Tennessee
Valley
Authority
United States
Environmental Protection
Agency
Research and Development
Office of Natural
Resources
TVA/ONR-79/11
Industrial Environmental Research EPA 600 7 79 036
Laboratory February 1979
Cincinnati OH 45268
Strip Mine
Drainage
Aquatic Impact
Interagency
Energy/Environment
R&D Program
Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-036
TVA/ONR-79/11
STRIP MINE DRAINAGE—
AQUATIC IMPACT ASSESSMENT
by
D. B. Cox, R. P. Betson, W. C. Barr, J. S. Grossman, and R. J. Ruane
Office of Natural Resources
Tennessee Valley Authority
Chattanooga, Tennessee 37401
Interagency Agreement EPA-IAG-D8-E721
Project No. E-AP 80 BDS
Program Element No. INE-625 B
Project Officers
Ron Hill, Don Mount, and Walt Sanders
U.S. Environmental Protection Agency
5555 Ridge Avenue
Cincinnati, Ohio 45268
Prepared for
OFFICE OF ENERGY, MINERALS, AND INDUSTRY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 30460
-------
DISCLAIMER
This report was prepared by the Tennessee Valley Authority and has
been reviewed by the Office of Energy, Minerals, and Industry, U.S.
Environmental Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the views and
policies of the Tennessee Valley Authority or the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
-------
FOREWORD
The future reliance on coal as a major energy source in the United
States will bring about a substantial increase in mining activity.
Because of the greater demand for coal and the stricter environmental
regulations governing mining activities, a procedure for predicting
the effects of proposed mining on aquatic communities is necessary.
The purpose of this investigation is to develop predictive models
for use by regulatory agencies in making planning decisions. The
models would be used in the design phase of mine planning to provide
regulatory agencies and planners with an easily accessible tool for
assessing the impact of policy decisions.
A description of the model components and a discussion of pre-
liminary data collected at area- and contour-mined sites in East
Tennessee for model calibration are presented in this report.
111
-------
ABSTRACT
This research program is aimed to demonstrate methodologies for
predicting the impact of strip mining on downstream biotic communities.
To accomplish this objective and provide data for model verifica-
tion, sampling programs were initiated at contour- and area-type mining
operations. These programs include streamflow and rainfall gaging at
both types of mines and surveys of fisheries, periphyton, and macrobenthos
at area-mined sites.
Preliminary findings of these field studies indicate that (1) drain-
ages from mined areas are alkaline rather than acid, and (2) calcium and
magnesium concentrations increase as a result of mining in almost every
instance. Furthermore, values for iron and sulfate increase in some areas,
but not in others, whereas values for trace metals are generally low in
all areas. The predominant fish in small Cumberland Plateau streams is
the creek chub (Semotilus atromaculatus), and its primary food source is
aquatic invertebrates, including midge larvae, springtails, and aquatic
mites.
Several model components have been developed, including a water
quality model for nonpoint sources, a continuous streamflow model, and
a storm hydrograph model. Other model components being developed or
evaluated include additional small-basin water quality models, water
quality and quantity routing models, a low-trophic-level stream biota
model, and a fisheries resource model.
This report was submitted by the Tennessee Valley Authority, Office
of Natural Resources, in partial fulfillment of Energy Accomplishment Plan
80 BDS under terms of Interagency Agreement EPA-IAG-D8-E721 with the U.S.
Environmental Protection Agency. Work was completed as of December 31, 1977
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables vii
Acknowledgments viii
1. Executive Summary 1
2. Introduction 3
Model design considerations 5
3. Study Watersheds 8
Area-mined sites 8
Contour-mined sites 11
4. Methods and Materials 16
Area-mined sites 16
Contour-mined sites 17
Summary 17
5. Project Studies 19
Water quality 19
Aquatic biota 39
Overburden analysis 42
Hydrological model studies 47
6. Model Components 62
Explanation of model components and linkages .... 62
Progress on model components 64
Bibliography 73
Appendixes
A. State of Tennessee strip mine regulations 77
B. Zoomacrobenthos collected in the vicinity of Jamestown,
Tennessee 81
C. Core logs for strata samples from project watersheds ... 85
-------
LIST OF FIGURES
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Sampling stations — Fentress County
Sampling stations — New River watershed
Alkalinity — Fentress County sites
pH — Fentress County sites
Acidity — Fentress County sites
Sulfate — Fentress County sites
Hardness- -Fentress County sites
Total iron — Fentress County sites
Total suspended solids — Fentress County sites
Lowe Branch loads and concentrations of iron, suspended
solids, and sulfate
Indian Fork loads and concentrations of iron, suspended
solids, and sulfate
Strata analyses — Fentress County sites
Strata analyses — New River Basin sites
Helton Branch at Greenwood, Kentucky — observed vs.
simulated flow
Cane Branch at Parkers Lake, Kentucky- -observed vs.
simulated flow ,
Results of TVA suspended sediment- storm hydrograph model .
Comparison of observed and predicted stream temperatures .
Results of TVA geomorphic model
Schematic diagram of model components
Generalized benthic module: Biological model, Jamestown
area streams
Fish module: Biological model, Jamestown area streams . .
TVA strip mine impact assessment model
Pagj?
9
13
22
23
24
26
28
29
31
33
34
45
46
50
51
56
58
60
63
69
70
71
-------
LIST OF TABLES
Number Page
1 Water Quality Analyses—Fentress County Sites 16
2 Summary of Site Information 18
3 Means and Ranges of Project Data from Fentress County
Sites
20
4 Means and Ranges of Project Data from New River Basin
Sites .......................... 21
5 Means and Ranges of Trace Metals Data from Project
Watersheds ....................... 3*>
6 Means and Ranges of Trace Metals from Points in the
Tennessee Valley .................... 37
7 Quantitative Monthly Macrobenthic Density Data
(March-October 1976)— Fentress County Sites ....... 40
8 Quantitative Macrobenthic Drift Data Collected Monthly
(May-October 1976) --Fentress County Sites ........ 41
9 Monthly Periphyton Data (April-October 1976) --Fentress
County Sites ...................... 42
10 Acid-Base Balance for Fentress County Strata ....... 48
11 Water Quality Simulation—TVA Nonpoint Source Water Quality
Model — Beaver Creek Watershed .............. 52
12 Test of TVA Background Nonpoint Source Water Quality Model
at Unmined Watersheds .................. 53
13 Tests of Universal Soil Loss Equation- -Upper Bear Creek
Watershed Data ..................... 55
14 Background Watersheds for Model Calibration Located in
Cumberland Plateau ................... 59
VII
-------
ACKNOWLEDGMENTS
Thanks are extended to EPA Project Officers Ron Hill, D. I. Mount,
and Walt Sanders. Appreciation is extended to J. E. Liner, J. R. Wright,
Jr., Harold E. Pratt, David B. Chastant, and Lynn B. Starnes for their
aid in the project.
This study was supported under Federal Interagency Agreement
Numbers EPA-IAG-D8-E721 and TV-41967A between the Tennessee Valley
Authority and the Environmental Protection Agency for energy-related
environmental research.
Vlll
-------
SECTION 1
EXECUTIVE SUMMARY
As demands for domestic energy increase and use of coal becomes even
more dominant as a part of U.S. energy policy, a procedure for predicting
the effects of proposed raining on the aquatic resource will be needed.
Development of such procedures is the objective of this investigation.
Field studies were initiated at nine small watersheds within the
Cumberland Plateau region of east-central Tennessee. Data from these
studies will be used for model development and verification. A few
generalizations can be drawn from these studies:
1. Acid mine drainage is the exception rather than the rule within
the Tennessee Cumberland Plateau. None of the mined watersheds
had acid drainages. In fact, pH values were somewhat higher
(by 0.5 to 1.0 unit) in mined watersheds than in unmined
watersheds nearby.
2. Alkalinity, calcium, and magnesium values were higher in mined
areas than in unmined areas. Leaching of these materials from
previously unweathered strata is the most plausible explanation
for an increase in pH at mined stations.
3. Values for sulfate and iron—an indication of acid information--
increase dramatically at some sites, but only slightly at
others, indicating that acid formation and subsequent neutrali-
zation are occurring in some areas, whereas little or no
formation is evident in others.
4. Trace metal concentrations were generally low when compared
with values encountered in many other mining situations. Con-
centrations did increase, however, at those areas in which
acid production was occurring.
5. Manganese concentrations increased with mining at all stations.
6. The most serious water quality problem encountered was erosion
and subsequent sedimentation.
7. Analysis of soil cores from the Fentress County sites indicated
that acid conditions will not predominate in drainage. This
conclusion is substantiated by the alkaline drainages and illus-
trates that overburden analysis can predict acid-alkaline
conditions before mining.
8. Natural streams in the area appeared to be dominated by the
creek chub (Semotilus atromaculatus), with smaller populations
of the fathead minnow (Pinephales promelas) and the blacknose
dace (Rhinichthys atratulus). Mining decreased both the
diversity and number of fish within the affected streams.
9. The primary fish food organisms, in decreasing order of impor-
tance, are midge larvae (Chironomidae), springtails (Collembola),
and aquatic mites (Arachnida).
-------
-2-
Model development has progressed concurrently with the field studies.
Much of the work has involved modification of existing models to account
for strip mine effects. Most of the modification has involved applica-
tion of the models to sites in southern Appalachia so that predicted
values can be compared with actual data, and model algorithms and
coefficients can be adjusted. Five tasks have been accomplished to date:
1. A water quality model for nonpoint sources under unmined con-
ditions that is based on hydrology and geology of the watershed
has been verified. Although quite simple, the model predicts
natural stream water quality so that changes to the water
quality as a result of mining may be assessed.
2. A continuous streamflow model, RUNOFF, which can be used inter-
actively, has been developed to predict instantaneous flows
when watershed characteristics and rainfall are known. The
model is now being modified after a largely successful
verification *run using project data.
3. A stream temperature model has been applied to four sites within
the Cumberland Plateau and verified.
4. A suspended solids model has been coupled with RUNOFF to predict
instantaneous solids concentrations. The model is a modification
of the Universal Soil Loss Equation. Verification has been
partly successful, but the model is being modified further.
5. A stream morphology model for use in routing flow and water
quality downstream has been developed and verified so that top
width, area, velocity, and depth can be simulated if stream
drainage area and slope are known.
-------
-3-
SECTION 2
INTRODUCTION
The environmental problems associated with mining coal are wide-
spread, serious, and often obvious. Much of the obvious degradation is
associated with old orphan mines that were left unreclaimed or poorly
reclaimed. In the last ten years, surface mining regulations have been
enacted in all the eastern coal-producing states which, along with more
stringent Environmental Protection Agency (EPA) regulations and Federal
laws, will ameliorate many of these problems. Coal mining will continue
to impact the environment, because any disturbance of the earth's surface
will alter the environment to some degree. Also, because of the increased
emphasis on coal in the Nation's energy program, there is still a poten-
tial for environmental degradation, despite more stringent legislation,
as a result of aggregated effects.
Objective assessment of the environmental impact of coal mining is
not simple. Most research done in the past has been directed toward a
study of the effects of past mining operations on water quality. Such
studies do not lead to methods for transferring the results to other
locations, for which the physical or geochemical conditions differ or
the mining or reclamation techniques vary. Those few studies that have
involved the modeling of mining effects usually required copious amounts
of data for model application, which severely limits their use. Also,
past studies have focused on only a few conventional water quality
parameters.
A problem arises from any generalized attempt to assess the environ-
mental impacts of coal mining: How does one equate the differences in
reported effects with variations in site characteristics and in mining
and reclamation techniques? In essence, no methodology exists for pre-
dicting cause-and-effect relationships. The topographic characteristics
of a.site to be mined and the chemical-physical characteristics of the
coalj overlying rock, and soil can cause considerable variation in the
effects of mining on the aquatic resource. These effects can be sub-
stantially modified by mining techniques, such as soil disposal and con-
tour vs. area, and by reclamation techniques, such as erosion and sedi-
mentation controls and revegetation alternatives. The effect on scale
must also be considered when evaluating the impact of mining on the
aquatic environment. Even with reclamation measures aimed at minimizing
the release of sediment and acidity, the quantity and quality of stream-
flow will change. Both the local streamflow regime and the water quality
immediately below any mine will be altered. Because downstream effects
diminish as a result of dilution from other areas, the impact of a
single mine with adequate reclamation of the aquatic environment tends
to be localized, often with the primary impact confined to a small
stream. However, when mining becomes extensive in a watershed, these
effects can accumulate and impact the biological productivity and use-
fulness of larger streams. No techniques are available for assessing
the cumulative effects on larger streams of mining extensive areas under
even controlled reclamation strategies.
-------
-4-
Mathematical models offer the best methodology for quantitatively
assessing environmental consequences that result from several factors
operating simultaneously. These models must consider the (1) physical
and geological characteristics of the mining site, (2) mining techniques
used, (3) reclamation techniques used, and (4) hydrology of the area.
The models must also be able to integrate the effects of multiple mining
sites at some point downstream from these operations. Finally, changes
in water quality and quantity must be translated into biological and
aesthetic impacts and, where possible, assigned costs. This applied
research project is designed to provide a composite model that can be used
when planning the development of surface mines. The model allows inves-
tigations of the potential effect of various mining reclamation techniques
(manipulative) or the (1) quality and quantity of local streams and (2)
indigenous aquatic biota. It also allows for prediction of cumulative
effects of pollutants from multiple mines with various "background" or
control scenarios.
The models are being designed for use by regulatory agencies in making
planning decisions. The models can be used to (1) investigate the suita-
bility of a site to be mined on the basis of geology and coal chemistry
involved, (2) evaluate the adequacy of reclamation measures being proposed
by a miner, (3) investigate the cumulative impacts of multisite mining
in a large watershed, or (4) assess the consequences to the aquatic
ecosystem that would result from changes in reclamation regulations or in
effluent limitation standards.
As a part of model development, information must be obtained not
only on the chemical and physical characteristics of mine drainage but
also on the effect of that drainage on the aquatic community. Several
types of information are necessary for evaluation:
1. Mining data, such as type of mining, total area disturbed,
approximate schedule of operations, number and type of sediment
control structures, depth of coal seam, length of contour cut,
and distance of the cut from the receiving stream.
2. Reclamation data, including the schedule for reclamation,
fertilization and liming rates, approximate slopes of reclaimed
areas, type of revegetation used, and success of revegetation
efforts.
3. Climatological information, such as hourly rainfall, storm
intensity and duration, and temperature.
4. Land cover, including land use, riparian land cover, and road
systems.
5. Watershed characteristics, such as topography, soils informa-
tion, drainage areas, any point sources of pollutants, and
drainage networks.
6. Overburden characteristics, including trace metal content and
the amount and strength of acid-forming and neutralization
materials.
-------
-5-
7. Stream-associated characteristics, including channel slope,
pool-to-riffle ratios, morphology vs. discharge relationships,
bed grain size, and dispersion coefficients.
8. Water quality, including both physical (pH, temperature, etc.)
and chemical (acidity, nutrients, minerals, and trace metals)
data.
9. Aquatic community characteristics, qualitative and quantita-
tive, of periphyton, benthos, and fish.
This type of information is essential to an understanding of cause-
and-effect relationships. Such data, however, are also needed for model
development; but in applications, the model will require only descriptive
information of the site.
The national goal of energy independence dictates that coal extrac-
tion will increase significantly in the near future, and because of the
far-reaching effects that are involved, policy decisions are becoming
more critical. Because conservation of natural resources is an agency
objective, the decision to construct a fossil-fueled generating plant
should include assessment of any potential mining costs to the environ-
ment and, consequently, to the social well-being of the people. Likewise,
a decision to enforce effluent limitations or stringent reclamation
standards by a regulatory agency should be based on the determination
that the environmental and social benefits to be derived are commensurate
with the economic costs involved. This kind of policy decision will
necessitate the use of models such as those being developed in this
project.
MODEL DESIGN CONSIDERATIONS
The model under development is being designed for easy use in
providing cause-and-effect relationships for planning applications.
These goals dictate several approaches that are incorporated into this
model.
Nonpoint Source Models
Nonpoint source models for application at ungaged, unmeasured loca-
tions will be developed. If the models are to be applied easily, lengthy
data-gathering to calibrate these models cannot be justified. Instead,
deterministic components must be used to the extent possible, with
regionalization of nondeterministic components. A deterministic model
component is one that defines the process so well that, when it is
applied, only site characteristics need be measured. When processes are
very complex, simplified models must often be used that imitate process
behavior. Typically, these models must be adjusted to data by using
optimization techniques to determine model parameters. If the parame-
ters determined by adjusting the model to many sets of data at different
locations are related to site characteristics in a second step, the
model is said to be regionalized, because the model parameters can be
predicted from the characteristics of ungaged sites (at least in the
general region in which the model was calibrated). By using this
-------
-6-
approach, data from many other locations on the Cumberland Plateau can
be used, thus minimizing the need for costly project data. In either
case, however, the result is a nonpoint source model component that can
be applied at any location on the basis of only the site characteristics.
First-Generation Model
First-generation model components are used to the extent possible
to simplify model applications. First-generation models were defined by
Betson (1975) as those that can be calibrated and used with published
data. Thus, application of first-generation, deterministic model compo-
nents would require only mapped data; calibration of a first-generation,
regionalized model would require only published data, and subsequent
application would require only mapped data. In other words, an objective
in developing this model is to minimize the need for gathering expensive
field information to apply the models. Although not an explicit goal of
this study, these models are being designed for use in an interactive
computer mode, provided suitable spatial land cover, soil, geology,
topography, and other information are stored on the computer. This
latter objective also requires minimization of the field data require-
ments for the model. (It also will necessitate interfacing these models
with models developed at other projects.)
Compartmentalization of the Model
The model will be compartmentalized to simplify modifications and
improvement. If first-generation models are to be used, then simplified
approaches will have to be taken because, as the complexity of a model
increases, the input information and data requirements generally increase.
A first-generation model is often adequate for planning applications for
which a model requiring much input information could not be economically
justified. However, if the model is compartmentalized, the model compon-
ents can be improved as necessary or when improved formulations are
developed. Thus, a model structure that can be adapted for other pur-
poses will be provided.
Linking of Mining and Stream Ecology
The model will link mining and stream ecology to allow investigation
of cause-and-effect relationships with a single model. Because changes
in water quality have impact only when they affect living organisms (man
included), a surface mine model must include provisions for considering
effects on aquatic biota. The component model compartment will be
interfaced in such a manner that, beginning with the mining and reclamation,
the effects will cascade through model components until the impact on
the biotic component is considered. Thus, rather than predicting only
physiochemical effects, a capability for evaluation based on the effects
on the entire ecosystem will be provided.
-------
-7-
Background Model Simulation Capability
A background model simulation capability will be provided. Any
impacts to streamflow quality or quantity or to the aquatic biota that
may result from mining must be quantified in relation to the background
or natural stream conditions. Impact assessment involves evaluating
the effects of changes that occur. Therefore, to measure change (impact),
the models must also be capable of predicting background conditions.
-------
-8-
SECTION 3
STUDY WATERSHEDS
Two study sites are located on the Cumberland Plateau area of
eastern Tennessee. Topography of the plateau varies from gently rolling
hills in the northwest to rugged mountains in the northeast. The stra-
tigraphy is Pennsylvanian in origin and is characterized by sandstones
and shales, with intermittent beds of coal. Overall, the sequence can
be divided into two principal components—an upper part, containing
thick beds of shale with thin, discontinuous sandstones; and a lower
section, consisting largely of thick conglomerates and sandstones,
separated by equal amounts of shale. In general, the upper sequence is
preserved only in the Cumberland Mountains, whereas the lower sandy
sequence caps the remaining flat-topped portion of the plateau (Johnson
1972).
The northern part of the plateau is drained by the Big South Fork
Cumberland River. The New River, North White Oak Creek, and Clear Fork
River are primary tributaries to the Big South Fork Cumberland River.
AREA-MINED SITES
The study sites chosen to reflect area-mined conditions are located
in Fentress County near Jamestown, Tennessee (Figure 1). The stratigraphy
reflects the lower sandy sequence of Pennsylvanian lithology found on
the plateau. Strata in the mined area consists of sandstone members of
the Rockcastle conglomerate overlying the Nemo coal seam. The mined
area is drained by Crooked Creek, a primary tributary to the Clear Fork
River.
Four water quality and three biological sampling stations are
located on Crooked Creek and its tributaries. Two water quality and
biological sampling stations are also located on streams draining
unmined areas. Rainfall and streamflow gages located at two mined sites
and one unmined site on Crooked Creek were installed early in July 1976.
Lynn Branch (Station LY 0.8)
Lynn Branch is a first-order stream (i.e., it has no tributaries)
located about 6.4 km ESE of Jamestown, Tennessee. The stream is a
tributary of Mill Creek, which drains into North White Oak Creek. The
drainage area is about 103 hectares (ha) at the sampling station, which
is located at river mile 0.8. The creek, 0.6 to 2.0 m wide, has sand,
gravel, and bedrock substrate, and its depth varies from 2 cm to 0.2 m.
Stream canopy is greater than 90 percent throughout the forested portion
(89 percent) of the watershed. Lynn Branch served as one of three
control stations.
Long Branch (Station LB 4.6)
A second control station is located on Long Branch about 2.5 km E
of U.S. 127 at Grimsley. Long Branch is a first-order tributary of
South Prong Clear Fork Creek, which joins Crooked Creek to form the
-------
-9-
STATIONS
I. CROOKED CREEK (CC 18.5)
2. CROOKED CREEK (CC 16.7)
UNNAMED TRIB. (UT 0.01)
CROOKED CREEK (CC 15.9)
5. LYNN BRANCH (LY 0.8)
6. LONG BRANCH (LB 4.6)
• FENTRESS COUNTY
A RAINFALL AND STREAMFLOW
GAGING STATION
O CORE DRILL LOCATIONS
10 kilometers
SCALE
Figure 1. Sampling stations--Fentress County.
-------
-10-
Clear Fork River. The drainage area is about 290 ha at the stream gage
and sampling station, which are located at river mile 4.6. The stream
is 0.9 to 2.1 m wide and 2.5 to 0.9 m deep at this point. The substrate
varies from areas of small gravel to bedrock to deep, decaying detritus.
Stream canopy was about 90 percent over the forested portion (89 percent)
of the watershed. The remainder was pastureland, with forest canopy of
about 80 percent.
Crooked Creek (Station CC 18.5)
A third control station is located at river mile 18.5 on Crooked
Creek about 2.4 km SE of Jamestown. The stream is 0.3 to 1 m wide and
0.15 to 0.6 m deep. The drainage area is about 180 ha and contains
about 31 percent forest land. Most (69 percent) of the watershed is
pastureland, with several small farm houses. A small part of the water-
shed drains Jamestown, Tennessee, and represents urban-residential land
use. The bottom substrate ranges from sand to decaying detritus.
Crooked Creek (Station CC 16.7)
A second station on Crooked Creek is located at river mile 16.7
about 2.9 km downstream from station CC 18.5 and represents an area that
is about 13 percent mined. The remaining land use is 31 percent forest
and 56 percent pastureland. Total drainage area at this station is
about 725 ha. The substrate is small gravel, bedrock, and sand, with
the percentage of sand cover increasing as a result of erosion from
strip-mined areas. The area sampled is rather homogeneous, 0.15 to
0.6 m deep and 3 to 4.5 m wide, with an undercut bank on both sides of
the creek.
Unnamed Tributary (Station UT 0.01)
About 50 m downstream from sampling station CC 16.7, a small tribu-
tary enters Crooked Creek. A sampling station, designated UT 0.01, was
established 0.01 mile upstream from the mouth of this tributary. The
drainage area is about 65 ha, about 43 percent of which is mined, 5 per-
cent forested, and 52 percent pastureland. The stream is 2.5 cm to
0.3 m deep and 0.3 to 1 m wide. The bottom substrate alternates between
bedrock and sand. Rainfall, streamflow, and water quality are measured
at this site.
Crooked Creek (Station CC 15.9)
A final sampling station is located about 1.3 km downstream from
station CC 16.7 on Crooked Creek at river mile 15.9. The drainage area
is about 950 ha, 13 percent of which is mined, 32 percent forested, and
55 percent pastureland. The creek in this locality is 1 to 6 m wide and
0.15 to 1 m deep; the substrate is mostly sand and silt, and a 200-yard
riffle exists below the Allardt Highway bridge. Above the highway
bridge are the remains of a silt dam, which was breached during the
winter of 1975. Since this occurred, the turbidity has increased
markedly at this station. The stream canopy is about 60 percent in the
vicinity of the sampling station. Rainfall, streamflow, water quality,
and biota were measured at this site.
-------
-11-
The study watersheds were first mined during the summer of 1975.
Mining was completed above station UT 0.01 about April 1976 and above
stations CC 16.7 and CC 15.9 during September 1976. Mining was con-
ducted using the standard open-pit method. Reclamation techniques
included topsoil segregation, return to approximate contour, mulching,
liming, and seeding. Several sediment traps and diversion ditches were
installed during active mining. Additional sediment control measures
were instituted after mining ceased. Sediment traps, generally of poor
design, consisted mostly of wooden barricades built across the stream
channel proper, which appears to be prevailing practice throughout much
of Tennessee. Because monthly water quality samples began in February
1976 at all six stations, no background information was collected in the
study watersheds.
CONTOUR-MINED SITES
Most of the contour mining in the State of Tennessee occurs in the
New River Basin in east-central Tennessee. The basin, about 989 km2 in
area, is located above a U.S. Geological Survey (USGS) stream gage at
New River and is contained wholly in Anderson, Campbell, Morgan, and
Scott Counties. The basin has been well mapped, including detailed
delineation of mined and unmined regions (Hollyday and Saver 1976).
The stratigraphy reflects the upper, shale-dominated sequences of
Cumberland lithology. The mined areas drain large vertical sections of
this lithology. The principal drainages are Smokey Creek, Brimstone
Creek, Paint Rock Creek, Buffalo Creek, Beech Fork, and Ligias Fork.
The intensity of the mining effort within the basin has moved many
groups interested in mining to develop various research or monitoring
programs. An interagency committee was established to eliminate dupli-
cation of efforts and interfacing agency programs, thus extending the
usefulness of the data. This group contains representatives from TVA;
USGS; the U.S. Army Corps of Engineers; the U.S. Department of Agricul-
ture, Soil Conservation Service; Oak Ridge National Laboratory; the
University of Tennessee; and the Tennessee State Planning Office,
Tennessee Division of Surface Mining, Tennessee Department of
Conservation, and Tennessee Wildlife Resources Agency.
The USGS operates all hydrological gaging stations in the basin.
TVA and all other agencies obtain rainfall and streamflow records through
the USGS, which also conducts a larger-scale program aimed at evaluating
the hydrological impact of mining in southern Appalachia. Of primary
interest are those studies dealing with storm-event sampling of sequential
downstream stations and sediment movement and storage.
The U.S. Army Corps of Engineers was directed by Public Law 93-251,
which created the Big South Fork National River and Recreation Area, to
formulate a comprehensive plan for the New River Basin for "programs to
enhance the environment and conserve and develop natural resources, and
to minimize siltation and acid mine drainage." The Corps retained a
consultant, L. Robert Kimball (1976), to assist in carrying out this
task. Kimball developed a 12-point program that includes an extensive
field reconnaissance survey, in which 300 water quality grab samples
-------
-12-
were collected to identify intensive sources of pollution. This part of
the program was completed, and a secondary phase of sampling was begun,
whereby a monthly grab sample is collected at 61 locations identified
during the reconnaissance survey. TVA will receive these data and has
made recommendations for trace metal analyses at several sites.
Most of the data produced for use in this project are collected as
part of a cooperative agreement with the University of Tennessee, whose
program is largely funded through the Tennessee State Department of
Public Health. TVA has agreed to fund rainfall and streamflow gaging at
three watersheds in the basin; in return, the University of Tennessee
has agreed to furnish TVA with their project data, which includes inves-
tigations of (1) the hydrological modifications resulting from watershed
disturbance by mining; (2) contaminant flows from mining spoil and their
dependence on spoil bank hydrology and mineralogy; (3) the behavior of
potentially toxic heavy metals in the streambed and suspended sediment
ultimately carried out of the basin; and (4) biological differences as
related to observational differences between mined and unmined basin
streams, heavy metal accumulations in stream biota, and ATPase ion pump
system in resident biota as an indicator of sublethal stresses on
organisms in disturbed watershed streams.
Those studies of greatest potential use to TVA include parts of
tasks 1, 2, and 3. Most of this work is carried out in six small water-
sheds within the basin (Figure 2), three of which represent contour-mining
operations and prevailing practice of reclamation; two areas represent
unmined conditions. Mining has recently begun on a third basin that was
originally selected to represent unmined conditions.
Topography of all watersheds is similar, with average slopes ranging
from 31.9 to 38.4 percent. Stream channels are nearly as precipitous,
with average channel slopes ranging from 7.4 to 8.4 percent for the
larger Indian Fork and Bowling Branch watersheds to 26 percent for Bills
Branch, the smallest watershed. Channels draining unmined areas consist
of alternating rapids with larger boulders to bedrock "waterslides" to
cobbles. Channels of mined watersheds exhibit sediment accumulations in
eddy areas, with heavy deposits at the mouth.
Lowe Branch and Bowling Branch
The two unmined basins, Lowe Branch and Bowling Branch, represent
drainage areas of about 240 and 760 ha, respectively. Lowe Branch is
located about 1.6 km S of Norma, Tennessee, and is tributary to the New
River. Bowling Branch is located 316 km SW of Smokey Junction and is
tributary to Smokey Creek. The watersheds are almost completely forested,
with limited logging operations in the Bowling Branch watershed. Lithology
of these watersheds ranges from the lower Red Oak Mountain formation to
the slatestone formation; these sequences are characterized by clayey to
sandy shales, light brown to dark gray in color with intercalated siltstone
and sandstone, and localized thin coal seams. The only coal reserve of
note is the Big Mary coal, 40 to 120 cm thick, which appears at an
elevation of about 2200 ft in both basins (Luther 1970) .
-------
-13-
SAMPLING STATIONS
I. INDIAN FORK
2. GREEN BRANCH
3. BILLS BRANCH
4.BOWLING BRANCH
5. ANDERSON BRANCH
6.LOWE BRANCH
0
•
RAINFALL AND STREAMFLOW
GAGING STATION
CORE DRILL LOCATIONS
SCOTT COUNTY
II kilometers
SCALE
Figure 2. Sampling stations--New River watershed.
-------
-14-
Indian Fork
The Indian Fork watershed represents an area of extensive mine
operations. The watershed represents an area of about 1119 ha, about 19
percent of which (210 ha) is mined. The sampling site is located about
0.5 km W of Braytown, Tennessee. Indian Fork is tributary to the New
River. The Big Mary, Walnut Mountain, and Pewee coal seams have been
mined on a larger scale for more than 20 years. Most of the coal has
been extracted by surface mining, but some underground mining has taken
place. The Big Mary seam (about 2250-ft elevation) is of marine origin,
as evidenced by a 7.6- to 10.1-cm zone of mudstone characterized by
marine fossils. The remaining overburden is primarily a homogeneous
mudstone, medium to dark gray. About 6 to 9 m above the coal, large
disc-shaped carbonate masses appear. These masses, ranging up to 1.2 m
in diameter, are distributed sporadically. About 21 m above the Big
Mary coal, the gray marine shale grades upward into interbedded silt-
stones, sandstones, and shales. The Walnut Mountain and Pewee coals
(about 2650- and 2600-ft elevation, respectively) are generally mined
simultaneously. The overburden consists of shales, sandstones, silt-
stones, and siderite nodules and beds. A carbonaceous layer, similar to
but thicker than that found above the Big Mary, is found immediately
above the Pewee coal (Carman 1975).
Reclamation of mined areas is varied. Strip mine regulations for
the State of Tennessee are reviewed in Appendix A. Because little or no
effort was made to stabilize mine spoils in the basin before the Tennessee
Strip Mine Law of 1967, many orphan mines exist. Also, mining of the
Big Mary seam was virtually completed before March 1972, when specific
regulation criteria were enacted in the form of amendments to the 1967
law. Active mining in the basin is limited mainly to the Walnut Mountain
and Pewee seams.
Bills Branch
Bills Branch, a second mined watershed, is tributary to Smokey
Creek, a primary tributary of the New River. The watershed is located
about 10 km SW of Smokey Junction, Tennessee. The Bills Branch watershed
represents an area of about 170 ha, of which 9.8 percent (16.7 ha) is
mined. Mining began in this watershed in late 1974 and was completed in
September 1975. The seam mined was the Big Mary coal. The area was
reclaimed according to the Tennessee Surface Mining Law of 1972.
Green Branch
The Green Branch watershed is separated from Bills Branch by an
extension of Bear Knob. Green Branch is also a tributary of Smokey
Creek. The drainage area, 357 ha, is about twice that of Bills Branch;
24 percent of the watershed is mined. Most of the mining has been in
the Big Mary and Pewee seams, but some has involved the Grassy Springs
and Jellico coals. Mining was initiated in the basin during 1972 and
completed in late 1975. About 60 percent of the reclamation was carried
out under 1972 regulations, and the rest complied with the 1974 law.
-------
-15-
Anderson Branch
The remaining small watershed, Anderson Branch, was originally
chosen as a control site;, however, mining of the Jellico seam began in
the spring of 1976 and will he completed during March 1977. The water-
shed divide is being used as a haul road for mining beyond the watershed.
Total disturbance is estimated to be 15.8 ha, or 7.5 percent of the 210-ha
drainage area. The Jellico coal adjoins the base of the Indian Bluff
formation. The major strata are shale, sandstones, and siltstones, the
most persistent and thickest sandstone member overlying the Jellico coal.
Anderson Branch, tributary to the New River, is located about 2.5 km S
of Norma, Tennessee.
-------
-16-
SECTION 4
MATERIALS AND METHODS
AREA-MINED SITES
Samples were analyzed for a wide variety of physical and chemical
properties (Table 1). Some experimentation necessary for developing
optimum analysis of dissolved, suspended, and total metals is reflected
in the inconsistency of early metals analyses. Analysis for pH, acidity,
and alkalinity and filtrations for dissolved metals were carried out in
the field whenever possible. However, in some instances samples had to
be transported to the laboratory. In these cases, the samples were iced
and transported immediately, and the analyses were carried out within
4 h of sample collection. Rainfall and streamflow gaging stations were
installed at stations UT 0.01, LB 4.6, and CC 15.9 during 1976. Auto-
matic water samplers were purchased and placed onsite at stations
UT 0.01 and CC 15.9 during the same year.
TABLE 1. WATER QUALITY ANALYSES—FENTRESS COUNTY SITES
Field parameters
Total minerals
Nutrients
PH
Alkalinity (total)
Acidity
Flow
Temperature
Specific conductivity
Solids
Hardness
Cl"_
S°~
Turbidity
Total phosphate
Orthophosphate
Total organic carbon
Total inorganic carbon
Nitrite and nitrate nitrogen
Total Kjeldahl nitrogen
Dissolved minerals
Metals
(total, suspended, dissolved)
Si02 Al
As
Caa
Co
Cr
Cu
DC
Fe
Hg
Ka
Mg
Mna
Ni
Pb
Se
Zn
Total metals analyses only.
Five stations (CC 16.7, CC 18.5, CC 15.9, LB 4.6, and LY 0.8) were
sampled monthly to determine the structure of lower-trophic-level biotic
communities. Qualitative and quantitative analyses of the zoomacrobenthos
and periphyton were performed. Additional water quality samples were col-
lected and analyzed for pH, temperature, dissolved oxygen, hardness, alka-
linity, acidity, and sulfate concurrently with collection of biological
samples.
-------
-17-
Four stations (CC 16.7, CC 15.9, LY 0.8, and LB 4.7) were sampled
monthly to determine fisheries populations. Fish sampling involved the
use of a Smith-Rout, 325-V, pulsating, DC, backpack shocker and a block
net at the head and lower end of the area to be shocked, which was 30 m
long.
CONTOUR-MINED SITES
The University of Tennessee began a sampling program in the basin
in January 1975. Water quality samples are collected weekly and analyzed
for many parameters: pH, sulfates, alkalinity, suspended solids, turbidity,
chlorides, total solids, and total and dissolved metals (aluminum, calcium,
cadmium, cobalt, chromium, copper, iron, magnesium, manganese, nickel,
lead, tin, and zinc). All analyses were performed by standard methods
(EPA 1974).
Continuous hydrological gaging of rainfall and streamflow was begun
at the onset of the study on Indian Fork, Bills Branch, and Lowe Branch.
Anderson Branch and Bowling Branch were provided with sharp-crested weirs
and staff gages for instantaneous measurements. By mid-December 1974, all
six basins were equipped with continuous rainfall and streamflow gaging
equipment.
SUMMARY
Sampling stations were established at mined and unmined basins in
two geologically dissimilar areas of the Cumberland Plateau. Mining and
land use information is summarized in Table 2.
-------
TABLE 2. SUMMARY OF SITE INFORMATION
Sampling station
Area
(ha)
Percent
mined
Seam
Dates of mining
Percent
forest
Percent
rural-resident
Fentress County
Lynn Branch (LY 0.8) 103
Long Branch (LB 4.6) 280
Crooked Creek (CC 18.5) 180
Crooked Creek (CC 16.7) 725
Unnamed Tributary (UT 0.01) 52
Crooked Creek (CC 15.9) 950
0
0
0
13
43
13
Nemo
Nemo
Nemo
4/75
9/75
4/75
9/76
9/76
9/76
89
89
31
31
5
32
11
11
69
56
52
55
New River Basin
Bowling Branch
Lowe Branch
Anderson Branch
Indian Fork
764
238
210
1119
Green Branch
357
0
0
7.5
8.9
1.7
2.8
7.4
2.2
—
-
Jellico
Big Mary
Walnut
Mountain
Pewee
Jellico
Big Mary
Pewee
Grassy
Springs
3/76 - 3/77
Late 50's
- 72
/ 52 - Present
/ 52 - Present
72 - 73
73 - 12/74
6/75 - 9/75
6/75 - 9/75
100
100
92
81
76
0
0
0
00
I
Bills Branch
174
9.0
Big Mary
12/74 - 9/75
90
-------
-19-
SECTION 5
PROJECT STUDIES
Preliminary findings from data collection activities are compared
for the Fentress County (area-mined) and New River Basin (contour-mined)
sites. Hydrological results are presented in the form of model tests
using data developed from similar studies. Test results for several
models are also presented.
WATER QUALITY
Water quality data for the six Fentress County sampling sites are
divided into three groups to facilitate presentation and interpretation
of the data from the standpoint of land use. Group M consists of the
three mined sites, CC 16.7, UT 0.01, and CC 15.9. Group F consists of
the two basins, LY 0.08 and LB 4.6, within which silviculture represents
the prime land use. Group A consists of data collected at station
CC 18.5, the only unmined basin with significant agricultural activity.
These data are presented in Figures 4 through 9 as monthly averages;
simple statistics from individual stations are presented in Tables 3
and 4.
pH-Acidity-Alkalinity
Although the term "mine drainage" typically connotes acid condi-
tions, more alkaline (pH > 6.0) conditions usually exist in the mine
drainages in the Cumberland Mountain area of Tennessee. As part of the
U.S. Army Corps of Engineers' reconnaissance survey of the New River
Basin, 319 grab samples were collected throughout the basin. Of these
samples, only 28 (9 percent) exhibited a pH of 6.0 or below; only 40
(13 percent) contained zero or negative net alkalinity. Data indicate
that mining increases pH and alkalinity. Figure 3 shows that, at the
Fentress County (contour-mined) sites, alkalinity is highest in agri-
cultural drainage, lower in mined drainage, and lowest in streams draining
forested areas. This reflects the extremely low buffering capacity
encountered in plateau streams.
Figure 4 indicates that pH is lowest in forested drainage and
slightly elevated in both mined and agricultural drainage, with no
appreciable difference existing between the two. Lower pH's in the
streams draining forested areas result primarily from dissolution of
atmospheric C02 into the unbuffered waters and from organic acids from
decaying vegetation. Increased pH's and alkalinity in the streams
draining mined and agricultural areas result primarily from dissolution
of substances from newly exposed and previously unweathered strata.
Acidity values (Figure 5) exhibit no consistent differences for the
three groups. These data indicate that mining increases both pH and
alkalinity, with little effect on acidity. They also suggest, however,
that similar effects are encountered for agricultural areas. Increases
in alkalinity and pH after mining have been documented by Connell (1976),
Curtis (1972), and Plass (1976). Acidity and pH appear to be time-
independent, except for a possible increase in pH in Group F during late
-------
TABLE 3.
MEANS AND RANGES OF PROJECT DATA FROM
FENTRESS COUNTY SITES
Constituent (mg/L)
Location
Lvnn Branch,
(LY 0.8),
a mni tied
Long Branch
(LB 4.6),
uomined
Crooked Creek
(CC 18.5),
unmined
Crooked Creek
(CC 16.7),
mined 13%
Crooked Creek
(CC 15.9),
mined 13%
Tributary to
Crooked Creek
(UT 0.01),
mined 43%
Statistic
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
PH
6.23
5.35-6.70
16
5.60
5.0-6.70
16
6.76
6.10-7.60
11
6.82
6.20-7.60
16
6.88
5.80-7.10
17
6.46
5.8-7.40
14
SO,
2
1-5
12
5
2-14
9
8
2-16
7
7
3-14
12
18
9-56
12
12
6-18
11
Total suspended
solids
7
1-39
7
11
1-33
7
26
11-50
5
54
2-300
7
77
4.0-400
7
50
3-180
6
Total hardness
(as CaO>3)
9
7-12
11
11
5-25
10
37
23-54
10
35
23-46
10
43
24-80
13
28
19-37
7
Alkalinity
(as CaO>3)
4.11
1-8
16
5.43
0.5-18
16
24.59
14-44
11
19.3
13-29
16
16.60
3.2-27
17
13.01
2-28
14
Acidity
(as CaCOjl
10
1-37
16
12
2-33
15
8
4-14
11
10
1-42
16
8
1-18
16
11
1-39
12
Total
iron
0.58
0.23-1.2
7
1.7
0.14-3.9
7
1.8
0.93-2.7
6
4.0
1.0-16
7
2.5
1.1-3.6
8
4.6
1.4-10
8
O
I
-------
TABLE 4. MEANS AND RANGES OF PROJECT DATA FROM
NEW RIVER BASIN SITES
Constituent (mg/L)
Location Statistic
PH
Anderson
Branch,
mmed 7 . 5%
LOVP Br
I-1
-------
-22-
40i
30-
to
O
o
O
o
20
o>
0
40
30-
K)
8
o
(O
o
0>
1 IOJ
0
40i
o
o
o
O
CO
^ 20-
10-
0
AGRICULTURE
JAN.
FORESTED
JULY
1976
JAN.
JULY
1977
\
JAN.
MINED
JULY
1976
JAN.
JULY
1977
/ D
JAN.
JULY
1976
JAN.
JULY
1977
Figure 3. Alkalinity--Fentress County sites.
-------
-23-
200n
200n
150-
IOCH
50-
200i
150-
o>
E
50-
AGRICULTURE
JAN.
JULY
1976
JAN.
JULY
1977
FORESTED
JAN.
Q.(350)
JULY
1976
JAN.
JULY
1977
MINED
JAN. JULY JAN. JULY
1976 1977
Figure 4. pH—Fentress County sites.
-------
-24-
50
to 40-
O
O
^20-
AGRICULTURE
JULY
(977
JAN.
MINED
JULY
1976
JAN
JULY
1977
JAN. JULY
1976
JAN.
JULY
1977
Figure 5. Acidity—Fentress County sites
-------
-25-
summer and early fall. Alkalinity exhibits a definite seasonality for
Group A, with a peak occurring in the fall. Groups M and F show no
similar trends.
Sulfate
Any production of acidity from oxidation of exposed pyrites will be
accompanied by associated production of sulfate ion (Singer 1970; Silver-
man 1967). As a result of this phenomenon, increases in sulfate concen-
tration are widely associated with mining activity. These increases may
or may not be accompanied by decreases in pH or increases in acidity,
depending on the amount and availability of strata with neutralizing
potential. Table 3 and Figure 6 contain sulfate data from the Fentress
County study sites; all sulfate values are relatively low. Mean sulfate
concentrations appear to increase with mining by about 10 mg/L, whereas
an increase of about 6 mg/L accompanies agricultural development.
These data indicate that little pyrite oxidation is occurring in
the disturbed watersheds, as evidenced by the near-background level sul-
fate concentration at the mined watersheds. This is not inherently
evident from pH, alkalinity, or acidity data because pyrite oxidation is
sometimes followed by neutralization. Exposed pyrites will possibly be
oxidized in the near future. A lag time between mining operations and
increases in sulfate production, previously observed by Dyer (1976) and
Plass (1976), is evident in data observed from the New River Basin
studies of the University of Tennessee and TVA (Minear 1976).
Data for sulfate concentrations from the New River Basin (contour-
mined) sites are presented in Table 4. In this case, sulfate concentra-
tions increase markedly with mining. Sulfate concentrations are highest
in Indian Fork (only 19 percent mined); lower, but still elevated by
about 130 mg/L, in Green Branch (24 percent mined); and only slightly
elevated (about 25 mg/L) in Bills Branch (10 percent mined).
The data in Indian Fork seem to be influenced by the existence of
deep mines located within the basin. These data are part of a University
of Tennessee study and are included here only for purposes of comparison.
The higher sulfate and iron concentrations indicate that pyrite oxidation
is occurring, but the relatively high alkalinity shows that neutralization
is occurring. Mining began in Indian Fork Basin in 1952, in the Green
Branch watershed in 1972, and in the Bills Branch watershed in late
1974. Short-term data (two years) indicate no such trends. Changes in
mining and reclamation techniques also influence drainage quality.
Bills Branch was largely mined and reclaimed under more stringent
regulations.
Total Hardness
Caruccio (1976) suggested that the factor for determining the acidic
character of mine drainage is the presence of calciferous materials in
the overburden material. Contact of precipitation and interflow with
newly exposed materials result in dissolution of soluble species such as
calcium and magnesium carbonates. Total hardness (as calculated from
calcium and magnesium) reflects this process.
-------
-26-
25-1
20-
en
6
15-
10-
5-
AGRI CULTURE
o o
D>
E
LU
15-
10-
5-
JAN.
FORESTED
JULY
1976
JAN.
JULY
1977
M\
1^
1/
JAN.
25
en
UJ
20-
'5
w 10-
MINED
JULY
1976
JAN.
JULY
1977
JAN.
JULY
197(5
JAN.
JULY
1977
Figure 6. Sulfate—Fentress County sites.
-------
-27-
Total hardness data from the area-mined sites are presented in
Table 3 and Figure 6. Waters from forested areas are extremely soft,
with hardness values (as CaCQ3) of about 10 mg/L. Waters draining
agricultural areas are measurably harder, but still moderately soft,
with hardness values of about 35 to 40 mg/L. Mined areas show similar
concentrations. A trend of increasing hardness is evident for the mined
and agricultural drainages.
Waters draining the forested areas in the New River Basin have
hardness values very similar to those found at the Fentress County
(area-mined) sites (Table 4); however, the similarity vanishes when
mined areas are compared. The hardness values range from a mean of 40.3
mg/L on Bills Branch, a soft-water system; to 152 mg/L (as CaC03) on
Green Branch, a moderately hard-water system; to 287 mg/L (as CaC03) on
Indian Fork, a hard-water system. As expected, increases in total
hardness roughly parallel increases in alkalinity in the Fentress County
data (Table 3). The absence of this relationship at the New River Basin
sites indicates that noncarbonate hardness is significant in these
waters (Table 4).
Iron
Concentrations of total iron in strip mine drainages can be highly
elevated, partly because of chemical and physical release of Fe 2 and
Fe 3 in the oxidation of iron pyrites. Increases can also occur as a
result of increased erosion and sediment transport.
Data from the area-mined sites are presented in Table 3 and Figure 8.
Means and ranges for the mined and agricultural sites appear significantly
higher than those for the control stations. No explanation of the two
extremely high values for iron at the agricultural sites was evident.
The one extremely elevated value for a mined site occurred during a
period of heavy rainfall. Samples collected during periods of moderate
to low flow, except for the two elevated values, suggest that mined and
agricultural areas show no discernible trends, whereas samples from
forested areas show an increase in total iron concentration during
autumn. This increase in forested areas may reflect the input from leaf
fall. During this period, water at the Long Branch control station,
which showed the highest iron values, was visibly discolored, probably
as a result of high concentrations of lignins and tannins. Iron values
were high, exceeding the criteria proposed by EPA (1976a) for freshwater
aquatic life (1 mg/L) 42 percent of the time.
Iron concentrations in streams draining the contour-mined sites
were considerably higher than those found at the area-mined stations.
Values presented in Table 4 reflect these elevated concentrations.
Again, as in the Fentress County stations, the control stations also
have higher iron concentrations, with values frequently exceeding the 1
mg/L criterion. Water at the station showing the highest concentrations,
Indian Fork, is frequently discolored by iron precipitates, part of
which probably come from abandoned underground mines in the basin.
Of the 319 samples collected for the Corps of Engineers in the New
River Basin, 48 (15 percent) meet or exceed the 1.0 mg/L criterion. Thus,
water quality in most parts of the basin may be limited as a result of
these elevated iron concentrations.
-------
-28-
6 On
to
O
^40
o
20-
AGRICULTURE
60-,
to
O
0 40-|
o
O
CO
o
20-
JAN.
FORESTED
JULY
1976
JAN.
JULY
1977
60-,
ro
8 40H
o
o
tn
a
en
£ 20H
JAN.
JULY
1976
JAN.
JULY
1977
MINED
JAN.
JULY
1976
JAN.
JULY
1977
Figure 7. Hardness—Fentress County sites.
-------
-29-
C7>
5.00
4.00-
' 3.00
1.00
4.00-
en
E
2.00
l.00:
AGRICULTURE O 12.00
5.00;
4.00^
^ 3.00 :
CD
E
aoo^
1.00 7
JAN. JULY JAN.
1976
FORESTED
A
^\
1 | ^-j-»-V*
JAN. JULY JAN.
1976
JULY
1977
A
v , •
JULY
1977
MINED
JAN.
JULY
1976
JAN.
JULY
1977
Figure 8. Total iron--Fentress County sites.
-------
-30-
Total Suspended Solids
Control of suspended solids is widely regarded as the greatest
water pollution problem in the United States (EPA 1976b). Suspended
solids represent substances, organic (algae, biological floe, leaves,
zooplankton, detrimental material) or inorganic (chemical precipitates
or eroded soils), that are transported within the water column. Solids
usually associated with mine drainages are inorganic. Iron floes,
during lower flows, can add significantly to concentrations of total
suspended solids in mine drainage. A total iron concentration of 10
mg/L, if all iron is suspended, typically represents 19 mg/L of sus-
pended solids [as Fe(OH)3], At station UT 0.01, suspended iron [as
Fe(OH)3] represented about 30 percent of the total suspended solids,
whereas sandstone sediments typically contained about 2.4 percent iron
(Degens 1965); the remainder represented iron precipitates. Many of the
TSS concentrations are below 100 mg/L. During higher flows, when signi-
ficantly more erosion and sediment transport occurs, the percentage of
iron flocculent solids decreases.
Generally, the suspended solids in strip mine drainage result from
rainfall washing disturbed soil into streams that drain the mine site.
This problem is obviously a flow-related phenomenon, with the most
serious events occurring during storm runoff.
Data presented in Figure 9 and Table 3 represent total suspended
solids concentrations for the Fentress County sites. During low- to
moderate-flow conditions (all samples except February), there is little
difference between the agricultural station and mined-area samples. The
samples representing forested areas are lower in all instances. Several
samples (four mined, three agricultural) exceeded the new source performance
standard of 70 mg/L for mine discharges.
These samples, however, reflect moderate- to low-flow conditions.
Over 250 grab samples were collected at stations CC 15.9 and UT 0.01
during nine separate storms occurring between March 1977 and January
1978. Suspended solids concentrations ranged from 12 to 4000 mg/L at
CC 15.9 and from 34 to 6300 mg/L at UT 0.01. Samples collected at the
contour-mined sites show a larger difference between mined and unmined
sites.
Almost 34 percent of the samples collected from streams draining
mined areas were found to violate the suspended solids criteria; two
undisturbed basins (Bowling Branch and Anderson Branch) also exceeded
these criteria. This higher incidence of excessive concentrations of
suspended solids probably results from the precipitous nature of the
stream channel.
The data were analyzed for a possible relationship between suspended
solids and turbidity (Minear 1976), but no significant relationship was
found. The indication is that turbidity measurements do not adequately
reflect suspended solids concentrations.
Because of the importance of rainfall-runoff relationships inherent
to strip mine drainage, a series of storm runoff events have been sampled
-------
-31-
200i
200n
150-
o>
100-
50
200n
50
1400)
AGRICULTURE
JAN.
JULY
1976
JAN.
JULY
1977
FORESTED
JAN.
Q.(350)
JULY
1976
JAN.
JULY
1977
MINED
JAN.
JULY JAN.
1976
JULY
1977
Figure 9. Total suspended solids—Fentress County sites.
-------
-32-
by the University of Tennessee (Dunning 1977). The first of these
events occurred on March 5-6, 1976. Samples were collected periodically
across the hydrograph and analyzed for various constituents—total
suspended solids, sulfate, alkalinity, turbidity, iron, manganese,
calcium, and magnesium. Plots of concentrations and loads for total
iron and suspended solids, along with the storm hydrographs for Indian
Fork and Lowe Branch, are presented in Figures 10 and 11. These results
are typical of other storms analyzed. Total flow for Indian Fork is
considerably higher than for Lowe Branch, as expected because of the
difference in watershed size. However, one would expect the hydrograph
of the smaller watershed to be more pronounced since channel slopes are
similar. This is not the case, indicating that mining may increase peak
flows or that rainfall may have been heavier at Indian Fork watershed.
Concentrations of total suspended solids reach a maximum during the
rise of the storm hydrograph, or about 2 h before the peak flow, and
then fall off rapidly. This phenomenon illustrates an initial flush of
readily available sediment, when the energy available for transport is
high, during the rising portion of the hydrograph, followed by a rapid
return to nearly normal concentrations. Loads of suspended solids are
similarly abrupt. About 38 percent of the suspended load was transported
during the hour surrounding the storm hydrograph peak; about 65 percent
was transported during the peak hour and the hour following the peak.
This illustrates the extreme transience of storm transport.
A different trend is observed for conservative substances such as
sulfate. Sulfate concentrations generally decrease during the rising
portion of the hydrograph and for some time past the peak, but essen-
tially level off during the recession and begin to return to prestorm
levels. This trend can be explained by simple dilution by surface
runoff. Loadings of sulfate show a peak similar to, but less abrupt
than, those for loadings of iron, showing a more gradual rise and
decline.
The storm hydrograph at Lowe Branch did not show a dramatic rise
and fall, as did that at Indian Fork. Also, sulfate concentrations were
somewhat elevated across the hydrograph, with the dilution phenomenon
never occurring. Loadings of sulfate reflected a definite peak with the
storm hydrograph peak, followed by an elongated tailoff.
The suspended solids relationships at Lowe Branch were similar to
those at Indian Fork in that concentrations during the storm runoff were
higher than those during normal flow. The main difference lies in the
elevated concentration obtained during early storm sampling at Lowe
Branch, which indicates that, for small storms, somewhat higher concen-
trations occur earlier during the storm on this small, unjoined watershed.
These results possibly indicate inadequate sampling.
Trace Metals
A topic of concern in recent years has been those substances,
especially trace metals, that may exert subtle, subacute effects on
aquatic organisms after exposure to relatively low levels for long
periods of time. Trace metals may affect organisms by uptake through
-------
-33-
90
to
~9 60
s
UJ
to to
30-
0
0 OIRON
A S04
O n S. SOLIDS
0°
D
O
Op 0 o
ODD
° 0
n o
&££ °°0ooooooo
•2000
^^
1500 -o
o>
1000 ~
z
o
500 -
O
12PM 2
8 10 12AM 2
8 O 12PM
30.0-
i —
O x
5 £200-
E ~
— UJ
§§
•*- w 10.0-
0 FLOW
/£j\ A S04
A D S. SOLIDS
An
^ A
AAA
n
CP
D
°8o0
C^ D°Oo°OOOoOO
. .D. .D. .D. ,n. .n. . . ,
o
o
X
150 5
CO
Q
1
•100 g
iPENDED
•50 ^
to
12PM 2 468
10 12AM 2
TIME (h)
8 10 12PM
Figure 10. Lowe Branch loads and concentrations of iron,
suspended solids, and sulfate.
-------
-34-
, IO.U
O
X
*.£
O "^
x ~IO.O-
^ en
*» 9
1 o
^ w 5.0
5* -^
u_ z
_l UJ
^ a
=>
en .
O OIRON
0 AS°4
Q D DS. SOLIDS
O _
o
°°o0 °o°o
D ° 00^0
o° °00oo0000
A A D
A ° D
D A A A A A A
$£ D D
i • > . .ni-'^-^'''-' —
-2.0
^^
a
\
a>
-1.0 ^
0
DC
•0
12PM 2 4 6 8 10 12AM 2 4 6 8 10 12PM
TIME (hrs)
.0
3.0-
O
X ^
5 | 2.0-
lO "~"
0^
0 0 0 FLOW
0 0 AS04
o00 as. SOLIDS
O
cP o o o n ^ _
00 °°o oo ooc
Sn
00
D
A
A A A A £
^
o>
B
CO
Q
_i
40 o
LJ
O
30 g
a.
co
ID
12PM 2 4 6 8 10 12AM 2 4 6 8 10 12PM
TIME (h»
Figure 11. Indian Fork loads and concentrations of iron,
suspended solids, and sulfate.
-------
-35-
the food chain (biomagnification), direct intake or absorption (bioaccumu-
lation), or influence on a segment of its food chain. In most cases,
the mechanism of toxicity, the chemical form of the material, and the
time and level at which chronic effects occur are unknown.
All these factors, along with the possibilities for synergism,
complicate the establishment of a "safe" level for trace metals in the
aquatic environment. Several attempts have been made at establishing an
"across the board" set of water quality criteria. Among the most notable
are Water Quality Criteria (FWPCA 1968), better known as the "green
book," developed by the Federal Water Pollution Control Administration
in 1968; the "blue book," Water Quality Criteria 1972 (NAS 1972), produced
by the National Academy of Science for the Environmental Protection
Agency; and finally, the "red book," Quality Criteria for Water (EPA
1976), published in 1976 by EPA as an alternative to the blue book
recommendations. All these have received both approval and condemnation
and remain controversial in their findings. Nevertheless, along with
the U.S. Public Health Service's Drinking Water Standards (USPHS 1962),
they are the most referenced and comprehensive sets of water quality
criteria available today. For this reason, the "red book" recommenda-
tions, where applicable, will be used as a basis for determining the
possible consequences of trace metals from strip mine drainages.
Selected trace metal data collected at the project watersheds are
summarized in Table 5. Where concentrations were below detection limits,
the detection limit was used for statistical purposes. Table 6 contains
values for these metals at four points—larger rivers selected because
of their large data bases rather than their similarity to the project
watersheds--in the Tennessee Valley; these values provide an acceptable
mode of comparison. The French Broad River at river mile 82.1 has a
drainage area of 2800 km2 that is largely forested. The river drains
the Cherokee and Pisgah National Forests; Asheville, North Carolina, is
the only city within the drainage area (Berthouex 1977).
The Holston River at river mile 118.4 is subject to significant
waste discharges upstream from industries at Kingsport and along the
North Fork Holston River. The drainage area at this point is 7450 km2.
The Tennessee River at river mile 602.3 is located at the tailrace of
the Fort Loudoun Dam. The drainage area at this point is 24,700 km2 and
is influenced by both the Holston and the French Broad Rivers. Two
cities—Knoxville, Tennessee, and Asheville, North Carolina—lie within
the upstream drainage area. The Tennessee River at river mile 471.02 is
the location of the tailrace of Chickamauga Dam. This point, well
downstream from any major point source, represents a drainage area of
53,850 km2.
Mean cadmium concentrations were similar for all streams, both
mined and unmined. The highest concentration, which occurred only once,
was 7 (Jg/L at Bills Branch. Only 1 of the remaining 102 samples exceeded
the detection limit of 1 \ig/L. These findings are consistent with
values encountered at the Valley-wide stations. The maximum values,
however, are above the "soft" water criterion of 4.0 |Jg/L for aquatic
life (EPA 1976). The higher values do not appear with any regularity,
and the means are all <2.0 pg/L.
-------
-36-
TAB1E 5. MEANS AND RANGES OF TRACE METALS DATA FROM PROJECT WATERSHEDS
Constituent ((Jg/L)
Location
Lynn Branch
(LY 0.4)
Long Branch
(LB 4.6)
Crooked Creek
(CC 18.5)
Crooked Creek
(CC 16.7)
Crooked Creek
(CC 15.9)
Tributary to
Crooked Creek
(UT 0.01)
Statistic
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Cadmium
1.0
<1. 0-1.0
10
1.0
<1. 0-1.0
10
1.11
<1. 0-2.0
9
1.0
<1. 0-1.0
9
1.17
<1. 0-3.0
12
1.27
<1. 0-4.0
11
Chromium
Area-mined
5.50
<5. 0-8.0
10
5.0
<5. 0-5.0
10
5.11
<5. 0-6.0
9
8.44
<5.0-33
9
5.1
<5. 0-6.0
12
5.45
<5. 0-9.0
11
Copper
Lead
Nickel
Zinc
Manganese
region — Fentress County
29.3
<10-90
14
22.9
<10-50
14
33.9
<10.0-80
13
27.7
<10-100
13
30.6
<10-90
16
26.7
<10-60
15
11.3
<10-22
17
11.2
<10-23
17
11.1
<10-22
16
12.2
<10-37
16
11.4
<10-32
19
11.3
<10-26
18
50.0
<50-50
1
50.0
<50-50
1
50.0
<50-50
1
50.0
<50-50
3
50.0
<50-50
2
16.4
<10-40
14
40.7
<10-170
14
21.5
<10-60
13
18.5
<10-40
13
30.0
<10-90
16
38.7
<10-120
15
20
<10-40
14
80
20-350
14
120
70-320
13
190
70-850
13
390
90-240
16
690
200-1000
14
Contour-mined region — New River Basin
Green Branch
Bills Branch
Indian Fork
Anderson
Branch
Lowe Branch
Bowling Branch
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
1.3
<1.0-20
71
1.1
<1.0-7
78
1.0
<1.0-4
64
1.0
<1.0-2.5
61
1.0
<1. 0-1.0
48
1.0
<1. 0-2.0
60
12
<1. 0-160
42
21
<1. 0-840
59
2.6
<1.0-10
9
2.0
<1.0-6.2
14
1.0
<1. 0-1.0
10
2.
<1.0-18
24
19
<1. 0-180
61
21
<1. 0-320
68
12
<1.0-69
55
5.7
<1.0-45
28
5.1
<1.0-48
38
1.0
<1. 0-1.0
24
4.2
<1.0-55
53
12
<1. 0-280
57
1.1
<1.0-13
55
2.8
<1.0-32
20
2.7
<1.0-40
23
2.6
<1.0-57
52
22
<1. 0-240
55
4.3
<1.0-50
67
29
<1. 0-110
58
2.7
<1.0-11
19
2.8
<1.0-25
21
10.9
<1. 0-100
28
65
<20-500
32
50
<20-900
33
46
<20-130
61
26
<20-190
52
20
<20-<20
28
20
<20-190
52
340
<40-1000
77
110
<40-1760
78
1200
<40-3700
80
50
<40-580
63
40
<40-70
60
40
<40-580
63
-------
TABLE 6. MEANS AND RANGES OF TRACE METALS FROM POINTS
IN THE TENNESSEE VALLEY
Constituent (MS/L)
Location
Holston River
Mile 118.4
French Broad
River Mile
82.1
Tennessee
River Mile
602.3
Tennessee
River Mile
471.02
Statistic
Mean
Range
N
Mean
Range
N
Mean
Range
N
Mean
Range
N
Cadmium
2.0
<1.0-12
160
1.0
<1.0-3
92
2.0
95
1.0
<1.0-9
95
Chromium
5.0
<5.0-<5.0
11
5.0
<5.0-<5.0
4
5.0
<5.0-<5.0
4
Copper
26.2
<10. 0-700
16A
36.0
<10. 0-570
92
29.2
<10. 0-840
95
49.1
<10.0-23
95
Lead
10.0
<10.0-32
160
11.0
<10.0-22
92
11.0
<10.0-21
95
11.0
<10.0-52
95
Nickel
50.0
<50.0-<50.0
7
50.0
<50.0-<50.0
4
50.0
<50.0-<50.0
4
Zinc
33.6
10.0-180
164
94.8
20.0-380
92
15.4
<10.0-90
95
26.0
<10. 0-320
95
Manganese
120
20.0-2900
171
72.7
<10. 0-1100
92
55.0
20.0-160
104
63.0
20.0-150
103
-------
-38-
Chromium concentrations for mined and unmined stations in the area-
mined region have similar means, with one value at CC 16.7 exceeding
10 |Jg/L. All values for the Valley-wide stations were below the detection
limit of 5.0 Mg/L. At Green Branch only two samples exceeded the criterion
for freshwater aquatic life (50 |Jg/L), whereas four samples exceeded the
criterion for domestic water supply (EPA 1976). At Bills Branch, one
sample exceeded the criterion for freshwater aquatic life, and three
samples exceeded the domestic water supply criterion. Copper concentra-
tions were also similar at mined and unmined sites in the area-mined
region. Copper concentrations appeared higher at Bills Branch and Green
Branch than at other stations. The mean values compare well with those
at Valley-wide stations. However, maximum values for these stations are
considerably higher than those from the study stations. All values were
below the criterion for domestic water supplies (1000 pg/L), except for
a single value at Tennessee River mile 471.02. This value, 2300 pg/L,
was about five times as high as the next highest observation and almost
50 times as high as the mean value (49.1 ug/L).
Mining had little effect on lead in the streams in the area-mined
regions. In the contour-mined New River Basin, Bills Branch, Green
Branch, and an unmined station, Bowling Branch, exceeded the criterion
for domestic water supplies (50 pg/L)--twice at the two mined stations
and once at the unmined station. Concentrations at the Valley-wide
stations exceeded the 50-|jg/L criterion only once.
As for the previously discussed metals, nickel concentrations were
not affected by mining in the area-mined (Fentress County) sites.
Indian Fork appeared to have elevated nickel concentrations as compared
with the remaining New River Basin stations. No concentrations at the
Valley-wide stations exceeded detection limits.
Mean values for zinc appeared largely unaffected by mining at all
sites. All values were well below the domestic water supply criterion
of 5000 pg/L.
Values for manganese were elevated at all mined stations, except
for Bills Branch. All stations, however, showed concentrations greater
than the 50-pg/L criterion for domestic water supplies, but were lower
than the 4-mg/L EPA effluent guidelines far manganese. Mean concentra-
tions for all stations, for Lynn Branch in the area-mined region and Lowe
Branch in the contour-mined areas, exceeded the 50-[Jg/L criterion. The
most severely affected stations were the tributary to Crooked Creek
(UT 0.01), station CC 15.9, Green Branch, and Indian Fork. Concentra-
tions at these stations were well above the Valley-wide stations.
In summary, concentrations of all trace metals except manganese
were largely unaffected by mining in the area-mined regions in Fentress
County. Mining in the contour-mined regions seemed to increase some
metal concentrations at selected sites, but the significance of these
slight increases is questionable.
-------
-39-
AQUATIC BIOTA
The fish populations in the streams around Jamestown were dominated
by the creek chub (Semotilus atromaculatus). This species was found at
every station, but its abundance was severely restricted at sites impacted
by siltation.
The control station at Long Branch (LB 4.5) had fewer fish than any
station sampled. The water at LB 4.6 was discolored, as if large amounts
of tannins were present. A reduced pH was also observed, with values
usually below 6, which could well account for reduced fish populations.
Several nonstream species, bluegill (Lapomis macrochirus), spotted
bass (Micropterus punctulatus), and largemouth bass (Micropterus salmoides),
were collected at the Crooked Creek stations. These fish probably
escaped from breached farm ponds when the area was mined. The collec-
tions to date do not indicate any reproduction for the introduced species.
The other control site, Lynn Branch (LY 0.8), has a good population
of creek chub and other minor stream species, including fathead minnows
(Pimphales promelas) and blacknose dace (Rhinichthys atratulus). Con-
tinued sampling at LY 0.8 will provide the baseline data for establishing
the transfer functions for developing the fisheries part of the ecosystem
model for the area.
Preliminary analyses of the contents of fish stomachs indicate that
the fish eat a wide variety of aquatic and terrestrial invertebrates.
The principal food resource is aquatic dipteran midge larvae (Chironomidae),
followed by springtails (Collembola) and aquatic mites (Arachnida).
These resources constitute 31.1, 18.8, and 7.4 percent, respectively, of
the individuals found in creek chub stomachs from the LY 0.8 site.
Other sites show similar trends.
These data indicate that fish assemblages in the streams differ in
the area, and these differences seem to be related to strip mine inputs
in the basiii.
Monthly biological sampling (qualitative and quantitative) of the
attached algal (periphyton) and macrobenthic communities involved col-
lection of 23 samples from each of five sites (CC 16.7, CC 18.5, CC 15.9,
LY 0.8, and LB 4.6). These 23 samples included three Surber square-foot
samples, collected from the dominant substrates present; one qualitative
sample, composed of subsamples collected from all the major substrates
present; three artificial basket substrates (4- by 4- by 2-in. cement
blocks); four drift samples (dawn, day, dusk, night); one light trap;
six sets of conservation webbing, an artificial substrate suitable to
quite a few macrobenthic organisms; and five Plexiglas slides, an
artificial substrate suitable for colonization by attached algae.
A summary of quantitative data collected with the Surber sampler is
presented in Table 7. These data indicate that the macrobenthic biomass
varies little among the five sampling sites. Station CC 18.5, a control
site on Crooked Creek, contained the most constant quantity of biomass,
which was attributed to the presence of a stable, homogeneous substrate
-------
TABLE 7. QUANTITATIVE MONTHLY MACROBENTHIC DENSITY DATA (MARCH-OCTOBER 1976)--
FENTRESS COUNTY SITES
Macrobenthic density (g/ft2) by Surber square
Station Surber No. 1 Surber No. 2 Surber No. 3 Station
•i
Insect Noninsect Insect Noninsect Insect Noninsect Insect Noninsect
Lynn Branch (LY 0.8)
Range 0.01-1.18 0.04-2.28 0 -1.78 0.04-2.23 0.02-1.65 0 - 0 -1.78 0 -2.28
Mean 0.387 0.521 0.403 0.650 0.304 0.005 0.364 0.392
Long Branch (LB 4.6)
Range 0 -0.204 0.05-0.29 0.01-1.94 0 -0.20 0.01-0.59 0.27-2.97 0 -1.94 0 -2.97
Mean 0.075 0.068 0.340 0.025 0.201 0.405 0.205 0.180
Crooked Creek (I8.5)b
Range 0.04-0.12 0 -0.05 0.21-1.30 0.22-0.72 0.05-0.19 0.01-0.73 0.04-1.3 0 -0.73
Mean 0.086 0.016 0.602 0.440 0.136 0.268 0.275 0.241
Crooked Creek (16.7)
Range 0 -2.28 0.01-5.45 0 -0.62 0.02-1.5 0 -0.53 0 -0.72 0 -2.28 0 -5.45
Mean 0.408 0.779 0.253 0.208 0.214 0.090 0.291 0.359
Crooked Creek (15.9)
Range 0 -0.23 0 -.076 0 -0.53 0 -2.18 0 - .60 0.04-7.10 0 -0.6 0 -7.1
Mean 0.046 0.010 0.092 0.492 0.154 0.899 0.097 0.467
-------
-41-
and a lack of siltation. Station CC 16.7, on the other hand, had insect
biomass quantities ranging from 0.0 to 2.28 g/ft2. This large variation
in biomass probably resulted from a shifting, unstable, eroding substrate
at this station. Station CC 15.9 contained the least mean quantity of
insect biomass, but the greatest mean quantity of noninsect biomass, as
a result of the large rubble substrate present. A species list of
zoomacrobenthos data is presented in Appendix B.
The quantitative drift data collected to date are summarized in
Table 8. Few trends are clearly discernible at this time. The greatest
movement (drift) of insects per unit flow occurred in Long Branch. This
conclusion may be biased, however, by the fact that flow data were
available from this station only in the spring and summer, when many
organisms were drifting. After August, the flow rate was too slow to
measure, and thus the biomass collected could not be related to flow.
Drift at station CC 18.5 may be biased in the opposite manner because
data were collected only for those months when flow was relatively high
and drift activity relatively low.
In general, the greatest amount of drift occurred at or around
dusk, although Lynn Branch recorded its greatest drift at night. Drift
was consistently greater at station CC 16.7 than it was downstream,
indicating that organisms may be colonizing areas between two stations.
TABLE 8. QUANTITATIVE MACROBENTHIC DRIFT DATA COLLECTED MONTHLY
(MAY-OCTOBER 1976)--FENTRESS COUNTY SITES
Station Dawn Day Dusk Night
Lynn Branch (LY 0.8)
Long Branch (LB 4
Crooked
Crooked
Crooked
Creek
Creek
Creek
(18
(16
(15
.6)
-5)
.7)
-9)
0
12
5
29
0
.24
.4
.7
.0
.58
6
35
25
1
0
.45
.0
.9
.3
.76
14
87
23
68
29
.4
.0
.3
.0
.5
32
9
3
16
6
.0
.0
.4
.0
.74
Data collected on the attached algal community (periphyton) and
associated heterotrophic aufwuchs are summarized in Table 9. Chlorophyll
a concentrations were used as a measure of algal biomass, although it
might more appropriately be called a biomass "estimate," because the
percentage chlorophyll per unit weight varies physiologically and taxo-
nomically. Nevertheless, chlorophyll was used to estimate relative
differences among similarly collected periphyton assemblages. The data
indicate that Crooked Creek station CC 18.5 contained by far the greatest
algal biomass per unit area, primarily as a result of two independent
factors: (1) Much less erosion occurred in this control area, resulting
in less suspended matter to scour away the algal communities; and (2) the
forest canopy was more open at this site, allowing for greater incident
radiation. Station CC 16.7 contained the least amount of algal biomass
as a result of increased scouring and canopy. Canopy and turbidity were
most likely the dominant factors limiting greater algal production at
station CC 15.9, whereas light and nutrients were probably limiting
production in the two small control streams, LY 0.8 and LB 4.6.
-------
-42-
Station CC 18.5 (control) also contained the greatest amount of
total attached aufwuchs (autotrophs plus heterotrophs). A large propor-
tion of the aufwuchs was autotrophic, as indicated by the low ratio
(0.48) of organic weight to chlorophyll.
Aufwuchs growing in surface waters that are relatively low in
organic pollution are usually dominated by algae (if, of course, no
other environmental factors are limiting). To obtain information on the
physiological condition of the algal periphyton, the pheophytin index
was also determined. Results of these determinations indicate that the
algal communities at station CC 18.5 were the most healthy (least percent-
age of pheophytin), whereas those at stations LB 4.6, LY 0.8, and CC 15.9
were the least healthy.
TABLE 9. MONTHLY PERIPHYTON DATA (APRIL-OCTOBER 1976)—
FENTRESS COUNTY SITES
Station
Lynn Branch (LY 0.
Long Branch (LB 4.
Crooked Creek (CC
Crooked Creek (CC
Crooked Creek (CC
8)
6)
18.5)
16.7)
15.9)
Chlorophyll
a
(mg/L)
0.36
0.69
8.82
0.29
0.53
Ash-free
organic
wt. (g)
0.014
0.012
0.040
0.011
0.026
Autotrophic
index
4.40
2.05
0.48
2.09
3.24
b
Pheophytin
index
1.18
1.17
1.55
1.31
1.18
3Ash-free organic wt. (mg/m2) divided by chlorophyll a (mg/m2).
bUnacidified OD663 divided by acidified OD663.
OVERBURDEN ANALYSIS
Chemical quality of drainage from strip-mined lands is a function
of the geology of the disturbed areas. Exposure of substrate below the
weathered zone results in chemical weathering of previously unavailable
materials.
The processes primarily responsible for establishing the acidic or
alkaline character of the drainage are (1) leaching of previously unex-
posed neutralizing substances, such as calcium or magnesium carbonates;
(2) exposure of mica or clay minerals, which serve as a buffer by reacting
with acid from pyrite oxidation; and (3) oxidation of pyrites to form
sulfuric acid and ferrous iron. Weathering also results in leaching of
materials such as iron, manganese, sulfates, and trace metals, from the
newly exposed and weathered overburdens. The amount of leached material
is a function of the amount of overburden exposed, the replacement
sequence of overburden, the amount and distribution of pyrites and trace
metals in the overburden, and the amount of water available for leaching
and transport.
-------
-43-
When pyrites are exposed to the atmosphere, they are chemically and
biologically oxidized. By-products of this oxidation include ferrous
iron, sulfate, and acidity. Oxidation is accomplished by the reaction
sequence,
2 FeS2(s) + 7 02 + 2 H20 ^ 2 Fe+2 + 4 H+ + 4 S04"2. (1)
This ferrous iron then undergoes oxidation to the ferric state in a
rate-limiting step:
4 Fe+2 + 0 + 4 H+ -> 4 Fe+3 + 2 H+ + 2 OH~ . (2)
Ferric iron then hydrolyzes to form insoluble ferric hydroxide, thus
producing more acidity,
Fe+3 + 3 H20 -» Fe(OH)3(s) + 3 H+, (3)
or oxidizes pyrite directly, thus producing more ferrous iron and
acidity,
FeS0(s) + 14 Fe+3 + 8 H 0 -» 15 Fe+2 + 2 SO,"2 + 16 H+. (4)
£* *• "
A preliminary indication of the character of the overburden is
illustrated by the concept of the acid-base balance developed by Smith
et al. (1974). The potential for a mine spoil to provide acidity can
be measured or predicted directly from pyritic sulfur concentrations.
The potential for a mine spoil to produce alkaline materials can also
be measured directly, but not predicted by other measurements. When
these measurements are expressed in similar units, they can be summed
to give the acid or alkaline potential for individual strata. These
strata potentials can be summed to give a similar potential for the
entire sequence.
Other characteristics important in determining the quality of water
draining the spoil are cation exchange capacity (CEC) and soil pH.
To collect representative data for individual strata and to facilitate
the comparison of data, cores were drilled in selected study watersheds.
These cores were obtained at four locations within Fentress County
(Figure 1) and at three locations within the New River Basin (Figure 2).
Cores JA1 and JA2 (numbers 1 and 2 in Figure 1), collected in Fentress
County, were in undisturbed land drained by Crooked Creek (upstream from
stations CC 15.9 and UT 0.01); core JA3 (number 3 in Figure 1) was taken
from mine spoil upstream from station UT 0.01; and the final core in
-------
-44-
Fentress County, Gl (number 4 in Figure 1), was collected near Grimsley
in an undisturbed area drained by Long Branch (station LB 4.6). All
cores were taken so as to represent the strata overlying the Nemo coal
seam, which is the primary coal mined in the Crooked Creek watershed.
The cores collected in the New River Basin were all taken from
within the Anderson Branch watershed. Each core represents a seam, or
group of seams that can be mined simultaneously, and contains 20 to 30 m
of overburden. One core, NR3 (number 1 in Figure 2), was taken to
represent mining disturbance from developing the Pewee and Walnut Mountain
coal seams. The core contains strata from elevations of 770 to 800 m
(770 m is the top of the Pewee seam, and the Walnut Mountain seam is
about 9 m below). The coals and the intervening strata were not recoverable.
A second core, NR2 (number 2 in Figure 2), was taken to represent
disturbance from developing the Big Mary coal seam and contains strata
from elevations of 670 to 701 m. The strata run from the coal seam
upward.
A final core, NR1 (number 3 in Figure 2), was taken to represent
disturbance from developing the Jellico coal seam and contains strata
from elevations of 478 to 519 m. The Jellico coal is about 30 cm thick
and occurs at 481 m below the existing surface.
After the core samples were logged (see Appendix C), composite
samples were taken to represent strata that were lithologically similar.
The characteristics used in determining similarity included color, tex-
ture, and mineral composition. Much of the core was unrecoverable.
This loss represents unconsolidated material that was "washed out" as a
part of the core drilling operation.
About 220 g of each composite was ground to pass through a No. 50
sieve, yet be retained by a No. 100 sieve, and was analyzed for total
sulfur, pyritic sulfur, cation exchange capacity (CEC), calcium carbonate,
soil pH, potential acidity with peroxide, and total neutralization
potential. Composite samples of about 2 kg, ground to pass through a
No. 10 sieve, were retained for leaching studies. The core logs are
presented in Appendix C.
Bar graphs depicting volume-weighted mean values for each core are
presented in Figures 12 and 13.
The cores taken from Fentress County are dominated by sandstones,
with intermittently occurring claystones and shales. Figure 12 contains
weighted mean values for the analyses performed on individual strata of
Fentress County cores. The means are weighted on the basis of depth of
the individual strata.
The CECs were similar for all cores, ranging from 7.4 to 14.0
milliequivalents (meq) per 100 g. Concentrations of calcium carbonate,
ranging from 0.13 to 0.46 meq per 100 g, indicate that the soils contain
a relatively high percentage of exchangeable magnesium, potassium, or
sodium. However, this soil content has not been reflected in water
quality data for which Mg:Ca, K:Ca, and Na:Ca ratios are comparable to
values obtained elsewhere.
-------
20
o
Q.
30-
0°
IM O
-j O
< 3 i.O'
tc *•
LU
32-
i24
cc
3.81
16-
rj
CO
8-
32
§24
u.
0
GRI
12.6
GRf
9.3
°-90
\&2L
10.0
rf
u
JA 1 JA 2 JA 3
CORE
JAI JA2 JA3
CORE
GRI JAI JA2 JA 3
CORE
I.Oi
80.5-
S!
Q. —
081
0.60
0.48
1
^
0.19
GRI JAI JA2
CORE
JA3
O
O 0.4-
« 0,3-
ro 02-
O
o
X O.I-
0.30
0.21
%
0.20
%
0.04
GRI JAI JA2
CORE
JA3
O
UJ O
s ~
if
o 5
12-
Z O
g 2
-------
32 -
5?
19.5
NR
NR
8.5
NR2
CORE
NR2
CORE
63
'/7s
r S S i
NR3
cc
£24-
P
6.3
yyy 2.0 i.e
/VV IX'-X'yj \S/S\
NR3
~ 5.0-1
o>
O 4.0
cr
2
io£OH
O
O
3 i-o
_ is-,
*
UJ Q
0 O
<
o
LU
pg
28
3.04
NR
71
NR
3.25
NR2
CORE
9.9
NR2
CORE
1.25
NR3
78
NR3
Figure 13. Strata analyses—New River Basin sites.
-------
-47-
Neutralization potentials are slightly higher than potential acidi-
ties for all cores, thus adequately reflecting the alkaline character of
the drainage. Total and pyritic sulfur values are very low, which
explains the low values for potential acidity. Pyritic sulfur makes up
50 to 75 percent of the total sulfur.
The data indicate that cores JA1, JA2, and JA3 are very similar and
that little weathering has occurred in the mine spoil (core JA3). Core
Gl contains significantly greater neutralization potentials.
New River Basin cores are dominated by siltstones and shales, with
core NR3 containing the only significant sandstone layers. Figure 13
contains results of chemical analyses on the New River Basin cores.
Values for neutralization potential and potential acidity are not avail-
able at this time. Unlike the Fentress County cores, the percentage of
exchangeable calcium to CEC is typical for many soil types. Significant
amounts of total or pyritic sulfur were not found in any cores.
Values for neutralization potential and for calculated and measured
potential acidities with peroxide were used to produce acid-base balances
for the Fentress County core. The results are presented in Table 10.
The only instance for which good agreement occurs is at core JA2. All
results except for the calculated balance for core JA1 indicate an
excess of neutralizing material. Water quality data suggest that an
excess should exist in all cores. If the coal were not included, the
calculated balance for this core would be only slightly acid at 0.06 ton
CaCOs equivalent per 1000 tons of spoil. Differences between individual
measured values and calculated values were quite large. A correlation
analysis produced an r value of 0.271. The mean value for the measured
values of potential acidity with peroxide was 0.70, with a standard
deviation of 0.33, whereas the mean and standard deviation were 1.90 and
3.51 for the values calculated from pyritic sulfur. This indicates
that, on the average, values predicted from pyritic sulfur contents are
higher than values measured in the laboratory. This is consistent with
the findings of Caruccio (1976), who has shown that framboidal (reactive)
pyrite, rather than total or pyritic sulfur contents, dictates the
amount of acid production. Values of excess neutralization potential
(Table 10), calculated by using data from the acidity with peroxide
procedure, indicate that all stations have a positive excess neutrali-
zation potential. This agrees with the findings of increased pH,
alkalinity, and hardness in the mined streams.
HYDRO-LOGICAL MODEL STUDIES
One advantage of the regionalized approach being taken with the
hydrological models is that data from other sources can be used to cali-
brate and verify the models. For example, a paper by Betson (1975)
describes simulations made at five quite different watersheds using
regionalized relationships developed for the continuous streamflow model
that is being used in this project. The following section describes
some work that has been done to date with off-project data.
-------
-48-
TAELE 10. ACID-BASE 1ALABCE H)E FENTRESS COUNTY STRATA
Excess
neutralization
potential
Neutralization
Core Strata potential
JA1 1
2
3
4
JA2 1
2
3
4
5
6
JA3 1
2
3
4
5
6
7
8
9
10
11
12
Gl 1
2
3
4
O.B
0.6
<0.3
1.5
<0.3
<0.3
1.8
0.8
2.3
O.B
0.5
0.8
1.3
1.5
0.3
0.4
<0.3
<0.3
<0.3
<0.3
1.1
1.1
2.2
7.0
10
1.6
Acidity
with
peroxide3
0.52
0.66
0.98
1.0
1.5
0.63
0.56
0.55
1.3
1.1
0.38
0.94
0.36
0.29
0.34
0.86
1.5
0.33
0.30
0.86
0.32
0.52
<0.05
0.95
0.62
0.69
Calculated
acidity3
0.62
0.31
11.9
2.81
6.87
0.31
0.3J
0.31
1.25
1.87
<0.31
<0.31
<0.31
<0.31
0.31
0.62
<0.31
<0.31
0.93
1.87
0.62
0.31
<0.31
0.31
13.75
1.87
Using
acidity
with
peroxide
0.28
-0.06
-0.68
0.5
-1.2
-0.33
1.24
0.75
1.0
-0.3
0.12
-0,14
0.94
1.21
-0.04
-0.46
-1.2
-0.03
0
-0.56
0.78
0.48
2.15
6.05
9.38
0.91
Hi
:ing
calculated
acidity
0.
0.
-11,
-1.
-6,
-0.
1.
.18
.29
,6
.31
.57
.01
.49
0.49
1.
-1
0.
0
0.
1.
-0.
-0.
0.
0
-0
-1
0
0
1
6
-3
-0
.05
.07
.19
.49
.99
.19
.01
.22
.29
.29
.63
.57
.48
.79
.89
.69
.75
.27
Individual
strata
depths
Measured
value
re)
39.2
17.0
7.3
12.3
3.9
45.5
29.4
2.7
4.4
2.1
8.3
8.3
8.3
8.3
8.3
8.3
8.3
8.3
8.3
8.3
8.3
8.3
45.2
9.2
4.6
14.2
Normalized
«)
51.
22.
9.
16.
4.
51.
33.
3.
5.
2.
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
8.
61.
12.
6.
19.
7
4
6
2
4
7
4
1
0
4
3
3
3
3
3
3
3
3
3
3
3
3
8
6
3
4
Corrected Smoaticm of
excess excess
neutralization neutralization
potential potential
Using
acidity
with
peroxide
0.145
-0.013
-0.065
0.081
-0.053
-0.171
0.414
0.008
0.050
-0.007
0.010
-0.012
0.078
0.100
-0.003
-0.038
-0.100
-0.002
0
-0.046
0.065
0.040
1.33
0.762
0.591
0.177
Using
Using acidity Using
calculated with calculated
acidity peroxide acidity
0.093
0.065
-1.11
-0.212
0.148 -1.16
-0.289
-0.005
0.498
0.015
0.053
-0.026
0.241 0.246
0.016
0.041
0.082
0.099
-0.001
-0.018
0.024
0.024
-0.052
-0.130
0.040
0.066
0.092 0.191
1.17
0.043
-0.236
-0.052
2,86 1.73
Tots CaCO, equivalent/1000 tons spoil.
-------
-49-
Model Testing Using Beaver Creek, Kentucky, Watershed Data
Musser et al. (1970) describe a pioneering study conducted in
Beaver Creek watershed in Kentucky that was aimed at quantifying the
effects of strip mining on hydrology. Some of the project data are
published, and most are available from the USGS (Louisville, Kentucky,
office). Streamflow recorder charts, precipitation abstracts for selected
storms, and daily rainfall data for several years were obtained from the
USGS for use in testing present modeling capabilities (water quality and
daily Streamflow data were published). Two watersheds were involved:
(1) Helton Branch, a 2.2-km2 control watershed that was 91 percent
forested; and (2) Cane Branch, a 1.7-km2 watershed, 10.4 percent of
which had been strip-mined by 1959.
Figure 14 shows observed Streamflow vs. that simulated with a con-
tinuous daily Streamflow model for the water year 1961-62 at Helton
Branch; model parameters predicted from previously developed regionalized
relationships were used. The observed volume of runoff for the water
year was 62.0 cm over the watershed, the simulated volume was 58.8 cm,
and the model explained 66 percent of the daily flow variance. These
same simulated values would have been obtained even if no stream gage
data had been available.
Figure 15 shows the results obtained at Cane Branch watershed. The
observed Streamflow totaled 65.5 cm, whereas the simulated Streamflow
totaled 57.4 cm, with the model explaining 76 percent of the variance.
Surprisingly, the results obtained with the model are almost as good at
Cane Branch as they were at the unmined Helton Branch. This indicates
that the effect of mining 10.4 percent of Cane Branch has not markedly
affected daily streamflows (beyond that accounted for by the model), and
that the existing regionalized relationships for this model are reason-
ably adequate. Whether these relationships will remain adequate as the
percentage of watershed mined increases or if the watershed is area-mined
remains to be tested.
The regionalized relationships that had previously been developed
for the storm hydrograph model being used at the project were also
tested with the Beaver Creek watershed data. This model had not previ-
ously been calibrated with data from watersheds in the Cumberland Plateau
physiographic province (where Appalachian coal is located). Results
indicated that the regionalized relationships are not adequate in this
province, even for unmined watersheds. Storm hydrograph data for addi-
tional watersheds have been obtained to improve the regionalized
relationships.
Tests of Background Nonpoint Source Water Quality Model
A regionalized, mineral water quality model for nonpoint sources
has been previously developed which can be used to simulate the back-
ground quality expected from natural areas (areas without major pollu-
tion sources). This model, described by Betson and McMaster (1975), was
extended to include additional constituents, as described by Betson
(1976). This model was also tested on the Beaver Creek watershed data.
Table 11 summarizes the observed data for many constituents for both
Helton Branch and Cane Branch along with the model simulations. The
-------
1.8
1.6-
1 -4
S 1-2-1
u
z
z 1-0-
*-*
i
a
sJo.i-
D-
_l
•_•
-------
0.0
0.0
Observed
Simulated
0.0
i
u
OCT NBV
DEC
JflN
FEB
MfiR
flPR
flflY
JUN
JUL
flUO
SEP
Figure 15. Cane Branch at Parkers Lake, Kentucky—observed vs. simulated flow.
-------
-52-
results obtained for Helton Branch indicate that the model is adequate
for most constituents. The results obtained for suspended sediment are
not good, but this regression-type sediment model could not be used in
this project in any case. In contrast, the results obtained at Cane
Branch are quite different when water quality in this mined watershed is
compared with the model that simulates water quality in background
(unmined) areas. The elevated concentrations of iron, sulfate, sedi-
ment, and dissolved solids, in particular, that result from mining will
be built into the project models.
TABLE 11. WATER QUALITY SIMULATIONS--TVA NONPOINT SOURCE
WATER QUALITY MODEL--BEAVER CREEK WATERSHED
Helton Branch
Cane Branch
Constituent
Si02, mg/L
Fe, mg/L
HC03, mg/L
SO,, mg/L
Cl, mg/L
TDS, mg/L
CaC03> mg/L
Specific conductivity,
PH
Suspended solids, tons
Observed
5.2
0.14
10.0
5.0
1.3
25.0
11.0
fjmhos 32.0
6.6
30.0
Simulated
5.6
0.02
15.0
5.8
1.3
33.0
18.0
48.0
6.8
107.0
Observed
-
5.4
1.0
280.0
-
370.0
130.0
662.0
3.6
1250.0
Simulated
5.7
0.02
16.0
6.2
1.4
35.0
20.0
57.0
6.7
83.0
a91% forest control watershed.
10.4% of area strip-mined.
°Simulated as unmined.
One other test was made of the background water quality model for
nonpoint sources. Initial data collected on the project indicated that
the iron and sulfate values in the control watersheds might considerably
exceed values simulated with the model. Although a few data from the
Cumberland Plateau province had been included in the model calibration,
they were insufficient. A compilation of watersheds located in the
Cumberland Plateau province was therefore obtained from STORET. The
watersheds were reviewed to locate those unaffected by mining; a total
of five (in addition to those used in the original model calibration)
were found. The model was then used to simulate average concentrations
for each constituent, which were compared with the observed data. The
results, shown in Table 12, indicate that the existing model seems to be
-------
TABLE 12. TEST OF TVA BACKGROUND NONPOINT SOURCE WATER QUALITY
MODEL AT UNMINED WATERSHEDS
Watershed
Obed River
near Crossville, TK
Observed
Simulated
Clear Creek
near Lansing, TN
Observed
Simulated
Greasy Creek
near Gobey, TN
Observed
Simulated
Bear Creek
at Ramsey, VA
Observed
Simulated
Gap Creek
near Arthur, TN
Observed
Simulated
Drainaee Constituent (mg/L)
area Percent Total Diss. Total Oiss. NH3 Org. Total
(km2) forest pH S04 Fe Fe TDS Cl Ka P04 P04 N N sioz Ma
31 98
7.0 5.2 1.1 0.7 58 3.3 1.4 0.07 0.04 0.08 0.4 2.0 0.40
6.8 5.2 0,6 0.02 30 1.1 1.1 0.1 0.06 0.02 0.2 5.5 0.06
396 85
7.0 6.1 0.3 0.08 - 3.1 1.0 0.02 0.01 0.02 0.2 2.6 0.02
6.8 6.2 0.6 0.02 - 1.4 1.3 0.1 0.06 0.02 0.2 5.7 0.06
36 100
7.0 8.3 0.6 0.14 21 1.9 1.4 0.03 0.02 0.02 0.2 4.7 0.02
6.7 5.1 0.6 0.03 30 1.1 1.1 0.1 0.06 0.02 0.2 5.5 0.06
35 99
7.2 169.0 - - - 7.3 7.5 0.8 - 0.62 1.1
6.8 5.2 - - 1.1 1-1 0.1 - 0.02 0.2
21 62
7.2 13.7 0.6 0.08 224 3.2 3.0 0.07 0.02 0.03 0.2 10.0 0.05
7.8 5.2 0.6 0.04 146 2.0 1.2 0.1 0.06 0.02 0.2 4.9 0.06
Ca Mg K
5.7 0.8 1.7
5.0 0.7 0.8
2.4 0.9 1.2
6.2 1.0 1.0
1.4 1.4 -
4.9 0.6 -
24.0 14.0
5.0 0.6
39.0 4.3 1.7
40.0 8.0 1.7
-------
-54-
adequate for simulating background conditions to be compared with after-
mining values. Observed sulfate values for Bear Creek considerably
exceed the simulated values, indicating that even this presumed back-
ground watershed may have been mined. Improvement in the accuracy of
simulated values for coal-associated constituents would require more
data, and probably some field data. As this model can now be applied
with only mapped data, no improvements to this background model are
planned.
Development of a Suspended Sediment Model
A suspended sediment model, which is coupled with the project's
storm hydrograph component of the streamflow model, has been developed
with the use of data from several other studies. Testing of the sedi-
ment model developed by Betson (1976) on the two Beaver Creek watersheds
(previously described) was not satisfactory. This multiple regression
model was not intended for use in this type of study, nor were the
results obtained by regenerating the calibration data particularly
encouraging. Attention, therefore, has focused on the Universal Soil
Loss Equation (USLE) and adaptations of it for several reasons: (1) A
considerable body of research and experience is associated with it, (2)
it includes the principal factors affecting erosion, and (3) it is
consistent with other components of the model in that it is simple and
easily applied.
As originally developed by Wischmeier and Smith (1965), the USLE
was used to estimate annual soil loss from agricultural lands on the
basis of rainfall energy, cover, and physical characteristics. The
equation was adapted by Williams (1975) to substitute the product of the
storm runoff volume times the peak discharge for the rainfall energy
term in the original equation, thus allowing for sediment to be determined
for individual storms. This adaptation makes the USLE attractive for
this project because (1) it can be interfaced with the hydrological
models, (2) it contains factors important in describing sediment generation
from mined areas, and (3) research is going on elsewhere to adapt the
technique to mining and construction sites. A test of the modified USLE
was, therefore, undertaken with the use of data from the Upper Bear
Creek experimental project located in northwest Alabama (TVA 1973). For
this test, a linear relationship between observed storm sediment loads
and the product term (storm runoff X peak discharge) was determined by
using several storms at each of six of the smaller watersheds in the
project. The modified USLE was then solved backward for the land cover
term "C." This latter term was then estimated directly from land cover
by using suggested values obtained from McElroy et al. (1975); Table 13
summarizes the results.
-------
TABLE 13.
-55-
TESTS OF UNIVERSAL SOIL LOSS EQUATION-
UPPER BEAR CREEK WATERSHED DATA
Sub-
watershed
SF1
SF2
SA1
SA2
SCI
SC2
Area
(km2)
6.42
36.0
5.18
18.6
2.25
4.97
Land
From data
0.011
0.002
0.152
0.052
0.187
0.140
cover factor "C"
Based on cover
0.002
0.016
0.164
0.098
0.135
0.129
Percent forest
100
85
22
42
25
29
Results of this test indicated that the modified USLE is capable of
accounting for a range of storm conditions and that the terms in the
equation can be estimated reasonably on a watershed scale. Williams (no
date) proposed further modification of the USLE which would permit
distribution of the storm sediment load across the storm hydrograph.
For this technique, an instantaneous unit sedograph (IUSG) is developed
on the basis of a relationship among the peak discharge, the maximum
value of source area runoff, and the time to peak. All these values are
available in the storm hydrograph model in such a manner that this sedi-
ment distribution algorithm is deterministic because no model parameters
are needed to use it. The algorithm was tested by using storm hydrograph
and sediment data from the Upper Bear Creek study (TVA 1973) and from an
urban hydrology study conducted in Knoxville, Tennessee (Betson 1976).
These tests indicated that this technique did allocate the storm sediment
reasonably well. The sediment distribution algorithm was also tested by
using data from the mined Cane Branch watershed in the Beaver Creek project.
Figure 16 shows a simulated storm hydrograph and sediment concentration
relationship vs. observed values obtained from Musser et al. (1970). For
this storm, suspended sediment totaled 96 tonnes from the 1.74-km2 watershed.
Development of a Water Temperature Prediction Model
Water temperature information will be needed in several model compo-
nents, such as the models for routing of suspended sediment and aquatic
biota. Obviously, because water temperature data would not have been
previously collected for a planning model application, a method for pre-
dicting these data must be devised. Water and air temperature data
collected at four sites in the Cumberland Plateau physiographic province
from Alabama to Kentucky were analyzed with a regression model that used
moving average air temperature as the independent variable. An average
equation was found to be capable of predicting water temperature:
WT = 2.4 + 0.86AT
3'
where
WT = water temperature, °C,
AT3 = 3-day average air temperature, °C.
-------
24
CANE BRANCH NEAR PARKERS LAKE, KY.
STORM OF 7/1/59
DRAINAGE AREA = 0.67 mi
10.4% STRIPPED
PREDICTED DISCHARGE
OBSERVED DISCHARGE
PREDICTED SEDIMENT
OBSERVED SEDIMENT
UJ
o
CO
ON
I
200O
TIME (h)
Figure 16. Results of TVA suspended sediment-storm hydrograph model.
-------
-57-
The relationship is a "hockey stick" function because, when
AT3 < 0°C, WT = 0°C. Figure 17 shows a year of observed and predicted
water temperatures for Carr Fork near Sassafras, Kentucky. The relation-
ship shown in the figure explained 92 percent of the variance. For appli-
cations involving stochastic, long-term simulations, water temperatures
will be approximated with a sinusoidal function.
Development of a Channel Geomorphology Model
Routing streamflow and water quality constituents downstream from
first-order streams to locations at which a viable aquatic ecosystem
might be impacted requires a model that will use channel measures.
Channel dimensions are particularly difficult and expensive to measure
in coal-mining areas because of the rugged terrain and dense vegetation.
Therefore, a channel morphology model was developed so that these channel
measures can be predicted with the use of readily available information.
In developing this model, stream discharge measurement data from
some 50 stream gage sites located in the Cumberland Plateau province
were obtained from Kentucky offices of the USGS. These stations were
thought to be virtually unaffected by mining in their watershed, and
for those stations located at a bridge, the cross section was not thought
to be significantly affected by the bridge. Further screening reduced
the number of stations used to 34. Relationships between instantaneous
discharge and channel width and area were then developed by regression.
The basic equations are
and
W = aQb
A = cQd,
where
W = channel width, m,
Q = stream discharge, m3/s,
A = channel cross-sectional area, m2,
a,b,c,d = coefficients.
From these two equations, two other channel measures can be derived:
HD = A/W
and
where
HD = hydraulic depth, m,
V = velocity, m/s,
-------
32
28
24
20
o
o
u l6
o:
6 '2
tr
UJ Q
OBSERVED
PREDICTED
WT= 2.0 -I- 0.92
i
11
X
I
CARR FORK NEAR SASSAFRAS, KENTUCKY
DEC. 1970 - NOV. 1971
JAN FEB MAR APR MAY JUN JUL AUG
SEP OCT NOV DEC
Figure 17. Comparison of observed and predicted stream temperatures
-------
-59-
The four coefficients in the equations were next used as dependent
variables in regression equations using, as independent variables, the
channel slope (SL) near the gage sites, measured from a topographic map,
and the drainage area (DA). The resulting equations are
a = 0.453 DA0'545 SL°'565,
b=1.134DA-°'239 SL-°'294,
c = 0.038 DA0'611 SI0'772,
d=1.507DA-°-0987SL-°-166.
The model was tested at several gage sites not included in the
calibration data set. Preliminary results indicate that the model is
adequate for routing purposes and, in some cases, provides a better
measure of general channel morphology than actual data at some stream
gages (which were not located with this purpose in mind). Figure 18
shows the results obtained at a test watershed.
Extension of the Storm Hydrograph Model Regionalized Relationship
As explained in the description of tests using data from the Beaver
Creek, Kentucky, watershed, the results obtained when storm hydrographs
at the two watersheds were simulated indicated that the regionalized
relationships were not adequate. Since the model had not previously been
calibrated in this physiographic province, this finding was not too
surprising. Therefore, storm hydrograph and rainfall data were obtained
for five stream gages located in this province from the USGS, Nashville
District. Because these streams drain unmined watersheds, the data will
be used to improve the regionalized relationships used to predict back-
ground storm hydrographs. Information on the stations, by Wibben and
Boyd (1975), is summarized in Table 14.
TABLE 14. BACKGROUND WATERSHEDS FOR MODEL CALIBRATION
LOCATED IN CUMBERLAND PLATEAU
Watershed
Bitter Creek near Camp Austin, TN
Raccoon Creek near Old Winesap, TN
Self Creek near Big Lick, TN
Forked Creek near Oakdale, TN
Byrd Creek near Crossville, TN
Drainage
area (km2)
14.3
3.94
9.84
6.32
2.85
Number of storms
analyzed
18
12
21
10
19
-------
10
T T
I — I I I I I
1 1—I I I I |
i—i—i i i i r
000
PREDICTED VELOCITY
PREDICTED WIDTH
OBSERVED VELOCITY
OBSERVED WIDTH
BITTER CREEK NEAR OAKDALE, TN.
DRAINAGE AREA = l?.6 mi2
SLOPE = 35 ft/mile
t i.o
o
3
LJ
100 X
O i
o
i
0.
t I 111
1 I
I 1 t I
10
1000
1.0
10
DISCHARGE (cfs)
100
Figure 18. Results of TVA geomorphic model.
-------
-61-
The relationships among the model parameters and the watershed
characteristics for these five background watersheds and the results
obtained from analyzing 21 storms at Helton Branch in the Beaver Creek
project and the three project background watersheds (Bowling, Lowe,
and Long Branches) will provide ample information for improving the
regionalized relationships.
-------
-62-
SECTION 6
MODEL COMPONENTS
EXPLANATION OF MODEL COMPONENTS AND LINKAGES
Figure 19 shows the components of the strip mine drainage model.
This schematic was developed to show the interrelationships among the
components, the principal linkages, and the location of the decision-
making nodes.
The top three component blocks in the schematic show the types of
basic site information that will be needed to operate the model. These
data provide information to the model that allows it to handle the
unique characteristics of each site. Much of the information in the
left block can be obtained rather easily from maps, published information,
and aerial photographs. Time-varying data, such as rainfall, streamflow,
or climatic information, will not be required for application of this
model, as are most hydrology-driven models of this type. Site data,
however, will be required to describe the overburden geochemistry (center
block) and the mining and reclamation techniques (right block). The
core samples can be obtained during site investigation drilling in those
locations for which no previous overburden analyses have been performed,
and information on the strip mine characteristics can be obtained from
the mine permit applications.
The first-order stream hydrology model is a nonpoint source model
that is used to simulate streamflow immediately below mine sites (drainage
area, 2 km2 plus) and from unmined or background areas. This model
component, which adjusts for differences in mining technique and handles
the effect of sedimentation ponds, drives the rest of the system.
The first-order water quality model, also a nonpoint source compo-
nent, is used to predict water quality constituent transport from the
mined and unmined areas of a watershed in conjunction with the stream-
flow model. Suspended sediment is a principal constituent of concern in
this model. The streamflow model provides, however, for transporting
other constituents in various pathways so that any constituent of environ-
mental concern can be handled.
A shortened version of the model, using the first-order streamflow
and quality components only, can be used to assess the capability of a
proposed mine to meet regulatory criteria as shown in the upper decision
node. In other words, the model could be used to simulate the effects
of a given mining scenario, but if the simulation would not meet quality
criteria, additional protection measures could be tried until the
simulations met the criteria.
The quality and quantity simulations from the first-order stream-
flow model components are next routed downstream on the basis of channel
morphology. At a point well downstream at which a viable fishery resource
could exist, the routed streamflow quantity and quality are input to the
aquatic biota and fisheries resource model components.
-------
LAND COVER AND
WATERSHED CHARACTERISTICS
* 1 i
LAND COVER \ PHYSICAL CHAR.
STREAM ASSOC. CHAR.
1
OVERBURDEN ANALYSIS
CORE DRILLING
^T^
PYRITE r WEATHERING
IDENTIFICATION \ EXPERIMENTS
CHEM. ANALYSIS
\
STRIP MINE CHARACTERISTICS
1 1
PHYSICAL CHAR. RECLAMATION
TECHNIQUES
! '
FIRST ORDER STREAM HYDROLOGY MODEL
REGULATORY AND
DESIGN INFORMATION
FIRST ORDER STREAM WATER QUALITY MODEL
SUSPENDED PHYSICAL DISSOLVED
MATERIALS MEASURES MATERIALS
WATER QUALITY/QUANTITY ROUTING MODEL
WATER QUALITY WATER QUANTITY
CONSERVATIVE NON-CONSERVATIVE
BIOTIC COMMUNITY MODEL
/
LOW TROPHIC LEVEL
STREAM BIOTA MODEL
FISHERIES RESOURCE
MODEL
i
ON
Figure 19. Schematic diagram of model components.
-------
-64-
These latter components are important features of this model because
they allow the linking of cause and effect, which will permit a decision-
making dimension not possible when only water quality criteria are
considered.
The policy evaluation node at the bottom of Figure 19 is located
there to indicate that, with the entire model in operation, the impact
of mining can be assessed in relation to (1) natural or background con-
ditions, (2) changes to the streamflow and water quality regime, and (3)
changes to the aquatic biota of the stream. Synergistic effects, such
as the burial of benthic habitat coupled with changes in water quality,
can be evaluated. Therefore, this linkage will allow evaluation of the
tradeoff between the costs of mining and reclamation controls and degra-
dation to the aquatic resource.
PROGRESS ON MODEL COMPONENTS
Overburden Analysis
This phase of the project is patterned after the work of Caruccio
et al. (1976), which has shown the importance of framboidal pyrite in
determining the acidity of leachate. Overburden core samples were taken
in the Fentress County study watersheds and above each of the three major
coal seams in the New River Basin. Prepared samples of these cores were
sent to the University of South Carolina, where Caruccio and his staff
are molding pellets that will, in turn, be used to identify the pyrite
forms under a microscope. Laboratory apparatus, based on the approach
of Caruccio, was set up in the TVA laboratory to artificially weather
the overburden samples to determine leaching rates for various
constituents determining water quality.
Watershed and Strip Mine Characteristics—
This is the basic information needed to operate the system models.
Most of the information on land cover and watershed characteristics is
obtained from existing maps or other sources. A few characteristics are
obtained from field measurements for use in studies with project data;
however, future applications will not require costly field measures.
Site-specific information on the mining technique, mine dimensions,
reclamation, erosion control measures, and sediment control measures
will be needed. In model applications, this information may be obtained
from the mine permit application. In the project watersheds, this
latter information is obtained from a map of the New River Basin prepared
from LANDSAT data by USGS (Hollyday and Sauer 1976) showing conditions
as of April 1973; from annual overflight photographs of these areas
taken by TVA; or from field measures. The State of Tennessee does have
summaries of mining permit information that are used.
First-Order Stream Hydrology Model--
Recent versions of the two components of the hydrology mode being
adapted in this study were described in a TVA report on an urban hydrology
study (Betson 1976). The two components are a continuous, daily stream-
flow model used to budget moisture and a coupled storm hydrograph model.
-------
-65-
Both models have been regionalized (to the range of conditions in the
Tennessee Valley) so that all parameters can be predicted within the
model from readily measured, mapped watershed characteristics.
Tests of the continuous, daily streamflow model, which uses data
collected in the Beaver Creek Strip Mine Study (Musser et al. 1970),
indicate that the regionalized parameter prediction equations for this
component model are adequate. Further testing will be required to
determine whether these relationships will remain adequate in areas in
which contour-mining is more extensive or in which area-mining methods
are used. Some additional testing with off-project data is planned;
however, most of this testing will be done with project data when
streamflow data for one to two full USGS water years become available.
Tests of the storm hydrograph model, which uses the Beaver Creek
project data, did not prove adequate when validation was attempted at
the Helton Branch control watershed. As noted earlier, some 100 storm
hydrographs from six background watersheds were analyzed with an analy-
tical version of this model, and the results will be used to improve the
regionalized relationships. Storm hydrographs from project watersheds
in the New River Basin (mined and unmined) have been analyzed with this
model at the University of Tennessee as part of various thesis projects.
These results will also be incorporated into the regionalization analyses.
A version of the coupled models was adapted for use on a computer
time-sharing terminal. Input to this version consists of the basic
descriptive watershed characteristics, mean monthly land pan evaporation,
and a measure of mean monthly rainfall and its standard deviation.
Total input is on ten lines. The model produces all rainfall data
stochastically and predicts total rainfall and streamflow for each year
and the rainfall, runoff, and peak discharge for each large storm that
will occur during the year. In limited validation tests (watersheds not
located in coal-mining areas), the model was found to be able to reasonably
reproduce the observed annual maximum flood flow frequency relationship
(25 years) with the use of regionalized relationships to predict model
parameters. Those discrepancies that did occur were attributed to the
rainfall generator, which has not been tested much for this type of
application. Further testing of this model application is planned
for watersheds located in coal-producing areas because a model version
comparable to this, in case of application, will be part of the final
model package.
First-Order Stream Water Quality Model—
The earlier section on model tests with Beaver Creek project data
described tests conducted with the background nonpoint source mineral
water quality model. These tests indicated that this background model
is probably adequate for simulating premining conditions for watersheds
in the Cumberland Plateau. No further development is planned for this
model, except to extend the capability to additional constituents, if
necessary, for the aquatic biota model.
-------
-66-
An earlier section also described the suspended sediment model that
has been developed. Algorithms for this sediment model have been incor-
porated into the storm hydrograph model. Earlier, the first-order
streamflow model had a capability for simulating daily sediment by using
a power function on the daily flows. Regression equations had been
developed to predict the terms in this equation (Betson 1976). Incor-
poration of the storm sediment algorithm eliminated the need for the
daily sediment model, resulting in a program change. Consequently, the
daily sediment load algorithm was applied to groundwater flow only to
provide a capability for simulating background sediment. Next the
groundwater and groundwater (background) sediment were incorporated into
the storm hydrograph component. With this change, the hydrograph model
can now be used during storm periods to simulate (1) groundwater plus
stormwater flows and (2) background plus storm sediment loads. Suspended
sediment concentrations thus can be simulated continuously. In limited
testing, the concentration distribution was found reasonable. It was
found, however, that at most locations storms are not adequately sampled,
and as a result, validation of these detail simulations is difficult.
The first-order streamflow model has provisions for simulating any
water quality constituents that are associated with leaching by using a
power function relationship with the daily groundwater flow or total
streamflow. Incorporation of a storm sediment model enables simulation
of the transport of various constituents associated with storm runoff by
using the concept of potency factors introduced by Donigian and Crawford
(1976). A potency factor is simply a ratio between the concentration of
suspended sediment and that of another constituent. Donigian and Crawford
found that this approach could be used to simulate the transport of BOD
and nutrients. Because of the high sedimentation rates associated with
strip mines, this approach should work for many constituents. As a
first cut at modeling water quality, correlations among the concentra-
tions of various constituents and suspended sediment were developed by
using project data collected to date. The results were presented in the
earlier section describing project data.
Water Quality-Quantity Routing Model—
A watershed model developed at the University of Illinois (Shane
1973) has been tested as a tentative routing model. This model has pro-
visions for inputting observed (or simulated) hydrographs at selected
locations; generating hydrographs at other locations on the basis of
drainage area ratios; routing water; routing conservative water quality
constituents; and calculating dissolved oxygen concentrations and tempera-
ture. The relationships used for the conservative water quality con-
stituents are the same as those used in our background water quality
model (log cone - log a + b log 0) (Dyer et al. 1973). Streamflow is
routed by using a similar relationship between velocity (V) and flow
(Q). The advantage of this approach is that, after the V vs. Q rela-
tionships are established, routing can be accomplished without field
data. The model was tested by routing a storm hydrograph through two
gaging stations in Upper Bear Creek, Alabama (TVA 1973). Velocity vs.
discharge relationships had been developed earlier for these watersheds
by Ardis (1974). The routing tests indicated that this approach and the
model were applicable at the project.
-------
-67-
A section in the previous chapter described development of a chan-
nel georoorphology model. With the geomorphologic model, the relationship
between velocity and discharge can be predicted at any location for use
in the routing model from channel slope and drainage area, both of which
are obtained from topographic maps. This eliminates the need for costly
field measurements of channel cross sections.
An equation for predicting sediment concentrations using the "unit
stream power" concept was presented by Yang and Stall (1976). The
equation uses data on velocity, depth, slope, temperature, and mean bed
material diameter, all of which are easily obtained. Several tests of
this equation were made by using historic storm data from two Upper Bear
Creek project watersheds (one with low sediment loads and one with high
loads). Results obtained for several occasions for which high discharge
measurements were made (so that velocity and depth were known) were
excellent, considering that the mean bed sediment grain diameter had to
be estimated. This equation could be used in a sediment routing model;
however, it would have to be coupled with algorithms to handle deposition
and scouring distributions and the effect of armoring. The suitability
of several sediment routing models, including those developed by the
USGS (Bennett 1974), Oak Ridge National Laboratory (Fields 1976), and
the Corps of Engineers (Thomas 1975), is also being investigated. These
latter models are all steady-state models; they also require detailed
information on cross section, and in some cases, they require that the
model be calibrated to obtain values for some parameters. Obviously,
such data requirements are difficult to justify for a planning model.
A field survey of a 1.3-km reach of Crooked Creek, between the
mouth of the unnamed mined tributary (site No. 3, Figure 1) and the gage
at site No. 4, was conducted in mid-1977. Data on the channel cross
section were obtained for use in validating various aspects of the
routing model in a strip mine environment. The geomorphic model was
found to be able to predict channel dimensions about as well as they
could be determined in the field. (Because the measures are related to
discharge, this had to be estimated by using Manning's equation.) The
model could not, of course, account for localized conditions caused by
obstructions in the channel or the consequences of localized aggregation
in the channel below mined tributaries. Tests of the Yang equation for
predicting sediment concentrations at measured cross sections with the
use of field-measured slopes indicated differences among the sections
related to the condition of the channel (i.e., the carrying capacity
would be low in reaches in which sediment was deposited). These data
will be used in hydrograph routing tests as soon as sufficient data on
suspended sediment are collected from the project watersheds to allow
determination of realistic sediment concentrations and flows from the
source areas.
Biotic Community Model--
Artificial channels were constructed and modified on the basis of
preliminary tests with aquatic biota taken from project streams in the
Fentress County area. These artificial channels will be used to quantify
the effect of expected changes in water quality resulting from mining on
productivity of indigenous biota. Toxic or subacute effects resulting
from increased levels of trace metals will also be studied.
-------
-68-
The aquatic biota model will consist of two components, one module
for low-trophic-level biota and one for fisheries. Figures 20 and 21
show the generalized flow charts for these components. The fluxes among
modules are shown, and output from the benthic module is input to the
fish module. Rather than identify individual species within the model,
functional groupings will be used. The model will be expressed as a
series of differential equations that, when integrated, will provide the
change within the compartment over time. The final model will be able
to address either numbers or biomass. Impacts resulting from mining
that may disturb the system can include temperature changes, toxic sub-
stances (trace metals), catastrophic discharges (floods, spates), and
sedimentation. Conceptual functional relationships expressing the
effect of these disturbances on the biotic system have been developed.
The transfer coefficients necessary to quantify fluxes among the
module components are being obtained from the field data collected at
watersheds in the Fentress County project (described earlier in this
report). Although these" data will be somewhat site-specific as a result
of using functional biotic groupings rather than species, generali-
zations can be made, and the model can be applied elsewhere by using
data published in the literature.
Decision Nodes—
Because the purpose in developing this model is to provide planners
with a tool for use in decision making (as opposed to developing a model
to study the effects of mining), the needs of two types of decision
makers were considered. Figure 19 shows a decision node termed "Regulatory
and Design Information" and another termed "Policy Evaluation." These
two decision nodes reflect the needs for two quite different levels of
decision makers who could use this model.
When water quality criteria are provided as a constraint, components
of the first-order stream hydrology and water quality models can be used
to simulate the effect of a proposed mining, reclamation, and sedimenta-
tion control methodology to determine whether regulatory constraints
will be met. In effect, this shortened version of the model can be used
to evaluate and, if necessary, improve the design.
Rigid water quality criteria, however, are not always applicable in
a given situation. The criteria may be more restrictive than background
conditions, or they may impose costly constraints that result in no
tangible benefits to the aquatic ecosystem. Mining impacts the aquatic
system only when it affects living organisms. Therefore, there is a
second decision node that uses the entire model. At this node, policy
decisions can be made to evaluate (1) the effect of allowing variances
to water quality criteria, (2) the applicability of the criteria, or (3)
the accumulative effect of extensive mining within a larger watershed.
Figure 22 was developed to depict the appearance of this model frora
the point of view of either decision maker. As shown, general basic
information describing the site to be mined will be required. The user
will not be concerned with model coefficients or parameters—the sorts
-------
ALLOCHTHONOUS DETRITUS
NATURAL (NON-PREDATORY)
DEATH
~~J
UJ
o
o
UJ
V)
Dl
T
ii
AUTO-
CHTHON-
OUS
12
IMMIGRATION
n
13
HERBIVORES
I
COLLECTERS
DC
HC
EXPORT (DRIFT)
I
FISH MODULE
FC
I I
PREDATORS
[PC
+-EGESTA
Figure 20. Generalized benthic module: Biological model, Jamestown area streams.
-------
O
I
Figure 21. Fish module: Biological model, Jamestown area streams.
-------
-71-
Input Information
Topographic measures
Mean monthly climatic data
Land cover
Overburden characteristics
Sediment grain size
Channel canopy and bank flora
Generalized species composition
Strip mine dimensions
Reclamation information
Decision information
Process Model
Decision Options
4,
Design
WQ
oo
e
Planning
10 y
Flow probability
(Good)
Trophic
State
Index
(Bad)
Pre-mine
system
state
r
River mile
Mine
site
Figure 22. TVA strip mine impact assessment model.
-------
-72-
of numbers that must be estimated by professionals or modelers. Output
from the system will be specific to his needs and will show, for example,
in the design node whether the proposed configuration will meet appro-
priate water quality criteria. If the entire model is used, the output
might depict the degree and extent to which a downstream river could be
impacted.
-------
-73-
BIBLIOGRAPHY
Ardis, C. V. 1974. A nonlinear channel routing model. In Proc. Part 1,
First Symposium on Flow, Its Measurement and Control in Science and
Industry, Instrument Society of America, Pittsburgh, Pennsylvania,
pp. 99-106.
Bennett, J. P. 1974. Concepts of mathematical modeling of sediment yield.
Water Resour. Res. 10.
Betson, R. P. 1975. First generation models for water quantity and
quality. In Proceedings, Contribution of Irrigation and Drainage
to the World Food Supply Speciality Conference, Biloxi, Massachusetts.
Betson, R. P. 1976. Urban hydrology—a systems study in Knoxville, Tennessee.
Tennessee Valley Authority, Water Systems Development Branch, Knoxville,
Tennessee.
Betson, R. P., and McMaster, W. M. 1975. A first generation nonpoint source
mineral water-quality model. Water Pollut. Control Fed. 47.
Caruccio, F. T.; Geidel, G.; and Sewell, J. M. 1976. The character of
drainage as a function of the occurrence of framboidal pyrite and
ground water quality in Eastern Kentucky. In Sixth Symposium on
Coal Mine Drainage Research, National Coal Assoc., Bituminous Coal
Research, Inc.
Connell, J. R., et al. Identification of factors affecting water quality
from strip mined sites. Final Report, OWRT Project A-055-VA.
Curtis, W. R. 1972. Chemical changes in streamflow following surface
mining in Eastern Kentucky. Presented at the Fourth Symposium on
Coal Mine Drainage Research, Pittsburgh, Pennsylvania, Apr. 26-27,
1972.
Degens, E. T. 1965. Geochemistry of sediments. Prentice-Hall, Inc.,
Englewood Cliffs, New Jersey.
Donigian, A. S., and Crawford, N. H. 1976. Modeling nonpoint source
pollution from the land surface. U.S. Environmental Protection
Agency, EPA-600/3-76-083.
Dunning, D. E., Jr. 1977. Water quality characterization and some
relationships to coal surface mining disturbance in six primary
East Tennessee watersheds. Master's thesis, The University of
Tennessee.
Dyer, K. I., and Curtis, W. R. 1976. Effect of strip mining on water
quality in small streams in Eastern Kentucky, 1967-1975. Presented
at Louisville, Kentucky, Oct. 19-20, 1976.
Dyer, H. L.; Walker, P.; Schregardus, D.; Tamblyn, T. A.; and Shah,
P. S. 1973. State of Illinois water quality management program,
Appendix E, the water quality model program BASIN. Argonne National
Laboratory Center for Environmental Studies, Argonne, Illinois.
-------
-74-
Environmental Protection Agency. 1976ji. Quality criteria for water.
EPA-440/19-76-023, Washington, D.C.
Environmental Protection Agency. 1976b. Erosion and sediment control—
surface mining in the eastern United States. Technology Transfer,
vol. 1, Planning; vol. 2, Design. EPA-625/3-76-006.
Federal Water Pollution Control Administration. 1968. Water quality
criteria. U.S. Government Printing Office, Washington, D.C.
Fields, D. E. 1976. LINSED: A one-dimensional multireach sediment
transport model. ORNL/CDS-15, Oak Ridge National Laboratory, Tennessee.
Carman, R. K.; Ferguson, C. C.; and Jones, M. L. 1976. Geologic map and
mineral resources summary of the Fork Mountain Quadrangle, Tennessee.
State of Tennessee, Department of Conservation, Division of Geology,
Nashville, Tennessee.
Hollyday, E. F., and Sauer, S. P. 1976. Mapping and measuring land cover
characteristics of New River Basin, Tennessee, using LANDSAT digital
tapes. U.S. Geological Survey, Water Resources Investigations Open
File Report, Nashville, Tennessee.
Johnson, R. C., and Luther, E. T. 1972. Stripable coal in the Northern
Cumberland Plateau area of Tennessee. Report of Investigations 34,
Tennessee Department of Conservation Division of Geology, Nashville,
Tennessee.
Kimball, L. R., Consulting Engineers. 1976. Second quarterly report,
October, November, December 1976, New River Basin comprehensive
study—phase II. Submitted to Study Participants.
Luther, E. T., and Avery, G. G. 1970. Geologic map and mineral resources,
summary of the Norma Quadrangle, Tennessee. State of Tennessee,
Department of Conservation, Division of Geology, Nashville, Tennessee.
McElroy, A. D.; Chiu, S. Y.; Nebgen, W. J.; Aleti, A.; and Bennett, F. W.
1975. Interim report on loading functions for assessment of water
pollution for nonpoint sources. U.S.-Environmental Protection
Agency, Project No. 68-01-2293.
Meinert, D. L. 1976. Water quality monitoring network. Memorandum to
A. H. Smalley, Nov. 15, 1976.
Minear, R. A., and Tschantz, B. A. 1976. The effect of coal surface
mining on the water quality of mountain drainage basin streams.
Water Pollut. Control Fed. 48.
Musser, J. J.; Collier, C. R.; Rickering, R. J.; et al. 1970. Hydrologic
influences of strip mining—chapters A-C. U.S. Geological Survey
Professional Paper No. 427, Q-383-348, U.S. Government Printing
Office, Washington, D.C.
National Academy of Science. 1972. Water quality criteria 1972. U.S.
Government Printing Office, Washington, D.C.
-------
-75-
Plass, W. T. 1976. Changes in water chemistry resulting from surface
mining of coal on four West Virginia watersheds. Green Lands
(Quarterly of The West Virginia Surface Mining and Reclamation
Assoc.), Winter, 22. 1962.
Public Health Service. 1962. Public Health Service drinking water
standards, 1962, PHS Pub. No. 956, Washington, D.C.
Shane, R. M, 1973. State of Illinois water quality management program—
appendix D, the hydrology package. Argonne National Laboratory Center
for Environmental Studies, Argonne, Illinois.
Silverman, M. P. 1967. Mechanism of bacterial pyrite oxidation.
J. Bacteriol. 94.
Singer, P. C. , and Stumm, W. 1970. Acidic mine drainage: The rate-
determining step. Science 167.
Smith, R. M.; Grube, W. E., Jr.; Arkle, T., Jr.; and Sobek, A. 1974.
Mine spoil potentials for soil and water quality. EPA-670/2-74-070,
West Virginia University, Environmental Protection Agency, Cincinnati,
Ohio.
Tennessee Valley Authority. 1973. Summary report on the Upper Bear Creek
experimental project. Knoxville, Tennessee.
Thomas, W. A. 1975. Using a scour and deposition model to determine
sediment yield. In Proceedings of the Sediment Yield Workshop,
Oxford, Mississippi, June 1975. U.S. Department of Agriculture,
ARS-S-40.
Ward, A. J.; Haan, C. T.; and Barfield, B. J. 1977. Simulation of the
sedimentology of sediment detention basins, University of Kentucky,
Water Resources Inst. Research Report No. 103, Lexington, Kentucky.
Wibben, H. C., and Boyd, E. B. 1975. Peak stages and discharges on small
streams in Tennessee, 1965-1975, basic data report No. 6. Open-file
report, U.S. Geological Survey, Nashville, Tennessee.
Williams, J. R. 1975. Sediment yield prediction with universal equation
using runoff energy factor. In Proceedings of the Sediment-Yield
Workshop, Oxford, Mississippi, June 1975. U.S. Department of
Agriculture, ARS-S-40.
Williams, J. R. No date. A sediment graph model based on an instantaneous
unit sediment graph. USDA-Agr. Res. Serv., P.O. Box 748, Temple,
Texas.
Wischmeier, W. H., and Smith, D. D. 1965. Predicting rainfall-erosion
losses from cropland east of the Rocky Mountains. Agriculture
Handbook 282, U.S. Department of Agriculture, ARS.
Yang, C. T., and Stall, J. B. 1976. Applicability of the unit stream
power equation. In Proceedings of the Third Federal Interagency
Sedimentation Conference.
-------
-77-
APPENDIX A
STATE OF TENNESSEE STRIP MINE REGULATIONS
-------
-79-
STATE OF TENNESSEE STRIP MINE REGULATIONS
Law
Regulation--major impacts and changes
No law prior to September 1967.
September 1967—Tennessee Strip
Mine Law of 1967. First State
provision for strip mine regu-
lation and control, permits
reclamation, enforcement.
No State strip mine control prior to
September 1967.
September 1967, Regulations Pertaining to
Surface Mining; General regulation of
complete backfill cover against light-
wall; placement and grading of spoil
banks and overburden to favor revegeta-
tion; drainage ditches; erosion and
sediment control structures; revegeta-
tion program; broad time-period con-
straints; $400/acre bond required.
March 1973, Regulations Pertaining to
Surface Mining; More specific regula-
tion criteria; water quality discharge
permit application required; major
mining constraint on >28° slopes;
drainage ditch and culvert criteria;
toxic materials segregation; maximum
solids, bench width/slope below coal,
outcrop criteria; 35° maximum spoilbank
slope; 30-foot maximum highwall exposure;
10-foot minimum fill above coal seam;
grasses and trees revegetation required;
general limitation on head-of-hollow
mining; $600/acre bond requirement.
March 1974--The Tennessee Surface June 1974, Regulations Pertaining to
March 1972--The Tennessee Surface
Mining Law, Amendments to 1967
Law.
Mining Law, Amendments to pro-
vide for more specific regula-
tion and control to minimize
injurious environmental surface
mining effects.
Surface Mining; More specific regula-
tion criteria and rules; newspaper
notice of permit application required;
water quality discharge permit required
prior to mining application; prospect-
ing permit required; specific rules on
access roads plans; 28° slope limita-
tion on mining; 125-foot downslope
spoil limit below cropline for slopes
<28°; 20-foot maximum highwall exposure
before July 1, 1975—no highwall after
July 1, 1975, on new cuts; landslide
prevention rules; no mining within 25
feet of wet-weather drainage center-
lines; no mining within 100 feet of
flowing stream centerlines; $1,000/
acre bond requirement.
-------
-80-
STATE OF TENNESSEE STRIP MINE REGULATIONS (Continued)
Law Regulation--major impacts and changes
December 1975, Addendum to Regulations
Pertaining to Surface Mining; Adoption
of drainage control handbook for con-
trolling mine haul road water runoff
and sediment; major revegetation changes,
including agricultural lime rates, mulch
requirements, and seeding rates.
July 1976, (Proposed) Addendum to Regula-
tions Pertaining to Surface Mining; 50-
foot downslope spoil limit for slopes
<28°; brush clearance area disturbance
and barrier below spoil—not to exceed
125-foot downslope from cropline; 30°
maximum spoilbank slope; 100-foot
distance limitation.
-------
-81-
APPENDIX B
ZOOMACROBENTHOS COLLECTED IN THE VICINITY OF JAMESTOWN, TENNESSEE
-------
-83-
Annelida
Oligochaeta
Crustacea
Cambarus sp.
Hyallela azteca
Immature decapod
Coleoptera
Hydrobius sp.
Ectoparia sp.
Gonielmis dietrichi
Helichus sp.
Optioservus immunis
Promoresia elegans
Oulimnius latisculus
Derovatellus sp.
Phycoetes sp.
Pelonomus obscurus
Collembola
Isotoma sp.
Smynthurus sp.
Diptera
Pilaria sp.
Antocha sp.
Unknown Chironomidae
Dicraaota sp.
Dixa sp.
Dolichopodidae
Hemerodromia sp.
Hexatoma sp.
Limonia sp.
Limnophila sp.
Ormosia sp.
Palpomyia sp.
Paradelphomvia or Pseudolimnophilia
sp.
Tipula sp.
Simulium sp.
Tabanidae
Tabanus sp.
Crysops sp.
Heleidae pupa
Bezzia or Probezzia sp.
Hemiptera
Gerris sp.
Microvelia sp.
Rhagovelia sp.
Ephemeroptera
Ameletus sp.
Baetis sp.
Ephemera sp.
Ephemerella sp.
E_^ excrucians
E_^ lutulenta
£_._ simplex
Hexagenia sp.
Leptophlebiidae
Stenonema sp.
Habrophlebia sp.
Habrophleboides sp.
Lepidoptera
Nymphula sp.
Megaloptera
Corydalus cornutus
Nigronia sp.
Sialis sp.
Pelecypoda
Psidium sp.
Sphaerium sp.
Odonata
Calopteryx (syn. Agrion) sp.
Boyeria sp.
Cordulegaster sp.
Gomphus (Gomphurus) sp.
Hetaerina sp.
Dromogomphus sp.
Plecoptera
Acroneuria sp.
Diploperla, Isoperla sp.
Hastaperla sp.
-------
-85-
APPENDIX C
CORE LOGS FOR STRATA SAMPLES FROM PROJECT WATERSHEDS
-------
-87-
TABLE C-l. LOG OF CORE, JA1
Depth
From
0.0
8.2
9.5
14.0
14.1
23.8
27.5
30.0
(ft)
To
8.2
9.5
14.0
14.1
23.8
27.5
30.0
34.2
Description
Overburden
White to yellow- tan, weathered, friable
sandstone, micaceous
Creamy white, slightly weathered sandstone
Dark- red hematite oxidized iron, Fe20« sandstone
Creamy white to yellow, slightly weathered sand-
stone, 2.1 ft recovered in 9.7 ft, assuming
core loss similar to recovered material
Dense gray to tan sandstone
Coal
Light to medium gray claystone with approximately
horizontal stratification, carbonaceous inclusions
Sample
number
—
JA1-2
JA1-2
—
JA1-1
JA1-1
JA1-3
JA1-4
n
All sandstones are slightly micaceous and show essentially horizontal
stratification. Strata B, C (JA1-2) and E and F (JA1-1) are all similar
sandstones, but show progressively less weathering with depth. This
might be reflected in chemical analysis.
-------
-88-
TABLE C-2. LOG OF CORE, JA2
Depth
From
0
7.0
21.4
23.0
36.4
41.0
43.8
52.4
53.0
55.3
56.5
(ft)
To
7
21.4
23.0
36.4
41.0
43.8
52.4
53.0
55.3
56.5
58.5
o
Description
Overburden
Cream to tan sandstone, crossbedded in part
(very poor recovery = 1.5 ft in 14.4 ft); this
indicates most of internal was highly weathered
and friable sandstones
Yellow to tan siltstone, yellow-like hydrous
oxides of iron
Tan, moderately weathered sandstone
Grayish tan, slightly weathered sandstone
Yellow tan, highly weathered, friable sandstone
Grayish tan, slightly weathered sandstone
Gray sandstone with thin carbonaceous strata
Coal
Dark gray carbonaceous shale
Gray stratified shale with carbonaceous
inclusions
Sample
number
—
JA2-3
JA2-4
JA2-2
JA2-2
JA2-3
JA2-2
JA2-5
JA2-1
JA2-6
JA2-5
Strata B, D, E, F, and G are all sandstones, but reflect varying
degrees of weathering.
-------
-89-
TABLE C-3. LOG OF CORE, JA3
Depth
From
15
18
21
24
30
33
27
9
12
6
3
0
(ft)
To
16
19
22
25
31
34
28
10
13
7
4
3
— . . , *. ,
Description
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Bag sample
Sample
number
JA3-1
JA3-2
JA3-3
JA3-4
JA3-5
JA3-6
JA3-7
JA3-8
JA3-9
JA3-10
JA3-11
JA3-12
-------
-90-
TABLE C-4. LOG OF CORE, Gl
Depth
From
0.0
6.4
6.6
6.7
11.5
12.0
15.6
17.3
18.9
20.0
20.5
(ft)
To
6.4
6.6
6.7
11.5
12.0
15.6
17.3
18.9
20.0
20.5
23.9
o
Description
Overburden
Alternating thin strata of clay and sandstone
(overburden)
Yellow tan, slightly weathered sandstone
(Core loss - assume similar to recovered material
11.5 - 12.0)
Yellow tan, slightly weathered sandstone
Light gray, dense sandstone
Medium gray claystone
Light gray sandstone
Coal (Gl-3)
Light gray claystone with considerable black
carbonaceous inclusions, fossil stems
Light gray siltstone with small amount of
carbonaceous inclusions
Sample
number
—
—
Gl-1
Gl-1
Gl-1
Gl-1
Gl-2
Gl-1
Gl-3
Gl-2
Gl-4
a
Strata C, D, and F are all sandstones, but reflect varying degree of
weathering.
-------
-91-
TABLE C-5. LOG OF CORE, NR1
Depth (ft)
From To
0.0 8.3
8.3 17.7
Overburden
Highly weathered
Description
siltstone interbedded with clay
Sample
number
Bl
soils; horizontally stratified
17.7 21.6 Relatively dense, unweathered siltstone B2
21.6 34.5 Slightly weathered siltstone B3
34.5 42.8 Dense sandstone with thin, essentially horizontal, Cl
carbonaceous strata
42.8 78.0 Shale with thin, horizontal, carboniferous partings Dl
78.0 94.4 Carboniferous shale D2
94.4 94.7 Coal (see E2) El
94.7 110.5 Carboniferous shale (see D» above) D2
110.5 124.0 Slightly carboniferous shale D3
124.0 125.0 Coal E2
125.0 135.0 No recovery
135.0 136.1 Siltstone with carbonaceous partings; insufficient Dl
for sample; similar to stratum DI above
136.1 Bottom of boring
-------
-92-
TABLE C-6. LOG OF CORE, NR2
Depth (ft) Sample
From To Description number
0 277.5 Overburden to top of rock
277.5 284.0 Medium gray, calcareous shale interbedded with NR3-1
thin bands of calcareous siltstone
284.0 285.5 Light to medium gray, fine, micaceous sandstone, NR3-2
slightly calcareous; several very thin shale
scams interbedded.
285.5 286.0 Medium to dark gray shale
286.0 286.4 Intermingled (mottled), dark gray shale and medium
to light gray silt and fine sandstone
286.4 288.0 Light to medium gray, micaceous, fine sandstone NR3-2
288.0 294.9 Medium to dark gray, slightly calcareous shale with NR3-3
band of light gray, fine sandstone interbedded
294.9 297.0 Dark gray, homogeneous siltstone
297.0 300.6 Medium to dark gray, silty shale with thin
interbeds of light to medium gray fine sandstone
300.6 317.0 Dark gray, homogeneous, noncalcareous silty shale NR3-1
317.0 338.8 Dark gray, noncalcareous shale with seams of black NR3-4
fissile oil shale, varying in thickness from a
few inches to 3.5 ft
338.8 342.4 Dark gray shale, mottled with carbonaceous
(coal-like) inclusions
342.4 372.0+ Interbedded medium to dark gray shale, silt, and NR3-3
fine sandstone, with one oil shale bed from 367
to 369 ft
372.0+ B.H. Coal
-------
-93-
TABLE C-7. LOG OF CORE, NR3
Depth (ft)
From To
Description
Sample
number
Top 23.0+
Rock
23.0 31.0
31.0 32.4
32.4 33.5
33.5 34.4
34.4 35.0
35.0 36.4
36.4 59.0+
59.0 60.0+
60.0 63.5+
63.5 65.7
65.7 75.6
75.6 78.9
78.9 82.5+
82.5 85.2
85.2 89.0
89.0 92.0+
92.0+ B.H.
Overburden
Medium gray, fine sandstone with thin beds of NR2-1
siltstone and shale interbedded
Reddish brown, weathered siltstone
Tan sandstone, porous M2-2
Filled-cavity, clayey silt from weathered shale
and siltstone
Grayish maroon, firm sandstone
Filled-cavity, maroon, silty sand
Tan to light brown, medium to coarse, porous, NR2-2
friable sandstone
Filled-cavity, weathered sandstone, clayed, sandy
silt
Dark gray, highly decomposed, shaley siltstone NR2-3
Tannish gray, fine sandstone NR2-4
Dark gray, fissile, oil shale, thin-bedded and NR2-3
highly decomposed
Medium gray, fine sandstone, interbedded with thin NR2-4
bands of gray siltstone
Dark gray, highly decomposed, silty shale M2-3
Tannish gray, fine sandstone NR2-4
Dark to medium gray, interbedded shale and NR2-3
siltstone, highly decomposed
Light to medium gray, interbedded shale, siltstone NR2-4
and fine sandstone
Coal
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-036
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
STRIP MINE DRAINAGE—AQUATIC IMPACT ASSFSSMENT
15. REPORT DATE
Jebruary 1979 issuing date
. PERFORMJNG ORGANIZATION CODE
7, AUTHOH(S)
D. B. Cox, R. P. Betson, W. 0. Barr, J. B. Grossman
and R. J. Ruane
B. PERFORMING ORGANIZATION REPORT NO.
TVA/ONR-79/11
. PERFORMING ORGANIZATION NAME AND ADDRESS
Division of Natural Resources Services
Tennessee Valley Authority
Chattanooga, Tennessee 37401
10. PROGRAM ELEMENT NO.
INE-625B
11. CONTRACT/GRANT NO.
EPA-IAG D8-E721
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Office of Research FT Development
Office of Energy, Minerals 8 Industry
Washington. PC 20460
13. TYPE OF REPORT AND PERIOD COVERED
6/75-12/77
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTES
project is part of the EPA-plafmed and coordinated Federal
Interagency Energy/Environment R£D Program.
16. ABSTRACT " 1 ~
The overall objective of this research program is to demonstrate methodologies for
predicting, on the basis of characteristics of the site to be mined, the impact of
strip mining on downstream biotic communities. To accomplish this objective and pro-
vide data for model verification, sampling programs were initiated at contour- and
area-type mining operations. These programs include streamflow and rainfall gaging
at both types of mines and surveys of fisheries, periphyton, and macwbenthos surveys
at area-mined sites.
Preliminary findings of these field studies indicate that on the Tennessee Cumberland
Plateau (13 drainages from mined areas are alkaline rather than acid, and (2) calcium
and magnesium concentrations increase as a result of mining in almost every instance.
Further, values for iron and sulfate increase in some areas, but not in others,
whereas values for trace metals are generally low in all areas.
Several model components have been developed, including a water quality model for non-
point sources, a continuous streamflow model, and a storm hydrograph model. Other
model components being developed or evaluated include additional small-basin water
quality models, water quality and quantity routing models, a low-trophic-level stream
biota model, and a fisheries resource model.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
tonpoint Source Model*
Stream Ecology,
Overburden, Tennessee,
Surface Mining,
Strip Mining,Extraction*
River Watershed,
[mpact Assessment,
c. COSATI Field/Group
Ecology
Mining
Water Duality
Aquatic Biota
Hydrology
Coal Mining
Hydrologic Models
Mine Drainage
8H
13B
SF~
7B
7D
Bee-logical
19. SECURITY I
Effects
8. DISTRIBUTION STATEMENT
Release to the Public
CLASS (ThisReport}
Unclassified
21. NO. OF PAGES
93
2O. SECURITY CLASS {Thispage)
Unclassified
22. PRICE
EFA Form 2220-1 IS-T3)
------- |