency remains high and
the trapped sediment is retained in the bare ground region.
This arrangement also lends itself to construction of filters
which may be cleaned as the sediment deposits reach unsatis-
factory depths.
Research has been conducted at the University of Kentucky
to develop relationships which may be used to predict the
amount of sediment trapped in a vegetative filter and the
quality of the outflow. In laboratory studies without
infiltration, the research indicated that mechanical settling
alone can trap the coarse silts and sand size particles, but
that the finer silts and clays pass through the filter. In
10
-------
field studies where infiltration was important, it was
apparent that the outflow sediment load was reduced by
infiltration, since the infiltrating water carried with it a
sediment load equal to the suspended sediment concentration
times the mass of water infiltrated. Thus, infiltration in
vegetative filters decreases the sediment load of the clay
and fine silt fraction, while the mechanical filtering
action of vegetation filters the coarse silts and sands.
Sediment Transport Within Vegetation —
Research conducted at the University of Kentucky sought
to quantify conditions within a vegetative filter as well as
downstream from the filter. The objective of this work was
to characterize outflow sediment load as a function of
readily available inflow parameters and/or filter dimensions.
Description of Sedimentation Process Within Vegetation —
A conceptual model of the sedimentation process as observed
by Barfield (1978) is presented in Figure 2. As sediment
laden flow impinges a grass filter, its velocity is retarded
and its transport capacity reduced. If the transport capacity
is less than the inflow sediment load, sediment is deposited
at the inlet of the filter media. This deposition causes
the channel slope to increase, with a resulting increase in
velocity and sediment transport capacity down the deposition
face.
As shown in Figure 2, the filter is divided into four
zones. The location of the transition from one zone to
another moves downstream with time; hence, the location of
each zone is denoted as a function of time. In zone A(t),
transport occurs along the top of the inundated media, with
essentially all of the incoming sediment being transported.
Sediment is deposited uniformly along the deposition front
in zone B(t), with the slope of this front being referred to
as the equilibrium slope. In zones C(t) and D(t), the
assumption is that the tractive force is less than the
critical value for the original channel bed. In section
C(t), sufficient sediment has been deposited on the original
channel bed so that all surface irregularities are filled,
allowing sediment to be transported as bed load. In zone
D(t), an insufficient amount of material has been deposited
on the bed,to fill the irregularities so that all sediment
reaching the bed is trapped. Since the location of the
zones change with time, both the transport and location of
each zone must be modeled after a given time of flow duration.
Infiltration effects are reflected by either an increase
or decrease in flow rate, depending on the quantity of flow
left after infiltration in zone D(t). The changed flow rate
11
-------
>S So.
z o
— OB.
I- a. k.
v kl S
Z 5; m
= O
±IC
-H
•O
0)
O
0)
(0
w
o>
o
o
c
o
•H
-P
•H
(0
O
a
0)
•o
•p
c
•H
•O
0
CO
(U
O
+J
cd
|
O
CO
CM
0)
N
12
-------
will in turn be reflected by a comparable change in mean
flow velocity and in transport capacity. This can be modeled
with one of the many relationships available for infiltration
along with the continuity equation.
The model development has evolved from a model for
steady-state flow/ with homogeneous sediment on uniform
slopes, to a model with non-steady-state flow, with non-
homogeneous sediment, on non-uniform slopes. Because of its
simplicity, a description is first given of the uniform
slope case and a subsequent description given for the more
complex case.
Steady-State, Homogeneous Sediment, Uniform Slope Model —
In order to understand the mechanisms illustrated in Figure
2, it is necessary to look first at zone D(t). In this zone,
insufficient material has reached the bed for bed load to
occur so that all material reaching the bed is assumed to be
trapped. Using probabilistic reasoning, Tollner et al.
(1976) showed that the fraction of sediment trapped in zone
D, (^sd'^so^/^sd' could be predicted from the effective
filter length, L(t), flow depth, df,' flow velocity, V,
settling velocity, Vs, and grass spacing, Ss, as given by
equation (1-2) in Table 1.
By assuming an analogy between flow in a filter medium
and that in a rectangular channel of width Ss and depth df,
Manning's equation was modified to calculate flow velocity
and depth of flow. Using these values in equation (1-2),
the value of the outflow sediment load qso can be calculated
for a given fall velocity, if the effective filter length
and sediment load into zone C(t), qsd, is known.
In order to calculate a value for qsa/ sediment transport
relationships were needed for flow in a filter. . Tollner et
al. (1978) modified Einstein's and Graf's transport and shear
parameters (Graf, 1971) to account for the presence of
vegetation, and also calibrated these equations for flow in
a filter medium. Either of these equations can be used to
predict a value for qsd. Einstein's equations were chosen
because of their widespread use.
The calculation of the effective filter length, L(t),
requires a prediction of how the deposition profile advances
with time. When a sediment-laden flow is introduced into
vegetation, the flow depth increases and velocity decreases
because of the retarding effect of the vegetative elements.
The decreased flow velocity results in decreased sediment
transport capacity. If the sediment transport capacity is
then sufficiently less than the input sediment load, deposited
13
-------
TABLE 1. SUMMARY OF THE FLOW AND TRANSPORT EQUATIONS FOR
STEADY-STATE, HOMOGENEOUS SEDIMENT
1
Zone D(t)
Sediment Trapping
qsd
Flow
W^s"
Zone C(t)
Flow - Same as Zone D(t)
Sediment Transport
A_=1.
Zone B(t)
Rate of Advance
. _ ^v
x(t,
-]
f =
(t"t*)
- sd
Se = set - Se
Sediment Transport
Ss^ ~SS-
10
-3
_ -0.28
t < t*
= 1.08
s_ a.
fs
-0.28
•' Vssdfs
s=i^/3
14
(1-2)
(1-3)
(1-4)
(1-5)
(1-6)
(1-7)
(1-8)
(1-10)
(1-11)
(1-12)
(1-13)
(1-14)
"(1-15)
(1-16)
-------
particles will form a triangular shaped wedge. If this
continues with constant inflow conditions, the deposition
wedge will approach the same height as the vegetation at
time t*. At this height, the wedge will flatten out, but
continue to advance downstream at approximately a constant
angle, as shown in zones A(t) and B(t) of Figure 2. By
performing a mass balance after the slope of the deposition
wedge, Sc, is known, the advance distance can be computed
using either equation (1-8) or (1-9), depending on the time.
The value of Sc can be calculated from the modified Einstein
and Manning equations mentioned previously. A summary of
the flow and transport equation is given in Table 1 for the
steady-state case.
Unsteady-State, Non-Homogeneous Sediment Solution —
For situations where the inflow rate and/or concentration is
changing, Hayes et al. (1979a) extended the steady-state,
homogeneous sediment solution by integrating the differential
equations defining the advance distance X(t) in zone B(t) of
Figure 2. The values of f, qsi/ an^ Se are averaged over a
time interval (tf - tjj to obtain t*. The incremental
change in distance is added to the initial advance distance
Xj_(t) to obtain the advance distance, Xf(t), after the time
interval. Equations (1-8) and (1-9) in Table 1 will be
replaced by these equations for unsteady-state flow or:
1/2
Xf(t) =
Xf(t) =
: qo-; \ 9
-5-^ I (tf - t.) + X±(t)2
be / ^
X
for t + Vnfe for t>t.
The other equations will remain the same as for the steady-
state case.
The design equations were further extended for the non-
homogeneous particle size distribution by subtracting a
portion of the particle distribution curve, based on what
has been trapped in previous zones, and utilizing the previous
unsteady flow equations. The representative particle size
for zones A(t) and B(t) is calculated by selecting the
average size of particles greater than 0.037 mm in diameter.
The smaller particles are not included, since observation
indicates that few of these particles can be trapped in
zones A(t) and B(t). After calculating the transport capacity
of zone C(t), the fraction trapped in zones A(t) and B(t)
can be found. The remaining fraction of these sizes is then
used as an input sediment load into zone D(t), and equation
(1-2) is used to calculate the fraction of these particles
15
-------
entering zone D(t) which are trapped. The settling velocities
for the 0.004 and 0.037 mm particles size range is also
calculated and substituted into equation (1-2) to obtain the
fraction of this size range trapped. It was also assumed
that those particles less than 0.004 mm would not be trapped.
Thus, by adding the portion less than 0.004 mm, the portion
remaining between 0.004 and 0.037 mm, and the portion remaining
greater than 0.037 mm, the total trapping efficiency of the
filter can be calculated.
Effects of Upstream Deposition on Sediment Load —r The
aforementioned analyses consider the transport capacity
within the vegetative medium, but do not consider the possible
effect upstream trapping may have on sediment load. Work
reported by Neibling and Alberts (1979) indicates that there
may be significant trapping in this area. To incorporate
this component into the design procedures, Hayes et al.
(1981) utilized a mass balance to calculate the sediment
trapped in this region. The depth of deposition, Y(t), has
previously been estimated by:
2 fq
Y(t) =
^sb~ et
1/2
The above equation is solved at two different times, tn and
t2 less than t*, while assuming that the deposition wedge
profile extends upstream horizontally until it intersects
the channel bottom; then the sediment load reaching the
filter, qsi, can be calculated by:
v 2 _ v 2 \ Y
Y2 *.l ' Ysb
2 se ^2 - *!>
where qs is the initial sediment load. The above equation
is the model for predicting the effects of upstream deposition
on the sediment load.
Accounting for the Effects of Infiltration — Where
infiltration is significant, the quasi steady-state system
shown in Figure 2 becomes a spatially varied flow problem.
The model was adapted for this situation by making the
following assumptions:
1. The infiltration throughout the filter length is
steady-state so that flow rate decreases linearly
from the upstream to the downstream end.
2. The sediment contained in the water infiltrated in
zones C(t) and D(t) is carried with that water and
trapped as it reaches the channel bottom or enters
16
-------
the soil profile (Figure 2).
3. The sediment contained in the water infiltrated in
zone A(t) and B(t) is not trapped because of the
higher flow velocity on the deposition wedge.
4. The sediment transport in each zone can be character-
ized by the average hydraulic and sediment conditions
which exist in that zone.
5. Upstream trapping was insignificant because of the
break in slope where the inlet connected to the
filter.
Accounting for Time Lag of Outflow — In order to
account for the time lag between inflow and outflow, outflow
conditions were lagged behind inflow conditions by a time
interval equal to the flow time through the filter.
Experimental Evaluation of Kentucky Model
Evaluations of the Kentucky Model were conducted in
laboratory flumes, using artificial media and real grasses,
and in field channels on real grasses. It is impossible to
control parameters such as media spacing and media size with
real grasses. For this reason, artificial media were con-
structed and used to develop the basic relationships and
calibration coefficients previously mentioned. These re-
lationships were first evaluated on real grasses in specially
constructed 15.2 meters laboratory flumes. For the labor-
atory studies, infiltration was not a factor. Results of
the laboratory tests indicated good agreement between obser-
ved and predicted results for four of the five laboratory
tests. A valid explanation could not be found for the poor
prediction of the fifth test. Complete results are presented
by Hayes et al. (1981).
Five field tests were also run on real grasses in which
infiltration was significant. Pertinent parameters are
shown in Table 2 for each of the tests run. Sediment was
generated from an erosion table on which shale overburden
from a surface mine in Western Kentucky was subjected to
simulated rainfall. Good agreement was found between pre-
dicted and observed results in all five of the tests.
Typical test results are shown in Figures 3 and 4 for Test
1. Complete results are available in Hayes et al. (1981).
The results of these previous experimental evaluations
indicate that the Kentucky Model can be used to predict the
performance of vegetative filters in the field. Barfield et
17
-------
Ill
I.OO
0.95
0.90
2 0.95
o
n.
aso
0.75
-e—r
®
6
ee
e
e
e
MEASURED
ESTIMATED
8 = O.O30
1 1 1 1
1 1 1 1
tO 4O 60 80 100 120
TIME (win.)
Figure 3. Fraction trapped versus time for field test 1
(Hayes et al 1981)
18
-------
io,ooo
"5500
5.OCK)
o
*
o
© MEASURED
ESTIMATED
S= (642M«/1
111! I
I 1
20 40 60 BO tOO IZO
T9ME (min.)
Figure 4. Outflow concentration versus time for
field test 1 (Hayes et al 1981)
19
-------
CO
EH
CO
EH
Q
H
H
fa
rt
0
co
H
H
g
<
P4
CN
TABLE
,£5
-P -P
rH SI'S
to (U —
O
•H
0) -P
ft fd
EH Q)
Q)
c
O
•H
-P -P
0) JH H
g -P\
•H G tn
rtf 0) g
(1) O —
CO G
O
0
"g
U
V,.
O
• —
£ CD oo
O -P g
rH fCi O
fa Pi ~
CJ1
(d -H g
•H O O
t! ^
^ \ ^x
00 00 OO rH rH
OO 00 00 f- r~
CN rH rH rH rH
Nl
Q)
J3
O = = = =
CO
Q)
o o o o o
o o o o o
O O CN O VD
W *. ^ K *.
VD OO "31 *5l< "31
^^ rH
CN
1 1 1 I 1
O O O O O
o o o o o
O O"i O O O
^ ^ •», *. «*
rH ^ 00 in CTl
00 CN 'S' rH
OO CTi -xf CN CN
OO OO C"i VD VD
OO OO CN OO OO
^D CO CO ^^ ^^
o^ r**- t"^- ro ro
rH rH rH iH tH
s.
1 1 Q^
O O CTi •
VO VD CTi rH
vo \ \ \ \
OO O O VO VD
CN OO OO CTi CT>
rH CN oo »* in
•
F* tl
U fi
"* CN
in ||
rH
11 1
& -p
4-> fd
-a M
•H 3
4J -P
O ^
rH 0)
pii £H
+J
C
(P
CO
Q)
CN
O
-P W
G -H
0) (d
g
tn CQ
Q) t3
CO <1>
CD
rH
(0
-p 3
G 0
-------
al. (1980) indicated that the following limitations should
be kept in mind when using the model:
1. Difficulties in estimating the exact value of the
roughness coefficient require that conservative
values of Manning's n be used for design purposes.
Conservative values of grass spacing should also
be selected for design purpose, since this char-
acteristic cannot be measured until the vegetation
is well established.
2. The depth of deposition in section D(t) at which
the vegetation ceases to operate as a filter has
not been determined experimentally. Further
research is needed to develop methods to predict
this variable.
Although the Kentucky studies indicate that vegetative
filters can operate to trap up to ninety-nine (99) percent
of sediment under carefully controlled conditions, additional
study was needed of the performance of vegetative filters
under natural conditions. Such a study is the subject of
the remainder of this report.
LEGISLATION, FUNDING, AND SITE SELECTION
Legislation
The Kentucky General Assembly created the Department
for Natural Resources and Environmental Protection in 1974,
enacting Kentucky Revised Statue 224. Responsibilities of
the Department include administration of functions relating
to conservation, maintenance, and preservation of land and
water resources, and the prevention, abatement, and control
of all water, land, and air pollution. The Department is
headed by a Secretary who is responsible for overall depart-
mental administration.
The Department's Bureau of Surface Mining Reclamation
and Enforcement (BSMRE) is vested under KRS 350 to regulate
and control the surface mining of coal in order to minimize
resultant safety and environmental hazards to the people and
resources of the Commonwealth. The Bureau, under the super-
vision of the Secretary and its Commissioner, has the follow-
ing authority and powers relative to this project study:
o To exercise general supervision and admin-
istration and enforcement of KRS 350 and all
rules, regulations, and orders promulgated
thereunder;
21
-------
o To encourage and conduct investigations,
research, experiments, and demonstrations and
to collect and disseminate information
related to strip mining;
o To adopt, without hearing, rules and reg-
ulations relative to the filing of reports,
the issuance of permits, and other matters of
procedure and administration;
o To examine and pass upon all plans and
specifications submitted by the operator for
the method of operation, backfilling, and
grading and for the reclamation of the area
of land affected by his operation.
Existing Sandards —
All surface mining operations in the Commonwealth of
Kentucky must be permitted through the Bureau of Surface
Mining Reclamation and Enforcement. Permit requirements are
described under KRS 350.060. Operators are also required to
submit a drainage plan along with permit applications. Di-
rectional flow of water, constructed drainways, natural
waterways used for drainage, and the streams or tributaries
receiving any discharges must be diagrammed. In addition,
KRS 350.090 requires operators to prepare and enact a recla-
mation plan for land affected by mining operations. KRS
350.090 reads (in part) as follows:
KRS 350.090 Method of operation and reclamation
plan — Waste on permit area onlyI(T)Under the
provisions of this chapter and regulations adopted by
the department, an operator shall prepare and carry out
a method of operation, plan of grading and backfilling
and a reclamation plan for the area of land affected by
his operation. In developing a method of operation,
and the plans of backfilling, grading and reclamation,
all measures shall be taken to eliminate damages to
members of the public, their real and personal property,
public roads, streams and all other public property
from soil erosion, rolling stones and overburden, water
pollution and hazards dangerous to life and property.
The plan shall be submitted to the department and the
department shall notify the applicant by certified
mail, return receipt requested or by registered mail
within thirty (30) working days after receipt of the
plan and complete application if it is or is not ac-
ceptable.
22
-------
The regulations which are applicable to KRS 350.090 and
also particularly relevant to this project are:
o 405KAR 1:160, Protection of the Hydrologic System,
requires that only minimum disturbances to the
hydrologic system are permitted and that all
surface openings be properly sealed.
o 405KAR 1:170, Water Quality Standards and Surface
Water Monitoring, identifies the following water
quality standards or effluent limitations eman-
ating from mine sites:
o pH between 6.0 and 9.0.
o Iron concentrations not in excess of
7 mg/1.
o If the pH is less than 6.0, or if the
iron concentration is greater than 10
mg/1., then total manganese must be less
than 4.0 mg/1.
o Total suspended solids must not exceed
70 mg/1.
In addition, 405KAR 1:170 specifies the methodology
for surface water monitoring and the reporting
procedures required by the operator.
o 405KAR 1:190, Diversion of Surface and Ground
Water Flows, states that:
o Diversion .of overland flow may be
allowed in order to minimize erosion and
to prevent or remove water from con-
tacting toxic producing deposits.
o Diversions shall be designed and constructed
to meet department specifications.
o Stream buffer zones shall be observed
within 100 feet of intermittent or
perennial streams.
o Discharge structures shall be constructed,
where necessary to reduce erosion and
prevent deepening of stream channels.
o Surface waters shall not be discharged
into underground mine working unless
department approval is granted.
o 405KAR 1:200, Sediment Control Measures,
specifies the various methods that may be
used to control sediment, including vegetative
sediment filters, as well as the design, con-
struction and maintenance of sediment ponds.
23
-------
Funding
This study was made possible through a cooperative
agreement from the U. S. Environmental Protection Agency
(EPA) to the Bureau of Surface Mining Reclamation and Enforcement,
Kentucky Department for Natural Resources and Environmental
Protection. The Bureau of Surface Mining Reclamation and
Enforcement (BSMRE) contributed five (5) percent matching
funds to the study.
The Environmental Protection Agency's authority to fund
the study is contained in Section 107 "Mine Water Pollution
Control Demonstrations" of the Federal Water Pollution Con-
trol Act Amendments of 1972, P.L. 92-500. Section 107
authorizes EPA to make grants for projects to demonstrate
valuable approaches eliminating or controlling mine water
pollution resulting from active or abandoned mining operations.
The Kentucky Bureau of Surface Mining Reclamation and
Enforcement is allowed to accept Federal grants and other
funds in accordance with the KRS 350.150 and 350.163.
All money received by the Kentucky Department for
Natural Resources and Environmental Protection through the
payment of fees, forfeiture of bonds, and Federal grants is
placed in the State Treasury. A general fund is appropriated
to the Department on a bi-annual basis as approved by the
Kentucky General Assembly. Funds are expended (under supervision
of the Bureau Commissioner) for the administration and
enforcement of KRS 350 and for the reclamation of improperly
reclaimed surface-mined lands.
Site Selection
The search for a suitable study site was initiated
immediately after award of the project contract.
Several mine operations in Eastern Kentucky were considered
and evaluated before on-site visits were made to investigate
and determine if a particular area and operation was suitable
to project study requirements. Site selection criteria
included the following:
o Availability of a study area adjacent to an area
disturbed by surface mining.
o Physical characteristics (i.e., slope and soil
conditions) typical of mined areas.
o Comparable time frames for mining land lease and
the scheduled project completion date.
o Complete cooperation from the mining company and;
o accessibility to the study site by project per-
sonnel under all weather conditions.
24
-------
The contour and mountaintop mining operation being
conducted by the Savoy Coal Company near Williamsburg,
Kentucky, met all site selection criteria. The project area
selected was directly below coal reserves planned for min-
ing. The area was under lease and would so remain for the
duration of the project study.
Additionally, both the physical characteristics of the
land and the existing soil conditions were calculated to
provide the parameters needed to obtain an accurate assessment
of the vegetative establishment procedures to be evaluated
in the project study.
The Savoy Mining Company was responsive, assured complete
cooperation, and was quick to supply needed preliminary
information.
DESCRIPTION OF PROJECT SITE
The study site chosen for the demonstration project
(Figure 5) is located on the western edge of the Appalachian
Plateau physiographic region in Southeastern Kentucky. This
specific demonstration site is approximately 12.9 Kilometers
south of Corbin, Kentucky, in an area locally known as
Buttermilk Hollow.
Drainage for the project area is in the Eaton Branch
watershed, which flows east for approximately 2.8 Kilometers
before entering Camp Creek. Camp Creek flows north to
northwest until entering the Laurel River 3.2 Kilometers
northwest of Corbin.
The Laurel River watershed area is characterized by a
dendritic drainage pattern and U-shaped valley bottom, with
well developed valley bottoms in larger streams..
The region is characterized by moderate to steeply
sloping hills, with elevations exceeding 762 meters in the
eastern portion of the region. In the immediate area of the
vegetative filter site, the slopes are moderate, with elevations
ranging from 310.9 meters north of Corbin to 518.2 meters
immediately north of the site.
The area surrounding the study site is typically composed
of woodlands and pasture contained on small farms. Although
the area is composed of these small farming operations, the
largest industry, by far, is the mining of coal.
The study area was located directly below a contour
strip bench and outslope created by previous mining activities.
This sparsely vegetated outslope area was highly suitable
because it provided the primary drainage for the project
site.
25
-------
KENTUCKY
LOCATION INDEX
CORaN-3.2 KILOMETERS
SCALE
O 1000 2000 FEET
O 3OO 610 METERS
EATON BRANCH WATERSHED
BOUNDARY
STREAM CHANNEL
VEGETATIVE FILTER SITE
O
Figure 5. Site location map
26
-------
PROJECT OBJECTIVES AND APPLICATIONS
Previous studies have shown that vegetative filters
have the potential to substantially reduce sediment yield
from disturbed areas. Their effectiveness had not been
evaluated under field conditions with natural rainfall and
repeated sediment loadings. In order to make this evaluation,
the project had as its objectives:
1. The establishment of a vegetative filter as close
as possible to a sediment source on a surface
mined area.
2. The collection of water quality and sediment data
at the filter for a one-year period.
27
-------
SECTION 4
PRELIMINARY SITE CONDITIONS AND VEGETATIVE FILTER DESIGN
GEOLOGY AND MINING OF THE PROJECT SITE
Geology
Stratigraphically, the vegetative filter site is located
in rock of the Breathitt Formation of the Middle Pennsylvanian
age. Typical of the Pennsylvania coal-bearing measures of
Kentucky, the strata consists of shales, sandstones, siltstones,
coals, ironstone, and clays. Individual lithotypes tend to
be discontinuous and irregular, both vertically and laterally,
often intertonguing with adjacent strata. The coal seams
generally are the most laterally persistent lithotypes and,
therefore, are often the only beds that could be considered
regional aquifers.
The specific rock types underlying and adjacent to the
project site are predominantly shales with sandstones, three
coal seams, and their associated underclays. In order of
proximity to the surface, mineable coal is found in the Moss
coal bed, the Jellico coal bed, and the Blue Gem coal bed.
All three coal seams are located above the elevation of the
project site. There are no major structural features in the
area. Localized structural highs and lows are.present, and
there is a slight dip of the strata toward the .southeast.
Mining
The area surrounding the study site was mined by both con-
tour and mountaintop removal methods. The abandoned area above
the study site was contour mined 20 years previously, but not
all the coal in this area was removed.
The primary coals of mineable thickness in the area are
contained in the Blue Gem, Jellico, and Moss coal beds. These
seams are mined extensively throughout the area. Mined areas
above the project site are shown in Figure 6.
SOIL CONDITIONS
Soils located in the vegetative filter area are contained
in the Muse-Wellston-Trappist association. The study area
28
-------
Q)
4->
•H
CQ
4J
U
0)
•n
O
H
0,
0)
(0
to
(0
QJ
M
(0
o;
•H
6
rH
0}
3
O
•H
0)
$-1
-------
is predominantly composed of natural soil, although some
spoil material pushed over the outslope is evident in the
upper portion of the area. Part of the area is devoid of
topsoil, due primarily to erosion of unvegetated areas.
This soil association consists of soils on broad,
narrow, rounded ridgetops and on convex side slopes. Likely
position for this soil is along temporary drainageways of
intermittent streams. The dominant material in the area is
shale, but some sandstone can be found.
Typically, the following percentages of each series are
found in the described areas:
o Muse - 60 percent
o Wellston - 20 percent
o Trappist - 10 percent
o Minor soils - 10 percent (Clymer, Dekalb
and Tate series)
This particular association of soils is very highly
dissected, with streams generally 70 meters below the ridge
tops.
The Muse and Trappist soils are found on side slopes,
narrow ridgetops, and small knobs on broad ridgetops. The
Wellston soils are commonly found on broad, smooth to slightly
convex ridgetops.
Muse, the major soil series found in the area, is
formed from residuum that weathered from interbedded acid
shale and thin sandstone, or colluvium transported down from
soils derived from shale.
The Muse soils have thick root zones, and are strongly
acid and moderate to low in natural fertility. 'In the study
area the profile consists of a surface five (5) centimeters
of very friable brown silt loam over 15 centimeters of
friable, yellowish-brown silty clay loam.
A typical profile 6f a Muse silt loam includes:
Al—0 to 5 centimeters, brown silt loam; weak, fine,
granular structure; friable; 1 percent of the
volume is shale fragments less than 1/4 inch
across; very strongly acid; abrupt, wavy
boundary.
A—5 to 20 centimeters, yellowish-brown light silty
clay loam; weak, fine, subangular blocky
structure that breaks to weak, fine granular;
friable; 1 percent of the volume is shale
fragments less than 1/4 inch across; very
strongly acid; clear, wavy boundary.
30
-------
Bl—20 to 33 centimeters, yellowish-brown silty clay
loam; weak, fine and medium, subangular
blocky structure, friable; 3 percent of the
volume is shale fragments 1/4 inch to 1-1/2
inches across; very strongly acid; clear,
wavy boundary.
B21t—33 to 51 centimeters, strong-brown heavy silty
clay loam; weak fine, subangular blocky
structure; friable; faint, patchy clay films
in pores; 5 percent of the volume is shale
fragments 1/8 inch to 1 inch across; very
strongly acid; gradual, wavy boundary.
B22t—51 to 76 centimeters, yellowish-red silty clay;
moderate, fine, subangular blocky structure;
friable, faint, patchy clay films in pores;
15 percent of the volume is shale fragments
1/4 to 1 inch across; very strongly acid;
clear, smooth boundary.
B3t—76 to 117 centimeters, yellowish-red silty clay;
common, fine faint varigations of yellowish
red and light reddish brown; weak, very fine,
subangular blocky structure; friable; faint,
patchy clay films in pores; 20 percent of the
volume is shale fragments 1/8 to 1/2 inch
across; very strongly acid; gradual, irregular
boundary. .
C—117 to 178 centimeters, yellowish-red silty clay;
common, fine, faint mottles of pinkish gray
and reddish yellow; massive; firm; 20 percent
of the volume is shale fragments 1/8 inch to
2 inches^ across; very strongly acid.
•^
HYDROLOGIC DESIGN CONSIDERATIONS
Engineering tasks performed during the project can be
grouped as: (1) development of a design storm; (2) design
using the Kentucky Grass Filter Model; (3) selection of a
filter design.
Development of a Design Storm
Runoff Hydrographs —
The development of a design storm involves the prediction
of a runoff hydrograph and a sediment pollutograph. A map
and cross-section of the contributing areas is shown in
Figure 7. A runoff hydrograph was calculated for a two (2),
five (5), and ten (10) year, twenty-four (24 hour storm
using the SCS TR 55 Method (SCS, 1975). Computation pro-
cedures are provided in Appendix I, and the hydrographs
plotted in Figure 8.
31
-------
c
o
•H
4J
U
a>
to
0}
to
o
n
u
•o
c
fO
m
s
r-
0)
32
-------
10-YEAR STORM
5 -YEAR STORM
2 - YEAR STORM
Figure 8.
125 I3X>
TIME(HRS)
Runoff hydrograph for 2, 5 and 10 year storm
33
-------
Pollutographs —
The development of sediment pollutographs is not as
straightforward as the development of runoff hydrographs and
no one computation procedure has gained widespread acceptance.
In order to predict a pollutograph, the sediment yield must
first be predicted. This yield is then distributed over a
period of time and used in conjunction with a runoff hydrograph
to develop a pollutograph.
Sediment Yields —
Several procedures- are available for computing sediment
yield. These procedures are summarized in Haan and Barfield
(1978). Since the sediment source area for this study is a
long, steep slope, a prediction procedure was needed to
account for the possibility of deposition which might occur
just prior to the grass filter. One procedure which takes
this into account is the Deposition Modified Universal Soil
Loss Equation (USLE) as given in Haan and Barfield (1978).
In order to make the sediment yield computations using
the Deposition Modified USLE, the source area was divided
into segments as shown in Figure 9. The sediment yield from
the slope is then calculated according to the following
procedure:
1. Determine the size distribution of sediment
aggregates which will be eroded from each
segment.
2. Determine the sediment detached from each seg-
ment, Djt.
3. Divide the detached material into bedload and
suspended load, DjB and DJS. Calculate the
material available for transport as that
entering the segment plus that detached from
the segment.
4. Calculate the bedload transport capacity of
water exiting each section, Tcj.
5. Compare the total bedload material available
for transport from each section with the
material available for transport. The actual
bed material exiting the section will be
less than the .transport capacity and the ma-
terial available for transport.
6. Assume that suspended load is unaffected by
narrow grass strips and mulches. Calculate
34
-------
CO
1
o
o
Ul
UJ
e
«D
fcl
O.
3
w
o
•H
4J
O
-H
•8
c
a)
•H
•o
0)
to
•o
0)
(0
tri
4J
ai
ft
Q)
0)
O
t-l
er>
(I)
!M
&
•H
35
-------
the fraction of suspended load trapped only
by long grass strips.
7. Repeat steps (1) - (6) for each segment.
A flow diagram of the computation process is shown in Figure
10.
Size Distribution of Eroded Sediment —
The probable size distribution of the eroded sediment
was determined by subjecting samples of the source area soil
surface material to simulated rainfall with an energy
content similar to that of a 10-year, 24-hour storm and
measuring the size distribution of the eroded sediment.
This was accomplished by using an apparatus in which simulated
rainfall was generated^rom a Spraying Systems 80-150 Nozzle,
operating at 6 psi, (.42 Kg/cm2) located at a height of ten
feet above the sample. This arrangement will give a rainfall
with an energy content similar to that "of a storm with an
average intensity greater than 0.6 in/hr (1.52 cm/hr.)
(Williams, et al., 1978, Meyer and Harman, 1977). At the
site selected for this study, the percent of 24-hour rainfall
greater than 0.6 in/hr (1.52 cm/hr.) is shown in Table 3
(using as SCS Type II rainfall distribution). At intensities
less than 0.6 in/hr (1.52 cm/hr.), the energy content of the
simulated rainfall will be greater than that of natural
rainfall; therefore approximately 37 to 40 percent of the
rainfall will have an energy content greater than natural
rainfall. The resulting effect on the size distribution of
eroded sediment should be such that the eroded aggregates
will be smaller than those generated by natural rainfall.
The result of this input into the vegetative filtration
model, to be discussed in a subsequent section, will be
conservative in that it will underpredict the trapping
efficiency.
The eroded sediment from the sample was directed into a
sieve stack, with sizes ranging from No. 16 to No. 230.
These sieves were gently washed five times to determine the
percent of sediment in each size range. The particles passing
the number 230 sieve were subjected to a pipette analysis to
determine their size distribution. The tests were replicated
four times and the size distribution plotted in Figure 11.
Sediment Detached from Slope Segment —
The sediment detached from any segment (j) during a
design storm was calculated from the equation (Haan and
Barfield, 1978):
36
-------
FLOWCHART FOR EROSION DETERMINATION
SUSPENDED
LOAD FROM
ttPSLOPE
•ED LOAD
FROM
ttPSLOPE
PLUS
DETACHED
SUSPENDED
LOAD FROM
SE8EMENT
• ED LOAD
TRANSPORT
CAPACITY OR
FLOW
1
PLUS
DETACHED
•ED LOAD
FROM
tEOEMENT
•EDUCE BY
FRACTION OF
SUSPENDED
LOAD TRAPPED
COMPARE AND TAKE
LESSER
TOTAL LOAD FROM
•E8EMENT
Figure 10. Flowchart for erosion determination
37
-------
I' I ' I' I ' I ' I ' I ' I ' j'
w
5
I S.
«•*»»!
JF f"S2
* K o. a- 2;
:
•B
W
10
(1)
JJ
(C
151
ID
IM
!Cr
D-i
O
•P
C
0)
a
•H
"O
0)
10
0)
•O
O
n
0)
«4-t
o
§
CO
•H
•O
Q)
N
•H
(0
•O
a)
4J
u
-H
•o
0)
S-l
0)
i-i
8 S S
JLMftlS* WI V3NU %
38
-------
TABLE 3. PERCENT OF 24 HOUR RAINFALL WITH AN INTENSITY GREATER
THAN 0.6 INCHES PER HOUR (1.52 CENTIMETERS/HOUR)
USING AN SCS TYPE II CURVE -'
Approximate Approximate
Duration of % of Rainfall
Frequency Rainfall 1.52 cm/hr greater than 1.52 cm/hr
2
5
10
yrs.
yrs.
yrs.
3.
3.
5.
5
5
0
hrs.
hrs.
hrs.
60%
65%
67%
Source: Haan & Barfield, 1978.
Dj = 0.0459 R KJ CPj Sj ( AjM+1-A jm+1) / (72.6)m
where Dj s the detachment from the segment in Ibs/ft-width;
R is the Universal Soil Loss Equation rainfall factor for
the storm; Kj is the soil erodibility for the exposed material,
CPj is the cover practice factor for slope segment j; A • is
the distance from the top of the slope to the bottom of the
segment; m is a coefficient depending on the slope; and
S-i = 0.43 + 30 sin0 + 430 sin2 6 _ ...
3 —— 67613 cos W
where 0 is the angle of the slope!/. Values for M, R, and
CPj '"are tabulated in Haan and Barfield (1978) and nomographs are
presented for determining KJ from soil size distributions and
permeability data. Detailed computations are provided in Ap-
pendix A.
Dividing the Detached Material into Bedload and Suspended
Load •—
The assumption was made that the detached material in
the silt size range and smaller (diameter < 0.61 mm) would be
transported as suspended load/ and larger particles would be
transported as bedload. Based on the size distribution shown
in Figure 11, 65 percent of the particles would be transported
I/ The term CosG was not included in the original Universal
~~ Soil Loss Equation, but is included here to correct for
the fact that steep slopes will not receive as much rain-
fall energy as flat slopes.
39
-------
as bedload and 35 percent as.suspended load.
The Transport Capacity for Each Segment —
The bedload transport capacity Tcj is the flow exiting
section j , is given by:
Tcj = r - R • T • A2 • (S1)6 ' Cro • CT
•*
where T • is the transport capacity for the flow leaving the
slope slgment j (Ibs/ft-width); r is a proportionality constant
which related the product of runoff volume and peak discharge
to Wischmeier's R factor; R is the return period rainfall
factor from The Universal Soil Loss Equation; and CT is the
transportability reduction factor given as a function of the
ratio^of velocity with treatment to that of a bare fallow soil.
Detailed graphs and charts for determining r, R, T, B, Cro, and
CT are given in Kaan and Barfield (1978)/ as well as Neibling and
Foster (1977). The computed values of Tcj, for each slope
segment, are shown in Appendix A.
The Final Sediment Yield —
The actual bedload sediment exiting each segment is the
lesser of the transport capacity, Tcj , and the material avail-
able for transport. The bed material available for transport is
the sum of that entering the segment and the bed material de-
tached from the segment. The bed material exiting the segment
is added to the suspended load from the segment to obtain the
total sediment load exiting. Detailed calculations are given
in Appendix A. The resulting sediment yield is shown in
Table 4.
TABLE 4. SEDIMENT YIELD FROM THE SOURCE
AREA TO THE FILTER AREA FROM A DESIGN STORM
Return Period Sediment Yield
Storm Tons kgs.
2 yrs. 40,260 18,300
5 yrs. - 70,125 31,875
10 yrs. 100,000 45,455
40
-------
Development of Pollutographs —
Pollutographs were developed under'two assumptions: (1)
that the concentrations of sediment remained constant with
time, and (2) that the concentration of sediment varied with
time. At the time the design computations were made, no data
were available from surface mines for checking these assumptions.
Should the concentrations vary with time, a relationship
has been proposed by Ward, et al (1978) which can be used to
predict peak concentrations:
Cp = 0.0636 (Yd/A0'74 (qp)"0'394 t^Q.Ill
where Cp is peak inflow concentration in mg/liter; Yd is the
sediment yield of the storm in tons; A is the area in acres;
q is the peak storm discharge in cfs; and t . is the storm
duration in hours. If the concentration is assumed to be con-
stant, the average concentration will be given by:
C = Yd/2000 xlO6
62.4V
where V is the runoff volume in cubic feet. Detailed computations
are shown in Appendix A and the results in Table 5.
TABLE 5. CALCULATED AVERAGE AND PEAK
CONCENTRATIONS OF STORM RUNOFF
I/
2/
Return
Period
2 yrs.
5 yrs.
10 yrs.
Concentration
Concentration
Peak!/
Cone
mg/liter
0.610
0.790
0.960
does not vary with time
varies with time
Avg.i/
Cone
mg/liter
0.089
0.103
0.124
The peak concentrations shown in Table 5 are very high,
whereas the average concentrations appear to be reasonable. At
the time the hydrologic design computations were made, data ex-
41
-------
ited to indicate that the average concentrations can be realized
and very limited data existed to indicate that the peak con-
centrations could be obtained.
Therefore/ it was assumed that the pollutographs should be
generated using the average concentrations. Thus, the resulting
sediment load can be calculated from:
where q_ is the sediment load; c is the average concentration;
and qw is the runoff hydrograph ordinates. The resulting pollu-
tograph for a 2.5 and 10 year storm is plotted in Figure 12.
Detailed computations are given in Appendix A.
The pollutographs ^cnd hydrographs for each storm frequency
are used as the inputs to the grass filter model presented in
Section 3.
Design Using the Kentucky Grass Filter Model
The Kentucky Grass Filter Model as given in Section 3 has
been computerized in an interactive format. Inputs required
for the model are given in Table 6.
TABLE 6. KENTUCKY VEGETATIVE FILTER
MODEL INPUTS AND OUTPUTS
Hydrologic
Inputs
Media
Inputs
Outputs
(Varying with Time)
Peak Inflow
Sediment Concen-
tration
Sediment Size Dis-
tribution
Filter Length
Slope of Ground
Media Spacing &
Size
Sediment Concentration
Sediment Size Distri-
bution
Fraction of Media Inun-
dated
Media Height
Filter Width
Infiltration Rate
Manning's Roughness
42
-------
1O-YEAR STORM
5 -YEAR STORM
2 -YEAR STORM
I2JD 125
TIME(HRS)
13.0
Figure 12. Predicted runoff pollutographs
43
-------
A complete listing is given in Hayes, et al (1981), as well as
graphical solution procedures. At the time the hydrologic de-
sign computations were made, the component, which accounted for
the effects of infiltration on the sediment mass balance, had
not bee developed. The results, therefore, should be conservative.
Model Sensitivity Analysis —
A sensitivity analysis was conducted, as part of the
hydrologic design, to determine the sensitivity of the model
to the input parameters of filter length, width, Manning's n,
media spacing, and media size. Standard values were input into
the model as a reference, and one by one, parameters were allowed
to vary over a range one might expect to encounter in the field
(referred to as a single parameter sensitivity analysis - Table
7).
TABLE 7. VALUES USED FOR SENSITIVITY ANALYSIS
OF KENTUCKY GRASS FILTER DESIGN MODEL
Parameters
Grass Spacing (cm)
Grass Diameter (cm)
Manning ' s n
Max. Deposition (cm)
Media Height (cm)
,-f.
Time Increments (sees.)
Slope (%)
Slope Length (meters)
Slope Width (feet)
Infiltration Rate I/
Standard
Value
1.6
0.1
0.012
0.75
12'.0
20.0
10.0
41.0
21.0
—
Ranges
1.0
0.08 -
0.009 -
0.300 -
8.0
5.0
3.0
10.0
20.0
^™
2.0
0.13
0.016
1.500
16.0
60.0
25.0
100.0
140.0
".
X"iL- UX1C= l_J-il«- ^SJ- »-*»^- --j ^ -. _,_.,, ..
been developed to account for the effects of infiltration
on the sediment mass balance.
44
-i—v •.w^t^cisma, »-<•«*•» *-.* i':-VV.-#'1*STWBw*~*!r"*'1 """."„"* ":'-''.r '"•" '*"' ". ""'.'?"'"'* "". : '. •".* •'-.*.* ,• f^' * •* .••' 1 •'r'. , • • ' ' : ^.f;'--"-'-^ ..„,,. _
-------
Runoff and sediment pollutographs from the ten, five, and
two year storm were input into the model, and outflow pol-
lutographs from the grass filter were predicted for each set
of conditions. Plots showing how the peak outflow sediment con-
centration varied with each input parameter are provided in
Appendix B. From a visual inspection of the plots the parameters
which are most important are slope, filter width, and filter
length. A reasonable estimate is sufficient for the other
input parameters.
Selection of a Filter Design
In the selection of a final design, the following criteria
were considered:
(1) Minimize earth movement,
(2) Keep slope as close as possible to the natural
slope of the ground,
(3) Allow for monitoring of inflow, and outflow from the
filter,
(4) Vegetative density and height must be such that
they can be rapidly obtained, and
(5) Maximum possible trapping efficiency for the site
selected.
Based on these criteria, a vegetative filter was selected with
the characteristics shown in Table 8.
TABLE 8. DESIGN VEGETATIVE FILTER CHARACTERISTICS
-- Slope 17%
Length 41.0 meters
Width ' . 21.0 meters
Spacing 1.6 cm
Type vegetation To be determined
Hydraulic parameters for
predicting model Kentucky 31 Fescue
Grass diameter .1 cm
Manning's n .012
Media height 12.0 cm
45
-------
.15 I
o"
•o **n
^s>
J.
I
§
to.
«
5 .05
/ \ -.- HIFLOW MYDI«0«*AI»H —
/ \ — — INFLOW 8EDIMENT8RAPH
/ *. OUTFLOW IEDIMENT8RAPH
/ \ _
1 \
™ • .
* &
* " ^Ml
i
,-^ I
• / \ \ —
§/ N •
* \
« \ •
I \ X _
" ! \ \
i \ •
* A \
! v^'x-
I
i \ \
i i \\
1 1 \ \ ~~
/ ' • ^ \
- / \\
// *'^
i i \ \
/ ' v •
// m \\
if ..,-" c ^Vx
\>/^f , 7""-— .^"^a*.-
24
22
20
18
6
4
2
10 tO SO 40
TIME (Mitt.)
50
60
* .
Figure 13. Predicted hydrograph and pollutographs for the
Figure xo des±gn filter with a 10 year storm
46
-------
.p
c
0)
E
•H
T3
0)
(0
tw
•a
c
10
•O
0)
+J
O
-H
«C
0)
^J*
f-l
Q)
47
- , w.^ .__,.- v-^.-T.5-.--.-.-r^c'.».-->.'-- ..7>=>vyn-r^
-------
.IS
^. .
•> .10
•x.
J,
W
i
s
•J
Ik
(J
fcl
£ «O3
^
— — INFLOW HYDROOtAPN —
— INFLOW SCDIMENT9RAPM
OUTFLOW SEDIUENTMAPH
—
r\
f \
t \
/ ' \
/ \
r \ -
/ \
/ x
/ —x \
1 / \ \
•'/ N^ \
// \ \
/ ' ' %* \
.1 V
- / 1 V \ "
' 1 • \ \
; i • \ • v -
/ / \ V
i I \ \
' 1 x \
/ / N \
. ' v V
i 1 x "v
/ / v \
. / • V v
1 / • \ X-
" / - • »•• • . ^ ^te
;/' ..-• f v-«>^
^ >: • • ^- • ^f—
24
22
20
8 •»
m
2
16 n
5
14 1
12 9
5
m
IO —
S
O ^fl
o
*^
6
4
2
OK) ID 30 40 80 «0 70
TIME &»!«•)
Figure 15
Predicted hydrograph and pollutographs for the
design filter with a 5 year storm
48
-------
I i-i • I • I
INOOM Ai H1N1J %
-------
.15
£ 'IO
.05
INFLOW HYDR08RAW
INFLOW SEDIMENTORAPH
OUTFLOW SEDIMENTORAPH
/
V \
\ x
//
V ^.
24
22
-20
18 •»
m
o
16 *
-4
12 3D
PI
10 £
o
CO
8 PI
o
;•
""" 4
K)
^0 SO 40 50
TIME (min.)
-^. ....
—!--••—- ""a
TO
60
Figure 17. Predicted hydrograph and pollutographs for the
design filter with a 2 year storm
50
-------
I . I . I « I . 1 . I
fi
||
to
c
o
X}
•H
M
•P
CO
-H
•O
a) C
N O
•H .p
10 (0
C rt
0) 0)
E >i
•H
T3 IN
0)
co m
o c
•H -t
o o
M U-l
CM
CO
rH
0)
8
9
1M9I1A M ttlNU %
51
-------
0)
.fi
+J
in
(C W
4J O
c •<->
0) (0
Sn
CO
•H iH
'S-a
ns c
l-i (0
in
O
0)
rt -H
4J 5
to
C M
•H 0)
4J
•O t-<
0) -H
4J (H
0) -H
V) (0
p, 0)
•0
i-l
(1)
8 8 8
Q3ddVMl 1N3WIQ3S %
&
•H
52
-------
en
a)
•O
o
X
+J
(0
C
0)
C O
O iH
C 13
O C
u m
> m
CM
-------
Design Sediment Predictions —
Predicted sediment inflows, outflows, and size distributions
are shown in Figures 13 through 20 for the design selected.
The predictions indicate that vegetation will drastically re-
duce the sediment load for all of the return period storms con-
sidered. All of the sand and the larger silt size particles are
trapped by the vegetative filter.
When the outflow sediment graphs are converted to con-
centrations, Figure 19, the resulting concentrations appear
to be quite large. However, when considered in light of the in-
flow concentrations, the fraction of sediment trapped is ob-
viously quite high. All of the instantaneous percentages are
greater than 85 percent as shown in Figure 19.
•-">
VEGETATIVE DESIGN
As is the case with almost all environmental activities,
planning is the key to the success of vegetative establishment.
Four major planning phases should be investigated before begin-
ning any field activities. These phases include:
o preliminary site and soil survey,
o vegetative species selection,
o fertilization program,
o seeding plan
Preliminary Site and Soil Survey
The first item for consideration was the soil reconnaissance,
Before any work on the site began, the soils were thoroughly
tested for a minimum of buffer pH, phosphorus and potassium.
Composite soil samples were taken from both the top and
subsoils to detect any toxic or acidic problems that might
be present.
In addition to soil fertility information, the quantities
of soil available become very important when selecting
vegetation. The topsoil on the site was removed, stockpiled,
and then replaced, with the total depth ranging from 10.2 to
15.3 centimeters. This determination limited the types of
vegetation which could be utilized, due to the acid conditions
in the subsoil. Only by careful evaluation of fertility
levels and soil characteristics can unsuccessful vegetative
stands be averted. Further knowledge of the area can be
gained by noting the existing vegetation and how well it is
adapted to the site, and whether the plant species present
are native or introduced.
54
-------
Vegetation Species Selection
The second phase in planning for vegetative establishment
is the actual selection of vegetative species to be seeded
in the filter area. The following discussion of selecting a
grass for this very specialized use is directed toward the
surface mining areas of the Eastern coal fields, where the
vegetative filter test site is located. This section is
not to be confused, however, with general revegetation
practices, although some are evident in the procedures
utilized. The selection of suitable vegetal media revolves
around the evaluation of ten major factors:
o Soil condition and depth
o Soil chemical analysis
o Characteristics of sediment to be trapped in
the filter
o Specific establishment procedures
o Maintenance requirements
o Duration of filter site use
o Selection of vegetation adapted to area climatic
conditions
o Current on-site vegetative species
o Major stresses encountered by the vegetation
o Physical site limitations
For the purposes of this report, discussion of these
factors will be directed to the specific needs of the study
area.
Soil Condition and Depth —
Different species of vegetation require different types
and depths of soil rooting medium. For example, a legume
such as alfalfa develops a taproot which penetrates deep
into the soil profile, while bluegrass, on the other hand,
develops a very shallow and fibrous root system. The veg-
etation should be selected to meet the soil limitations
found on the site. In the case of the study area, the
subsoils are very acid and only 12.7 centimeters of topsoil
were available for use as a growth medium. A root growing
into this very acid subsoil would definitely limit the
plant's uptake of nutrients and growth, unless a very high
acid resistance was exhibited by the plant. The main limiting
factor in this category, in view of the above considerations,
is the rooting depth of the plant species.
Soil Chemical Analysis —
Information gathered during the initial soil testing
program is utilized to match the soil fertility to the
nutrient requirements of the plant. As an example, some
plants can tolerate very acid conditions and continue almost
55
-------
normal growth when no other limiting factors are present.
If the study area is plagued by acid conditions and very low
phosphorus levels are indicated, this should be taken into
consideration when selecting vegetation.
Characteristics of Sediment to be Trapped —
Surface testing of the material to be trapped by the
filter should also be considered. Since eroded material,
when diverted into the filter zone, would be trapped and
deposited on the soil surface surrounding the vegetation.
Characteristics of this trapped material could dramatically
effect the survival of the plants. The acid conditions on
the study site prompted the application of 1200 pounds (544
Kilograms) of additional agricultural lime to the top 3.05
meters of the filter, to counteract the high acid concentrations
expected in the inflow. Chemical testing for pH should have
given good indictions of problems that might be encountered
from inflow containing high concentrations of toxic micro-
nutrients.
Specific Establishment Procedures —
When considering a vegetative species for planting,
evaluate how the planting would be accomplished and specifically
what equipment would be used. In most situations, the size
of the area involved would determine the equipment that
could be used efficiently. Of course, equipment already
available could be adapted for use in seed-bed preparation
and seeding. This takes into consideration the fact that
the seed should be covered for good germination and quick
erosion control.
Maintenance Requirements —
Different grass species may require totally different
levels of management. Some species require high fertilization
levels and disease control precautions. Low maintenance was
a key factor in the selection of vegetation for this project.
By placing the vegetation under abnormal stress conditions,
some maintenance will inevitably be required. Consequently,
it is desirable to select a species which may require very
little maintenance under normal conditions. •
Duration of Filter Site for Use —
In the development of a mine plan, estimations are
based on the amount of time an area will be disturbed by
the mining process. The same type of evaluation should be
made in the design of the vegetative filter zone. Some
vegetative species are characterized by an annual growth
habit and some are perennial; the choice of which type to use
56
-------
being dependent on criteria set on deviation of use and site
conditions. The monitoring period established for this
study was one year. The growth cycle would need to be
perennial in order to perpetuate the vegetation through the
life of the project. The post-mining land use of the area
also merited consideration so that double handling of the
soil material would not be necessary.
Selection of Vegetation Adapted to Area Climatic Conditions —
In the Southeastern Kentucky study area, several
species of vegetation are well adapted for normal growth.
In other areas of the country, a completely different set of
vegetation species may be required for good adaptions. Both
cool and warm-season grasses are found in the Eastern coal
fields, but the predominant species are cool-season. The
growth habit of the cool-season species denotes high growth
activity between the 10 and 24 degrees centigrade temperature
range. Optimum growth period for Kentucky cool-season
grasses are in the spring and fall of the year. This growth
period matches well with the times of high rainfall and the
runoff in the project area. Depending on the individual
case, the warm-season species might be more applicable to
.other areas of the country.
Existing On-Site Vegetative Species—
Observation of existing plant species under natural
"conditions may indicate potential site problems. The most
evident visual assessment, which is easily distinguished, is
an acid toxic condition in the soil. Stunted vegetation, or
in some cases, an absence of existing vegetation, can in-
dicate problems with pH or toxic elements.
Major Stresses Encountered by the Vegetation —
Normal field conditions introduce the natural forces of
rainfall and drought conditions, but these stresses are
greatly amplified when the vegetation is utilized as a
sediment control device. Flowing water creates problems by
not allowing the vegetation an upright growth habit. Grass
species with weak or low-growing stems can lay over under
the forces of flowing water and create a grass waterway.
This is detrimental, since erect growth is essential for the
effective trapping of eroded sediment. Physiological stresses,
such as wet or dry conditions, are also placed on the plant.
The plant species selected for use must exhibit a tolerance
to the above mentioned stress conditions.
Physical Site Limitations —
Steep slopes;can create varied problems in vegetation
establishment, one of which may be averted by choosing a
57
-------
plant species with a quick establishment growth habit. The
: faster a species establishes itself, the less the chance of
erosion or loss of seed. At least one species in the seeding
I mixture should germinate and grow very quickly.
: Plant Descriptions —
Based on the examination of the ten factors affecting
i the vegetative selection, two vegetative species were chosen
' for seeding in the study area. The following descriptions
reflect the objectives of the search for the proper vegetation.
Tall Fescue - (Festuca Arundinacea) — Tall fescue is a
deep-rooted, long-lived perennial whose growth habit is
1 typically a bunch grass, although some short underground^
; stems do occur. The roots are tough and coarse and contribute
to good sod production under proper seeding conditions.
This variety of fescue is adapted to a wide range of
climatic conditions and is grown from Florida to Canada.
I Due to its cool-season growth habit, the best adaptation
' occurs in the central and northern transition regions of its
i range (Figure 21). However, tall fescue is the only cool-
season grass which will persist through the hot summers and
i cool winters, year after year, in the southeast.
Tall fescue is very tolerant of poor drainage, particularly
: in the winter months. It is capable of thriving in many of
the poorly drained areas in the U.S. and Europe. This
', plant's tolerance to poor drainage is exceeded only by its
' drought tolerance. A deep, fibrous rooting system enables
i ' this fescue to withstand lengthy periods of very dry weather.
' Soil pH ranges best suited to this species are from 5.5
i to 6.5 although pH's from 4.7 to 8.5 can be tolerated. The
i tolerance of tall fescue to saline and alkaline soil conditions
' is better than most other cool-season turf grasses.
*#• i
Perennial Ryegrass (Lolium perenne) — The ryegrasses
are considered to have a bunch-type growth habit,_with no
i record of a creeping nature. Perennial ryegrass is commonly
'! used in pasture mixtures and in a high quality forage. This
; species is not long-lived and persists for only three (3) to
i four (4) years before thinning begins.
; Ryegrass is grown west of the Sierra Nevada and Cascade
Ranges and in the southern humid regions of the U.S. Its
i use also extends up the Atlantic Coast (Figure 22)i Ryegrass
' is less winter hardy than certain other forages such as
Timothy and Orchardgrass.
58
-------
Figure 21. Tall fescue producing areas of the U.S. (37)
Figure 22. Ryegrass producing ^reas of the U.S. (37)
59
-------
The ryegrasses are well-adapted to most soil types
although they do not respond well to low fertility. Best
growth occurs on neutral to slightly acidic soils of medium
to high fertility. Tolerance to wet conditions is good, as
long as there is adequate surface drainage.
These two grass species meet the criteria for revegetative
selection, as previously outlined in both their physical and
chemical characteristics. In addition to the parameters for
selection, one other important factor was taken into considera-
tion. The computer model developed at the University of
Kentucky had for one of its inputs the diameter of the
vegetative species seeded in the filter. Since one of the
subordinate objectives of the study was to compare actual
field data with information predicted by the model, the
species chosen are of approximately the same diameter at
maturity. By design, tliis would reduce the inputs into the
computer program and make the validation of the model more
reliable.
Fertilization Program
The third step in the establishment of a vegetative
cover is the development of a fertilization program. The
recommendations discussed in this section are based on
information available in the University of Kentucky Extension
Publication, AGR 40, which describes the fertilization of
mine spoils.
Before a final decision was made on a fertilization
plan, the optimum conditions for growth of the chosen vegetative
species were considered. Tall fescue can tolerate high or
low levels of fertilization as well as acid conditions.
Perennial Ryegrass is more adapted to high levels of fertilization
for optimum growth. The following discussion of relative
amounts of nitrogen, phosphorus, and potassium nutrients are
tailored to soil requirements and optimum growth of the
plant species selection.
Soil samples collected during the preliminary site and
soil survey were utilized to calculate the amounts of fertilizer
needed for vegetative establishment. The results of these
tests are shown in Table 9.
Based on the soil test results, the following recommenda-
tions were made on lime, nitrogen, phosphorus and potassium
for use as soil amendments.
60
-------
TABLE 9. SOIL TEST RESULTS
Topsoil Subsoil
pH (water) 5.6 5.2
pH (buffer) 6.5 5.9
Phosphorus 1 4
Potassium 156 121
Lime —
The test results show a buffer pH reading of 6.5 and
the corresponding lime rate from AGR 40 is three tons/acre.
Due to the possible stress of acid materials washing into
the filter from the outslopes, the lime rate was increased
to five tons/acre. This increased rate would buffer the
soil against acid drainage and further amend soil pH to
acceptable levels for vegetative growth. Argicultural lime
was applied in pelleted form. The five ton requirement
stated above translates to 37.00 pounds (1678 Kilograms) of
lime for the filter area.
Nitrogen —
Nitrogen amonants utilized for initial vegetative establishment
are typical of tKe new seeding practices of any grass species.
Thirty pounds (13 Kilograms) of actual nitrogen per acre
would be applied at the time of seeding. This translates to
33 pounds -(15 Kilograms) of ammonium nitrate for the filter
area- When the vegetation reached a height of four to five
centimeters, an additional five to ten pounds (2 to 4 Kilograms)
of nitrogen would be added to optimize growth. Maintenance
fertilization with nitrogen through the course of the project
would be based on the characteristics of the grass species.
Phosphorus —
The levels of phosphorus in the soil at the time of
seeding can be one of the most limiting factors to germin-
ation and growth. Phosphorus levels in the area were very
low and could cause various problems with establishment of
the grasses. One-hundred fifty pounds (68 Kilograms) of P2
Os/acre would be required to raise the phosphorus level to
an acceptable point. This translates into 55 pounds (15
Kilograms) of a 0-0-60 potassium fertilizer.
61
-------
Seeding Plan
Seeding rates were chosen to meet both the design
criteria for sediment trapping and the overall requirements
of the study. In common terms, the seeding rates used are
approximately half those used for seeding a lawn. These
rates are used for generating thick stands of dense vegetative
cover.
From the preliminary design calculations, a grass
spacing of 1.6 centimeters was considered necessary for the
filter area. This spacing interval met the requirements of
the flow introduced into the filter area for monitoring.
In order to achieve the required density, 75 pounds (34
Kilograms) of tall fescue seed and 30 pounds (14 Kilograms) of
perennial rye were seeded on a per acre basis. Reasons
governing the use of the two grasses were covered under the
vegetative selection section. Rates of seeding were governed
by the density requirement of the grass as outlined in the
design criteria.
MONITORING SYSTEM DESIGN
The original task description of the project outlined a
hand sampling program to monitor inlet and outlet sediment
concentrations. The monitoring program was altered considerably
when the University of Kentucky, Department of Agricultural
Engineering agreed to cooperate in the project. The University's
interest in the project stemmed from their previous work in
vegetative filters. During earlier vegetative filter design
work, the computer model discussed in this report was
developed from laboratory data. -This project allowed the
university an opportunity to test the validity of the model
in the field. The data required by the university was more
extensive and thus required a more sophisticated monitoring
system. The monitoring equipment discussed below was used
in the study primarily because of its availability at the
start of the project.
The monitoring system revolved around several major
components which were combined to provide a complete system.
These components include those listed in Table 10.
The H-Flumes were designed and built for the purpose of
handling a ten-year storm event. Designing the flumes for
a two-year event was considered, but discounted on the
premise that should a large event occur, it would not be
effectivley monitored. On the other hand, if the upper limit
62
-------
TABLE 10. MONITORING INSTRUMENTATION
Parameter Instrument Manufacturer
Water flow rate H-Flume and water Stevens (Model 71
stage recorder recorder)
Sediment Concen- Automatic pumping Isco (Model 1680)
tration
Rainfall Weight type rain- Weathermeasure
gauge
were set on a ten year event, any storm of a low return period
might not be detected. The 1.5 foot H-Flume was selected
because it better fit the flow characteristics expected on
the site. The flumes were constructed from sixteen gauge
sheet metal by the University of Kentucky.
Following construction, the H-Flumes were calibrated at
the University of Kentucky engineering water laboratories.
The methods of calibration were:
o high flow - flow rate measured by calibrated
venturi meter, and
o low flow - flow rate measured by dye tracer tech-
niques.
(calibration curves are shown in Figure 23).
In addition to the flumes, a metal flow divider was also
designed and constructed. The divider was designed for
location directly below the inlet flume. As the flow of
water passed through the flume, it dropped into a splash
box, and then into the flow divider network. The metal flow
divider consisted of a row of metal plates positioned to
separate the total discharge into fourteen equal flows.
Attached to the metal plates were the two-by-four dividers
which would distribute the flow equally across the width of
the filter area. Design details of the inlet flume installation,
metal flow divider, outlet flume installation and H-Flume
plans are shown in Figures 24 and 25.
The Stevens water stage recorder, with chart drive,
operated from a float activated by a stilling well on the
flume. The details regarding installation of this
equipment are discussed in the water monitoring installation
section.
63
-------
• 10
- 5
Q«.052«4-64H
CALIBRATION CURVE ..
H- FLUME ^'
DYE TftACER FLOW MEASUREM
CALIBRATED VENTUKI FUJW MEAS
ME AD (ft.)
UJ
Figure 23. Calibration curves for the H-Flumes
64
-------
0)
•O
•H
>
•H
•o
•o
0)
I
I
ffi
I
JJ
CO
>1
(Q
c
•H
n
o
4J
•H
i
n)
rH
0)
•P
•H
(0
M
O
W
m
0)
o
0)
65
-------
03
-P
C
0)
I
o
CO
iH
I
-P
0)
C
(0
4J
0)
iH
C
•H
O
14-1
0)
OT.
C
•H
5
n
•o
C
o>
•H
n
CJi
Ci
IT)
w.
0)
66
-------
The Isco pumping sampler was selected because it could
be set on either a time mode or a flow mode, depending on
the application desired. Sampling under this system could
take place automatically — initiated by flow, and the
samples could be collected later by project personnel. The
thirty centimeter raingauge was equipped with a clock drive
unit set for monthly recording.
The equipment was totally self-sufficient; battery
power was used by all instruments. All instruments were
housed in corrugated steel containers to avoid theft and
damage by weather.
Instrumentation such as that described above was utilized
to meet the data collection and research needs of the project.
Actual field use of a vegetative filter by the mining industry
would not require the use of an extensive monitoring system such
as the one described.
67
-------
:il^:.ij3a&aiB*jd=&3&irjJfu. ^ai^: mi^Ji.JJa&J.'S&av.'.J^'aVSr^ **"•' r S&frJPuSU.•Ji»~_££a C'jjab i>M'JM. ?JJ J'i«fci_-.'«Jji-.ii^^..,^,_.'» . •
SECTION 5
INSTALLATION OF VEGETATIVE FILTER
General topics to be discussed in this section include:
site alterations and soil handling, vegetative establishment,
water monitoring system installation and costs for filter
installation. The overall goal of the techniques utilized
in the filter construction was to simulate actual mining
techniques for soil handling. Site specific conditions were
carefully considered and design criteria established before
work began on actual installation of the,filter area.
SITE ALTERATIONS AND SOIL HANDLING
The initial phase of site alterations consisted of
setting grade stakes for use by the bulldozer operator as
guidelines. These stakes were placed to leave an area 41.0
meters in length, 21.0 meters in width on a seventeen (17)
percent slope. The rationale describing the design criteria
behind the seventeen (17) percent slope was discussed in
detail under the design considerations section. In order to
achieve the desired slope, soil had to be removed from the
rise and spread to form the level area for the filter. The
depth of removal would exceed the 12.7 to 15.2 centimeters
topsoil depth and extend into the subsoil. Since the
subsoils in this area of Kentucky are highly acid (pH 2.0 to
4.0), the topsoil would be removed (Figure 26) and redeposited
after subsoil leveling. The topsoil material was stockpiled
beside the area for redistribution after grading.
Rough grading of the subsoil material took place between
the four grade stakes placed on the corners of the rectangular
filter area as reference marks (Figure 27). Accuracy needed
for purposes of the study demanded 3.1 centimeters level
across the subsoil surface. If this level were not maintained,
channelizing could take place and erosion would adversely
affect the results of the field tests.
As previously discussed, the subsoil has an inherent
fault which could greatly inhibit plant growth. This deficiency
is a very low pH, which can severely limit nutrient uptake
in plants. To the best of our knowledge, the technique
utilized in this study to correct the problem had never been
attempted before this experiment. The lime requirement
68
-------
^4^2*SW2U$
C
•H
o
o
4-)
CO
M
o
M-l
O
(0
to
>
i
(1)
M
3
DI
-H
CM
69
-------
W~^">'VS"- *%&/.? r.?#
a. i':: £f?*A± ;i. W^sr - '•'"fe-Vfe -«^'.;<'
^Irafe''.*^
*I*«€I''V1':/;;.-
o
(Q
to
0)
A
4J
c
-H
•a
s
0)
h
3
tn
-H
fa
70
-------
after soil testing specified an application rate of agri-
cultural lime of five tons per acre. This translated to
about 3700 pounds (1678 Kilograms) of lime for the filter
area. Instead of incorporating the entire amount of lime
into the topsoil, approximately 1200 pounds (544 Kilograms)
was applied between the topsoil and subsoil, thus providing
an "alkaline barrier". In theory, this barrier might lessen
the effects of the strongly acid subsoil on vegetative
growth.
In the initial stages of leveling the area on the
required slope, approximately 1720 cubic meters of material
were moved. The following replacement of topsoil involved
the movement of 573 cubic meters of the original 1720 cubic
meters for a second time.
Topsoil was replaced over the entire area to a depth of
from 10.2 to 15.3 centimeters (Figure 28). A level surface
was obtained by driving stakes into the subsoil to a previously
marked position and stretching a nylon string between the
stakes 12.7 centimeters above the subsoil. This method was
utilized every 6.1 meters, over the 41 meter length of the
area. As the string was positioned, the bulldozer operator
filled or cut to the string. Although the process may
appear time-consuming, the work was accomplished quickly.
One very important factor often overlooked in soil
handling practices is the weather constraints. The weather's
influence on the success of grading operations is greatly
intensified if the slope area is steeper than two to five
percent. Under spring storm conditions, when rainfall
intensities are sometimes very high, the loosely regraded
topsoil may be lost to erosion. Planning the installation
of a vegetative filter should always include planning for
optimum weather conditions. During the course of the in-
stallation of the filter, pollution by erosion was controlled
by several methods. One of these methods was the use of a
straw bale dike (Figure 29).
VETETATIVE ESTABLISHMENT AND PERFORMANCE
Fertilization and seeding activities were planned to
begin immediately after the bulldozers completed the required
grading. The only fertilization done prior to this was the
application of 1200 pounds (544 Kilograms) of lime between
the topsoil and subsoil to serve as an "alkaline barrier".
An additional 2500 pounds (1134 Kilograms) of lime was
applied to the surface, along with the recommended amounts
of phosphorus and potassium. The fertilizer was broadcast
with cyclone seeders and all three materials were disced
71
^^
-------
t Si&lil siifc
1 *:i?,i'^fi, i.,Si'*- '•'»;,•.«HK
\-:»*=m?K«t
?«***r*a.-«*»<-.V---,'k"«t »••-» * -• "ijfflfcfif-*
: -"^ . v, v:.j--^»c:;-- 1L • ••-.•,.*i^.-i
•.•.-•.^.-.,V!i^ .;•.! \j*«r. Mb jii-TiHr'^
,- i,v-,.c'*t*^.- m ;"7^-1%ii
; i>-j f^.j;- V,*" '. -w .*5i«» ,/«4. • «P'!«-^-* .
• * /^'^T^^J^SW** »SE:^ 4
7f . ^~^&^^Sg -
N^-^^P^^.:
-«,.«>. i»j«sS»*Bt»«M-.. *y<>.'
if., f V, ''C „ '* ;#•'? J ^'"«"^
fe>4'..'-;?t -;«.,--jF^,:. -»;,•"•-•%»«
^:j^.;**{? ^l^^f,
tet!-;«- j,t?|j .j^.- f,'- ".srf*|^
S?* f^si *4£»;£&
^i*~^*5m&
0)
o
0)
•H
O
to
(X
O
-p
•o
0)
U
O
4J
to
00
72
-------
Figure 29. On-site erosion control using a straw bale dike
73
-------
twice into the topsoil with a straight disc to a depth of
10.16 centimeters.
The seeders criss-crossed the area two times from
different directions to ensure accurate seed distribution.
The topsoil was then lightly disced to cover the seed and
promote good soil seed contact. In addition to increasing
germination, covering the seed was intended to reduce seed
loss through erosion, a distinct possibility on the 17
percent slope. After the seed was covered, a heavy rye-
straw mulch was applied to the area for erosion control.
From the time of seeding on May 15, 1979, the vegetative
growth was checked weekly to accurately monitor progress.
Heavy rainfall made adequate establishment questionable, but
because of a diversion cut across the top of the study area
to divert flow away from the filter, no major damage resulted.
On May 31, 1979, when the vegetation reached a height of
7.62 centimeters, an additional 10 pounds (5 Kilograms) of
actual nitrogen was applied to the site. This is a common
practice in the management of turfgrass and, with the increased
seeding rate, the additional nitrogen was needed to speed
the vegetative establishment.
The first problem in establishment became evident in
early June. Covering the seed with the disc increased
germination, but resulted in some rows of dense vegetation
throughout the area. The fact that these rows ran up and
down the slope increased the nature of the problem. It was
evident that the area had to be reseeded to establish a
filter medium of uniform density. After.scarifying the
surface, tall fescue seed was broadcast and 30 pounds (14
Kilograms) of nitrogen applied. -After seeding and fertilization,
1200 pounds (544 Kilograms) of lime was applied to the acid
area at the top of the study site. Following the second
seeding, a mowing management routine was imposed to reduce
competition with the new seedlings. This work corrected the
problem, resulting in vegetation of the high quality shown
in Figure 30,
Extreme drought conditions in the area necessitated in-
stallation of an irrigation system to relieve the water
stress. On August 17, 1979, Savoy Coal Company's water
truck was made available to transport water from a nearby
sediment pond to the filter area. Plastic (PVC) pipe was
attached to the truck with a dam-lock coupling and the water
pumped 192.9 meters down the outslope, through Rainbird 40
sprinklers, and onto the area. Approximately 1.27 centimeters
of water was applied. The irrigation relieved the stress
and ensured vegetative success.
74
-------
£
(
f.
'? 1F31WZ&&
- fi&-'.'''-\>*3°>J.~
i .:.-*•' ":•.. '• ;
C '•! ••'-,'.::
HI^-%
"., t.v.i, "•'"•>
•^s ;^"*
-f&:': -':: '••-
^M ' ""•£'.
.V*-*,i*r ' •' '"^.
':«•
;•
^-
Hr-'^'i
-,»i-33a
IS v •,•:•*•' v ;--6-^t^"iTi
i:;l^^ST>
•I- ' • r . . '" . . ' *,.: «.>^> '^ •, ~
. •• >• • - '"., * -t --•••• V ^,*il. .J5 ^3»-i- v C*/*^ ^*'«
'^.•'i'f>±'^m^^fj?ijf:.->
v •'^f'm:~M^^mm*$.
,.- •" .•-. :;$.-* 5 " -' - -.^"(f v* 'tr T'^S'*'«*»-'^t '
•: - -• > •-.-• ,.-. - •• ivs-i •-.*«-• •-i>-»i.a>.~.»?»-^-<. •
6-*v*:. - .
-ftek
I
0)
4J
«W
nt
to
(Q
m
J-l
Cn
•c
c
m
Q>
s
o
0)
0)
fa
o
n
Q)
75
-------
Vegetative Maintenance
During the course of the monitoring period, the degree
of maintenace required was governed by vegetative growth.
The only major nutrient problem encountered was a nitrogen
deficiency. The characteristic yellowing of the vegetation
gave good indication that nitrogen was insufficient. This
deficiency was corrected by broadcasting nitrogen, on two
separate occasions, at a rate of 50 pounds (23 Kilograms)
per acre. The acid conditions that- had existed at the top
of the filter due to acid inflow were corrected with the
previously discussed application of lime. To ensure erect
growth of the vegetation, a mowing plan was established.
Under the stress of high flow conditions, the vegetation
would lay over and create a grass waterway effect. Three
mowings of the filter were sufficient to keep the grass in
an erect growth habit.
Vegetative Performance
Overall, performance of the vegetation during the
project time frame was excellent. The key to such success
is to plan as thoroughly as possible before going into the
field. As previously noted, the two major problems encountered
were soil and runoff acidity during establishment and nitrogen
deficiencies after vegetative maturity. Other than these
two problems, little difficulty was experienced relative to
vegetative performance.
Deposition of fine-grained sediments in the filter area
were concentrated in the upper areas, although some deposition
did occur throughout the filter. During one storm, which
exceeded a 100 year event, rocks 7.62 centimeters in diameter
were washed 12 meters into the filter area. Even under
these extremely high flow conditions, the vegetation was not
uprooted.
Erosion within the filter area was a real concern
throughout the course of the project. This concern was
alleviated with the installation of a flow divider network
to spread the flow across the top of the filter to prevent
channelization. Without this flow divider, the concentrated
flow on the inlet area would have caused severe damage. The
following section will discuss both the details of the
design and installation of flow dividers and level spreaders.
WATER MONITORING SYSTEM INSTALLATION
When the established vegetation reached a density suf-
ficient to resist erosion and support the stresses of vege-
tative filtration, the monitoring system was installed.
76
-------
Installation was accomplished through a cooperative effort
between Hittman Associates, Inc. and the University of
Kentucky, College of Agricultural Engineering.
Section four presented the preliminary design considerations
for the filter area and the monitoring system. This section
applies the design criteria, described in section four, to
the field study site to create a complete, functional system.
The monitoring system installation process was broken down
into several distinct steps, including:
o preliminary site alterations,
o location of footers for flume implacements,
o installation of flume implacements,
o installation of flow divider and level spreader network,
o placement of equipment housings,
o monitoring equipment installation, and
o system calibration
Preliminary Site Alterations
Before any work could begin, the positions of the
inflow and outflow flume supports had to be marked in the
study area. The positions were marked by measuring the
halfway -point in the area with a metal tape. The positions
for the level spreaders would also be utilized to achieve an
equal distribution of flow across the filter area.
Once preliminary measurements were completed, the
actual site alterations began. A backhoe was utilized to
dig the trenches for the footers previously marked. The
depth of each hole was checked to ensure that the flumes
would be at the correct height once they were installed.
The final site alteration involved cutting diversions with
the bulldozer on the contour to divert the water into the
inlet flume. These diversions were cut to a .46 meter depth
and an 2.44 meter width (Figure 31). In addition to the
diversions, overflow channels were cut to divert water
around the filter area during large storm events. These
overflow channels were designed to divert runoff greater
than a ten-year event and divert them into a straw bale
dike.
Placement of Footers for Flume Implacements
In order to avoid problems with future maintenance, the
flume implacements were set on a 30.5 centimeter concrete
footer. Concrete delivered to the site proximity in a
commercial concrete truck was transported to the site in a
backhoe shovel. The footers for each of the flume implacements
were poured to the same level position by alignment with
77
-------
^mi^s-awe** ',V.--i-2.\
;k-:..;$# ^
V .Sr i?*
• :<.<.f,«^ri>vv^---v-x<«
•:• ^:/^^Mt%" ^:*JV;W;j *,;; i
V.Y ^v»rM-i'5..^ XI,.,' '
-------
previously surveyed stakes set in the bottom of the trenches
(Figure 32).
Installation of the Flume Implacements
After allowing the concrete footers to dry and set up
the concrete block, flume supports were constructed for both
inlet and outlet monitoring stations. Detailed plans for
these supports are shown in the third section, under water
monitoring system design. Concrete block were used to
construct two support walls on both ends of the study area.
During the construction of the supports, the H-flumes were
positioned and mounted as shown in Figure 33.
Installation of Flow Divider and Level Spreader Network
The specialized flow divider was installed directly
below the inlet flume spreader box (Figure 34). The function
of the spreader box was to spread the flow before the water
dropped into the flow divider mechanism. Both the spreader
box and flow divider were anchored in place by 1.27 centimeter
angle iron set in the subsoil.
A level spreader was installed approximately three
meters below the metal flow divider. The spreader extended
the full 20.73 meter width of the filter. Initially, the
3.05 meter lengths of wooden spreaders were connected to-
gether with truss-plates to form one large section. One
side of the board was coated with a tar sealant and four (4)
mil. plastic sheet was attached with roofing nails (Figure
35). The entire section was then placed in the trench and
the plastic pulled downslope to expose the 3.05 meters of
bare soil. ^
This bare soil area is where the 2x4 inch (5 x 10
centimeter) flow dividers were to be anchored. The
base plates of the flow divider were laid out radiating from
the central metal flow divider and soil was filled in be-
tween the dividers (Figure 36). The plastic connected to
the level spreader was then pulled up the slope to cover the
plates. The second set of treated flow dividers were measured
and nailed to the lower plate, leaving the plastic in between.
Metal strips were used to connect the wooden dividers with
the metal flow divider. The entire flow divider, level
spreader network was anchored to the subsoil using steel
rebar stakes, 46 centimeters in length. The completed flow
divider network is shown in Figure 37.
A plastic lined collection system was installed at the
downstream end of the filter to convey water from the filter
to the outlet flume (Figure 38). A level spreader, similar
to that used with the flow divider network, was used to
79
-------
Figure 32. Leveling concrete for flume supports
80
-------
K* •-•;«^--
(0
0)
H
K
O
c
•H
C
O
-H
•P
•H
to
O
A)
ro
CO
Q)
81
-------
0)
•O
•H
-H
'O
r
rH
(0
-P
Q)
C
0)
I
u
It
rH
Pi
0)
n
3
DI
•H
82
-------
;__igfl.--*^^fc1 •» jj -~ .*•••—i- •> • ••** * ~ - •• w "V.
^S^S^f^
Figure 35. Attachment of 4 mil. plastic to level spreader
83
-------
to
a
•P
-H
84
-------
••>!•«•' .•.*•»•TTt--""K.-V- •*.r:?^_ -ir"
i^i
-.W %v'*vr,u > >&:T-~ * • •/•
w^&^M^:^' '^&r ,u ? ?. .:-;-*'sf*:?-/-w;;r-.^-^. • • ^ •
^14? -Sii,- 1 %'^^.fe--^%-;.:- • *- '^e*1 ,*'. If:- ,•'-'—"-aarv •,- =->v,-,^,i. ,, .-&. ~'f~-'«•• — -•««•"
^i^-Sl^^Q3&fej;i-/'- •.'•-..-?Se, ;'|J^ '';.-.r-.i^t T'.v?*-v>-p:"•''.-,-- V :' ' i';
^j^S^^fl^'i" ':'-:W:- '-:||^|:,''y% "/*-'s?^,;^. :5. :: .j;
•••• rfiF/f/f, .-.:--£-=^^>.?-^f<^T
r»l Ji '? >••- -^g;-:---Uv^-^-.»:;
iif^ fij fer"--
?"'4^^ W
?^r/K:^^^^f^;.
' WC/ M ^^f^^l1^^^
^is" wb Sli^^^?^^^??'"'
j^ "V. "^/"lf" 'v'fc^ . .^ ;^^S^3?S*-^ *£'
ft • . (^ -( '» ! *!» • * *^,' ' --.-•„!*•' • " ,_. ._J«2iJ'-^*'SI<*r^_-*imi>',''':*^'- :..•
f .v ^:^,:^,.;, :;;^^^^|:
.,, "".',:.f^:";--:,-,« •.•••ii" J-;.l: '''r5%'^*1^;*^
i^^- fSr^^'^-1 it'^^'
*.;M^ ^^"sS^,^,-^^
0)
c
I
•H
•o
E
•o
(1)
4J
0)
iH
Ou
§
u
-------
;
^fiSiiS3W~==r^r:
'ii-n . •'^'«.?'<£'J?>''"r-T. ^.-*- •'j»*""Mii .«,,,.^ _
-i*^'^ "f&f «*£,•-•*.••.'.-.- —•'««
i«S;-3"^i">'*ty*J-'*l" ~i- —•*-.
*r-v ^Pft^^^^^^BSi;
s^^^t%;#^:'isi:'itr
Figure 38. Plastic diversion leading to lower flume
86
-------
anchor the plastic lining at the downstream edge of the
filter. The outlet level spreader and plastic diversion
were constructed to funnel the flow into the outlet flume
(Figure 38).
Placement of Equipment Housing
The equipment housings installed to protect the instru-
mentation from weather and theft consisted of corrugated
steel pipe modified to contain the monitoring system equip-
ment. These enclosures were set in place by anchoring the
bases in the subsoil and compacting soil around the pipe.
Monitoring Equipment Installation
The following discussions of the equipment installations
are applicable to both the inlet and outlet monitoring
stations.
The Isco automatic samplers were installed in small
culvert housings. From the sampler, the collection tube
extended to a point directly below the flume, where a special
collection device was installed. This device is shown in
detail in the design drawings in Section 4.
To maintain an accurate record of sample collection, an
event marker was installed on the Stevens water stage recorder.
The event marker was actuated by an electrical connection
between the sampler and the Stevens recorder and placed a
mark on the chart paper when the sampler was activated.
„ The activation of the samplers was controlled by electric
water level sensors mounted in stilling wells installed at
the side of the H-flumes. When the stage reached a pre-
determined level in the flume and stilling well, the water
level sensors activated the samplers. The samplers continued
to collect samples throughout the storm event until the
water level again dropped below the sensors.
The Stevens water stage recorder was also activated by
the changing water levels in the stilling well attached to
the flume. A tube attached to the stilling well governed
the level of a float in an additional well in the base of
the equipment housing. The Stevens recorder operated with a
standard float mechanism and spring-wound chart drive.
A Weathermeasure 30 centimeter raingauge was set in
concrete at the base of the filter area. The completed
inlet and outlet monitoring stations installed on the filter
site are shown in Figure 39.
87
-------
I v«jf;j?••j$v*,t*iF'*i&---*?.
|fe •;\v--»>:.'>i».-\; i&'?,'*.••{•':•. 14'• *
™ . -1 -"-j« Jf ^ '"'Sr 5- > 3r •
* '•:?;•£•.: <%A. ;. '>iv''«' -' : "•
^ >£-w, .!.^^^*" **' ^-'^ ."V^'J
.-r. ^ - ' ^V.-f • V.?T ^ ' "^* -V,^ '^ -."^<»
;.*% ^".^M-St ^*fe^>*f'^> -*^-Vi-r '* *f-\. .
to
c
o
to
I
+J
O
iH
4J
3
O
•O
4J
0)
•o
0)
+>
0)
«H
t
u
a)
88
-------
System Calibration
In order for the system to function properly, the time
of inflow and outflow sampling had to be coordinated. The
delay time of inflow into the filter and the outflow was
very important for accurate sample collection. The sampler
activators for inflow were set higher than those for outflow,
so that the inflow sampler would not run out of sample
bottles before the outflow sampler in the event of long
duration rainfall.
The water stage recorders were set to zero by priming
the connection hoses with water and adjusting to zero.
Initial testing of the inflow and outflow stations was
accomplished by simulation of a storm event. .According to
plan,, the automatic samplers were then programmed to sample
every ten minutes during a storm event.
TIME AND MATERIAL REQUIREMENTS OF FILTER INSTALLATION
Table 11 lists the approximate number of labor hours
required to install the vegetative filter used during the
demonstration project.
TABLE 11. APPROXIMATE NUMBER OF MANUAL AND EQUIPMENT HOURS
REQUIRED FOR FILTER INSTALLATION
Activity
Topsoil Removal
Grading
Topsoil Replacement
Soil Preparation
Fertilization
Seeding/Mulching
Follow-up Maintenance
Manual
Hours
4
18
4
4
4
10
4
Equipment Total
Hours Hours
2— / 6
6i/ 24
2-L/ 6
2— / 5
4
10
4
TOTALS 48 11 59
I/ Small bulldozer
2/ Small Tractor and Disc
89
-------
TABLE 12. TYPE AND AMOUNT OF MATERIALS
REQUIRED FOR FILTER INSTALLATION
Material Type " Amount/Applied
Lime 3,700 Ibs. (1678 Kg.)
Nitrogen 30 Ibs. (15 Kg.)
Phosphorus . 55 Ibs. (25 Kg.)
Potassium 34 Ibs. (15 Kg.)
Seed ,„ 70 Ibs. (32 Kg.)
Mulch 3 tons/acre (6804 Kg./hectare)
These figures were applied to create a level filter area,
41.0 by 21.0 meters on a 17 percent slope.
Table 12 lists the type and amount of materials used. The
hours shown in Table 11 reflect only the.time required for filter
installation, and do not reflect the time required_for related
activities such as engineering design, surveying, installing
the monitoring network, constructing the access road, etc. Al-
though the figures in Table 11 and 12 may provide a general in-
dication of time and material costs for a vegetative filter, they
apply specifically to the study site and should not be used for
cost estimates on vegetative filters of varying sizes and/or
uses.
90
-------
SECTION 6
SEDIMENT AND WATER QUALITY MONITORING
SAMPLING PROCEDURES
The original project proposal called for the installation
of one H-flume and water stage recorder at the inlet to the
filter to measure runoff. In addition, bi-weekly grab samples
were to be used to measure sediment concentrations. These
sampling procedures would not yield enough data to validate
several available computer models pertaining to the sediment
filtration capacity of vegetation. Since researchers at the
University of Kentucky were interested in further evaluation
of their models, an informal agreement was reached whereby
automatic samplers available to the University could be used
at the inlet and exit of the filter to measure concentrations
and flow rates. In conjunction with periodic hand sampling,
it was hoped that adequate data could be collected to validate
the University of Kentucky computer model as well as evaluate
the effectiveness of the vegetative filter.
As discussed in the description of the water monitoring
system in Section 5, two automatic pumping samplers were
used, along with .two H-f-lumes to monitoring inflow and
outflow water rates and sediment concentrations. In order
to obtain a complete data set for any storm, both samplers
had to activate and run properly and both water stage recorders
had to operate and ink properly. This did not happen in all
of the storms over the one-year monitoring period. Siltation
of the H-flumes was also a problem, particularly in the
inlet flume. This siltation caused a shift in the calibration
curves, therefore, the flow rates calculated from water
stage data are subject to some error. For future experimen-
tation, the use of a drop box weir or a sloping H-flume
would be recommended. For the reasons mentioned, hand
sampling was also accomplished for some storm events.
Hand sampling consisted of the use of project personnel
at the inlet flume and the outlet. Upon collection of each
sample, a reading was taken from a point gauge mounted in
the flume (Figure 40). Sampling continued through the
duration of the storm event.
91
-------
•H
m
•o
S3
si
e
-H
0)
IQ
o
IM
O
SC
c
•H
•d
0)
i
0)
D>
s
4J
c
-H
O
04
o
«*
0)
M
s
01
92
-------
LABORATORY ANALYSES
The University of Kentucky, College of Agricultural En-
gineering, performed all analyses of the water and sediment
samples collected during the course of the demonstration
project. This entailed testing for total sediment concentrations
of inflow samples, outflow samples, and performing periodic
size distribution measurements of the trapped sediment.
Results of the monitoring and testing program are depicted
graphically in the following section of the report. The
actual methods for testing are shown in Appendix C.
All soil testing for fertility required in vegetation
establishment was performed by the University of Kentucky
Soil Lab. Soil samples from the filter study area and spoil
samples from the drainage area were tested for pH, phosphorus •
and potassium.
Quality control on the analysis of ^the water and sediment
samples was achieved by sending duplicate samples to the
Hittman Laboratory in Columbia, Maryland. No problems with
the laboratory results surfaced over the course of the
project.
93 .. 1
-------
SECTION 7
DESCRIPTION OF RESULTS
"*.
BACKGROUND AND CALCULATIONS
The filter was constructed in the late spring of 1979.
The flow was diverted away from the filter area while vegetation
was being established. Flow was diverted into the filter in
the late fall of 1979 and monitoring equipment installed.
During the months of January and February, the monitoring
equipment was removed to prevent freeze damage. During that
period, storms were hand-sampled whenever possible. Monitoring
equipment was reinstalled in the spring of 1980, and monitoring
continued until November, 1980. During the mentoring period,
several storms of sufficient magnitude occurred to cause
runoff and flow through the filter. One storm, with a
duration equal to two hours, had a return period of over 100
years. The data from these storms are presented in the
following sections.
Calculations utilized in the storm data evaluations are
discussed below:
o Trapping efficiency, Tr, can be estimated by:
Tr = M-M
.
where M^ and Mo are the total mass of sediment in the
inflow and outflow.
o The mass of inflow on outflow can be predicted from:
n
M± =
where Ci-i , qij , and At j are the sediment concentrations ,
flow rate, and time interval of the jth element of in-.
flow. M0 can be calculated from:
n
= z coj 90j
94
-------
where coj and <3oj are the concentration and flow rate
of the jth element of outflow. For some of the data
sets, one set of the flow rates was either missing or
unreliable, due to excessive sedimentation or failure
of the water stage -recorders to properly ink.
For those cases where either inflow or outflow rates were
missing, a very conservative estimate of trapping ef-
ficiency can be obtained from:
TT = f•—C
lr *- ^-
Ci
Where C^ is the average inflow concentration CQ is
the average outflow concentration. This equation is
typically conservative, since infiltration causes the
outflow volume to be less than the inflow volume, unless
the rainfall rate exceeds the infiltration rate.
The study data from the individual storm events are
presented in the following sections.
STORM EVENT DESPCRIPTIONS
Storm of January 11, 1980 - An unusually warm period in
January brought rains and runoff at the filter site. All
equipment and records had been removed for the winter but the
flow was hand sampled for a thirty minute period for sediment
concentration only. No flow rates were measured. The results
are plotted in Figure 41.
The storm did not generate massive amounts of sediment; the
peak, concentration was only 6,750 mg/1. Outflow concentrations
were typically 500 to 600 mg/1, with the exception of one sample
that was 1600 mg/1. Based on equation 7-4, ninety-two percent
of the sediment was trapped in the filter area, indicating that
the filter was performing well.
Storm of April 3, 1980 - On April 3, 1980, a complete data
set was taken by hand sampling the inflows and outflows and by
measuring flow depths in the H-flumes with point gauges. The
results are shown in Figures 42 and 43. The size distributions
of the sediment in the inflow and outflow are shown in Figure 44
The peak inflow and outflow rates were 0.11 and 0.03 cfs
(3.11 and .85 I/sec) respectively, indicating a significant
attenuation of the flow rate, due primarily to infiltration.
The infiltration volume was 54 percent of the inflow volume, as
shown in Table 13.
95
-------
BX^OOO
10,000
1,000
r
i
100
K>
1
-!!
I
t
\ /
INFLOW
OUTFLOW
1
1
1
1
O
Figure 41.
10 CO 10 40
TIME AFTER START OF RUNOFF (»««.)
Inflow and outflow sediment concentrations
for storm on January 11, 1980
96
-------
B0400
10,000
ft
f \
v» \ —
MFLOW
OUTFLOW
•—e*
1
1
1
1
Figure 42,
to to 10 «o » «o
TIME AFTER *TART OF RUNOFF (win.)
Inflow and outflow sediment concentrations
for storm on April 3, 1980
97
-------
INFLOW
OUTFLOW
.09
10 20 30
TIME AFTER START OF
Figure 43. Inflow and outflow hydrograph
for storm on April 3, 1980
-------
AM013* *• M3MJ %
99
-------
The peak and average inflow sediment concentrations were
71,000 mg/1 and 21,850 mg/1 respectively. The peak and aver-
age outflow concentrations were 240 and 98 mg/1, yielding
a trapping efficiency greater than ninety-nine percent. Again,
in this storm, the vegetation did an excellent job of filtration.
The size distribution of the inflow sediment (on an
aggregate basis), as shown in Figure 44, indicates that
99 percent of the particles were silt size or larger. This
coarse size distribution explains the excellent trapping per-
formance.
Observations were made of the flow patterns during the
storm event. The flow was highly non-uniform, with flow
fingering occurring at several locations across the filter.
The fingering was a result of the longitudinal planting
orientation of the filter up and down the slope. If the
planting orientation and resulting roughness elements had
been perpendicular to the slope, the flows would have been
slower and the trapping efficiency should have been better.
Storm of April 11, 1980 - Runoff from the storm of
April 11, 1980, was automatically monitored. The flow
lasted for 105 minutes. Flow rates are shown in Figure 46.
The peak inflow and outflow rates were 0.170 and 0.013 cfs.
(4.81 and .36 I/sec) respectively, making the storm a small
event. The volume of infiltration accounted for 79 percent
of the inflow volume.
The inflow and outflow sediment concentrations are
shown in Figure 45. Peak and average inflow concentrations
are 6380 and 4800 mg/1. Peak and average outflow concentrations
are 240 and 160 mg/1. The percent trapped is in excess of
99 percent. Again, the vegetation did an excellent job of
filtering the sediment.
Storm of May 17, 1980 - Runoff from the storm of May 17
was monitored with automatic equipment. The flow lasted for
120 minutes, but flow through the filter apparently lasted
only 30 minutes or less. Data plots are given in Figures 47
and 48. Peak inflow and outflow rates are 0.038 and 0.012
cfs. (1.08 and .35 I/sec). As a result of infiltration, eighty-
five percent of the runoff did not appear as outflow.
Peak and average inflow sediment concentrations were
30,000 and 15,120 mg/1. Peak and average outflow concen-
trations are 4600 and 2967 mg/1. The trapping efficiency of
the vegetation was 97.1 percent, below that recorded in
earlier tests, but still highly acceptable.
100
-------
00,000
10
Figure 45.
120
•0*°*°
TIME AFTER START OF RUNOFF (mln.)
Inflow and outflow sediment concentrations
for storm on April 11, 1980
101
MX)
-------
K>
1.0
.1
JOI
MFUDW
OUTFLOW
IOO
80
10 w
o
0
ao co to ito
TWE AFTER START OF INFLOW W*)
Figure 46. Inflow and outflow hydrograph
for storm on April 11, 1980
102
-------
10,000
^ tpoo
***
mt
o
£
fc
K>0
MFLOW
OUTFLOW
I
I
I
I
I
Figure 47.
BO «0 90 MO 180
TWE AFTER START OF HUMOFF(»*O_
Inflow and outflow sediment concentrations
for storm on May 17, 1980 f
103
-------
10
I
i
o
TI*E AFTER «TART OF
Figure
and outflow hydrograph
104
-------
Storm of May 24, 1980 - The storm of May 24, 1980, was
monitored by automatic equipment. The ink in the water
stage recorder on the inflow H-flume smeared badly, making
it impossible to determine an inflow hydrograph. Since
outflow rates were very low, it may be assumed that inflow
rates were also low.
The inflow and outflow concentrations are shown in
Figure 49. Peak and average inflow concentrations were
67,000 and 37,402 mg/1. Peak and average effluent concentrations
were 911 and 522 mg/1. The trap efficiency, based on inflow
and outflow concentrations was 98.3 percent. Again, the
vegetation adequately performed the task of filtering the
sediment.
Storm of July 10, 1980 - The runoff from the July 10
storm was sampled by automatic equipment. Inflow and'outflow
hydrographs are shown in Figure 51. Peak inflow and outflow
rates were 0.070 and 0.047 cfs. (1.98 and 1.33 I/sec). In-
filtration accounted for 35 percent of the inflow.
Inflow and outflow concentrations are shown in Figure
50. Peak and average inflow concentrations were 3788 and
1420 mg/1. The trap efficiency, although flows were low,
fell to 97.1 percent. The vegetation as of July 10, continued
to do an excellent job of filtering sediment from the flow.
Storm of August 29, 1980 - The storm of August 29 was
sampled by automatic equipment. The ink pens for the water
stage recorders did not work properly, therefore only the
concentrations can be calculated from the data. The con-_
centrations are shown in Figure 52 and size distribution in
Figure 53. The trapping efficiency of the vegetation was
only 78.3 percent during this test, which was the lowest
value for the filter to that .date. By this time, however,
considerable deposition had occurred throughout the filter.
The aggregate size distribution of the eroded sediment
was such that 99 percent of the sediment was silt size or
larger, indicating the filter continued to provide good
trapping.
Storm of September 3, 1980 - The storm of September 3,
1980, was monitored by automatic equipment. According to
the rainfall charts, over 9.0 cm of rain fell. This repre-
sents an intensity which has a return period of over 100
years (Haan and Barfield, 1978). An inspection of the site
after the storm indicated that massive amounts of sediment
had washed in during the storm, plugging the stilling wells
and partially plugging the inlet to the pumping sampler.
Sediment had also deposited in the middle of the filter with
numerous boulders, four inches (10.2 cm) in diameter, washed
105
-------
100,000
IQpOO-
IB
O
i-
ttO 840
TIME AFTER tTART OF RUNOFF (mm.)
Inlow and outflow sediment concentrations
for storm on May 24, 1980
106
-------
OO/XX)
X
o
K
I-
X
111
o
X
o
o
IOO
1O
&•"
INFLOW
OUTFLOW
I
I
I
I
Figure 50.
tS 10 45 «0
TIME AFTER START OF RUNOFF (win.)
Inflow and outflow sediment concentrations
for storm on July 10, 1980
T6
107
-------
MFLOW
OUTFUDW
~JO 9O ttO »0
TIME AFTER START OF *UNOFF(»I«.)
Figure 52. Inflow and outflow sediment concentrations
for storm on August 29, 1980
109
-------
I ' I ' I ' I ' I ' I ' I ' I
,1.1.1.1.1.1
o
CO
cr\
rH
CM
•P
CD
J3
tn
O
•P
to
M
O
Q)
e
-------
as far as 40 feet (12.2m) into the plot.
Approximately 20 percent of the filter was completely
inundated with sediment and large amounts were deposited in
the vegetation.
These amounts of sediment would indicate a large concen-
tration in the inflow, although a plot of the data as shown
in Figure 54 does not substantiate this conclusion. Con-
centrations for this storm were lower than for most of the
other small events. The aggregate size distribution is
found in Figure 55.
STEADY STATE TESTS
Because of the difficulty in coordinating a trip to the
site to coincide with runoff, it was decided to conduct at
least one series of tests with an artificial water supply.
By directing water over bare soils, sediment could be eroded
and the sediment-laden water directed into the filter. A
coal company water truck was made available for water supply
and a test conducted on November 18, 1980. Prior to "dredging"
sediment, clear water was directed onto the plots in order
to satisfy some of the initial infiltration capacity. Water
was directed onto the plots at approximately 0.14 cfs. (3.96
1/min.) for a 20 minute duration. At this point, the water
truck had to be refilled.
Flow through was observed after four truckloads of
clear water had been directed into the filter. At this
time, sediment-laden flow was directed onto the plots and a
total of three steady-state tests were conducted.
The results of the steady-state tests are shown in
Figures 56 through 59. In each test, there was a large
infiltration component, as evidenced in the difference in
inflow and outflow hydrographs in Figure 59. This infil-
tration component is helpful in trapping the fines in the
flow.
The inflow and outflow sediment concentration, along
with percent trapped, are shown in Figure 56 through 58.
The trapping percentages are quite high, considering that
much of the filter had been inundated with sediment in the
100+ year storm that occurred over the site on September 3,
1980. Much of the trapped sediment was loose and readily
available for erosion in subsequent storms.
SUMMARY
In summary, the vegetation (fescue) worked well as a
filter for most of the year. The storms evaluated are
111
-------
INFLOW
OUTFLOW
ICO
TIME AFTER
Figure 54. Inflow and outflow sediment concentrations
for storm of September 3, 1980
112
-------
CM
113
-------
TtST-1
INFLOW
OUTFLOW
0,000!
-------
taopoo
Tt3T-t
1
INFLOW
OUTFLOW
opoo
Figure 57.
10 tO
TIKE AFTER START OF TEST (•!«.)
Inflow and outflow sediment concentrations
for steady-state test 2
115
-------
loopoo
TE«T-»
IMFLOW
OUTFLOW
«poo r
Figure 58,
to to so
TIME AFTER START OF TEST
Inflow and outflow sediment concentrations
for steady-state test 3
•0
116
-------
117
-------
summarized in Table 13. The 100+ year storm which occurred in
September 3, 1980, caused significant deposition in the
filter. Subsequent tests showed a decrease in trapping
efficiency. The trapping efficiency in storms early in the
monitoring period were very high, (95-100 percent), but
toward the end of the period trapping efficiencies had
decreased approximately 75 percent.
The size distribution of the eroded sediment was such
that most of the aggregates and primacy particles were silt
size or larger.
118
-------
SC
¥
I'
fc'
VI'
I
W.
O
E-t
to.
D
CO
w;
.3 !
pf\
H
•[TBjUTea
fet
BSEW
T/&JI
I/Bui
•ouoo
auntQA
•oas/i
-oas/i
B^EH
CM \D CO
CM CM CM
oo CM ro o
at en at
S m
o o *r »n
p- V£>
ro* in* in
§ §
rH
O
1
S S I S ! I
te in u>
| S S S
ts oo m
S I S 1 i
' 3
1 ill
O T -O
-to n \£>
»H CO V
CM
ID O
S o
i 2 S 2 i 2 i i ill
CM cn
O *H
g s
w
S rt
-------
SECTION 8
COMPUTER MODEL VALIDATION
INTRODUCTION
One of the desired objectives of the project was further
experimental evaluation of the Kentucky Grass Filter Model
described in Section 3 . Data from the storms in which only
automatic samplers were used proved insufficient for model
evaluation, so it was decided to use only the hand sampled
data sets. The data selected came from the April 3, 1980
data set and the steady-state data sets. The parameter
evaluated was the ability of the model to predict trapping
efficiency.
As used in this section, trapping efficiency is calculated
by:
Where Ci,j and co,j refer to the inflow and outflow concentration
of the j , the elements qij and qoj refer to the inflow and outflow
rate of the j . The element Tr j is the trapping efficiency
the the j element.
In order to make predictions, it is necessary to have
the inputs required for the model. The inputs not directly
measured as a part of the project were taken from Hayes et
al (1981) . A summary of the data inputs used is given in
Table 14.
RESULTS
April 3, 1980 Storm - Predicted and observed trapping ef-
ficiencies for the April 3 data set are shown in Figure 60.
The agreement is excellent. At this time, the filter was es-
sentially free of sediment deposited in the vegetation, there-
fore, the soil surface tended to serve as an absorbing barrier,
trapping all particles which struck it.
120
-------
115
too
S
5 »o
TRAP EFFICIENCY
VS.
TIME
CTORM OM 4-»-tO
85 ESTIMATED
—- OBSERVED
80
5 10 15 20 » 30
TSME (win.)
Figure 60. Sediment trapping efficiency vs time for
storm on April 3, 1980
121
-------
TABLE 14. INPUTS NECESSARY FOR RUNNING
THE KENTUCKY GRASS FILTER MODEL
Filter Length = 41 meters (measured)
Filter Width = 21 meters
Filter Slope = 17 percent (measured)
Spacing = 1.43 - 1.70 cm I/
Grass Diameter = .128 cm _'
Manning's n - 0.161
Inflow rate (specific for each test)
Inflow sediment concentration (specific for each test)
Inflow size distribution (specific for each test)
I/ From Hayes et al (1981)
Steady-State Tests
Predicted and observed trapping efficiencies are shown in
Figure 61 through 63. The agreement is poor. Most of the lack
of agreement can be attributed to the massive amounts of sediment
deposited in the filter during the September 3 storm. By visual
inspection, one could observe that most of the deposition occurred
over the center sections of the filter and less amounts in the
outer edges. This unequal deposition caused flow patterns
to be irregular and unpredictable. If the deposition had
been uniform across the slope, it could have been accounted for
in the model by adjusting the slope length.
One would not normally expect the sediment to deposit in
the uneven manner described in the previous paragraph. In
the storm leading up to September 3, deposition was uniform
across the slopes. However, in the storm of September 3,
flows were so large that the H-flume acted as a jet, causing
the flow to "gush" over the flow dividers out onto the filter.
From there, the flow and deposition spread radically in all
directions.
It should be pointed out that the potential exists for
making the predictions fit the observed data by optimizing
the parameters. This was not done, since it was believed that
the optimization served no useful purpose.
122
-------
I
«-
*
90
40
30
TRAP EFFICIENCY
VS.
TIME
TEST-1 t-30-80
ESTIMATED
OBSERVED
Figure 61.
Sediment trapping efficiency vs time
for test 1
123
-------
100.-
80
TO
IL
tt-
4f
1C
50
40
A.
TRAP EFFICIENCY
VS.
TIME
TEST-* 9-50-80
ESTIMATED
OBSERVED
IO
15
25
TIME (miit)
Figure 62.
Sediment trapping effiiency vs time
for test 2
124
-------
loop
o
t
TRAP EFFICIENCY
VS.
TIME
TCtT-3 9-30-80
ESTIMATED
OBSERVED
TIME (win.)
Figure 63. Sediment trapping efficiency vs time
£O3T tGSt «3
125
-------
Conclusions from Model Evaluation
In conclusion, the Kentucky Grass Filter Model did an ex-
cellent job of predicting the trapping efficiency in the storm
data. For the steady-state tests, the agreement was fair to
poor. The probable reason for this lack of agreement was the
poor condition of the filter resulting from the large amounts of
deposition which occurred during a previous extreme event.
One can conclude from these tests, and previous studies
cited in Chapter 3 that the Kentucky Grass Filter Model may be
used to adequately predict trapping efficiency for filters
that are in good condition.
126
-------
REFERENCES
1. Barfield, B.J., E.W. Tollner and J.C. Hayes, Hydraulics of
Erect Vegetal Filters Used for Sediment Control. ASAE
Paper No. 79-2025, American Society of Agricultural En-
gineers, St. Jospeh, Michigan, 1979a, 13 pp.
2. Barfield, B.J., J. Hayes, and R.I. Barnhisel. The Use of
Grass Filters for Sediment Control in Strip Mine Drain-
age, Volume II. Predictions based on the theoretical studies.
IMMR 39-RRR4-78, Institute for Mining and Minerals Re-
search, University of Kentucky, Lexington, KY, 1978.
3. Bingham, S.C., M.R. Overcash, and P.W. Westerman. Effec-
tiveness of Grass Buffer Zones in Eliminating Pollutants
in Runoff from Waste Application Sites. Paper No. 78-2571,
American Society of Agricultural Engineers, St. Joseph,
MI., 1978.
4. Dallaire, G. Controlling Erosion and Sedimentation at
Construction Sites. Civil Engineering 46(10): 73-77,
1976.
5. Doyle, R.C., G.C. Stanton, and D.C. Wolf. Effectiveness
of Forest and Grass Buffer Strips in Improving the Water
Quality of Manure Polluted Runoff. Paper No. 77-2501,
American Society of Agricultural Engineering, St. Joseph,
MI., 1977.
6. Environmental Protection Agency. Erosion and Sediment Con-
trol, Surface Mining in the Eastern U.S., Volume I. Plan-
ning. EPA-625/3-76-006, U.S. Environmental Protection
Agency, Washington, D.C. 1976.
7. Fenzl, R.N., and J.R. Davis. Hydraulic Resistance Rela-
tionships for Surface Flow in Vegetated Channels. Trans.
A.S.A.E. 7(1): 46-51, 55, 1964.
8. Gourlay, M.R. Flow Retardance in Vegetated Channels. Jour-
nal of the Irrigation and Drainage Division, A.S.C.E. 96 (IR3)
351-357, 1970.
127
-------
9. Graf, W.H. Hydraulics of Sediment Transport. McGraw-
Hill, New York, NY, 1971.
10. Hayes, J.C., B.J. Barfield, and R.I. Barnhisel. Evalua-
tion of Grass Characterizatics Related to Sediment Filtra-
tion. ASAE Paper No. 78-2513, American Society of Agricul-
tural Engineers, St. Joseph, Michigan, 1978, 21 pp.
11. Hayes, J.C., B.J. Barfield, and R.I. Barnhisel. Fil-
tration of Sediment by Simulated Vegetation - II. Un-
steady Flow with Non-Homogeneous Sediment. Transactions
ASAE 22(5): 1063-1067, 1978a.
12. Hayes, J.C., B.J. Barfield and R.I. Barnhisel. Evaluation
of Vegetal Filtration for Reducing Sediment in Surface
Mine Runoff. Proceedings of Symposium on Surface Mining
Hydrology, Sedimentology and Reclamation, University of Ken-
tucky, Lexington, KY., 1979b.
13. Hayes, J.C., B.J. Barfield and R.I. Barnhisel. The Use of
Grass Filters for Sediment Control in Strip Mine Drainage
Vol. III. Laboratory and Field Evaluations on Real Grasses.
Institute for Mining and Minerals Research, University of
Kentucky, Lexington, KY., 1981, (In Press).
14. Haan, C.T. and B.J. Barfield. Hydrology and Sedimentology
of Surface Mined Lands. Office of Continuing Education
and Extension, College of Engineering, University of Ken-
tucky, Lexington, KY., 1978.
15. Kao, D.T.Y., B.J. Barfield, and A.E. Lyons, Jr. On-Site
Sediment Filtration Using Grass Strips. National Sym-
posium of Urban Hydrology and Sediment Control, Univer-
sity of Kentucky, Lexington, KY., 1975, 73-82.
16. Khaleel, R., G.R. Foster, K.R. Reddy, M.R. Overcash, and
P.W. Westerman. Paper No. 79-2003, American Society of
Agricultural Engineering, St. Joseph, MI., 1979.
17. Kouwan, N., T.E. Unny, and H.M. Hill. Flow Retardance
in Vegetated Channels. Journal of the Irrigation and
Drainage Division, A.S.C.E. 95 (IR2): 329-342, 1969.
18. Li, Run-Ming and Hsieh W. Shen. Effect of Tall Vegetation
on Flow and Sediment. Journal of the Hydraulics Division,
HY5: 793:814, 1973.
19. Meyer, L.D. Overview of the Urban Erosion and Sediment
Process. Proceedings National Symposium on Urban Rainfall
and Runoff and Sediment Control. UK BU 106, College of
Engineering, University of Kentucky, Lexington, KY., 1974,
pp. 15-23.
128
-------
20. Meyer, L.D. and W.C. Harman (1979). Multiple Intensity
Rainfall Simulation for Erosion Research in Row Sideslopes.
Transactions ASAE, 33:100-103, 1979.
21. Neibling, W.H. and E.E. Alberts. Composition and Yield
of Soil Particles Transported through Sod Strips. ASAE
Paper No. 79-2065, American Society of Agricultural Engi-
neers, St. Joseph, MI., 1979.
22. Neibling, W.H. and G.R. Foster. Estimating Deposition and
Sediment Yield from Overland Flow Processes. Proceedings
International Symposium on Urban Hydrology, Hydraulics,
and Sediment Control, UK BU 114, College of Engineering,
University of Kentucky, Lexington, KY., 1977.
23. Norman, D.A. Design Criteria for Grass Filter Areas,
Paper No. 78-2573, American Society of Agricultural Engi-
neering, St. Jospeh, MI., 1978.
24. Ohlander, C.A. Defining the Sediment Trapping Character-
istics of a Vegetative Buffer - Special Case: Road
Erosion. Proceedings of the Third Federal Inter-Agency
Sedimentation Conference, 2/77 - 2/81, 1975.
25. Ree, W.O. Hydraulic Characteristics of Vegetation for
Vegetated Waterways. Agricultural Engineering 39(4):
184 - 189, 1949.
26. Schwab, G.O., R.K. Fevert, T.W. Edminster and K.K. Barnes.
Soil and Water Conservation Engineering. John Wiley and
Sons, New York, 1966.
27. Soil Conservation Service. Urban Hydrology. Technical Re-
lease 55, USDA, Soil Conservation Service, Washington, D.C.
1973.
28. Thompson, G.T. and J.A. Roberson. A Theory of Flow Resis-
tance for Vegetated Channels. Trans. A.S.A.E. 19(2): 288-
293, 1975.
29. Tollerner, E.W., B.J. Barfield, C.T. Haan and T.Y. Kao.
Suspended Sediment Filtration Capacity of Rigid Vegetation.
Transactions ASAE 19(4): 676-682, 1976.
30. Tollner, E.W., B.J. Barfield, C. Vachirokornwatana, and C.T.
Haan, Sediment Deposition Patterns in Simulated Grass
filters. Trans. A.S.A.E. 20(5): 940-944, 1977.
31. Tollner, E.W., J.C. Hayes and B.J. Barfield. The Use of
Grass Filters for Sediment Control in Strip Mine Drainage-
Vol. I. Theoretical Studies on Artificial Media. IMMR35-
RRR2-78, Institute for Mining and Minerals Research, Univer-
sity of Kentucky, Lexington, Kentucky, 1978.
129
-------
32. Vanderhom, D.H. Design of Vegetative Filters for Feedlot
Runoff Treatment in Humid Areas. Paper No. 78-2570, Amer-
ican Society of Agricultural Engineering, St. Jospeh, MI.,
1978.
33. Ward, A.D., C.T. Haan, and B.J. Barfield. The Design of
Sediment Basins. Transactions of ASAE 23, 1980.
34. Williams, R.G., B.J. Barfield, and R.I. Barnhisel. De-
sign Report for Development of a Rainfall Simulator. De-
partment of Agricultural Engineering, University of Ken-
tucky, Lexington, KY., 1978.
35. Wilson, L.G. Sediment Removal from Flood Water by Grass
Filtration. Trans. A.S.A.E. 10(1): 35-37, 1967.
36. Young, R.A. Effectiveness of Nonstructural Feedlot Dis-
charge Control Practices. Paper No. 78-2572, American
Society of Agricultural Engineering, St. Joseph, MI.,
1978.
37. Heath, M.E., Metcalf, D.S., and Barnes, R.F. Forages -
The Source of Grassland Agriculture. The Iowa State Uni-
versity Press, 1973.
130
^^
-------
METRIC CONVERSIONS
For use in the Appendices as well as the body of the report,
TO CONVERT
Acres
Cubic Feet
Feet
Miles (Statute)
Inches
Pounds
Tons
TO
Hectares
Cubic Meters
Meters
Kilometers
Centimeters
Kilograms
Kilograms
MULTIPLY BY
4.047 x 10"1
2.832 x 10~2
2.048 x 10"1
1.609
2.540
4.536 x 10"1
907.2
131
-------
APPENDIX A
COMPUTATION PROCEDURES FOR RUNOFF HYDROGRAPHS
A. CALCULATION OF RUNOFF HYDROGRAPH
1. Area (See plot plan)
TABLE Al. AREA
Segment
AB
BC
CD
TOTAL
2 2
Acreage ft mi
.3291
.3129
.4709
1.1129 48477 .0017
2. Curve
Number
Select: Hydro-logic Soil Group C (Haan and Barfield,
1978, Table 2.19,
Soils with slow infiltration Pg. 67)
rate when wet
Curve No. = 88
(Haan and Barfield,
1978, Table 2.20,
Pg. 63)
132
-------
3. Runoff Volume
TABLE A2. RUNOFF VOLUME
Return Period
(Yr.)
24 Hour Precip.
(in.)
(Haan & Barfield,
1978, Appendix 2A-
2, pg. 107)
Runoff
(in.)
(Haan & Barfield,
Fig. 2.26, pg. 70)
2
5
10
3.1
3.9
4.5
1.8
2.7
3.2
4. Time of Concentration
Average slope = 95 ft./279 ft. = 34%
Velocity
= 1.5 ft./sec.
Distance
Tc
Use
(Haan and Barfield,
1978, Figure 2.34,
pg. 32, Between
forest and trash
fallow conditions)
= 230 ft.
= 230 ft./I.5 ft./sec. = 2.55 min.
= .1 hr. minimum value
5. Hydrographs (Soil Conservation Service, 1975, pg. 5,7)
T =0.1 hr.
c
= 0.04 hr. = 0
133
-------
TABLE A3. HYDROGRAPHS
Time
11.0
11.5
11.7
11.8
11.9
12.0
12.1
12.2
12.3
12.4
12.5
12.6
12.7
i
i 12.8
12.9
13.0
; 13.2
13.5
14.0i
15.0
Ordinate
(CSM/in)
24
51
299
991
746
477
233
152
132
121
111
85
74
70
68
65
52
38
39
29
Area
(mi2)
2 yr.
1.8 in.
.0017 0.07
0.16
0.91
3.03
2.28
1.46
0.71
0.47
0.40
0.37
0.34
" 0.26
0.23
11 0.21
0.21
0.20
ti
ii
0.12
0.09
Hydrograj
Storm
5 yr.
2.7 in.
0.11
0.23
1.37
4.55
3.42
2.19
1.07
0.70
0.61
0.56
0.51
0.39
0.34
0.32
0.31
0.30
0.18
0.13
>h
10 yr.
3.2 in.
0.13
0.28
1.63
5.39
4.06
2.59
1.27
0.33
0.72
0.66
0.66
0.46
0.40
0.38
0.37
0.35
0.21
0.16
134
-------
B. COMPUTATION OF SEDIMENT YIELD TO FILTER
1. Site Data
TABLE Bl. SITE
Segment Slope Slope Ground Drainage
Length % Cover Area
X ....... % (Ac)
(ft) (estimate) (From topo
Map)
Width
(ft)
AB
BC
CD
• DE
65
80
155
206
23.4
57.4
26.8
15.8
40
40
70
70
.3291
.3129
.4709
190
180
165
TOTAL
1.1129AC
48,477 ft:
2. Return Period Values of Precipitation and Erosion Index R
TABLE B2. RETURN PERIOD VALUES OF PRECIPITATION AND EROSION INDEX R
Return Period
Years
24 Hr. Precip. (in.)
(Haan & Barfield,
1978, pg. 123)
Single Storm Rainfall
Factor -R-
(Haan & Barfield, 1978,
pg. 188)
1
2
5
10
3.
—
3.1
3.9
4.5
Size Distribution of Eroded Sediment
28
46
80
114
135
-------
4. Determination of Factors in Deposition Modified USLE
(a).
(b).
(c).
(d).
(e).
(f).
(g).
r = 14 x 10~8
Assume: Runoff potential
for bare fallow
soil, mod - high
R = 46 for 2 yr. storm
30 for 5 yr. storm
114 for 10 yr. storm
T— "3 "7 v i n {• ~ sec.
• j. / x j.u i £. - ;
(Haan and Bar field,
1978, Fig. 5.19,
pg. 210)
(Using data for
Lexington, Kentucky)
(Using DSQ - 0.1 from
Fig. 1A-1)
A B =65' X __ (cumulative) = 65' (From Table
X BC - 80> X AC
X CD - 155' X AD
A DE = 2°6' AAE
S1 for segment AB, sin 6 =
BC, "
CD, "
DE, "
B = 1.7
C C Segment
ro ro ^
.56 AB
.56 BC
.50 CD
.50 DE
= 145'
= 300'
= 506'
.228
.498
.259
.156
(Using particle dia-
meter of 0.1 mm &
SG = 2.65)
(Haan and Barfield,
1978, Table 5.8,
pg. 209)
1)
136
-------
(h). Use Medium Runoff Rates (Haan and Barfield,
_ . . . ^ n 1978, Fig. 5.20b)
Particle D = 0.1 mm
SG = 2.00
For Segment AB & BC Velocity Ratio =0.85
Assume treatment is poor
grass with rough surface
depression
For Segment CD & DE Velocity Ratio =0.50
Assume treatment is good
grass
TABLE B3. SLOPE SEGMENT DATA
Segment Velocity % Slope
Ratio
AB 0.85 23.4
BC 0.85 57.4
CD 0.50 26.8
DE 0.50 15.8
Relative
Transport
Capacity
(Table
5.9)
.55 x
.65 x
.06 x
.06 x
Ratio of Relative C
Transport to
Transport of
Medium Sized
Aggregates,
SG = 2.65
(Fig. 5. 20)
.96 = .53
.98 = .64
.62 = .04
.65 = .04
(i). K. From Size Distribution Curve Figure 1A-1
% Sand ( >.lmm)
% Very Fine Sand
() .1-. 062mm)
% Silt (.062-. 004mm)
% Clay ( <. 004mm)
50
16
29
5
100
% Organic Matter 0%
% Silt = % VFS 45%
K = 0.38
(Haan and Barfield, 1978,
Fig. 5.90, pg. 189)
137
-------
(j)
Segment
AB
BC
CD
DE
CPj - Cs x Cr X Co
% Cover
40
40
Disced, Raked,
Bedded Weed
Cover WC Soil
Condition Poor
.19
.19
Appreciable Brush
(2m fall ht)
Canopy Cover
Cover - W
70
70
.06
.06
(Haan and Barfield,
1978, Table 5.5,
Factor for Mechanic-
ally Prepared Wood-
land Sites)
Surface C
Roughness s
Factor
.65
.65
.11
.11
(Haan and Barfield,
1978, Table 5.6,
Factor for Permanent
Pasture)
(k).
Segment
AB
BC
CD
DE
_ .043
—
Summary
.11
.11
.06
.06
30x.
_J
430x.
J
6.613
x = tan 6 = % slope/100
Segment x
(Haan and Barfield, 1978,
egn. 5.9, p. 193)
AB
BC
CD
DE
.234
.574
.268
.158
S.
2.8
11.4
4.0
2.0
138
-------
(1). m exponent dependent on slope
for slopes _> 5%
m = 0.5 (Haan and Barfield,
1978, pg. 191)
(m). Tabulation of Factors
TABLE B4. TABULATION OF FACTORS
139
^
-------
CO
0
u
fe
fe
o
gv
0
H
EH
^
j-j
m
<
9
^
oa
w
! •]
CQ
EH
e
•n
CO
•f-1
04
o
«"""
EH
U
0
$_|
O
n
«.
CO
g 4J
••< 3 H-i
"O *-*
•^
(1)
W CM
EH 1 -P
,Q "4-1
rH
»— '
.
CO
f*°* i [
rH ><
fV^ *
W
in M
to
CM V<
-^
edois
"I
o
oo
•
CM
,_,
i—]
•
O
oo
00
o
m
m
•
0
\0
in
•
o
,— i
oo
CM
CM
•
in
VD
m
o
i — |
X
.
n
^*
r- 1
l-l
O
00
VD
**
OO
o
"x
2
CSS
^Jl ^^ <^^
* * •
tH ^3< CM
rH
rH VD VD
rH 0 O
• • •
O O O
•* rj< «3<
VD O O
• • •
O O 0
VD O O
in in in
• • •
O 0 0
E = E
OO O*i VD
cr\ in in
••a1 CM rH
• • •
in o VD
.•«* o o
rH 00 in
E E E
SEE
ESS
E E S
S E S
U Q W
PQ C.} C3
140
-------
^
5. Calculation of Transport Capacity and Material Detached
(a) . Transport Capacity
T = r
c
R . T . X
cro . CT
(Haan and Barfield,
1978, Eq. 5.18)
(b)
T (Ib/ft-width)
Segment
AB
BC
CD
DE
Detached Soil
2 yr.
Storm
242
5489
431
5 yr .
Storm
421
9546
750
10 yr
Storm
600
13603
1069
D. =
D
D.
AB
BC
CD
6 . Routing
0.0459 . R . K.
(Ib/ft-width)
15
151
78
of Sediment
. CP. .
26
263
136
o nm+1
• sj • (xj
38
375
193
Particles > 0.062 mm will be transported as Bed load - 65%
Particles < 0.062 mm will be
The lesser of T and D.
c D
exiting each segment.
transported as Suspended
Load - 35%
is taken as the bedload sediment
141
-------
TABLE B5. ROUTING OF SEDIMENT
Detached Material (D . )
(Ibs/ft-width) 3
Transport
Capacity
(Ibs/ft-width)
Segment Total Bed I/ Suspended I/
Load Load Load
2
AB
BC
CD
5
AB
BC
CD
10
AB
BC
CD
YEAR STORM
15 I/ I/
151
73
YEAR STORM
26
263
136
YEAR STORM
33
375
193
242
5489
431
421
9546
750
600
13603
1069
Material
Transmitted
from Segment
(Ibs/ft)
15
166
244
26
289
425
38
413
606
I/ Since T > D. , it is not necessary to divide
D
j into bed load and suspended load.
142
-------
7. Sediment Load to Filter Area (See topo map Figure 7.)
TABLE B6. SEDIMENT LOAD TO FILTER AREA
Segment Average Sediment Total Sediment
Width Production Production
(ft.) (Ib/ft.-width) (Ib)
2 YEAR STORM
AB
BC
CD
190
180
165
15
166
244
2850
29330
40260
5 YEAR STORM
AB 190 26 4940
BC 180 289 52020
CD 165 425 70125
10 YEAR STORM
AB
BC
CD
190
180
165
38
413
606
7220
74340
99990
143
-------
C. DEVELOPMENT OF POLLUTOGRAPH
1. Average Concentration
Return Sediment Runoff Area
Period Yield (in.) (ft2) (Ib7ft ) /??? Concentra-
(yr.) (Ibs.) tion
(Ib/lb)
2 40,260 1.8 48477 62.4 453,745 0.089
5 70,125 2.7 " " 680,617 0.103
10 99,990 3.2 - " 806,657 0.124
2. Peak Concentration (Ward, 1978)
Peak concentration can be predicted by:
Cp = 0.063577 (Yd/A)0'74 (gp)~°'394 (tgt)0.177
where:
C = peak inflow concentration (mg/l)
Yd = sediment delivery for storm (tons)
A = area (ac)
g = peak runoff rate (cfs)
t = storm duration (hrs.)
Return
Period
(yr.)
2
5
10
(Ton)
20
35
50
A
(ac)
1.1129
"
n
(cfs)
3.03
4.55
5.39
(hr.)
24
24
24
C
(g/g)
0.61
0.79
0.96
144
-------
APPENDIX B
PLOTS OF SENSITIVITY ANALYSIS OF KENTUCKY
GRASS FILTER MODEL
145
-------
40 r-
80
£
>c
I
>c
*w^
25
20
c>
ac
«c
K
10
IO-YEAR STORM
5-YEAR STORM
2-YEAR STORM
\
I
I
20
40 60
FILTER LENGTH(M)
00
100
Figure B-l. Sensitivity analysis for length
146
-------
10-YEAR STORM
6-YEAR STORM
2-YEAR STORM
10
eo
WIDTH (M)
90
Figure B-2. Sensitivity analysis for width
147
-------
tO-YEAR STORM
5- YEAR STORM
— 2-YEAR STORM
25
Figure B-3. Sensitivity analysis for slope
148
-------
50
25
I
X ~
. -~" ZU
O
o
10
IO-YEAR STORM
5-YEAR STORM
2-YEAR STORM
I
I
LO
1.4 IS
«RASS SPACING (cm)
2JO
Figure B-4. Sensitivity analysis for spacing
149
-------
35
SO
£5
I
I
* EO
10
JOB
IO-YEAR STORM
6- YEAR STORM
2-YEAR STORM
I
I
09 JO .11
GRASS DIAMETER (en)
J3
Figure B-5. Sensitivity analysis for diameter
150
-------
I I I
i!
8
I
•
I
•
I
•
I
•
I
IL
10
5
M
5
o
5
S3
CQ
tn
C
•<-{
G
C
(0
S
to
-H
10
(C
-H
•H
4J
•H
CO
C
0)
CO
vo
CQ
0)
NOI1VH1N30NOO
151
-------
APPENDIX C
ANALYTICAL PROCEDURES UTILIZED FOR PROJECT LABORATORY TESTING
All methods utilized during the course of the study for
laboratory testing are contained in the following descriptions.
Analyses performed on both water and soil are detailed in
the information presented in this appendix.
WATER AND SEDIMENT TESTING METHODS
Pipette Analyses - Particle Size Distribution
Apparatus:
1 - 2000 ml graduated cylinder
1 - 1000 ml graduated cylinder
1-25 ml pipette
Filter papers (that trap .3 microns & larger)
Vacuum filtering stand
1 - thermometer
Drying oven (set at 105°)
Theory:
That pipette analysis of particle size distribution
is based on Stoke's Law for fall velocity. This equation
gives the equivalent spherical diameter for a given
settling depth and time.
L = 1 g(d)2(SGp-SG£)
t 18 v
L = distance settled (cm)
t = time to settle (sec)
2
g = gravity (1000 cm/sec )
152
-------
d = equivalent spherical diameter (cm)
S.G.p = specific gravity of particles (assume 2.65)
S.G.f = specific gravity of fluid used to suspend
the particles (hour case water is used).
v = kinematic viscosity (temperature variable)
(cm2/sec)
Percent finer is calculated from
^T x 100
Ci
where:
p
T = concentration of 25 ml sample
C
i = initial concentration
Procedure:
Part I. Determination of sample concentration
1) A sample is thoroughly mixed by shaking the
bottle (2-3 min.).
2) Immediately; a 25 ml sample is pipetted from
the shaken sample.
3) The 25 ml are filtered through the filter paper.
The pipette is rinsed with distilled water into
the filter.
4) The filter paper (previously weighed) is dried
at 105°C for 20-30 min. or until completely dry,
5) Concentration is calculated from:
(Filter & Sediment - Filter _ SED(g)
25 ml ~ 25 ml
Part II. Pipette analysis
1) Weigh a full sample bottle.
2) A #270 sieve is placed in a funnel above a
graduated cylinder.
153
-------
3) The runoff sample is poured through the sieve.
Distilled water is used to rinse the sample
bottle.
4) Reweigh the sample bottle.
5) Wash the sieve residue into a previously weighed
beaker. Put in drying oven until water is driven
off. Weigh and record weight of beaker and sieve
solids.
6) Wash the funnel with distilled water into the
graduated cylinder. Fill the cylinder to a
convenient readable level with distilled water.
7) Mix contents of cylinder by inverting continuously
for one minute.
8) Set cylinder down (where it won't have to be
moved) and start a stop watch.
9) Sample with a 25 ml pipette at depths and times
that correspond to the particle size wanted.
10) Record the temperature of cylinder with distilled
water as a control.
Example:
Time Depth Size (dia.)
hr. min. sec
0:00:44 10 cm .048 mm
4:26 10 .019
1:11:00 10 .0048
6:56 10 .002
7:10 5 .0014
11) Filter the 25 ml through a previously weighed
filter paper, dry the filter paper, weight and
record.
12) Determine 25 ml sample concentration (C_) from:
CT = (Filter + Sed.) - Filter
25 ml
13) Determine percent finer from (C /C.) x 100 =
% finer T x
154
-------
SEDIMENT CONCENTRATION TESTING METHODS
Apparatus:
1 filter paper and filter pan per sample
1 beaker per sample
1 pipette
1 wash bottle
vacuum filter flask apparatus
drying oven
Procedure:
1) Filter paper is placed in filter pans and put
in drying oven for approximately 15 minutes.
2) Bottles containing samples are weighed.
3} Beakers are numbered and weighed.
4) Pans and filters are removed from drying oven,
numbered, and weighed.
5) Sample is allowed to settle until sediment is on
the bottom of the bottle and the water above
is relatively clear.
6) "Clear" water above sediment is removed with the
pipette and poured through vacuum filter flask
apparatus. Pipette is rinsed with distilled
water into the filter flask.
7) Sediment remaining in sample bottle is poured
into a beaker, and the bottle is rinsed with
distilled water. Washings are added to the
beaker of sediment.
8) Filters and beakers are placed in drying oven
until all water has evaporated. When dry,
they are weighed.
9) After the empty sample bottles have been allowed
to dry out some, they are weighed.
155
-------
10) Concentration is calculated from:
(Filter & Sediment) - Filter + (Beaker & Sediment) - beaker _ sed(g)
(Bottle & Sample) - bottle sample (g)
SOIL TESTING METHODS
Standard EPA methods were utilized for all soil testing:
o pH (paste pH method)
o buffer pH (titration against a standard solution)
o phosphorus (extractable P)
o potassium (extractable K)
Source: EPA Field and Laboratory Methods Applicable to
Overburdens and Minesoils, EPA-600/2-78-054,
March, 1978.
156
-------
GLOSSARY
acid soil: A soil with a pH below 7.1, but for practical
-•purposes a soil with a pH of 6.6.
annual: Plant that completes its life-cycle in one year
from seed.
cool-season: Grass species adapted to rapid growth during
the cool-moist periods of the year.
flow hydrograph: A flow rate plotted opposite a time
increment.
flow pollutograph: Concentration of pollutants versus time.
flow transport capacity: Mass of sediment which a flow
will carry.
Manning's roughness coefficient: Values assigned to dif-
ferent materials to denote friction.
mg/ls Abbreviation for milligrams per liter which is a
weight volume ratio commonly used in water quality
analysis. It expresses the weight in milligrams of
a substance occurring in one liter of liquid.
outslope: The slope formed on the outer edge of a spoil
disposal area.
perennial: Plant that persists year after year without
reseeding.
pH: The negative logarithm of the hydrogen-ion activity
which denotes the degree of acidity or a basicity
of a solution. Acidity increases with decreasing
values below 7 and basicity increases with increas-
ing values above 7.
point gauge: Device for manually measuring depth in
a flume or weir.
reclamation: The procedures by which a disturbed area
can be reworked to make it productive, useful, or
aesthetically pleasing.
157
-------
regrading: The movement of earth over a surface or de-
pression to change the shape of_the land surface.
sediment: Solid material settled from suspension in a
liquid medium.
storm return period: Probability of a given rainfall
to be equaled or exceeded during a given time
period.
warm-season: A grass species adapted to rapid growth
during the warmer parts of the year.
158
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
3. RECIPIENT'S ACCESSION-NO.
. TITLE AMD SUBTITLE
Use of a Vegetative Filter Zone to Control Fine-
Grained Sediments from Surface Mines
5. REPORT DATE
MARCH 1981
6. PERFORMING ORGANIZATION CODE
. AUTHOFUS)
Steve C. Albrecht
Billy J. Barfield
8. PERFORMING ORC
. PERFORMING ORGANIZATION NAME AND ADDRESS
Hittman Associates, Inc.
Lexington, KY 40504
10. PROGRAM ELEMENT NO.
CBBN1G
University of Kentucky
Lexington, KY 40506
11. CONTRACT/GRANT NO.
CS-805632
2. SPONSORING AGENCY NAME AND ADDRESS
Energy Pollution Control Division
Industrial Environmental Research Laboratory
Office of Research and Development
US EPA, Cincinnati, OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA 600/12
5. SUPPLEMENTARY NOTES
6. ABSTRACT :
This project was initiated with the specific purpose of conducting a field test on
vegetation as a viable sediment trapping medium. From the onset, the project was wholly
designed for a field evaluation under typical mining conditions. The filter area was
constructed directly below an abandoned surface mine bench, on typical soil types found
in mined areas of Eastern Kentucky. The outs lope located above the filter was the
primary area from which sediment-laden drainage was to be diverted to the inlet
monitoring station. Sediment-laden water samples were collected at the inlet flume for
comparison with samples collected at the outlet flume to permit evaluation of the
sediment removal capability of the vegetative filter.
Results of the monitoring efforts revealed that a dramatic reduction in sediment
load was achieved by vegetative filtration for particle sizes larger than clay. Based
on results of this study, it is concluded that vegetative filters are an effective
control for reducing the quantity of sediment transported in surface streams and rivers
from disturbed mined lands.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.!DENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Coal Mines
Soil Stabilization
Surface Mining
Water Quality
Sediment Control
Pollution Control
Eastern United States
Water Pollution Control
Kentucky
Pollution Abatement
Vegetative Filters
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
171
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 12220-1 (9-73)
159
------- |