REVISED DRAFT, .DECEMBER 1. 1991
HANDBOOK FOR CONSTRUCTED WETLANDS
RECEIVING ACID MINE DRAINAGE
Edited by
Thomas Wildeman
Department of Chemistry and Geochemistry
Colorado School of Mines
Golden, Colorado 80401
(303) 273-3642
Contract No. CR 815325
SITE Emerging Technology Project
Project Officer
Edward R. Bates
Office of Research and Development
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
REVISED DRAFT, DECEMBER 1f 1991
-------
ACKNOWLEDGEMENTS AND AUTHORS
The Big Five Project team is responsible for the research that went into generating this handbook.
The organizations and participants include:
The Chemistry and Geochemistry Department, Colorado School of Mines. Larry Chang, Dr.
Ronald Klusman, Pat Keller, Leslie Laudon, Steve Machemer, Mahani Mamat, Leslie Moe, Dee Mohamed,
Nits Mohdnoordin, Tom Oliver, Scott Plummer, Chris Sellstone, Dr. David Updegraff, Dr. Thomas
Wilde man
The Environmental Sciences and Engineering Ecology, Colorado School of Mines. Wafa Batal,
Judy Bolis, Dr. Ronald Cohen, Dr. John Emerick, John Dietz, Dr. E. A. Howard, Peter Lemke, Julie
Reynolds.
Denver Knight Piesoid, Denver Colorado. Dr. John Gormley, James Gusek, Dr. Lorraine Fiiipek.
Camp, Dresser, and McKee, Denver Colorado. Dr. Rick Chappell, Susan Morea, Dr. Roger Olsen.
Region VIII, U. S. Environmental Protection Agency. Holly Fliniau
Risk Reduction Engineering Laboratory, U. S. Environmental Protection Agency, Cincinnati, OH.
Dr. Edward R. Bates, James Kreissl, Dr. James Lazorchek, Mark Smith
Colorado Department of Health, Denver, CO. Rick Brown, Jeff Deckler, Jeb Love
The authors of the various sections are:
Dr. Thomas Wildeman: Sections 1,2,3,4,5, and 7
James Gusek: Sections 8,9,10,11,12,13,14,15, and 16
John Dietz: Section 6
Susan Morea: Sections
II
-------
CONTENTS
Acknowledgements and Authors i
Figures vii
Tables xi
1. Introduction 1*1
SECTION A: THEORETICAL DEVELOPMENT
2. Chemistry of Acid Mine Drainage 2-1
Pyrite Oxidation 2-3
Incongruent Weathering 2-4
Mineral Acidity 2-5
Central City Drainages 2-6
Hydrology Related to Acid Mine Drainage 2-8
Summary 2-10
3. Removal Processes in Constructed Wetlands 3-1
Overview of Removal Processes 3-1
Constructed Versus Natural Wetlands 3-3
Exchange of Metals onto Organic Matter 3*5
Sulfate Reduction Catalyzed by Bacteria 3-9
Oxidation and Oxyhydroxide Precipitation Catalyzed
by Bacteria 3-15
Adsorption of Metals by Oxyhydroxides 3-17
Uptake of Metals by Plants 3-18
Other Processes 3-19
Summary 3*19
4. Big Five Wetland: Design, Construction,
Operation, and Results 4-1
introduction 4*1
Design and Construction of the Pilot Treatment System 4-3
Transplanting Vegetation to the Demonstration Site 4-5
General Operation of the System 4-7
Basic Structure of System 4-7
Mine Drainage Distribution System 4-7
Vegetation 4-7
Initial Performance 4-8
Subsequent Modifications 4-15
Cell A Modification 4-15
Plug Flow, Upflow, and Down!low Cells 4*15
Cell B Modification 4-19
Cells D and E Design, Construction, and Operation 4-26
Operations During the Winter 4-27
Guidelines for Winter Operation 4-27
Conclusions 4-28
5. Evidence 1or Sulfate Reduction 5-1
Introduction 5-1
Immediate Operation Even During the Winter 5-1
Laboratory Adsorption Studies 5-2
III
-------
CONTENTS CONTD.
5. Evidence for Sulfate Reduction Contd.
Field Evidence for Adsorption Versus Sulfide Precipitation 5-5
pH Increase of the Effluent 5-12
Sulfur Forms in the Substrate 5-14
Form of Heavy Metals in the Substrate 5-20
Six Step Extraction Sequence 5-21
Five Step Extraction Sequence 5-22
Metal Extraction Summary 5-22
Decrease of Sulfate in the Effluent 5-35
Sulfate Reduction Evidence 5-35
Changes in Sulfate Concentration with Flow 5-36
Summary 5-37
6. Ecophysiological Considerations of Plants at the Big Five
Constructed Wetland, Idaho Springs, Colorado 6-1
Introduction 6-1
Quantification of Biomass 6-1
Effect of Rant Respiration on Eh Values 6-2
Evapotranspiration Measurements 6-4
Elemental Analysis of Plants 6-6
Procedure 6-6
Results
Other Considerations 6"13
Changes in Vegetation 6-13
Physical Effects 6-13
Aesthetic Considerations 6-13
Effects on Fauna 6-14
Conclusions 6-14
7. Area Requirement and Loading Factors 7-1
Discussion of Units 7-1
Volume 7-1
Concentration 7-1
Flow 7-1
Loading Factor 7-2
Surface Flow Systems and Subsurface Flow Systems 7-4
Review of Loading Factors for Municipal Systems 7*4
Review of Loading Factors for Mine Drainage Wetlands 7-7
Early Concepts on Loading Factors 7-7
Area-Adjusted Loadings and Removals 7-8
Loading Factors lor Sulfate Reducing Wetland Cells 7-12
The Limiting Reagent Concept 7-12
Volume Loading Factors 7-13
Recent Example of the Use of the Volume Loading Factor 7-14
Summary 7-16
SECTION B: DESIGN CONSIDERATIONS
8. Regulatory Impacts 8-1
Pre-Construction Issues 8-1
Operation and Decommissioning Issues 8-1
RCRA Waste 8-2
Iv
-------
CONTENTS CONTD.
8. Regulatory impacts Contd.
Water Quality Discharge Standards and the Clean Water Act 8-5
Technology Based Limitations 8-6
Water Quality Base Limitations 8-7
Individual Control Strategies for Point Sources
Causing "Toxic Hot Spots" 8-8
Floodpiains and Wetlands Considerations 8-8
EPA's Wetlands Protection Policy 8-9
State Water Rights 8-9
Endangered Species Act 8-10
Reclamation Bond Release (Post-Mining Land Use) 8-10
Superfund Act 8-10
9. Site Considerations 9-1
Sources 9-1
Flow Rate Variability 9-1
Fluid Collection Alternatives 9-3
Underground Impoundments 9-3
Portal impoundments 9-5
Rock/Pipe Galleries 9-5
Open Ponds 9-6
10. ConstructabiBty-Methods and Materials 10-1
Substrate from Natural Sources 10-1
Hydraulic Conductivity 10-1
Particle Size Distrfoution 10-7
Bio-Compatibility 10-11
Offensive Smells and B.O.O. Considerations 10-11
Organic Content 10-12
Carbonate Sources 10-12
Substrate from Synthetic Sources 10-12
Containment Structures 10-13
Modular Unit Concepts 10-13
11. Conveyances/Flow Control 11 -1
Open Channel Flow Conveyances 11 -3
12. Wetland Design Methodologies 12-1
Area/Flux Method 12-1
Adjusted Loadings and Removal 12-4
Mass Loading Method 12-4
Volumetric Loading Method 12-17
Volumetric Biomass Accumulation Method 12-19
SuKate-Reducing Stoichiometry Method 12-22
Evapotraaspiration Losses 12*23
13. Design Configurations 13-1
General Configurations 13-1
Natural/Conventional Configuration 13-5
Stacked Plate Configuration 13-5
Detailed Configurations 13-6
Flow Directions 13-6
Conceptual Cell Design 13.6
v
-------
CONTENTS CONTD.
14. Instrumentation/Performance Evaluation 14-1
Wetland Cell Instrumentation 14-1
Sampling Substrate for Physical Testing 14-2
Baseline Sampling of Candidate Substrate Materials 14-2
Sampling In-Situ Materials 14-2
Physical Testing 14-4
Specific Gravity of Solids 14-5 -
Bulk Density of Substrate/Water Mixtures 14-5
Material Classftcatbn Content 14-6
Carbon Content (Ultimate Analysis) 14-7
Material Classification/Size Distribution 14-7
Volume Weights, Water Holding Capacity
Air Capacity of Saturated Peat 14-7
Hydraulic Conductivity, Laboratory Methods 14-8
Hydraulic Conductivity, Reid (In-Situ) Methods 14-9
15. System Operation and Maintenance 15-1
Substrate Maintenance 15-1
Maintenance of Conveyances and Flow Controls 15-2
Pipeline Maintenance 15-2
Surface Conveyance/Wetland Containment Maintenance 15-3
16. Cost Estimating 16-1
Capital/Construction Costs 16-1
Preliminary Engineering and Testing 16-1
Environmental Baseline Studies, Permits 16-2
Land Acquisition 16-3
Rights of Way Access 16-3
Final Engineering Design and Construction Specifications 16-3
Construction 16-3
Operating Costs 16-5
References R-l
vl
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
FIGURES
Paoe
Minerals In the ore zones of the Central City
Mining District 2-7
Diagram ol a typical free surface flow wetland 3-2
A Langmulr isotherm for Iron adsorption onto wetland
peal 3-8
Eh-pH diagram tor sulfur species in water 3-10
Plan view and cross section of the Big Five Tunnel
wetland 3-13
Typical location of plant species in each cell 4-6
Iron removal in the Big Five Cells over 1987-88 4-9
Copper removal in the Big Five Ceils over 1967-68 4-10
Zinc concentration versus loading factor for June & July 1S88 4-11
Iron concentration versus loading (actor for June & July 1988 4-12
Present Big Five Site conliguration 4-1S
A cuts-way diagram of the Cell A redesign 4-17
A cut-a way diagram of the Ceil B redesign in the
downflow mode 4-20
Removal of contaminants in Cel B Upflow over 1969-90 4-21
Removal ol contaminants in CeH B Downflow over 1989-90 4-22
Removal of corrtaminants in Cell E over 1989-90 4-23
Adsorption of Mn. Fe, Cu, and Zn versus concentration
in the mine drainage 5-4
Removal ol contaminants in Ceil B Downflow over
the first four months of operation 5-9
Removal of contaminants in Cell E over the first
four months of operation 5-TO
vll
-------
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
FIGURES CONTD.
Eaofi
Effluent pH for Cells B Upflow, B DownfJow, and £
over the first four months of operation. 5-11
Extraction sequence for the forms of sulfur
determination. 5-16
Changes in sulfur content and forms within the substrate
in Ceil A over the first 10 months. 5-18
Changes in sulfur content and forms within the substrate
in Cell B over the first 10 months. 5-19
Six step sequential extraction lor metal speciation in
substrate samples. 5-23
Six step manganese speciation in original mushroom compost
and in substrate from the top of Cell A after 10 months. 5-25
Six step zinc speciation in original mushroom compost
and in substrate from the top of Cell A after 10 months. 5*26
Six step iron speciation in the original mushroom compost
and in substrate Irom tie top of Cell A alter 10 months. 5-27
Six step copper speciation in the original mushroom compost
and in substrate from the top of Cell A after 10 months. 5-28
Five step sequential extraction for metal speciation in
substrate samples. 5.29
Five step manganese speciation in the original mushroom compost
and in substrate from the top of Celt A after 10 months 5-31
Five step zinc speciation in the original mushroom compost
and in substrate from the top of Cell A after 10 months. 5-32
Five step iron speciation in the original mushroom compost
and in substrate from the top of Cell A after 10 months. 5-33
Five step copper speciation in the original mushroom compost
and in substrate from the top of Cell A alter 10 months. 5-34
Sulfur balance in Cell E in October, 1989. 5-38
Change in sulfate concentration versus flow in Cell A. 5-39
vlll
-------
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
FIGURES CONTD.
Page
Change in Eh versus decrease in sulfate
concentration in Cell A. 5-40
Diurnal variation in evapotranspiration from Ceil C over
a 24 hour period, August 21 / 22,1989. 6-5
Comparison of Cu in C. AquatiOs roots from wetland sites
not impacted by mine drainage and from Cells B and C. 6-8
Comparison of Zn in C. Aquatilis roots from wetland sites
not impacted by mine drainage and from Cells B and C. 6-9
Uptake of Cd into roots and leaves of the plant species in
Cell C from October 1987 through October 1989. 6-10
Uptake of Pb into roots and leaves of the plant species in
Cell C from October 1987 through October 1989. 6-11
Uptake of Mn and Fe in Typha roots and leaves in July and
August, 1988. 6-12
A diagram of a typical subsurface flow wetland. 7-3
Different possibilities for modular wetland configurations. 7-6
Decrease in copper concentration in Cell A
versus Flow for 1989. 7-9
Decrease in iron concentration in Cell A versus Flow for 1989. 7-10
Area adjusted removal factor (gdm) for sulfate versus
flow in Cell A for 1989. 7-11
Changes in used substrate disposal alternatives with time
and concentrations of metais. 8-4
A diagram of a typical downfbw laboratory scale
permeameter 10-3
A diagram of a typical upflow laboratory scale
permeameter 10-4
A diagram of a bench scale permeameter 10-5
A cross-section view of a wetland cell
flow oontrol system 11-4
Ix
-------
53
54.
55.
56.
57.
58.
59
60
61
62
63
64
65
66
67
FIGURES CONTD.
Paas
A cross-section view of a constant-prime siphon 11-5
Changes in loading and head losses as depth varies in substrate.
The values plotted are from Table 31. 12-6
Changes in hydraulic gradient with varying permeability.
The values plotted are from Table 32. 12-8
Changes in flow with area keeping alt other variables constant.
The values plotted are from Table 34. 12-11
Decrease In permeability by one percent for each 0.1 feet of depth
of substrate. The values plotted are from Table 35. 12-13
Changes in flow, permeability, and loading with increasing depth.
The values plotted are from Table 36. 12-15
A schematic plan view of a conventional
wetland configuration 13*2
A schematic cross-section view of a stacked
wetland configuration 13-3
A schematic construction detail of a downflow
wetland cell 13-4
A schematic plan view of a downflow
wetland configuration 13-8
A schematic cross-section view of a
downflow wetland installation 13-9
A prototype substrate sampling device 14-3
A downflow laboratory permeameter modified
for fufl saturation 14-10
A typical spool arrangement used to monitor
internal pipeline conditions 15-4
A diagram of the pipe cleaning valve arrangement
used at tee* or wye-intersections 15-5
x
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
TABLES
Page
Concentration of environmentally important
constituents in acid mine drainages and in coal. 2-2
Dissolved constituents in mine drainages of the
Central City Mining District. 2-8
Differences between diffuse and conduit aquifers. 2-9
Metals in various fractions of peat from four separate wetlands. 3-7
Bacterial populations in the Big Five Wetland ceils
during the first year of operation. 3-12
Bacterial populations in the top 1 cm in the Big Five Wetland cells
over 1988 - 89. 3-14
Bacteria populations in the Big Five wetland cells over 1989 - 90. 3-14
A chronological list of activities at the Big Five Wetland. 4-2
Analyses performed on waters, substrates, and plants. 4-3
Contaminant concentrations in the effluents of the
Big Five Cells over 1987-88. 4-13
Contaminant concentrations versus loading factor
in June and July, 1988. 4-14
Contaminant concentrations in the effluents of the
Big Five Cells over 1988-89. 4-18
Contaminant concentrations in the effluents of the
Big Five Cells over 1989-90. 4-24
Results of the adsorption of metals from mine
drainage onto mushroom compost at a pH of 4.5. 5-3
Contaminant concentrations for Cells B-Upflow, B-Downfiow,
and E over the first four months of operation. 5-8
Acid-base characteristics of substrate materials. 5-14
Description of substrate sample locations for the
forms of sulfur sequential extractions. 5-17
xl
-------
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
TABLES
Paflft
Forms of sulfur in substrate samples and NBS
standard coal 1635. 5-17
Results of soluble sulfide titrations on well water
samples from Big Five Cells. 5*17
Results of the six step metal extraction sequence on
substrate samples from the top of Cell A after 10 months
of operation. 5-24
Results of the five step metal extraction sequences on
substrate samples from the top of Celt A after 10 months
of operation. 5-30
Sulfate concentrations and pH in the Cell C effluents
over the winter of 1988-89. 5-36
Results of Eh measurements taken on soil growing Typha
within a controlled growing chamber. 6-3
Conversions of loading factors 7-2
Hydraulic loading rates for the preliminary design of
constructed wetlands for treating municipal wastewaters. 7-5
Constituent concentrations In mg/L in the Quartz Hill Tunnel
mine drainage and in effluents from bench scale tests. 7-14
Humification effects on coefficient of seepage values
of different kinds of peat. 10-6
Typical values of permeability coefficients. 10-8
Permeabilities and sol size fractions from Cell A mushroom compost. 10-10
Estimate of pressure drop across an upf low or downflow
wetland eel using Daisy's Law. 12-3
Estimate of pressure drop and metal loading across an
upfbw or dowrrflow wetland cell of various depths 12-5
Modification of Table 31 to determine minimum
permeability 12-7
Modification of Table 31 to a depth of 3 feet 12-9
xll
-------
TABLES
Uumher Eaas.
34 Modifications of Tables 31 and 33 allowing flow
and flux to vary. 12-10
35 Modification of Table 34 to allow for variation of
the permeability with depth 12-12
36 Modification of Table 35 to allow for changes in flow,
depth, and permeability with varying depth 12-14
37 Projected wetland volume requirements base on void space
availability tor metal sulfide precipitate formation 12-18
38 Effects of biomass accumulation and sulfide precipitation as sources
and sinks of void space on wetland cell design life 12-20
39 Projected wetland volume requirement based on the
stoichiometry of the sulfate-reducing bacteria reaction 12-21
40 Net evapotransplration losses at a hypothetical
constructed wetland site 12-24
xlli
-------
NOTICE
This document is a preliminary draft. It has not been formally
released by the U.S. Environmental Protection Agency and should not
at this stage be construed to represent Agency policy.
It is being circulated for comments on its technical
and policy implications.
SECTION 1
INTRODUCTION
In 1988, two milestone conferences were held on mine drainage and constructed wetlands. In
April of 1988 a conference on Mine Drainage and Surface Mine Reclamation was held in Pittsburgh. The
program chairman, R. L P. Kleinmann assembled an excellent group of mine drainage and constructed
wetlands papers that were published as U. S. Bureau of Mines Circular 9183 (1). In June of 1968, the
International Conference on Constructed Wetlands for Wastewater Treatment was held in Chattanooga,
TN. Dr. Donald A. Hammer served as the program chairman for this conference. The proceedings of this
conference were published in late 1989 (2). The groups from the U. S. Bureau of Mines and the
Tennessee Valley Authority should be considered among the founders in the use of constructed
wetlands for mine drainage.
In the summer of 1987, a pilot constructed wetland was built at the Big Five Tunnel in Idaho
Springs, CO. This is among the first pilot systems to receive metal-mine drainage. Accounts of the first
year of operation are contained in recent publications of the wetlands research group at the Colorado
School of Mines (3,4,5, 6,7,8,9,10,11,12). Currently, the operation of this wetland is being funded by
the U. S. Environmental Protection Agency under the Emerging Technologies Program. One of the
objectives of this current EPA project is to publish a practical handbook on the theory, design and
construction of wetlands for receiving mine drainages.
The monographs by Kleinmann and Hammer have made the production of this handbook quite a
bit easier. Much of what is contained in the following pages is an assimilation of the individual papers in
those two monographs. The primary contribution from the Big Five Study is to put into practice on a metal-
mine drainage the ideas that were developed during those conferences. Also, we find that our study has
generated much fundamental research on wetland processes and design, and these results are
integrated into this handbook. Finally, many triumphs and pitfalls have been encountered during this
project and it Is hoped that our experiences will smooth the route for the others who are considering
constructing a wetland.
What is contained in this handbook is evolutionary. The comments and criticisms of others in the
field of wetlands research and construction are most appreciated.
1-1
-------
1-2
-------
SECTION A
THEORETICAL DEVELOPMENT
-------
SECTION 2
CHEMISTRY OF ACID MINE DRAINAGE
Acid mine waters are not new. Their production was noted in Roman times, and their possible
toxicity was reported by Agricola in Qe Re Metallica (Nordstrom, U. S, Geological Survey, personal
communication, 1989). Research on the refinement of the causes for acid mine drainage production is
also not new. Most of the important ideas on the mechanism of production were generated in the 1960's
and 1970's. The primary reactants are pyrite, water, and ultimately oxygen; and important catalysts are
bacteria, particularly Thiobaciiius ferrooxidans. Since many of the ideas on the cause of acid mine
drainage were established about a decade ago, they can be found in texts and monographs that are often
easier to locate than the primary literature sources. This review will draw extensively on these secondary
sources so that the reader can more readily augment this paper. For each section, the useful monographs
will be cited.
To define the subject, Table I shows the concentrations of constituents that are routinely
determined in coal mine drainages, the constituents in a comparable metal mine drainage, the abundance
ranges of these elements in coals, and the maximum contaminant levels for public drinking water. The
references to the data are noted at the bottom of Table 1. For the coal information, the monographs by
Bouska (13) and Valkovic (14) are useful. Manahan (15) gives a good explanation of the environmental
effects of each constituent.
Although, from different regions and geologies, it is reasonable to consider that the drainage
chemistries of the waters in Table 1 are similar. For the coal mine drainages, the concentrations of the
major contaminants are quite similar in the Illinois and Kentucky coai regions and these compare well with
the values for the whole United States. In the EPA document on effluent limitations for coal mining (16),
tests were made on whether drainages from Western U. S. coal mines and anthracite mines should be
separate categories, and no case could be made for subcategories. The Big Five Tunnel (3) is a metal-
mine drainage and the concentration of most of the constituents fall well within the ranges for United
States coal-mine drainages. Consequently, there appears to be reasonable cause to group all acid mine
drainages together rather than split the waters Jnto a number of categories. If this is done, then
differences from the usual chemistry can be more successfully investigated. Prodan, Meie, and Schubert
(17) give means, standard deviations, minimum values, and maximum values for 110 effluents from
abandoned coal refuse sites in Illinois and the numbers give a good indication of how far waters range from
the median values and the ranges reported in Table 1.
For the metal and ooal mine drainages, the concentrations of Fe, Mn, Al, and SO4- are In the same
range. For a coal seam, the possibility of large abundances of Cu, Zn, Cd, Pb, and As exists. However,
other than the EPA document (16), data on concentrations of these constituents in effluents are difficult
2-1
-------
to find. These heavy metals do exist in metal mine drainages, and Wildeman (20) reviews the possibility of
heavy metals in coal mine drainages. In this section the geochemistry of the weathering of pyrite will be
developed and this will be related to the weathering reactions of the minerals responsible for the other
Table 1. Concentrations of environmentally important constituents in add mine
drainages and in coal. For waters, the concentrations are in mg/L; for coal, in ppm.
Substance
Coal Mlna nrainane Drinking
Unted Big Five Water
States Illinois Kentucky Tunnel Standards
Coal
Al
—
37
—
18
—
14000
Fe
. 0.6-220
57
50-500
50
0.3
16000
Mn
0.3-12
6.4
—
32
0.05
100
Cu
0.01 -0.17
—
—
1.6
1.0
19
Zn
0.03-£2
—
—
10.
5.0
39
Cd
0.01-0.10
—
—
0.03
0.01
13
Pb
0.01-0.40
—
—
0.01
0.05
16
As
0.002 - 0.20
—
—
0.02
0.05
15
PH
32-75
3.0
1.8-35
2.6
6.5 - 8.5
—
SO4-
- 1300
500-12000 2100
250
Ref.
6
1. The 10 to 90 % concentration range of 23 acid drainages from coal mines throughout the United
States taken from the EPA effluent limitations document (16).
2. Median of 110 drainages from coal refuse disposal sites in Southern Illinois compiled by Proudan,
Mele, and Schubert (17).
3 • Regional estimates from Caiuccio and co-workers (18).
4. A typical metal mine drainage from the Front Range Mineral Belt of Colorado
collected by Wildeman and Laudon (3).
5. Compiled from the U. S. Code of Federal Regulations (19). For mine drainages, effluent limits in
mg/L are: Fe, 7.0 daily maximum, and 3.5 monthly average; Mn, 4.0 daily maximum, and 2.0
monthly average; pH between 6.0 and 9.0 at all times. For the other substances in the table,
there are no written restrictions (16).
6. Average of United States coal compiled by Valkovic (14).
2-2
-------
contaminants in mine drainage. The chemistry of mine drainages from the Central City Mining District in
Colorado will be used as an example of how the weathering reactions are interrelated. Finally, the role of
hydrology in the production of acid mine drainage will be reviewed.
PYRITE OXIDATION
In coal mining situations, pyrite is the mineral that is responsible for acid drainage problems. This
same mineral is also the cause of the problems in metal mining situations. Understanding how pyrite
weathers is essential to understanding the causes of the problem and the relations between coal and
metal-mining pollution problems. Stumm and Morgan (21) review the chemistry of pyrite weathering and
the following description is summarized from their text. The overall stoichiometric reactions are:
FeS2(s) + 7/2 02 + H20-> Fe*- + 2S04- + 2H+ [1J
Fe2+ + 1/4O2 + H+ —> Fe* + 1/2H2O [2]
Fe3+ + 3H2O —> Fe(0H)3 + 3H+ [3]
FeS2 + 14 Fe3* + 8 H2O —> 15 Fe2* + 2 SO4* + 16 H+ [4]
The accepted reaction path for the dissolution of pyrite is:
Fe(ll) + S2*
a' + 02(a)
a
FeS2 + 02(g) SO4- + Fe(ll) + H+
+02(g) b d +FeS2(s)
stow fast
c
Fe(lll) <—>Fe(0H)3
Key features of the stoichiometry and reaction path are:
1) Weathering is by oxidation. Since pyrite formation only occurs in a reducing environment, oxygen
2-3
-------
gas from outside the deposit is the ultimate oxidant.
2) Hydrogen ions are produced by the oxidation. For every mole of pyrite oxidized, two moles of H+
are produced by the oxidation to sulfate (Reaction 1), and two moles of H+ are produced upon the
precipitation of ferric hydroxide (Reactions 2 and 3).
3) Since ferric hydroxide is so insoluble, pyrite oxidation is among the most acid producing of all
weathering reactions.
4) The slow step in the reaction path is oxidation in solution of Fe(ll) to Fe(lil). Sulfur oxidation is
relatively rapid.
5) Once the weathering has produced Fe(III). this species can rapidly oxidize pyrite as shown in
Reaction 4 and Step d of the reaction path. Therefore, Fe(lll) cannot persist in the presence of
pyrftic minerals.
6) Step a and step d can be separated in time and space to enable the production of acid drainage
from different environments.
Microorganisms can significantly catalyze the rate of Steps a and d in the mechanism. The
monograph by Erlich (22) is a good review of how weathering reactions can be mediated by bacteria.
Thiobaeillus ferrooxidans can accelerate the rate of Step d by orders of magnitude. Thiobaeillus
thiooxidans can catalyze Step a. Bacteria are necessary to increase the rate of pyrite weathering to the
extent that pollution problems will occur (23,24).
Recent studies on the stable isotope geochemistry of the sulfate in acid mine drainage have
added some refinements to the pyrite weathering mechanism (23,24). Reaction 4 is found to be a major
cause of sulfide oxidation and this reaction does not directly use molecular oxygen. Therefore, flooding
mine workings to eliminate air-pyrite contact may not necessarily stop pyrite weathering. Weathering could
continue by bacterial mediation of Reactions 1 and 2 in the unsaturated zone in the soil, and then by
reaction 4 in the flooded workings. In addition, where pyrrhotite is present along with pyrite, production of
acid drainage is apparently more widespread (Kalin, Boojum Research Ltd., personal communication,
1988).
INCONGRUENT WEATHERING
The concept of congruent and incongruent reactions is important to pyrite weathering and to
reactions that form other constituents in acid mine drainage (25). To demonstrate incongruence, consider
manganese in coal which exists as rtiodochrosite, MnCOa (14). Below pH 4, the MnC03 will reset
accordingly:
MnCOa + 2 H* —> Mn2* + HjO + COg (g) [5]
2-4
-------
C02 gas can escape since it is slightly soluble in water and if this occurs, MnC03 can not be reprecipitated
in an acidic solution. This is an example of incongruent weathering. Some othenreaction or severe altering
of solution conditions is necessary to cause reprecipitation of the reaction products. Reaction 5 is the
basis for how Mn exists in coal mine drainages as Mn2+. Reactions 1 through 4 show that Fe(lll) and SO 4-
in mine drainage cannot be changed back to pyrite through the reversal of a simple reaction.
Other sulfide minerals can weather by congruent reactions. A possible reaction for the
weathering of sphalerite, ZnS, is:
ZnS(s) + H2S < > Zn2+ + HS- + OH- [6]
However if ferric ion is present, it can oxidize the bisulfide ion in the same way as in the dissolution of
pyrite:
8 Fe3* +HS- +4 H2O —> 8 Fe2+ +SO4- +9H+ [7]
In contact with an acid mine drainage solution, ZnS will also be weathered in a manner that cannot be easily
reversed.
Tables 1 and 2 list the chemistry of some acid mine drainages. Fe, Mn, and SO4- dominate the
constituents in coal mine drainages; Reactions 1-5 explain their presence. In drainages from metal and
coal mines, Cu, Zn, Cd, Pb, and As are often present in amounts detrimental to the environment.
Reaction 6 explains the presence in solution of these base metal cations. The presence of Al in mine
drainages is best explained by acidic solutions causing the dissolution of clays (26). Groundwater
hydrology, fluctuations in rainfall, and the manner of ore deposition can also affect mine drainage
chemistry (27). The model for the chemistry of the Argo Tunnel is an example of what these other factors
can do (26). However, Reactions 1-7 are basic to the system and the other factors cause secondary
changes in the rate and extent of these reactions.
MINERAL ACIDITY
Of ail the environmental problems related to mine drainage, the low pH is the most troublesome.
Not only do the pHs of the drainages shown in Table T fall far out of bounds from the drinking water
standards, but also increasing the pH to within the drinking water standards is necessary for long term
removal of all the other pollutants. Consequently, most every pollutant removal method relies on raising
the pH (3, 27). In addition, for acid-base stability of most natural water, buffering by the carbonic acid-
bicarbonate-carbonate system is the most Ifcely method (21). This begins to occur at a pH of between 5
and 6. Any effluent that is released into natural surface waters should be at a pH above 6 to insure that it
will not harm the existing ecosystem.
2-5
-------
However, the low pH is not just caused by the presence of H+ ions. Examination of Reaction 3
shows that Fe(lll) hydrolyzes forming Fe(OH)3 precipitate and H+. Fe(lll) should be considered an acid the
same as H+. Ai(lll) and Mn(IV) will also strongly hydrolyze forming H+. Because the pH of an acid drainage
depends on all these chemical factors, the term mineral acidity or just acidity is given to the situation.
Acidity is operationally defined by how the analysis is conducted (29). In the analysis, hydrogen peroxide
is added, the solution is boiled, and then titrated with standard sodium hydroxide to a pH of 8.2. If the
water contains appreciable concentrations of Fe, Mn, and Al, the solubility products of the these metal
hydroxides will determine which constituents will hydrolyze by a pH of 8.2 and contribute to the acidity.
Using solubility product data from Lindsay (30), Al(lll), Fe(lll), and Mn(IV) will completely hydrolyze
by this pH of 8.2, but Fe(ll) and Mn(ll) at the concentrations encountered in mine drainages will still be in
solution . However, Fe(ll) can oxidize according to Reaction 2 and Mn(ll) can do likewise. These
constituents should be considered potential contributors to the acidity. Since the acidity analysis calls for
addition of hydrogen peroxide and boiling, it is certain that Fe(ll) and Mn(ll) are oxidized to some extent
and counted in the measure of mineral acidity.
Step b in the reaction path shows that the oxidation of Fe(ll) to Fe(lll) is slow and experience with
the treatment of mine drainages shows the oxidation of Mn(ll) to be even slower (28, 31). This slow
oxidation implies a long time release of mineral acidity that can cause the reversal of some treatment
methods that rely on hydroxide precipitation (31). Slow oxidation is also responsible for the persistence
of mine drainage conditions long after the water has breached the surface. For example, the red and roily
nature of surface waters associated with mining is caused by the slow oxidation of Fe(ll) and its
subsequent precipitation as Fe(OH)3.
CENTRAL CITY DRAINAGES
The studies by Wildeman and co-workers (26, 27, 32) on the Central City Mining District in
Colorado give some perspective on how pyrite affects the concentrations of contaminants in mine
drainages. This district is a typical example of a zoned hydrothermal deposit of gold and base metal ores
(33). The distribution of minerals from the high temperature Central Zone to the lower temperature
Peripheral Zone is shown in Figure 1. The chemistry of drainages emanating from mines in the various
zones is summarized in Table 2. The striking feature about the chemistry of these waters is that Cd, Zn,
and Pb are in lowest concentration in the Peripheral Zone even though the ore minerals for these metals
are in highest abundance in that zone. The concentration of all the contaminant metals in the drainages
correlates with the abundance of pyrite in the ore. Fe(lll) and H* in the groundwater catalyze the
dissolution of the other sulfides to such an extent that they become important constituents in the
drainage from a metal mine even though the base metals may be in low abundance in the deposit.
2-6
-------
QUARTZ
PYRITE
CHALCOPYRITE
* TENN ANTITE
ENARGITE
SPHALERITE
GALENA
CARBONATES
BARITE
GOLD
SILVER
Figure 1.
-------
Table 2. Dissolved Constituents in Mine Drainages of the Central City Mining
District from Wildeman et al (1974) and Wildeman (1981). All concentrations
are in mg/L except pH; n. d. means not detected.
MINE DRAINAGES
Constituent Central Zone Intermediate Zone Peripheral Zone
Al
25-100
n. d.
ad.
Fe
200 - 700
2-170
05-4
Mn
90-120
20-40
1.0-5.0
Cu
6-60
0.1-5
<0.01 -0.11
Zn
60-400
7-100
0.3-8.0
Cd
0.2-2.9
< 0.01 - 0.20
< 0.01 - 0.04
Pb
0.1 - 0.5
< 0.01 - 0.20
< 0.01 - 0.06
As
0.2-2.7
< 0.001 - 0.01
<0.001
SO4-
2300 - 4000
900 • 1300
240-800
PH
2.1-2.7
4.0-6.0
5.4 - 6.9
When coal deposits are considered, all the heavy metals listed in Table 1 are associated with the
pyrite and other sulfide minerals in coal and associated overburden (13,14). Consequently, when the
pyrite weathers, the products of that weathering are highly likely to release trace heavy metals from the
coal. Other than the EPA document (16), it is difficult to find information on concentrations of heavy
metals in coal drainages. However, Watzlaf (31) gives some insight into why trace heavy metals have
generally not been measured. In the treatment of acid drainages, manganese'is the most difficult metal to
remove. Investigations by the Environmental Protection Agency (16) found heavy metals in untreated
coal mine drainages. However, it was found that if Mn was reduced to 2 mg/L in the effluent, the heavy
metals were also reduced to acceptable levels. Therefore, limitations on these metals were not
promulgated, and a limitation on Mn of 2 mg/L was established. Watzlaf (31) has determined that this
guideline is reasonable. However, the sludge produced is quite unstable and subject to resolubilization.
HYDROLOGY RELATED TO ACID MINE DRAINAGE
Although the presence of pyrite is definitely the key factor that determines mine drainage quality
issuing from underground adits, there have been some studies that show how groundwater hydrology is
involved. In a long term study of the Argo Tunnel drainage in Idaho Springs, Colorado, Wildeman (26)
found that the chemistry of the water varied little with the seasons and precipitation events. To explain the
2-8
-------
findings he used aquifer models developed to explain the chemistry of carbonate springs in Pennsylvania
(34, 35). Two simple models for groundwater systems are generated: the conduit flow system and the
diffuse flow system. These two models can be treated as the end members of all recharge systems. The
properties that distinguish the two systems are given in Table 3. In both aquifers, recharge is from the
Table 3. Differences between Diffuse and Conduit Aquifers from Wildeman (1983).
diffuse
CONDUIT
1. No response to climatologicai change. 1.
2. Little fluctuation in flow. 2.
3. No suspended solids in the water. 3.
4. Water temperature may not change 4.
throughout the year.
5. Parameters indicative of concentration 5.
such as conductivity (umho/cm) and
hardness do not change with the
climate.
6. Specific concentrations of ions show 6.
little change with the climate.
7. Residence time of months for the water 7.
in the aquifer.
Responds to climatologicai changes.
Obvious fluctuations in flow.
Carries suspended solids at times of
high runoff.
Water temperature changes with the
seasons.
Parameters indicative of concentration
such as conductivity (umho/cm) and
hardness show obvious changes with
storms and runoff.
Specific concentrations of ions show
obvious changes with storms and
runoff.
Residence time of days for the water in
the aquifer.
surface, through the soil vadose zone, and down to the ground water table (26). The Argo Tunnel
drainage is an example of a primarily diffuse aquifer with some characteristics of a conduit aquifer.
A surprising characteristic of a diffuse aquifer is that when annual recharge occurs in spring
and the flow of water does rise slightly, some constituents in the water will increase in concentration.
Wildeman (26) found that all the metais associated with pyrite dissolution increased in concentration
during the spring recharge. He suggested that pyrite weathering is a slower reaction than
carbonate and silicate weathering. If the weathering products are retained in microfaults in the
vadose zone above the water table, then the reaction is most favored. This water is then released
from the faults during spring recharge.
2-9
-------
Many of the adits in Colorado that Wiideman studied are regional systems that serve to
lower the ground water and expose deeper deposits. Some adits in Eastern United States are of
this type, but the greater concern is with overburden exposed during strip mining and with coal
refuse piles. Caruccio and coworkers have done extensive research related to this problem (18,36,
37).
During studies on how overburden related to acid mine drainage, Caruccio (36) noted that
the weathering reactions that produced acidity appeared to be much slower than those that
produced alkalinity. The slow rate suggests that pyrite oxidation is kinetically controlled; whereas
the faster rate for carbonate dissolution suggests it is controlled through equilibrium processes. He
suggested that situations that produced frequent flushing intervals of the overburden should
reduce acid mine drainage. Recently, Snyder and Caruccio (37) tested a comparable hypothesis on
two surface coal mine backfills. Through careful monitoring of the water budget, they were able to
separate the shallow subsurface flow that is associated with rapid recharge from the slow, deep
ground water recharge. They found that the baseflow associated with deep ground water carries
the acidity. This water is associated with the spring recharge which sustains the acid mine seeps for
the rest of the water year. The results of the hydrology studies in Colorado and the Eastern United
States con-elate quite well.
SUMMARY
Whether from coal- or metal-mining situations, the nature of acid mine drainage production
is the same. Pyrite is weathered through oxidation by oxygen with water being a necessary
reactant. Even though base metals such as Cu, Zn, Cd, Pb, and As may be in relatively low
abundances in the deposit, the pyrite oxidation catalyzes the weathering of the sulfides with which
these base metals are associated. As a consequence, environmentally significant concentrations
of heavy metals often occur in acid mine drainages.
The stow kinetics of pyrite dissolution dictates certain environments where the problem will
be most severe. If the pyrite zone is in an unsaturated overburden that contains low amounts of
carbonate minerals, chances for an acid mine problem are significant. Also, if the hydrology is
dominated by long term base flow as opposed to short term recharge, then the possfcility for acid
mine drainage is increased.
2-10
-------
SECTION 3
REMOVAL PROCESSES IN CONSTRUCTED WETLANDS
This section reviews the removal processes that can operate in a wetland. Much of the
information is edited from the reviews by Klusman and Machemer (28) and Wildeman and Laudon (3).
Previous reviews of the removal mechanisms operating in wetlands suggested that removal by humic
material adsoption or through uptake by plants subsequently harvested could be important metal removal
processes (38, 39, 40). Recently, it's been suggested that removal through microbial activity, both
aerobic and anaerobic, may be the dominant removal mechanism (3,41). Whatever the mechanism, there
are reasons why a constructed wetland may remove metals better than a natural one.
OVERVIEW OF REMOVAL PROCESSES
Figure 2 is a model of a typical wetland. Low cost immobilization of pollutants for long time periods
is the goal of using wetlands for mine drainage treatment. Klusman and Machemer (28) list the removal
processes operating in a wetland in the following sequence of decreasing priority :
1) Exchange of metals by an organic-rich substrate, which is usually peat in natural wetlands. 2) Sulfate
reduction with precipitation of iron and other sulfides.
3) Precipitation of ferric and manganese hydroxides.
4) Adsorption of metals by ferric hydroxides.
5) Metal uptake by living plants.
Others would add the following to this list (3,42):
6) Filtering suspended and colloidal material from water.
7) Neutralization and precipitation through the generation of NH3 and HCO3- by bacterial decay of
biologic matter.
8) Adsorption or exchange of metals onto algal materials.
The first five processes will be considered in detail. Filtration is a physical process associated with
wetlands used for polishing treatment. Neutralization is certainly an important process in wetlands used
for municipal treatment (42). It is an unknown factor in wetlands used for mine drainage. There is growing
evidence that algae do remove metals from mine drainages (43,44,45). The significance of processes 7
and 8 needs further study. However, all these processes should also be examined with regard to how a
constructed wetland can be used for mine drainage cleanup.
Geochemical study of the metal removal suggests that removal processes 2,3,4,7 and 8 should
be made dominant. This suggestion is based on what happens to a wetland over geologic time (21,30,
46,47,24), on recent wetland studies (41,48,67), and on recent experience at the Big Five Tunnel site
(8). The basis for this suggestion is explained below.
3-1
-------
DAM
ANAEROeiq
TYPICAL WETLAND ECOSYSTEM
Figure 2.
Diagram of
a typical free surface flow wetland.
-------
If a wetland were buried, upon diagenesis, it would eventually become a bog deposit, coal, or
black shale. (46, 47). Reviewing metals occurrence in these sediment types that have undergone early
diagenesis may identify the metal forms with long term stability. The rationale is that mineral forms for
manganese, iron, and the other base metals in these sediments represent the most thermodynamically
stable phases of these elements. In sediments formed by chemical precipitation, the stable iron minerals
are hematite (Fe^), pyrite (FeS2), or siderite (FeCOa); stable manganese minerals are pyrolusite (Mn02),
and rhodochrosite (MnC03) (21,30,46,47). Trace elements such as Co, Ni, Cu, Zn, Ag, Cd, Au, Hg, and
U occur as sulfides, oxides, and carbonates. The same is true in lignite and coal deposits. With the
possible exception of V and Ni, metals are not retained by the organic fraction in organic-rich reducing
sediments (14,21,46).
The importance of these observations lie in determining the role of organic material in a wetlands
system. If the above observations do indeed point to sulfides, oxides, and carbonates as the most stable
form of trace element precipitates, then immobile organic forms of these elemfnts are intermediate
products that will eventually undergo diagenesis to inorganic precipitates. This implies that the strategy
for optimizing a wetlands system is to concentrate on the formation of inorganic precipitates and use the
organic portions of the system to develop conditions that promote the formation of inorganic precipitates.
Removal processes 2,3, 4, 7, and 8 are those that promote inorganic precipitate formation. Chemically,
this approach to metals removal by a wetland amounts to reversing Reactions 1 through 7 listed in
SECTION 2 and making an ore deposit adjacent to the mine portal.
In the Big Five study, emphasis has been made on the formation of sulfides and oxides. Part of
this section will review the progress made in emphasizing this one process. Other studies that have also
concentrated on specific removal processes in natural and constructed wetlands will be reviewed where
appropriate.
CONSTRUCTED VERSUS NATURAL WETLANDS
There are a number of reasons why use of a natural wetland for mine drainage treatment is not
preferred. It's quite likely that a natural wetland is not available to receive mine drainage. Even if a natural
wetland is available, it may hdve been receiving mine drainage for such a long period that it is close to
saturation (49). In natural wetlands that have peat as the primary substrate, the flow is primarily across the
surface and transmission of water through the substrate is limited. Surface flow diminishes the
possibilities of the anaerobic processes. Also, a natural wetland may be rich in humic acids that limit the
capability to neutralize the acid drainage (28,42, 50). Finally, there's the possibility of destroying the
natural ecosystem by the addition of contaminated waters (39,48). Although natural wetlands have been
used for removal of metal pollutants (51), a constructed system offers more promise for treatment of
heavily contaminated water.
3-3
-------
In the last decade, engineers began to use wetlands for the removal of contaminants from water
(2,52). In some instances, natural wetlands were used. However, a natural system will accommodate all
the above removal processes and probably will not operate to maximize a certain process. If a wetland is
constructed, it can be designed to maximize a specific process suitable for the removal of certain
contaminants from water. Engineering as well as ecological reasons lead to the choice of constructing a
wetland for contaminant removal rather than using an existing natural ecosystem.
As an example of constructing a wetland to maximize specific removal processes, consider the
bacterial processes that are items 2,3, and 7 in the above list. Typical microbial mediated reactions that are
possible in the aerobic zone of a wetland include:
4 Fe*+ + O2 +IOH2O —> 4 Fe(OH)3 +8H+
2O2 +H2S -> SO4- + 2 H+
2H2O +2N2 +502 -> 4NO3- +4H+
Typical microbially mediated reactions that are possible in the anaerobic zone of a wetland include:
4Fe(OH)3 + CH20 + 8H+ ->4Fe2+ + CO2 + 11 H2O
3CH2O ~ 2N2 + 3H2O -> 4NH3 + 3CO2
SO4- +2 CH20 -> HjS + 2HCO3-
In these reactions, "CH2O" is used to symbolize organic material in the substrate.
It is apparent that the anaerobic reactions are approximately the reverse of the aerobic reactions.
Both zones exist in a wetland. If removal involves aerobic processes, then the wetland should be
constructed so the water remains on the surface, if removal involves anaerobic processes, then the
wetland should be constructed so the water courses through the substrate. In a natural wetland, the water
typically remains on the surface. Also, note that the aerobic reactions generate hydrogen ions and the
anaerobic reactions consume hydrogen ions. In the important area of microbially mediated removal, the
wetland must.be constructed to maximize removal reactions and minimize competing reactions. In the
case of removing contaminants from acid mine drainage, it is clear that removal processes should
consume hydrogen ions, consequently anaerobic processes are emphasized (3,41). The research and
development at the Big Five Tunnel site in Idaho Springs, Colorado has concentrated on understanding
the chemistry and ecology involved in removal and designing structures from readily available materials
that maximize these processes.
3-4
-------
EXCHANGE OF METALS ONTO ORGANIC MATTER
Exchange of dissolved metals with the humic and fuivic acids in the substrate is a likely mechanism
whereby the metals are temporarily retained in a wetland (50, 53). Humic and fuivic acids are poorly
characterized natural organic materials that are found in large abundances in wetland substrates,
especially peat. Under acidic conditions, humic acids will be insoluble in water, fuivic acids will be soluble.
Both groups of compounds have carboxyl and phenolic acid groups attached to a larger organic molecule.
The organic acid groups can also be attached to larger humin particles. Since they have an organic acidic
nature, the humic and fuivic acids will dissolve in basic solutions, humin will not. The exchange with metals
is primarily controlled by these acid functional groups and can be described by the following reactions
(28):
The acid portion is represented by the carboxylic group, -COOH, that dissociates to the carboxyl
ion, -COO-, and hydrogen ions (Reaction 1). The R- represents the inert, organic portion of humic or
fuivic acid or humin. Upon dissociation, the carboxyl ion can react with metal ion M* forming a complex.
The reactions are comparable to how lactic or citric acid react with metals in solution. The double arrows in
the two reactions signify that these are equilibrium reactions that can easily be shifted by changes in the
concentrations of substances.
There are a number of factors important to the operation of this system in nature. They all are to
be connected to the concept that the reactions are an equilibrium system. The pKa for acid dissociation of
humic materials averages approximately 4.2. In a mine drainage with a pH of 3, the dominant species in
solution will be carboxylic acid which will not complex the metal ion. Efficient complexation begins
between a pH of 4 and 6 depending on the metal ion (53). Some other process in the wetland is required
to raise the pH of the acid drainage to a more neutral situation. At pH 4.7, the following order from 100 %
complexedtol0%wastound(53): Hg-Fe-Pb«Cu-AI-Cr>Cd>Ni-Zn>Co>Mn. Since Reaction
2 is also an equilfcrium situation, two consequences are possible when a mine drainage interacts with
humic materials. If the peat had sufficient ionization of the acid groups, then when it came in contact with a
solution laden with metal ions, Reaction 2 would be strongly shifted to the complexed metal product. On
the other hand, if a peat that had high concentrations of metals came in contact with an acidic solution,
reaction 2 could be reversed, releasing the metals.
in the study of the interaction of mine drainage with natural organic humic materials, the works by
Weider and co-workers are quite important (40.48). In one study, they performed sequential extraction
procedures on peat from four wetlands (40). The results are listed in Table 4. Red Lake is a peatland that
RCOOH <—> RCOO- +H+
2 RCOO- + M3f < > M(RCOO)2
(1)
(2)
3-5
-------
receives virtually all its mineral matter from precipitation. Buckle's and Big Run Bogs are intermediate but
receive appreciable mineral matter from precipitation. Tub Run Bog receives most of its mineral matter
from an acid mine drainage. In Table 4, the pyrophosphate extractable step releases the metals that are
bound to the humic and fulvic acids. For Fe, Al, and Mn in all four peats, the majority of the metal is
associated with the organic acids regardless of the ecosystem.
Since the humic add is a solid, Reaction 2 can be represented by a Langmuir adsorption isotherm.
Weider and Lang tried this for iron on three of the peats and the results are shown in Figure 3. Use of an
adsorption isotherm allowed them to calculate the maximum amount of iron retained on the peat and this
ranged from 42 to 88 micro moles of Fe per gram of dry peat mass. Upon Fe adsorption, Ca, Mg, Na. K,
and H+ were desorbed (48).
Two important ideas come from these adsorption studies: 1) Adsorption of a metal ion also
implies desorption of another metal or hydrogen ion. 2) There is a limit to the amount of metal adsorption
by humic materials. Nevertheless, Weider suggests that in a man-made Sphagnum wetland, adsorption
onto organic matter will be the dominant removal process followed by formation of amorphous iron oxides
(48).
These studies and others generate a number of conclusions and speculations on removal of
contaminants by adsorption onto organic materials in a constructed wetland:
1) Adsorption works best at pHs that are higher than those encountered in the mine drainages in
Table 1. Some other process and not the adsorption process has to raise the pH.
2) A large mass of humic acid acts as a tremendous acid-base buffer adjusting any incoming water to
»
a pH of about 4. This is probably responsible for the rise in pH when acidic water encounters peat.
However, this also implies that raising the pH to drinking water limits using a peat wetland is
difficult.
3) If conditions change, then Reaction 2 can be reversed, desorbing the metals.
4) Removal of Mn, Zn, and Cd (metals often associated with mine drainages) by organic adsorption
will be difficult.
Currently, it's suggested that organic exchange sites could serve for temporary retention of the
metal cations on the substrate of the wetland (7, 28). This increases residence time for microbially*
mediated metal removal processes to operate. An organic-rich substrate also produces nutrients for
microbes and reducing conditions necessary for sulfate reducing bacteria. However, without additional
processes, the capacity of a wetland to increase pH and retain metals would soon be saturated.
3-6
-------
Table 4. Sequential extraction results for Fe, At, Mn, and S and organic matter
concentration surface (0-20 cm) peat; means ± standard errors (from Weider (40)).
Red
Buckle's
Big Run
Tub Run
Lake. MN
Bog, MD
Bog, WV
Bog, WV
Fe(wimol/g dry mass)
Total (HCI extractable)
45±8
101 ±17
300 ±46
869 ±96
Pyrophosphate extractable
23 ±6
47 ±10
269 ±44
484± 44
Oxalate extractable
8.4 ±1.1
17±2
45 ±9
299 ±42
Dithionite extractable
8.7 ±1.8
24 ±4
17±2
186 ±29
FeSa
2.4 ±0.2
2.1 ±4
6.1 ±0.5
7.6 ±0.8
FeS
1.0 ±0.3
0.3 ± 0.04
3.2 ±1.0
3.1 ±0.3
Al (wimol/g diy mass)
Total (HCI extractable)
65 ± 11
109 ± 16
159 ± 26
243 ±12
Pyrophosphate extractable
35±5
61 ±8
140 ± 24
109± 12
Oxalate extractable
6.2 ±1.5
11.2 ±0.9
12.4 ±2.2
31.9 ±1.9
Dithionite extractable
2.9 ±1.3
7.4 ±2.0
9.3 ±0.7
20.4 ±1.6
Mn (mimol/g dry mass)
Total (HCI extractable)
1.5 ±0.3
5.3 ±0.8
2.5 ±0.4
4.9 ± 2.0
Pyrophosphate extractable
1.3 ±0.2
4.7 ±0.7
2.2 ±0.3
2.5 ± 0.8
Oxalate extractable
0.4 ±0.04
0.7 ±0.2
0.2 ± 0.05
1.2 ±0.6
Dithionite extractable
0.06 ± 0.03
0.6 ±0.03
0.2 ±0.02
1.1 ±0.3
S (mimol/g dry mass)
Total
48.6 ±2.5
87.2 ± 5.4
134±8
108±6
Organic
41.2 ±2.1
77.2 ±5.4
114±8
83.9 ±4.7
FeS2
4.9 ±0.5
4.3 ±0.8
12.2 ±1.0
15.1 ±1.5
FeS
1.0 ±0.3
0.3 ± 0.04
3.2 ±1.0
3.1 ± 0.3
SO42-
1.5 ±0.3
5.5 ±0.6
5.2 ±0.5
5.8 ± 0.5
Organic matter (%)
84.1 ±1.4
86.3 ± 2.3
89.6 ± 2.3
55.9 ±2.8
3-7
-------
80
u
09
BUCKLE'S BOG
BIG RUN BOG
TUB RUN BOG
1000 2000 3000
FILTRATE Fe" CONC. (ftmol/L)
Figure 3. A Langmuir isotherm for iron adsorption onto wetland peat.
-------
SULFATE REDUCTION CATALYZED BY BACTERIA
Generally, microorganisms survive in nature by catalyzing chemical reactions that are far from
equilibrium and that can release energy to the organism upon reaction (15). For example, the formation of
acid mine drainage is significantly promoted by bacteria that subsist on the energy generated by the
oxidation of pyrite (21,23). The distribution of sulfur species with redox conditions (better known as an
Eh-pH diagram) is shown in Figure 4 (21). For sulfate reducing bacteria to operate, Eh and pH conditions
have to be maintained in the field in Figure 4 where sulfide species are stable. This implies acidic waters
that are reducing are most favored, just the conditions that occur in a wetland. The presence of decaying
organic matter in the peat substrate of the wetland rapidly depletes the oxygen and creates acidic soil
waters. The sulfate reduction can be schematically written as follows:
The bacteria most capable of carrying out the catalysis are in the Desulfovibrio family (54). They
need an organic nutrient and this is symbolized by CHaO in the reaction. Lactic acid and cellulose material
are the best for this (54). C ler than this requirement, the bacteria are quite hardy. They will tolerate
temperatures below -5 °C and above 50 °C, and pH's below 5 and to 9.5. Note that by reactions 3 and 4
the anaerobes create their own microenvironment. If the pH becomes too low, H2S and CO2 will exsolve
limiting the pH decrease. If the pH is too high, H2S and H2CO3 will neutralize the base. To limit the
environment from becoming too reducing, a source of Fe is sometimes necessary. Precipitation of FeS
and FeS2 prevents excess sulfide buildup. The only environment the bacteria cannot tolerate is long
periods of aerobic conditions. If the pH fails below 5, the activity of sulfate-reducing bacteria is severely
repressed (54).
In a wetland receiving mine drainage, sulfate reducing bacteria are helpful in two ways. Reaction 3
consumes hydrogen ions, so if the water is highly acidic, the loss of H2S raises the pH. This results In the
rotten egg smell sometimes associated with wetland bottom sediments. If the microenvironment is less
acidic, then reaction 4 generates HS- and this will form highly insoluble sulfide precipitates with Cu, Pb,
Zn, Cd, and Fe. It should be noted that MnS is more soluble and not as easily removed by sulfide
precipitation.
The study of sulfide precipitation in reducing environments is an important field in sedimentary
geochemistry (21, 22, 25, 46). Much of the research focuses on the formation of pyrite. This can occur
directly:
S04«+2CH20+2H+ _Jafleria_>HjS + 2H2O + 2CO2 (pH<7.0)
SO4- + 2 CH2O Jadaoa^> HS- + 2 HCO3-+H+ (pH>7.0)
(3)
(*>
Fe2* + S» + H2S <—> FeS2 + 2 H+
or indirectly through iron monosulfide:
(5)
3-9
-------
1.20
1.00
V
0.80
HSO/
0.60
0.40
S04;
> 0.20
LLi
0.00
:0.20
-0.40
-10
-0.60
0
2
4
6
pH
Figure 4. Eh-pH diagram for sulfur species in water.
3*10
-------
Fe*+ + H2S <—> FeS + 2H+
FeS + S° <—> FeS2 (6)
There is an alternate mechanism that produces highly reactive framboidal pyrite (8). Examination
of Eh-pH diagrams of Fe - S systems show that pyrite is the most stable iron sulfide in typical reducing
environments (21, 28). This is corroborated by how common pyrite is in sediments formed in reducing
environments.
In a review of sulfate reduction, Laudon (8) made these observations:
1) Pyrite formation is limited by the rate of sulfate reduction or by iron availability. Sulfate reduction in
turn is limited by the supply of sulfate or useable organic material. In marine systems, reduction is
limited by organic matter and in freshwater systems, by sulfate availability.
2) Since ail mechanisms for pyrite formation require elemental sulfur, pyrite formation is also affected
by the availability of slightly oxidized sulfur. Because of this, prime environments for pyrite
formation are at the oxic-anoxic interface and around the oxidizing root zones.
3) In freshwater wetlands and salt marshes, seasonal variations in sulfide formation and sulfate
retention are observed. As sulfate becomes a limiting reagent or as redox conditions change,
sulfides oxidize and become a source of sulfate.
Laudon (8) and Hedin (41, 65) came to the same conclusions on how sulfate reduction and
sulfide retention should be favored in a wetland receiving mine drainage. Water levels and chemistries do
not fluctuate throughout the year, and a deep reducing zone is maintained. The supply of sulfate and
reactive Fe(ll) from the acid drainage is abundant as well as the supply of organic matter from the substrate.
Also, the availability of sulfate reducers does not seem to be a problem. As shown in Table 5, in all four
substrate materials used in the Big Five wetland, sulfate reducers were present even after the material was
dried and stored for three months (5).
In the study of a constructed wetland receiving coal mine drainage, Hedin and co-workers (41,66,
67) found all the properties of sulfate reduction operating. The wetland was made of mushroom compost
with hay bales used to increase the path length of the water. In various spots in the wetland and especially
behind the hay bales, black areas showed high pH and low sulfate and iron concentration. In the soil,
pyrite and elemental sulfur were present. Also, in the water from these areas, Mn and Al were greatly
reduced relative to the mine drainage. Hedin concluded (41, 65,66) that in a wetland, sulfate reduction
and sulfide retention have important advantages over other contaminant removal processes.
In the Big Five wetland study, sulfate reduction has been extensively studied (3,4,5,7). Figure 5
is a diagram of this site. Three different substrates were used; Cell A contains mushroom compost, Cell B
contains a blend of equal portions of peat, aged steer manure, and decomposed wood product, Cell C
contains six inches of limestone cobbles overlain by the same substrate as in Cell B. Mine drainage
3-11
-------
Table 5. Bacterial Populations (X10s bacteria per gram) in the Substrates in the Cells
in the Big Five Wetland during the first year of operation.
Type of Pop. of Pop. of Pop. of Pop. of
Substrate Iron Oxidizers Sulfate Reducers Iron Oxidizers Sulfate Reducers
Initial Components, October 1987
Aged Manure 0
0.9
Wood Product 0
0.03
Mushroom Compost
0
5
Peat
0.002
0.03
Peat/Manure/Wood
0
02
Big Five Drainage
0.002
0
Cell
1 Wei
Cel
Wsl
15 cm depth January, 1988
90 cm depth, January 1988
A
3
0.2
10
A
3
0.3
100
A
5
10
30
A
5
6
100
B
3
0.05
10
B
3
0.1
50
B
5
0.08
10
B
5
0.02
20
C
1
0.4
10
C
1
0.04
80
c
4
0.05
9
C
4
0.2
90
15 cm depth, June, 1988
90 cm depth June, 1988
A
3
0.01
10
A
3
0.01
20
A
6
0.02
5
A
6
0.02
20
B
1
0.04
4
B
1
0.01
4
B
6
0.01
2
B
7
0.008
10
C
1
0.03
8
C
1
0.008
10
C
5
0.01
2
C
5
0.01
10
15 cm depth, August, 1988
90 cm depth,
Augus
A
3
0.08
1
A
3
0.008
20
A
6
0.02
5
A
6
0.01
70
B
1
0.01
6
B
1
0.002
20
B
6
0.004
80
B
6
0.02
80
C
1
0.01
30
C
1
0.02
80
C
5
0.002
200
C
5
0.01
50
15 cm depth, November. 1988
90 cm depth, November, 1988
A
3
2.7
2.7
A
3
4
150
A
6
13
7
A
6
2
50
B
1
.4
8
B
1
.4
4
B
6
.4
60
B
6
.03
33
C
1
1.4
20
C
1
.2
110
C
5
.37
61
C
5
.4
100
-------
OUTLETS
PORTAL
- TOP OUTLET
u /
/&¦:
s-JNLET
i^lLr-uiiu Si
HYPALON
LINER
SAMPLING
WELLS
ROCK BOX
AT INLET
Figure 5. Plan view and cross section of the Big Five Tunnel wetland. Sampling location wells for
Tables 5,6, and 7 are shown for each cell.
-------
Table 6. Bacterial Populations (x 10*5 bacteria per g) in the top 1 cm of depth in the Big Five
Wetland cells over 1988-89.
Pop. of Pop. of Pop. of Pop. of Pop. of Pop. of
Cell Well Oxidizers Reducers Cell Well Oxidizers Reducers Cell Well Oxidizers Reducers
August, 1988 November, 1988 January, 1989
A
3
2
0.02
A
3
190
29
A
A
6
0.3
40
A
6
2
8.8
A
B
1
0.5
0.5
B
1
190
35
B
1
86
23
B
6
0.02
0.6
B
6
1.7
8.3
B
6
6.6
18
C
1
0.6
20
C
1
6.6
6.6
C
1
0.74
0.2
C
5
0.2
20
C
5
24
18
C
5
14
7.4
Table 7. Bacterial Populations (x 10-5 bacteria per g) in the Big Five Wetland cells over 1988-
90. For the B Ceils in 1990, (n) is the north part of the cell, and (s) is the south.
Pop. of Pop. of Pop. of Pop. of
Cell Wei Iron Oxidizers Sulfate Reducers Cell Well Iron Oxidizers Sulfate Reducers
15 cm depth, January, 1989
90 cm depth, January, 1989
B 1
.5
.77
B
1
.22
3.6
B 6
.13
.89
B
6
.1
1.4
C 1
1
21
C
1
5.9
2.2
C 5
1.1
4.6
C
5
.65
1.2
15 cm depth, July,
1989
90 cm depth, July, 1989
A 3
<1.7
190
A
3
20
1300
A 6
<1.8
190
A
6
1.8
370
B 1
16
270
B
1
.12
105
B 6
-
160
B
6
-
140
C 1
<1.6
260
C
1
1.1
270
C 5
<1.5
250
C
5
.45
97
15 cm depth, January, 1990
90 cm depth, January, 1990
A 6
2.5
140
A
6
5.6
3.3
B(n) by inlet
5
5.3
B(n) by inlet
.18
4.6
B(n) by outlet
.038
300
B(n) by outlet
.15
3.3
B{s) by inlet
1.6
110
B(s) by inlet
.55
1.2
B(s) by outlet
.13
.63
B{s) by outlet
.3
3.6
C 5
.53
620
C
5
.43
330
E middle
.11
340
E middle
1.9
28
15 cm depth, August, 1990
90 cm depth, August, 1990
A 6
1.1
280
A
6
5
39
B(n) by inlet
<.016
80
B(n) by inlet
.42
2.3
B(n) by outlet
<.014
26
B(n) by outlet
.17
9.7
B(s) by inlet
<.014
1
B(s) by inlet
.11
1.7
B(sj by outlet
.04
.43
B(s) by outlet
.098
2.7
C 5
.88
37
C
5
.083
16
E middle
<.014
31
E middle
.19
7.4
3*14
-------
started flowing through the system in October of 1987, and almost immediately, Cell A became more
efficient at removing contaminants than the other two cells. Since the system started in the winter
when the plants were dormant, many possible processes were eliminated. During the first five months
of operation, the water soil interface in Cell A changed from one that was oxidizing and orange in color
to one that was black and reducing. Table 5 show the levels of bacteria in the three cells through the
first year of operation (5,10); the positions of the wells are noted in Figure 5. The population of sulfate
reducers is high in all three ceils and higher in Ceil A in the first few months of operation, in all cells,
sulfate reducers are ubiquitous throughout the substrate; however, Cell A turned anaerobic
immediately as seen by the high concentration of sulfate reducers in the first six inches of the substrate
and, in August of 1988, in the top 1 cm of the substrate (see Table 6). Also, it should be noted in Table
7 that the populations of sulfate reducers is maintained in 1989 and 1990, whereas the population of
iron oxidizers decreases from the levels of 1988.
The Big Five wetland will be discussed in greater detail in the next section. However, sulfate
reduction is certainly occurring in these constructed sites and it appears that this process is
responsible for the better performance of Cell A.
In conclusion, sulfate reduction and sulfide retention are processes that do operate in wetlands
and generate preferred modes of contaminant removal. Sulfate concentration is reduced, pH is
increased, and the metals are removed by the formation of highly insoluble precipitates.Since this
process has only recently been investigated in a formal manner in constructed wetlands, the complete
nature of the removal process is still uncertain. Two key factors that.will determine the success of this
process are the need for the drainage to flow through the anaerobic portion of the wetland and the
need for continuous flooding of the wetland so that the products of reduction are not oxidized.
OXIDATION AND OXYHYDROXIDE PRECIPITATION CATALYZED BY BACTERIA
In SECTION 2, it was noted that the mechanism for pyrite oxidation favors the use of Fe(lll) for
further oxidation and thus, mine drainage has a significant concentration of Fe(ll). Also, manganese is
in the +2 oxidation state instead of 44. The hydroxides of Fe(ll) and Mn(ll) are far more soluble than
those of the higher oxidations states (30). At pH ¦ 8, both the +2 ions have solubilities of greater than
1000 mg/L. On the other hand, Fe(lil) and Mn(IV) are highly insoluble. At pH - 5, their hydroxide
solubilities are less than 1 mg/L. Efficient removal of Fe and Mn by hydroxide precipitation requires
oxidation. In a wetland, this is most readily achieved by microbial catalysis in the aerobic zone.
The geomicrobiology of iron and manganese is reviewed by Erlich (22). There are a host of
bacteria that oxidize iron. The most important of these is the Ferreoxtdans family that starts the
oxidation of pyrite. It is one of the few bacteria that can tolerate a pH less than three and can exist on
the inorganic nutrients of Fe(ll), JO2. and NH4+. As shown in Table 5, Ferrooxldans were found in the
3-15
-------
Big Five drainage and in the peat which had recent contact with natural waters. As shown in Tables 5,
6, and 7, there was a high original concentration of Ferrooxidans in all three cells at the Big Five site.
These bacterial populations were maintained, even at a depth of 90 cm into the substrate for the first 6
months. However, as seen in Table 7, populations started to fall in 1989 as the cells turned more
anaerobic, by 1990, it was difficult to find Ferrooxidans at depths of 90 cm. When the pH of a system
reaches 5, a large number of species of bacteria are capable of oxidizing Fe(ll). These bacteria require
organic material as a nutrient source.
For manganese, oxidation is more difficult. Mn(ll) is stable to higher pHs than Fe(ll) (21); and
thus, bacteria that can use the oxidation of Mn(ll) as a source of energy have to operate in systems with
higher pHs. Bacteria that directly use the oxidation of Mn as a source of energy do not seem significant
in acidic wetland environments. Bacteria that indirectly oxidize Mn(ll) are most readily found in
environments above a pH of 7 {22). Further reasons for the resistance of Mn (II) to microbial oxidation is
the high activation energy for the reaction and the complexity of the mechanism.
The Tennessee Valley Authority currently is operating seven and planning eight constructed
wetland systems for treating coal mine drainage (68,69). The primary design for these wetlands is a
surface treatment system using whatever substrate is readily available. The systems are large, typically
contain a number of stages, and closely resemble a natural wetland. In these systems it is suggested
that microbially catalyzed oxidation processes are the primary removal processes. Five of the operating
wetlands have produced effluents meeting all discharge limits. These five systems are associated with
moderate inflow quality (Fe ¦ 11-69 mg/L and Mn - 5-14 mg/L), relatively large Mn to Fe ratios (average
Mn/Fe - 0.44), and significant inflow alkalinity as evidenced by the pH of the inflow being above 5.5
(68). As explained in SECTION 7, the treatment area requirements for these systems range from 0.6 to
3.4 nftrng/min.
Reactions 2 and 3 in SECTION 2 show the oxidation and precipitation of iron. The comparable
reactions for Mn are:
Mn2* +1/2 02 +2 H+ <—> Mn4+ + HjO (7)
Mn*- + 2 H2O <—> Mn02 +4H+ (8)
As in the case of iron, precipitation generates hydrogen ions, and this limits the extent of oxidation and
precipitation.
The importance of oxidation and precipitation in wetlands is considerable, especially for Mn. In
natural wetlands where the water is primarily flowing across the surface, Weider considers it just as
significant as adsorption by organic material for the removal of iron (48). However, the inefficiency of
the process for manganese removal is one of the great frustrations in the use of wetlands for
3-16
-------
contaminant removal. Klusman (28) reviewed this frustration. In active mine drainage treatment
systems he states that although the rates of oxidation of Mn and Fe are both pH dependant, iron
oxidation occurs at an acceptable rate at pH values near 7. The rate of Mn oxidation cannot be brought
into an acceptable range without raising the pH above 9.0.
Based upon these ideas, some conclusions can be made concerning the use of oxidation and
precipitation in wetlands.
1) Since the oxidation and precipitation sequence generates hydrogen ions, the reactions will be
self limiting. Some other process has to be operating to raise the pH.
2} In peat wetlands, where sofl waters remain acidic, there is little possibility of Mn removal.
3) Sulfate reduction and metal oxidation are mutually exclusive processes. One requires
anaerobic conditions; the other aerobic. In constructed wetlands, this implies the design
should be a staged system. Since sulfate reduction raises the pH, the stage promoting this
process should be first.
4) Even though it is difficult to remove Mn by oxidation, this process has the best chance of
success since MnS is relatively soluble. However as shown in SECTION 7, it does appear that
manganese can be removed in an anaerot sell. It is hypothesized that MnCOs is being
formed.
5) Most of the area in a constructed wetland will be dedicated to the removal of Mn. Brodie and co-
workers (55) suggest the area needed per milligram of Mn is about 2.5 to 3.5 times that
needed per milligram of Fe.
ADSORPTION OF METALS BY OXYHYDROXIDES
When Fe(lll) and Ai precipitate as hydroxides, the solid is quite gelatinous. The scavenging
properties of these precipitates have long been used in wastewater treatment (15). In addition, these
two metal hydroxides as well as the Mn(IV) oxyhydroxide have a strong capability of adsorbing other
metal ions onto their surface (21). Consequently, if the process of oxidation and precipitation occurs, a
side benefit of further removal of contaminants by the precipitates will also occur.
When these precipitation reactions occur, the products are not simple crystals. For iron, the
reaction sequence is roughly as follows (30):
Fe3* +3 OH- <—» Fe(OH)3 (amorphous) (9)
The amorphous hydroxide is the first form that precipitates. It has some polymeric properties as does
aluminum hydroxide and this causes the gelatinous appearance. As the hydroxide ages it turns into
Fe(OH)3 <—> FeOOH +H2O (goethite)
2 Fe(OH)3 <—> Fe2Q3 +3 H2O (hematite)
(10)
(11)
3-17
-------
crystalline hematite in dry conditions or goethite in moist situations. The hematite and goethite have
better capabilities of adsorbing other trace metals (21). There are a number of theories of how the
surfaces of oxyhydroxides adsorb metals (21). Basically, the surface operates as a weak acid which
attracts hydroxide ions making a negative surface. The negative surface of the particles attracts the
positive metal ions. In this explanation, the hydroxide surface changes from positive to negative as the
pH increases. The pH where there is no surface charge can be measured and this is called the pH of
zero point of charge (zpc). For manganese (IV) oxyhydroxides the pH of zpc ranges from 3 to 7; for
aluminum oxyhydroxides the pH of zpc ranges from 5 to 9; and for iron oxyhydroxides the range is from
6.5 to 8.5 (21). Since the surface of the manganese oxyhydroxide turns negative at lower pHs, it is
generally a better absorber of cations than the other two solids. This is verified in studies of trace
element relations with Mn and Fe oxyhydroxides (31,56,57).
The process of metal adsorption can be used in a wetland in the polishing and buffering
stages. In the final portions of a wetland where hydroxide precipitation may be significant, these
hydroxides will help to coagulate suspended material in the water. At the same time, these precipitates
will help remove the final metal contaminants.
A common feature seen in mine drainages is the accumulation of oxyhydroxide precipitates by
algae. There is a question of whether this is assimilation or preferred precipitation but nevertheless
some strains of algae become quite orange with Fe(lll) hydroxide coatings. Some projects have been
carried out to investigate how algae can help the removal process. In Ontario, Kalin is using algae to
increase surface areas for better Fe(lll) precipitation and is also studying whether algae assimilate
metals. In Missouri, Wixon (43) has used algae to polish mine waters that are discharged from a settling
pond. Kepler (45) is studying whether algae ponds are better than cattail ponds for removing Fe and
Mn from coal mine drainage. Further work is needed to determine the relative amount of assimilation
compared to precipitate accumulation. However it appears that algae are helpful in accelerating the
final removal processes that are associated with oxyhydroxide precipitation.
UPTAKE OF METALS BY PLANTS
In the earlier studies using wetlands to treat add mine drainage, It was presumed that uptake by
plants was an important process (38,39,41). However, recent results have shown that uptake by the
stems and leaves of plants account tor only 1 to 5 % of metal accumulation (49,58,59). Metal removal
by the roots and rhizomes of plants may be significant. The roots themselves do not seem to
accumulate metals (49,58,59), but they do generate mteroenvironments that promote the reduction
and oxidation processes. In one case it was observed that the soil below Tvpha roots was more
reducing (58). Also its been found that oxygen can be respired down the stems of Tvnha causing the
precipitation of Fe(lll) around the roots (60). This property of oxygen transpiration to the soil could be
used to advantage in wetlands that rely on oxidation and precipitation.
3-18
-------
Another significant role of the plant material in a wetland is to provide the biomass necessary
for the other processes. Decayed plant material produces the organic matter that will be capable of
removing metals by adsorption and exchange. The cellulose in plants provides the nutrients for the
sulfate reducing microbes. Also, even a constructed wetland should be considered an ecosystem
where the plants provide the long term nutrients and vegetative cover for the substrate which is
providing the treatment.
The studies on plants in The Big Five wetland Is summarized in SECTION 6.
OTHER PROCESSES
Are there any processes still to be discovered that will be important to the contaminant removal
process? Based on the explosion of studies on water treatment by wetlands, it's certain that new and
significant processes will be uncovered. Also, since much ol the recent research is on microbial
processes, it's quite likely that new microbial processes will be discovered. In the area of metals
removal, an important process that is still uncertain is the increase in pH. Earlier, it was suggested that
sulfate reduction is responsible for this pH rise. However, that has not been verified. It is highly likely
that microbes are involved in this process.
Two microbial processes deserve mention as possible candidates. In the microbial
degradation of protein, NH3 is generated which will hydroiyze to NH4OH. This process win raise the pH.
The proteins would be part of the substrate materials. Another more speculative microbial process is
suggested as part of the microbial reduction process. Apparently, microbes exist that generate H2gas
from hydrogen ions (Or. D. R. Updegraff, Colorado School of Mines, 1989, personal communication).
These bacteria live in concert with sulfate reducers and methane generators which use the H2 as a
nutrient. If these microbes do indeed exist in a wetland, their use in raising the pH is obvious.
SUMMARY
Upon review of the possible wetland processes some guidelines do become clear. Among
the most important is that almost all removal processes are associated with the wetland substrate.
However, little is known about that substrate. In particular, knowledge about the substrate as a nutrient
for growth of microbes important to metals removal processes is sparse. Also few advances have been
made in determining how to use the substrate as a vehicle for increasing the pH.
On the question of which processes are or are not important, the answer is all processes are
important. The wetland design problem then becomes how to develop stages that will take advantage
of the group of processes that will best do the specific metals removal process. The design of stages
and separation of processes is readily divided into those that operate in an anaerobic system and those
that best operate in an aerobic system. This question of design of stages is being investigated at the
Big Five site and preliminary results will be developed in the next section.
3-19
-------
SECTION 4
BIG FIVE WETLAND: DESIGN, CONSTRUCTION, OPERATION, AND RESULTS
INTRODUCTION
The Idaho Springs-Central City mining district in the Front Range has massive waste rock dumps,
mill tailings piles, and abandoned'mine shafts and tunnels from precious metal ore production (33).
Tunnel drainage typically has low pH and high metal concentrations that affect the regional aquatic
resources. Table 1 is typical of the chemistry of the drainage water. The Clear Creek site, which includes
the Big Five Tunnel, is on the National Priorities List under the Comprehensive Environmental Response,
Compensation, and Liability Act of 1980 (CERCLA or Superfund). When CERCLA was amended in 1986,
Congress placed a special emphasis on the use of alternative treatment technologies for cleaning up
hazardous waste sites. As part of the Feasibility Studies required by the Superfund Program, it was
recommended that passive treatment by a constructed wetland be considered as a cost-effective option
for treatment of acid mine water associated with the Clear Creek site (70). To assess the feasibility of
wetland treatment, the Big Five Tunnel Pilot wetland site was constructed in the summer of 1987.
From June 1987 through September 1988, study of the wetland was funded through Region VIII
of the EPA. Beginning in October, 1988, funding of modifications and monitoring of the Big Five site was
through the Emerging Technologies Program (ETP) of the U.S. EPA Superfund Innovative Technology
Evaluation Program (SITE). This handbook was written as part of that (ETP) project. Table 8 is a
chronology of activities at the Big Five Tunnel. Table 9 is a list of the analyses performed during the
project. In the 1987-88 portion of the project, the analyses were performed under the Contract
Laboratory Program (CLP) of Superfund. Under the ETP Project, a Quality Assurance Project Program
(QAPP) specific to the project was established. The analyses were performed at the Colorado School of
Mines and at the laboratories of U.S. EPA Region VIII. A separate report on Quality Assurance/Quality
Control that includes all the data on water and soils is available.
During 1987-88, monthly routine sampling of the wetland effluent.?, quarterly sampling of cell
wells, and six month sampling of cell soils was performed. Under the ETP project, routine sampling of cell
effluents was performed bimonthly and soils were sampled twice a year. The analyses performed on a
routine basis are listed in Table 9. In addition to the routine studies, a number of special research studies
were performed. During the summer of 1988, a study was conducted of how flow affects concentrations
of metals in the effluents was performed. Laudon (8) studied the forms of sulfur in the substrates in the
wetland cells. Batal (10) studied the changes in microbial populations in the cells. Lemke (9) studied how
hydraulic properties of the substrates affect the operation of the wetland cells. Machemer (7) has been
studying the role of sulfate reduction in the operation of the wetland cells. Dietz has been studying the
role of plants in the wetland cells and these results are presented in Section 6. In this section, the
significant results from routine sampling are presented. The results of the special studies are also
summarized.
4-1
-------
Table 8. Chronological list of activities at the Big Five Wetland Site.
1987
Jun. 15-Aug. 15: Sampling, selection, and analysis of candidate substrate materials.
Selection of sites to secure plants.
Aug. 1*22: Preparation of site, installation of plumbing, cells, and substrates.
Aug. 25-Sep. 30: Substrate soaked with municipal water. Transplanting and sampling of
sedges, rushes, and cattails.
Oct. 13-15: Sampling of mine drainage, cell effluents, and cell wells to establish
baseline conditions.
Oct. 25: Flow of mine drainage into cells initiated.
Nov. 3: Routine sampling of effluents, wells, and substrates initiated.
13M
Jan. 1: Routine sampling under the Region VIII project continued.
Jun. 15-Jul. 20: Special flow rate versus effluent concentration study conducted.
Jul. 1-Aug. 31: Special sampling and analyses for forms of sulfur in the substrates
conducted.
Oct. 1: ETP Project begins.
Nov. 1: Routine sampling under the ETP project initiated.
Dec. 1-15: Reconstruction of Celt A carried out.
1989
Jan. 1: Routine sampling and analyses under the ETP project continued.
Jan. 1: Studies on hydrologic properties of substrates initiated.
Jun. 15-Aug. 1: Special studies on the role of plants conducted.
Aug- 1*31: Reconstruction of Cell B and construction of Cells D and E carried out.
Sep. 1-Dec. 1: Special studies on substrate processes conducted.
1990
Jan. 1: Routine sampling and analyses under the ETP project continued.
Jun. 1-Sep. 15: Special study of the sulfate reduction process conducted.
Jun. 15: Bench scale studies on Quartz Hill and National Tunnels initiated.
4-2
-------
Table 9. Analyses Performed on Waters, Substrates, and Plants.
During 1987-88 under the CLP Program
Waters
|
Substrates
Plants
Field
RAS*
nh3
RAS*
RAS*
PH
CO32-
TSS
B
B
Eh
HCO3-
TDS
NH3
cond.
cr
B
Ntot
flow
F"
TOC
CtOt
temp.
S042*
acidity
Ptot
NO37NO2"
Sforms
During 1988-90 under ETP Project
Waters CSM
Waters EPA
Substrates CSM Substrates EPA
Plants
Field
Mn
RAS*
Mn
RAS*
RAS*
PH
Fe
cr
Fe
NH4
B
Eh
Cu
r
Cu
Ntot
P
cond.
Zn
S04-
Zn
ptot
Flow
SO4-
N03/N03-
S
temp.
NH4+
*RAS (Routine Analytical Services) includes: Al, Sb, As, Ba, Be, Cd, Ca, Cr, Co, Cu, Fe, Pb, Mg, Mn, Hg,
Ni, K, Se, Ag, Na, Tl, V, and Zn.
DESIGN AND CONSTRUCTION OF THE PILOT TREATMENT SYSTEM
The completed design of the pilot treatment system was a reinforced concrete structure with
dimensions of 0.61 m (2 ft) in depth, 3.05 m (10 ft) in width, and 18.3 m (60 ft) in length. For the initial
investigations, the structure was divided into three 6.1 m (20 ft) sections, with provisions to divide the box
into six 3.05 m (10 ft) sections at some later time if this were to be desired (Figure 5).
The concrete sections were separated by walls constructed from 5 x 15 cm (2 x 6 in.) treated
wood. Aluminum channels were grouted into void tubes in the concrete walls to allow the addition of
lumber to form sidewalls and endwalls of adjustable height. In the initial study, the walls were built up to a
height sufficient to allow the total depth of the cells to be 1.22 m (4 ft).
4-3
-------
Each cell was fitted with two drains, one active and one reserve. The reserve drains were installed
so that the number of cells could be changed from three to six if desired, and to drain the cells at the end
of the study. The drains were built using 15 cm (6 in.) i.d. polyvinyl chloride (PVC) pipe, and the active
drains consisted of standpipes initially set at a depth of about 1 m (3 ft}. The drains deliver the overflow
water to an existing runoff pond. A 0.76 mm (30 mil) Hypalon™ liner was used to line the cells so that they
would be separated from one another and to prevent chemical reactions between the treated wood,
concrete or aluminum channels and the organic substrates and mine drainage.
Rock baskets were constructed at the upstream end of each of the cells to allow the mine
drainage to contact as much of the upstream cross-section of the organic substrate as possible. These
baskets, approximately 30-45 cm (12-18 in.) thick, were built using expanded plastic fence and extended
to the full depth and width of each of the cells. Washed 10-15 cm (4-6 in.) river rock was used to fill the
baskets. Plastic curtains were suspended from supports just above the substrates on the downstream
side of the rock baskets. These curtains extended down to 1/2 to 2/3 of the total depth to force the flow
downward into the cells. Six access wells were installed in each cell to allow sampling of interstitial water,
the location and the number of the wells is shown in Figure 5. These sample wells were made from 15 cm
(6 in.) i.d. PVC and completed to allow water to enter from the lowest, middle, and the upper 30 cm (1 ft) of
the nylon screen to prevent clogging with the organic material. Two wells of each completed depth were
placed in each of the 3 cells, for a total of 18 sample wells.
A small concrete dam was constructed just inside the tunnel portal to provide enough head to
distribute water to the system cells 50 feet away and 2 feet downgradient. Water was piped from the portal
to each of the cells through 2 inch diameter PVC lines, reduced in size through the system, and fitted with
valves to control the total flow and the flow to each individual cell. Due to the harsh winter climate of the
location, all plumbing had to be insulated. Water is distributed across the entire width of each cell by
allowing it to flow into the rock baskets through 10 cm (4 in.) i.d. PVC perforated drain pipe, which extends
from one side of the cell to the other. Excess water from the tunnel is allowed to drain into an adjacent
pond, which percolates into nearby Clear Creek.
Once the sample wells were placed and the rock baskets were completed, the sections were filled
with the organic substrates to a depth of about 1 m (3 ft). The first cell was filled with fresh, unused
mushroom compost, which consisted of approximately 50 percent animal manure and 50 percent barley
mash wastes from a local brewery. The second cell received a mixture of equal parts of peat, aged steer
manure, and decomposed wood shavings and sawdust. The third cell was filled with the same mixture as
the second cell, except that the third cell was filled to a depth of 15 cm (4-6 in.) with 5-8 cm (2-3 in.) of
limestone rock before the cell was filled with the organic mixture. Initially, the organic substrates were
saturated with municipal water to reduce stress on the transplanted vegetation.
4-4
-------
TRANSPLANTING VEGETATION TO THE DEMONSTRATION SITE
The transplanting of cattail, sedge, and rush species was initially envisioned to be a moderately
easy process of taking samples of the plants about 30 cm in diameter and transplanting them to the
treatment beds. In practice, however, separating the root systems around and underneath a section of
plants to detach them proved to be very difficult since the root systems of these species are extremely
intertwined and the soil quite heavy in some cases. The sizes of the vegetation samples were ultimately
determined by the weight that two people could lift without causing the sample to disintegrate. The
practical method was to cut the intertwined roots around a section of plants and to lift the edges of the
sample while other workers separated the sample from the underlying soil or gravel. Once free, the
samples were lifted onto plastic tarpaulins to prevent them from fading apart and to facilitate lifting them into
the trucks. Vegetation samples were then hauled to the mine drainage demonstration site and placed in
the treatment cells, which had been previously saturated with municipal water.
Several different species of aquatic plants were transplanted into the treatment cells. Cattails
fTvpha ananstifolia. T. latifolia). and sedges ftorm iitririilata, C. aouaiilisl were transplanted from an area
of similar elevation near the northeast shore of a mountain lake in Grand County, Colorado. Since this site
was on public land, a required permit was obtained from the U.S. Forest Service before vegetation was
transplanted from this site. Four light trucks were used, six workers were involved, and the work took
about ten hours. This effort included the travel time, and extracting and transplanting the cattails and
sedges. About 25-30 percent ol each of the treatment beds were covered as a result of this work.
On September 5, sedges IC. aonatilisl and rushes (Juncus arcticusi were transplanted from a
wetland about 6.5 km (4 mi) up Stanley Road from the demonstration site. A one-ton stake bed truck with
a hydraulic lift was used, which greatly facilitated the loading and unloading ot the vegetation samples.
About 50 percent of the treatment beds were covered at that time.
The remaining transplanting was done on September 18. Cattails were transplanted from a
foothills wetland in Mount Vernon Canyon, and sedges were transplanted from a wetland adjacent to the
demonstration site. The effort took about eight hours using two trucks and four workers. As a result of
this transplanting, about 85 percent of the treatment beds were covered with species of either cattails,
sedges or rushes. Figure 6 shows the general location of the different species of vegetation in the
treatment rails.
After the transplanting was completed, municipal water was mn through the system until October
25, 1987 when acid mine drainage was diverted into each of the cells. The initial flow rate was 3.8
liters/min (1 gpm) which was equivalent to 200 ft^/gpm.
4-5
-------
outlet
l C.aq.(2)
X
C.aq.(4)
C.ut.(1)
T.lat.(1)
T.lat.
&
C.aq.(3)
J.arc.(2)
C.ut.(1)
T.lat.(3)
C.aq.(2)
T.lat.
&
C.ut.(1)
inlet
box
Figure 6. Typical location of plant species in each cell.
-------
GENERAL OPERATION OF THE SYSTEM
Basic Structure of System
The basic structural components of the Big Five system appear to be functioning as designed.
The concrete structure and the separating walls held by the aluminum channels are performing well. The
30 mil HypalonR liner used to line the cells remains intact and no leakage from the system has been
observed. Recently, the liner withstood a severe hailstorm that shredded 10 mil polyethylene.
Mine Drainane Distribution System
The dam and plumbing constructed inside the Big Five adit continues to function adequately in
diverting the desired portion of flow to the cells, although the area behind the dam is slowly filling with
metal hydroxide sediment. The PVC lines that distribute the mine drainage to the individual cells are
insulated adequately, since no freezing of the inflow water has been observed through two winter
seasons. The standpipe drains continue to work well in all three cells and clogging has not been
observed.
Metal hydroxide precipitates, however, occasionally clog the mine drainage inflow lines. Even
though taken from the surface of the impoundment inside the portal, the incoming flow still retains
enough metal hydroxide sediments to clog the lines. The lines have to be periodically flushed to remove
the sediments in order tcmaintain the desired flow rates to each of the cells. The reason for the clogging
appears to be a trap created by the requirement for installing a vertical section of pipeline to lift the flow to
the top of the cells. The problem could be alleviated by avoiding sharp turns and vertical sections in the
incoming lines. The clogging of the lines turned out to be serendipitous, however, in that much larger
increases in pH values are found when the flow is reduced. This observation led to further studies to
quantify the metal removal effitiences and increases in pH values under different flow rates and retention
times.
The rock baskets fill with metal hydroxides after a few months of operation and possibly reduce
the opportunity for the mine drainage to contact the entire cross-section of the substrate. The layer of
limestone rock in Cell C appears to be somewhat more effective in distributing the flow through the lower
part of the substrate. This may not continue, however, as the interstices in the limestone layer become
clogged with sediment.
Some of the nylon screens covering the openings in the sample wells become clogged with
organic matter, resulting in very slow recharge once the wells are pumped down. These wells are pumped
first during sampling to allow time for recharge. The clogging was due to the substrate and not to a buildup
of precipitates. In general, metal sulfide precipitates cause much less clogging than metal hydroxide
precipitates.
The method of delivering wastewater to the treatment cells depends on the overall system
configuration, objectives, and costs. In general, a simple and inexpensive system is preferred over
complex plumbing and pumping arrangements.
Vegetation
The vegetation transplanted in the fall of 1987 has recovered well in Cells B and C. The plants in
4-7
-------
Ceil A do not appear to be quite as vigorous, possibly due to the differences in the substrate and
microbiological processes believed to be occurring in this cell. The speculation is that the level of
ammonia was initially too high in the fresh mushroom compost. The health of the plants in Cell A appears
to be improving as the decomposition processes continue. Metal uptake by the plants is measurable
(around 1%) but remains insignificant when compared to metal removal through the activity of bacteria
present in the organic substrate.
Some channelization of surface water is caused by the hasty placement of the transplanted
vegetation with respect to maintaining appropriate water levels in the system. Channelization may reduce
contact between the acid mine drainage and the organic substrates and thereby reduce the efficiency of
the system. Careful placement of the different species of vegetation may be more effective in reducing
the channelization. However, by the beginning of the second growth season, the cattails dominated the
wetland and prevented channelization.
Thus, the presence of vegetation appears to be more important for stabilization of the substrate,
reduction of channelization in the surface flow, and continual additions to the biomass of the system than
for metal uptake. Metal uptake by plants was also found to be insignificant in comparison to metal removal
through other processes by Sencindiver and Bhumbia (1988). The choice of the species of vegetation,
therefore, is not of primary importance, as long as they are able to tolerate the conditions of the acid mine
drainage and local climate. Since the concept is to emulate a natural ecosystem, complexity may be
favored rather than simpler systems (Hammer and Bastian 1988).
INITIAL PERFORMANCE
From the beginning, removal of heavy metals occurred in all cells, and Cell A with mushroom
compost was most effective in removing contaminants and raising the pH. For the first year of operation,
selected values for the mine drainage input and the cell outputs are shown in Table 10. Mn, Fe, Cu, and
Zn are the primary metal contaminants and also give excellent indication of the removal processes that are
operating. Figures 7 and 8 plot the concentrations of Fe and Cu respectively in the mine drainage, Cell A
effluent, and Cell B effluent for the first year of operation. In the first year an attempt was made to hold the
loading rate constant at 200 square feet/gallon/minute.
During June and July of 1988, a study was made on how removal changed with flow rate. Since
the size of the cell is fixed at 18.6 m2, (200 ft2), the flow is inversely proportional to the flow rate. The
results of that study are given in Table 11. Changes in concentration for Zn and Fe plotted against the
loading factor in square feet/gallon /minute are plotted in Figures 9 and 10. Note that for copper in Cell A,
100 % removal began at about 400 square feet/gallon /minute. The best results during this period were
for Cell A at a loading rate of 600 square feet/gallon/minute. Removal of Cu and Zn was 100 %, removal of
Fe was 63 %, and pH increase was from 3.0 to 6.2. Mn was not removed. The removal patterns and
results from other experiments gave convincing evidence that the important removal process was bacterial
reduction of sulfate dissolved in the mine drainage to hydrogen sulfide and subsequent precipitation of
the metals as sulfides. The case for sulfate reduction will be presented in SECTION 5.
4-8
-------
-J
~o>
E
o
z
O
(J
&
~ OPA *
NOV
JAN
MAR
MAY
1987
1988
"JUlY
Figure 7.
iron removal in the Big Five Ceils over 1987-88. In the figure, OPA is effluent from Cell A,
OPB is effluent from Cell B, and MD is the mine drainage
-------
1.25
~OPA*
•OPB
wO.75
:0.50
NOV JAN MAR
1987 I 1988
MAY
JULY
Figure 8. Copper removal in ihe Big Five Cells over 1987-68. In the figure, OPA is effluent from
Cel A, OPB is effluent from CeN B, and MD is the mine drainage
-------
Zn REMOVAL
800
600
£
O
-j
Li.
u 400
o:
<
200
8
10
PPM Zn
Figure 9. Zinc concentration versus loading factor for June & July 1988.
-------
00
5 600
O
-j
lL
<
LU
^ 400
<
Fe
200
REMOVAL
0
O
20
30
40
PPM Fe
Figure 10. Iron concentration versus loading factor for June & July 1988.
4-12
-------
TABLE 10. Concentrations (mg/L) of metals, percent reduction of metals, pH, and flow rates (liter/minute)
in the Big Five Mine Drainage and wetland cell output waters during 1987-88.
The area of Cells A, B, and C is 200 ft2.
% % % % flow
water Sample. Mn red. Fe red. Zn red. Cu red. pH rate
November 3, 1987
Mine Drainage
35
33
9.6
0.98
2.8
Cell A
40
-14
27
18
8.4
12
0.81
17
3.1
3.8
CellB
33
6
26
21
8.7
9
0.86
12
2.8
3.8
CellC
34
3
26
21
8.3
13
0.82
16
3.1
3.8
December
11, 1987
Mine Drainage
32
33
10.6
1.07
2.8
Cell A
27
16
18
44
7.8
26
0.44
59
4.6
3.8
Cell B
33
•3
24
27
9.8
8
0.89
17
3.1
3.8
CellC
34
¦6
22
33
9.6
10
0.91
15
3.3
3.8
February 13, 1988
Mine Drainage
28
28
8.2
0.86
3.3
Cell A
27
3
18
35
5.9
28
0.14
84
5.2
3.8
CellB
28
0
23
18
7.6
7
0.77
10
3.6
3.8
CellC
28
0
25
11
7.9
4
0.87
0
3.6
3.8
March 9,
, 1988
Mine Drainage
30
32
9.5
0.93
2.8
Cell A
29
3
19
41
6.8
28
0.21
77
4.2
3.8
CellB
30
0
24
25
8.6
9
0.83
11
3.2
3.8
CellC
30
0
26
19
8.8
7
0.85
9
2.9
3.8
April
2,
1988
Mine Drainage
29
34
9.1
0.88
3.2
Cell A
29
0
19
44
6.7
26
0.43
51
4.2
3.8
CellB
29
0
24
29
8.4
8
0.80
10
3.1
3.8
CellC
28
3
25
26
8.2
10
0.79
10
3.2
3.8
May
31,
1988
Mine Drainage
25
44
"8.1
0.75
3.0
Cell A
26
-3
28
36
5.5
32
0.02
97
4.3
3.8
CellB
25
0
17
61
7.4
10
0.64
15
3.0
3.8
CellC
25
0
21
52
7.7
5
0.68
9
3.0
3.8
June
27,
1988
Mine Drainage
27
43
8.4
0.85
3.0
Cell A
27
0
22
49
CM
o
96
0.06
93
4.9
3.8
CellB
28
-3
25
42
7.6
7
0.75
12
2.9
3.8
CellC
27
0
23
46
7.6
7
0.69
19
3.0
3.8
July
29,
1988
Mine Drainage
26
37
8.1
0.91
2.9
Cell A
25
3
20
46
0.03
100
0.17
81
5.5
1.9
CellB
26
0
16
57
6.4
21
0.57
33
3.3
0.90
CellC
25
3
11
70
5.8
28
0.38
58
3.4
1.3
4-13
-------
Table 11. Loading factor (ft^/gal/min) versus concentration in mg/L at the Big Five wetland cells
in June and July 1988. Area / Flow is in square feet/gallon/minutes
Date
Cell
Area/Flow3 pH
Mn
Fe
Zn
Cu
6-23
A
933
6.65 —
—
—
—
6-23
B
6600
6.25
,6-23
C
8400
6.70
-Changed Flows on A, B, & C on 6-23-
6-27
A
200
4.95
6-27
B
200
3.95
6-27
C
200
3.0
6-27
MD
3.0
-Changed Flows on A *
Bon 6-27-
6-30
A
400
6.25
26.
20.
<0.05
<0.03
6-30
B
400
3.20
29.
18.
7.3
0.54
6-30
C
200
3.0
29.
23.
7.5
0.73
7-4:
A
400
6.1
26.
15.
<0.05
<0.03
7-4:
B
460
3.1
29.
18.
7.4
0.58
7-4:
C
200
3.1
29.
20.
7.3
0.79
7- 7:
A
380
6.05
26.
16.
<0.05
<0.03
7-7:
B
450
3.05
28.
19.
7.5
-
7-7:
C
200
2.95
28.
22.
7.2
0.75
7-7:
MD
—
3.05
28.
40.
8.6
0.95
-Changed Flows on A, B,
* C on 7-7-
7-8:
A
200
4.6
27.
25
0.8
<0.03
7-8:
B
600
3.10
27.
18.
6.4
0.49
7-8:
C
400
3.10
28.
16.
6.3
0.56
7-9:
A
200
4.65
27.
25.
1.3
<0.03
7-9:
B
600
3.15
27.
16.
6.3
0.49
7-9:
C
400
3.15
28.
16.
6.3
0.56
7-10
A
200
4.50
27.
29.
2.3
<0.03
7-10
B
600
3.05
27.
16.
6.7
0.45
7-10
C
450
3.15
28.
13.
7.0
0.56
7-14
A
200
4.6
26.
23.
1.4
<0.03
7-14
B
600
3.25
28.
19.
6.1
0.46
7-14
C
500
3.35
28.
15.
6.3
0.37
7-14
MD
2.95
-Changed Flows on
B & C on 7-14; On A on 7-17-
7-18: A
400
5.95
26.
21.
<0.05
<0.03
7-18
B
800
3.55
29.
20.
6.1
0.32
7-18: C
650
4.30
28.
13.
7.4
-
7-21
A
600
6.20
25.
16.
<0.05
0.04
7-21
B
900
3.90
28.
21.
6.3
0.54
7-21
C
800
4.05
28.
9.7
5.5
0.31
4-14
-------
In Tables 10 and 11, there are large variations in removal even though the flow is constant. This is
because the loading capacity of the cells, as determined by the amount of sulfate that can reduced, was
exceeded. How loading affect removal will be developed in SECTION 7.
SUBSEQUENT MODIFICATIONS
As pointed out in Table 8, a number of modifications and additionswere made to the pilot plant cell
in 1988 and 1989. Figure 11 is a diagram of the current configuration of the Big Five site. The changes
and the results of those changes are discussed in this section.
Cell A Modification
The first redesign concentrated on the issue of increasing the contact of the drainage with the
substrate, especially in the anaerobic zone. This was accomplished by the addition of: a) Two walls
running the length of the cell to increase the flow path length by a factor of three, and b) Six redistribution
baffles to collect water flow Irom the top surface and redistribute it to the bottom of the substrate. This was
done on Ceil A and a cut-a-way view of the redesign is shown in Figure 12. Essentially, Cell A was
redesigned to be a six segment plug-flow reactor (70).
Although the initial structure was amenable to major changes, the results of the redesign were
discouraging. The desired plug flow (ideally, horizontal flow at all depths in the substrate) from segment to
segment through the lower part of the substrate was not achieved. Considerable water flowed across the
top of cell segments and leaked from one segment to others. When the cell was modified, the original
mushroom compost was removed, stockpiled, and returned to the cell after the remodeling. It was
speculated that through this handling, substrate permeability significantly decreased. Subsequent
experiments on new and used mushroom compost from Cell A verified that permeability decreased from
3.0 x 10"3 to 9.2 x 10'5 cm/sec (9). Selected values on removal and changes in pH from the cells during
this phase are shown in Table 12.
As a result of this setback, a laboratory and bench scale program was developed to determine
how well typical soil tests could be adapted to this highly organic substrate. Especially important were the
development of methods to determine hydraulic conductivity that could give reasonable indications of
what to expect in a constructed wetland (9). Other tests on substrate materials included the determination
of specific gravity, bulk density, size fractions, and percent moisture. The processes that cause the
permeability of the substrate to change with time will be discussed in more detail in SECTION 10.
Plm-Flow. tJoflow. and Downflow Calls
The initial constructed wetland designs for the Big Five Cells used the concept of plug flow.
Basically, plug flow is similar to pipe flow in that the water is meant to travel through a cross-sectional area
that is small relative to the length of travel in the substrate. In the case of wetlands, plug flow was meant to
be essentially horizontal flow throughout the entire thickness of the wetland, as shown in Figure 12. In
operation, however, most of the flow ends up being at or near the surface, due to the compaction-caused
rapid decrease of permeability with depth in the wetland.
4-15
-------
CELL E
CELL C
CEL
TO PORTAL
SCHEMATIC OF TREATMENT CELLS
BIG-FIVE PILOT SYSTEM
SURFACE AREAS square feet square meters
Cell A 200 18.6
Cell B - Upflow 100 9.3
Cell B - Downflow 100 9.3
Cell C 200 18.6
Cell D -100 -9.3
Cell E -100 -9.3
Figure 11. Present Big Five Site configuration.
4-16
-------
CUT-A-WAY VIEW
WETLANDS PILOT SYSTEM
OVERFLOW
DISCHARGE
DISTRIBUTION
BAFFLE
END BAFFLE
Figure 12.
A cut a way diagram of the Cell A redesign.
-------
TABLE 12. Concentrations (mg/L) of metals, percent reduction of metals, pH, and flow rates (liter/minute)
in the Big Five Mine Drainage and wetland cell output waters during 1988-89.
The area of Cells A, B, and C is 200 ft2.
%
%
%
%
flow
Water Sample
Mn
red.
Fe
red.
Zn
red.
Cu
red.
PH
rate
December 18, 1988
Mine Drainage
29
38
9.2
0.80
3.0
Cell A
28
3
31
18
8.6
6
0.62
22
3.4
4.5
Cell B
28
3
30
21
7.8
15
0.74
8
3.2
4.3
CeUC
26
3
28
26
7.7
16
0.69
14
3.3
4.7
January 21,
, 1989
Mine Drainage
25
31
10.9
0.91
2.8
Cell A
33
-32
32
-3
10.8
0
0.61
33
3.3
4.6
CellB
34
-36
26
16
10.8
0
0.60
34
3.1
3.6
CeUC
33
-32
25
19
10.5
3
0.60
34
3.1
3.8
February 21, 1989
Mine Drainage
23
38
9.0
0.72
3.0
Cell A
28
-22
19
50
4.9
46
<0.05
100
5.1
1.0
CellB
31
-34
39
-2
5.6
38
0.48
33
3.4
12
CeUC
25
-10
31
18
7.2
20
0.26
64
3.5
1.2
March 19,
1989
Mine Drainage
22
40
8.3
0.73
3.0
CeHA
24
-9
23
42
0.28
100
<0.05
100
6.5
0.21
CeflB
29
-32
7.3
82
3.9
53
0.21
71
5.7
0.36
CellC
28
-27
7.3
82
ZO
76
<0.05
100
6.3
0.60
April 16,
1989
Mine Drainage
29
39
9.8
0.70
2.7
Cell A
32
-10
31
20
8.5
15
0.47
33
3.9
2.1
CellB
32
-10
18
54
6.6
33
0.34
51
3.6
0.94
CeUC
30
-3
40
•2
9.6
2
0.66
57
3.9
1.4
May 21,
1989
Mine Drainage
44
48
10.5
0.85
3.3
Cell A
41
7
31
35
9.4
10
0.58
32
3.8
2.3
CellB
37
16
0.46
100
0.22
100
<0.05
100
5.7
2.2
CellC
35
20
24
50
8.9
15
0.62
27
3.2
2.3
June 16,
1989
Mine Drainage
29
48
7.6
0.76
3.0
Cell A
34
-17
26
46
4.4
42
<0.05
100
4.6
1.7
CellB
29
0
30
38
7.9
-4
0.60
21
3.0
2.2
CeUC
35
-21
26
46
5.7
25
0.46
39
3.0
2.4
July
14, 1989
Mine Drainage
30
41
6.9
0.72
3.0
Cell A
30
0
17
59
0.25
100
<0.05
100
5.7
1.2
CellB
33
-10
17
59
7.1
20
0.36
50
3.3
1.1
Ce»C
33
-10
13
68
2.2
75
<0.05
100
5.1
1.0
4-18
-------
Plug flow design is suitable for aerobic wetlands because the aerobic layer is at the surface. In
anaerobic wetlands, it is desireable to maximize flow through the subsurface anaerobic layers. Therefore,
upflow and downflow cells were designed to maximize vertical flow, as shown in Figure 13. In upflow and
downflow designs, which are similar to conventional trickling filters in concept, the cross'sectional area of
substrate perpendicular to flow is large compared with the length of travel of fluid in the substrate.
Kali B Modification
Results of laboratory and bench-scale permeability experiments led to the modification of Ceil B
into upflow and downflow cells to monitor and evaluate permeability at the pilot scale. The original cell was
divided into two lined, identical cells so individual variables could be tested. All features needed to
determine soil permeability in the cell were included. A special feature in the design was inclusion of a
plenum beneath each subcell for even distribution of drainage in the upflow configuration and even
collection when used as a downflow cell. Each cell could be operated in the upflow or downflow
configuration. Figure 13 is a cut-a-way diagram of the downflow operation.
There are a number of features in this modification that should improve contact with the substrate.
When operating either as an upflow or downflow cell, the mine drainage is forced through the substrate
before discharge. This configuration is comparable to a trickling filter process instead of the plug flow
reactor design in the Cell A modification (70). Also included in the modification, was the addition in series
of two 150 gallon stock tanks before the inlets of the subcells. Figure 13 shows the placement of one of
the tanks. The tanks serve the purpose of completing the precipitation of ferric hydroxides before
drainage enters the subcells so that plumbing, plenums, and filter fabrics do not become clogged. In both
the upflow and downflow parts of the new Cell B, plants were not used. Instead, the substrate was
covered with 25 cm of hay and then 6 mil black plastic to provide insulation during the winter.
Remodeling of Cell B was completed in August, 1989. The two subcells were filled with fresh
mushroom compost to a depth of 0.61 m. This substrate material was selected because it appeared to
provide the best metal removal performance among the three substrates used in the original cells. Flow of
mine drainage through the subcells was initiated on September 1. One subcell (Cell B-North, or B-N) was
operated in the upflow configuration, the other (Ceil B-South or B-S) downflow. The positions of these
cells is shown in Figure 11. Upflow permeability measured 1.3 x 10'3 cm/sec, and downflow permeability
measured 3.1 x 10"4 cm/sec (9). The average permeability of substrate in cells operating in the upflow
mode is expected to be higher than permeability of substrate in cells using downflow or plug flow. The
increased permeability probably occurs because the upward driving force required under upflow
conditions counters the downward compressive force due to the effect of gravity on overlying layers of
substrate, so that compaction of the substrate would be less. The development of laboratory and bench-
scale methods for predicting permeability in actual constructed wetlands has proven successful.
4-19
-------
MINE DRAINAGE
RESERVOIR
SURFACE
DISTRIBUTION
SYSTEM
INLET
SUBSTRATE £$$$&
' -m -m V V V *.• '
*.v 'a/.x'v/a/.v ' \ '
vWvv'
'.vvA
riii.¦<'.i • \-a•: JC^YVVVVVYv a"a-.w A^/a.'.V*
\'\>
,A-W
CELL B
SUBSURFACE
DESIGN
DISCHARGE
FLTER FABRIC
PLENUM
Figure 13. A cut-a way diagram of the Cell B redesign in the downflow mode.
-------
CELL B UPFLOW REMOVAL TRENDS
H
3
Q.
3
Q.
»-
to
rvA
Mn B-U
S04 B-U
DAYS
— Zn B-U
FeB-U
Cu B-U
Figure 14. Removal of contaminants in CeN B Upflow over 1989-90.
-------
CELL B - DOWNFLOW REMOVAL TRENDS
M
M
CL
z
H
3
CL
Mn B-D
ZnB~D
DAYS
S04B-D
Fe B-D
—~— CuB-D
Figure 15. Removal ol oonlaminanls in CeR B Downltow over 1-989-90.
-------
CELL E REMOVAL TRENDS
1.0
0.8
0.6
Mn E
S04 E
DAYS
Zn E
—~— Fe E
—o Cu E
Figure 16.
Removal of contaminants In Cell E over 1989-90.
-------
Table 13. Concentrations (mg/L) of metals, percent reduction of metals, pH, and flow rates (liters/minute)
in the Big Five Mine Drainage and wetland cell output waters during 1989-90*.
%
%
%
%
flow
Water Sample
Mn
red.
Fe
red.
Zn red.
Cu red.
PH
rate
October
3, 1989
Mine Drainage
35
46
9.9
0.66
3.2
Cell A
35
0
31
33
2.2 78
0.07 89
5.1
0.98
Cell B-Up
34
3
39
15
9.3 6
0.59 11
3.5
1.1
Cell B-Down
7.7
78
23
50
0.76 92
<0.05 100
6.5
0.60
CeilC
36
-3
29
37
7.8 21
0.22 67
3.5
0.79
Cell D
32
8
24
48
<0.05 100
<0.05 100
6.1
0.32
Cell E
24
31
0.5
100
<0.05 100
<0.05 100
6.3
0.71
November
5, 1989
Mine Drainage
32
38
8.7
0.61
2.9
Cell A
32
0
27
29
8.2 6
0.49 20
3.5
2.0
Cell B-Up
31
3
17
55
8.4 3
0.48 21
3.6
0.83
Cell B-Down
20
38
0.36100
6.0 31
<0.05 100
5.9
0.72
CellC
32
0
27
29
8.2 6
0.52 15
3.5
1.4
Cell D
29
9
16
58
0.77 91
<0.05 100
5.7
2.6
Cell E
20
38
0.29 100
<0.05 100
<0.05 100
6.5
0.42
December
3, 1989
Mine Drainage
30
43
9.0
0.58
3.0
Cell A
30
0
21
51
1.3 85
<0.05 100
6.0
1.1
Cell B-Up
30
0
34
21
9.0 0
0.44 24
3.5
1.4
Cell B-Down
25
17
7.1
83
6.8 24
0.21 64
4.1
1.1
CellC
30
0
32
26
8.5 5
0.52 10
3.4
1.3
Cell D
30
0
17
60
0.07100
<0.05 100
6.2
0.57
Cell E
26
13
2.6
94
<0.05 100
<0.05 100
6.3
0.42
January 13, 1990
Mine Drainage
31
33
9.0
0.60
2.9
Cell A
31
0
26
21
7.4 18
0.36 40
3.3
1.5
Cell B-Up
30
3
33
0
8.9 0
0.49 18
3.2
0.79
Cell B-Down
27
13
12
64
8.1 10
0.44 27
3.2
0.76
CellC
42
35
31
6
5.5 39
<0.05 100
6.0
2.1
Cell E
28
10
9.8
70
<0.05100
<0.05 100
6.0
.38
Mine Drainage
February 6, 1990
31
36
9.0
0.53
2.8
Cell A
33
-6
20
44
4.8 47
0.17 68
3.8
0.57
Cell B-Up
30
3
36
0
7.0 22
0.12 77
3.3
0.45
Cell B-Down
31
0
18
50
9.1 0
0.44 17
3.2
0.49
Cell E
32
-3
14
61
<0.05100
<0.05 100
5.8
0.76
March 10, 1990
Mine Drainage
29
60
9.0
0.52
3.2
Cell A
28
-3
36
40
2.0 78
<0.05 100
5.4
0.87
Cell B-Down
25
14
4.9
92
7.6 16
<0.05 100
6.0
NA
Cell E
28
-3
17
72
<0.05 100
<0.05100
6.4
0.83
The area of Cells A, B, and C is 200 ft^; the area of Cells B-Up, B-Down, D and E is 100 .
4-24
-------
Table 13. continued
%
%
%
%
flow
Water Sample
Mn
red.
Fe
red.
Zn red.
Cu red.
PH
rate
April 6, 1990
Mine Drainage
31
55
8.7
0.48
3.0
Cell A
30
3
31
44
1.5 83
<0.05 100
5.2
0.91
CeiB-Up
27
13
11
80
1.2 86
<0.05100
6.6
0.26
Cell B-Down
26
16
2.9
95
7.7 11
<0.05 100
6.0
0.15
Cell C
29
6
16
71
0.08 100
<0.05 100
6.6
1.30
Cell D
22
29
9.9
82
<0.05 100
<0.05 100
6.8
0.68
CeltE
29
6
7.1
87
<0.05100
<0.05 100
6.2
0.41
May
7, 1990
Mine Drainage
30
45
9.5
0.54
2.9
Cell A
28
7
33
27
1.5 84
0.05 100
5.4
0-83
Cell B-Up
19
37
5.1
89
0.08100
<0.05 100
6.8
0.19
Cell B-Down
20
33
6.3
86
2.8 70
<0.05 100
6.3
0.30
CellC
28
-6
7.0
84
6.7 29
0.30 44
4.1
0.76
Cell D
20
33
5.9
87
<0.05 100
<0.05 100
6.6
0.23
Cell E
25
17
7.0
84
<0.05 100
<0.05 100
6.4
0.42
June
11, 1990
Mine Drainage
32
40
8.8
0.59
3.0
Celt A
33
-3
36
10
2.5 72
<0.05 100
4.9
0.79
Cell B-Up
14
56
0.42100
0.08 100
<0.05100
6.6
0.45
Ceil B-Down
33
-3
3.9
90
5.0 43
<0.05 100
5.5
0.41
CeKC
33
-3
NA
-
4.6 48
NA
3.8
0.15
Cell E
31
5.2
<0.05
87
<0.05 100
<0.05 100
6.0
0.68
July
3, 1990
Mine Drainage
30
36
9.4
0.55
3.1
Ceil A
29
3
26
31
5.0 45
<0.05100
4.0
1.0
Ceil B-Up
18
40
1.8
95
<0.05100
<0.05 100
7.0
0.26
Cell B-Down
27
10 *
11
71
0.78 92
<0.05100
6.3
0.15
CeKC
32
¦6
13
66
6.3 32
0.07 87
4.0
0.79
Cell D
26
13
4.8
87
<0.05100
<0.05100
6.2
0.94
Cell E
28
7
1.8
95
<0.05100
0.08 85
6.2
0.11
August
13, 1990
Mine Drainage
34
47
9.2
0.54
3.0
Cell A
35
•3
19
60
2.6 72
<0.05 100
5.0
0.87
Cell B-Up
25
26
3.9
92
<0.05 100
<0.05100
6.8
0.57
Cell B-Down
45
32
21
55
3.7 60
<0.05 100
6.7
0.34
CeKC
36
-6
8.2
62
6.0 35
<0.05 100
5.0
0.76
Cell D
34
0
6.0
87
<0.05 100
<0.05 100
6.4
0.53
Cell E
35
-3
2.0
96
<0.05100
<0.05100
6.2
0.53
'The area of Cells A, B, and C is 200 it2; the area of Cells B-Up, B-Down, D and E Is 100 ft2.
4-25
-------
Values on removal of Mn, Fe, Cu, and Zn and increases in pH for Cell B-Upflow and Cell B-
Downflow are given in Table 13. Just as for the original Cell A design, removal has been impressive from
initiation of flow. Figure 14 shows output/input over time for Ceil B in the upflow mode; Figure 15 shows
the same for Cell B downflow.
Initially, removal from the downflow system was better than from the upflow, however, flows
through the cells have not always been equal. Also, the pattern of removal of Zn and Mn with little or no
removal of sulfate, indicates that the initial removal in the downflow cell is by organic complexation instead
of sulfate reduction (7). After 3 months, removal in the downflow cell is not as good as during the first
month. This is particularly the case for Mn and Zn. This implies that the sites for organic complexation are
being saturated after about three months (7). Organic complexation compared with sulfate reduction will
be discussed further in SECTION 5. In the upflow cell, poor initial removal is attributed to ponding of
water on the surface of the substrate and an excessive loading factor. In March of 1990, after about 180
days of operation these problems were rectified. Since then, removal of heavy metals from the upflow cell
has been excellent. That the adjustments worked gives important evidence that sulfate reduction
processes can recover if loading factors are not exceeded. Also, to keep an upflow cell anaerobic,
effluent has to be taken from the cell before it breaches the surface.
Cells D and E Etesian Construction, and Orwallon
When Cell B was remodeled, Cells D and E were constructed using the original substrate from Cell
B. Their positions and configuration in the site are shown in Figure 11. Cell D was designed to polish
discharges from anaerobic cells by using aerobic processes. Features of the design include a shallow
depth (0.50 m) and a length to width ratio of 10. Substrate and plants from the original Cell B were used. It
has been receiving the discharge from Cell A. Removal of Cu, Zn, and Fe is completed in Cell D and the
pH Is raised to above 6.
Cell E was designed to operate as a downflow, subsurface wetland. Construction was completely
accomplished with materials found locally. It is approximately 9.3 m2 and the substrate is 0.61 m deep. As
in Cell D, substrate and plants from the original Cell B were used. Subsurface flow is achieved by flow
through landscape fabric into 2.5 cm gravel and subsequent discharge into a tube on the downflow end.
For this system, results have been excellent.
Flow of mine drainage through Cell E was initiated on September 1,1989. From the beginning,
removal of Cu, Zn, and Fe has been 100 %. pH has increased to 6.5, and Mn removal has averaged 25 %.
Removal results are given in Table 13 and are shown in Figure 16. Laboratory experiments have
confirmed that sulfate reduction with subsequent precipitation of metal sulfides is the predominant
removal process in Cell E (7).
The removal successes with Cells D and E confirm that it was not the chemical or microbiological
characteristics of the peat/manure/decomposed wood substrate that caused Cell B to operate more
poorly than Cell A. Rather, the poor performance was caused by the low permeability of the substrate.
Products of microbial reactions occuring within the substrate couldn't come in contact with the metals in
the mine drainage.
4-26
-------
OPERATIONS DURING THE WINTER
The last four winters have allowed observations on how well wetland cells operate during the
winter. Not all of the cells have kept operating during the winter. Two key factors allow winter operation:
The mine drainage water is about 12 to 15 °C the year round, and portions of the site are in winter sun
throughout the day. Below, the winter success of each cell is described.
Cell A has operated continuously through all four winters. This is the case even when the flow
was cut back from 4 L/min to 1 L/min. Three reasons can be given for the winter success of Cell A:
1. It is continuously in the sun.
2. More of the water flows through the substrate rather than across the surface.
3. The inlet is small and insulated so the energy within the water is not lost.
When the flow into Cell C was cut back to 1 L/min, it has frozen over the last two winters. This cell
is more shaded during the winter. Also the dense growth of Typha in Cell C inhibits solar radiation from
reaching the substrate-water interface. In addition, Cell C still has the original rock box inlet. This inlet
allows much of the water energy to be lost.
Cell 0 was built in the summer of 1989 and in both winters it has frozen. This is primarily a surface
fk»w cell and the depth of the substrate is only 0.5 meter. Also it is not well insulated.
Cell E has worked well over the last two winters. It is shallow, but is a subsurface flow cell. The
important feature that keeps Cell E operating during the winter is that much of the water that enters the cell
flows across the surface and over the spillway. This excess water insulates the substrate.
The modified B Cells have operated over the winter primarily because the surface of the cells was
insulated with hay and plastic. In fact, the temperature of the outlet only dropped 4 to 5 °C during January
and February. On Cell B-Upflow the new outlet installed in the spring of 1990 was not insulated. It froze in
December, 1990 and the cell had to be turned off for the winter. Cell B-Downflow did have an insulated
outlet, and it has continued operations throughout both winters.
Guidelines for Winter Operation
From these observations, a number of guidelines can be established for insuring the operation of
wetland systems in cold winter climates.
o Use the thermal energy within the mine drainage water to best advantage. Insure that delivery
systems are insulated. Keep inlet structures small and insulated,
o Place wetland cells so they receive winter sun. If this cannot be completely achieved, at least
insure that outlets are in winter sun.
o Insulate the top of the cell with excess mine water as in Cell E or with hay and plastic as in the B
Cells.
o Insulate the outlet and allow a method for the effluent to flow away from places where it could
cause freezing problems. This is especially important if it is planned to sample water during the
winter.
4-27
-------
o If possible, design subsurface flow systems sucJi as Cells B and E. The thermal energy within the
substrate will aid operation, whereas in a surface flow system the waters are exposed to the
elements.
CONCLUSIONS
Using constructed wetlands for wastewater treatment is still a developing technology. However,
the results from the Big Five Pilot Wetland that was funded by the Emerging Technology Program (ETP)
of the U. S. EPA shows promising removal of heavy metals and increase of pH for acid mine drainage.
Conclusions from the project include:
1. Toxic metals such as Cu and Zn can be removed and the pH of mine drainage can be increased
on a long term basis.
2. The major removal process is sulfate reduction and subsequent precipitation of the metals as
sulfides. Exchange of metals onto organic matter can be important during the initial period of
operation.
3. A downflow, trickling filter style of configuration achieves the best contact of the water with the
substrate.
4. Removal efficiency depends strongly on loading factors. In the Big Five wetland, factors above
1,000 feet2 per gallon/minute are needed for reasonable removal.
5. Permeability of the substrate is a critical design variable for successful operation. Using
laboratory and bench-scale tests, a good indication of the soil permeability in a constructed
wetland can be determined.
6. Solutions to problems such as plugging of plumbing by ferric hydroxides and freezing of
discharge lines during winter have to be designed and constructed into the passive nature of
wetlands to achieve long term operation.
7. Eventual removal of precipitated metal sulfides for metal recovery or disposal must be included
in the operating plan. Estimates of how long the substrate will last are included in SECTIONS 8
and 12.
4-28
-------
SECTION 5
EVIDENCE FOR SULFATE REDUCTION
INTRODUCTION
Using wetlands to treat acid mine drainage has only been studied for a little over a decade (71).
During the first few years of this research, biologists and ecologists dominated studies (39). As a result,
constructed wetland studies concentrated on the surface ecosystem as the key to removal processes.
Key wetland features in such a constructed wetland are:
1. If an ecosystem is needed, the smallest pilot system should be about 200 square feet.
2. Plants are necessary in a constructed wetland.
3. Aerobic processes are the key to removal.
4. Peat as a substrate would be desirable.
Typical constructed systems that were built using these guidelines are those in the Tennessee
Valley (72), the Simco #4 wetland (73), the natural wetland along Peru Creek in Colorado (74), and the
Tracy wetlands in Montana (75).
In some cases, removal of contaminants was achieved (39, 72). However, In most cases Fe was
only partially removed (72,73) and manganese and sulfate were generally not removed. In some cases
the pH increased (72,73), but just as often the pH decreased (72, 74). In some cases (74) it was feared
that the drop in pH would release metals from the wetland system. From the discussion of wetland
removal processes in SECTION 3, the drop in pH and release of metals might be expected. Aerobic
processes oxidize ferrous to ferric and its precipitation definitely lowers the pH of the effluent. If the peat
removes metals through some organic adsorption process, then increasing the concentration of
hydrogen ions definitely shifts the adsorption equilibrium to release metals.
At the American Society for Surface Mining and Reclamation Meeting in 1988, the first
suggestions that sulfate reduction may be an important process were made (41,76,77). In the same year,
these suggestions were expanded at the Constructed Wetlands Conference sponsored by the TVA (3,
65). Since then, research efforts on sulfate reduction in constructed wetlands have centered on the
studies done by the Colorado School of Mines group and by the U. S. Bureau of Mines group (41,65,66,
67). This section explores the evidence from the Big Five Pilot Wetland for sulfate reduction and sulfide
precipitation as a removal process in a constructed wetland. The evidence follows the thought
chronology while the project developed. It includes: Wetlands functioning immediately even during the
winter; the pH of the effluent increasing; sulfide in the substrate; the metal removal pattern; and finally,
sulfale decreasing in the effluent.
5-1
-------
IMMEDIATE OPERATION EVEN DURING THE WINTER
Design and construction of the Big Five Wetland was met with a number ot delays because the
site was on the national CERCLA (Superfund) list. Transplanting of vegetation took place in September of
1987 and mine drainage first flowed through the wetland starting on October 21, 1987. There was
concern that, because of the late starting time, no immediate removal would occur and plants would not
survive the winter. Surprisingly, removal of contaminants and Increase in pH occurred within the month.
Table 10 and Figures 7 and a document this immediate removal. Celt A containing mushroom compost
was particularly effective In contaminant removal. In Cell A, the surlace of the wetland turned from rust
colored to gray-black within five months. This was a strong clue that anaerobic processes were
responsible for contaminant removal. In wetlands thai were effective in Pennsylvania, Hedin noticed the
same change to anaerobic conditions (41).
During the initial operation of Cell B-Downflow and Cell E contaminant removal patterns and
changes in pH were closely monitored (7). Both these systems operated as downflow subsurface
wetlands so comparisons should be on the chemistry and not the physical design. There are differences
in the substrates. Cell E substrate was taken from the original Cell B and was peat/manure/and
decomposed wood product laden with active sulfate reducers. In Cell B, new mushroom compost was
used and mine drainage was Immediately applied to the dry compost with no presoaking. In this case, the
activity of sutfate-reducfng bacteria should be low.
Laboratory Adsorption Studies
Examining the list of removal possibilities in SECTION 3 reveals that removal of contaminants by
adsorption processes onto the organic or inorganic oil sites could occur immediately. Since the substrata
is highly organic, it Is assumed that adsorption by humic acids would be the predominant process. If this is
the case, then the ideas in SECTION 3 concerning the role of pH and the order of metal adsorption should
reveal whether adsorption is occurring. In particular, it has been found that Fe and Cu are more strongly
adsorbed than Zn and Mn (53).
First, a laboratory adsorption test was conducted. To determine the extent o( the adsorption ol
metals from the mine drainage by organic material tn the substrate, a set of experiments was conducted
that tested the adsorbing capability of fresh mushroom compost using the Big Five Tunnel mine drainage
collected on October 19,1969. First, 10 mL of a 1000 mg/L solution of SeO*- was added to 0.100 g of
the substrate used in Cell B-Dowrtftaw to inhibit sulfate reducing bacteria (54). Second, various volumes
of mine drainage <10,20, 30,40 mL) were added to shaker tubes containing the substrate and SeQ<-.
Next, the pH was adjusted to 4.5 with HNO3 and NaOH, and the volume of the solution was ad$usted to 50
mL with deionized water. The tubes wen shaken for three days and the pH checked and adjusted to 4.5
each day. Next, the concentration of Mn, Fe, Cu, and Zn was determined in each of these solutions. The
results are given in Table 14. A blank containing no mine drainage showed no metals extracted from the
5-2 -
-------
Table 14. Results of the laboratory experiment testing
adsorption of metals onto a substrate of fresh
mushroom compost during equilibrium with mine drainage
at a pH of 4.5. 1
Volume in mL of mine drainage in 50 mL total volume
10 20 30 40
Initial mmol/L of
Mn, Fe, Cu, & Zn
0.32
0.63
0.95
1.3
Mn mg/L initial
6.9
14
21
28
Mn mg/L final
1.1
11
18
25
Mn micromol adsorbed
5.3
2.9
2.9
2.3
Fe mg/L initial
8.9
18
27
36
Fe mg/L final
0.04
0.2
0
0
Fe micromol adsorbed
8.0
16
24
32
Cu mg/L initial
0.12
0.25
0.37
0.50
Cu mg/L final
0
0
0.04
0.08
Cu micromol adsorbed
0.10
0.20
0.26
0.32
Zn mg/L initial
16.9
3.8
5.8
7.7
Zn mg/L final
0.22
2.3
4.6
6.7
Zn micromol adsorbed
1.1
1.2
0.89
0.75
Note: In this experiment, the mine drainage was tested at four different dilutions. Hie initial and final
concentrations of Mn, Fe, Cu, and Zn in solution are given along with the sum of these four metal
concentrations in the initial solution. The amount of Mn, Fe, Cu, and Zn adsorbed onto 0.10 gram
of fresh mushroom compost is given in micromoles. The error tot these data is less than 10
percent.
original compost. The test which contained 20 mL of mine drainage was run in duplicate and the results
were confirmed to within 10 %. Similar experiments were also conducted with 0.1 g of substrate and an
adjusted pH of 5.5, and with 1.0 g of substrate at pH's of 4.5 and 5.5 and essentially all of the metals were
adsorbed onto the substrate under these conditions.
Data from this experiment show that competition ior sorption sites on the substrate is significant.
Figure 17 shows the ratio of Mn, Fe, Cu, and Zn concentrations in the final solution to those in the initial
solution versus the sum of the concentration of the four metals in the original solution. These results are
similar to those from Kerndorf and Schnitzer (53) where, at a pH of 4.7 and 5.8, the amount sorbed for Fe,
Cu, Zn, and Mn is as follows:
Fe > Cu » Zn - Mn
An important difference between the results of Kerndorf and Schnitzer (53) and this experiment is that the
5-3
-------
SORPTION ONTO SUBSTRATE AT pH = 4.5
0.32
0.63
0.95
MMOLE / L OF HEAVY METALS
Figure 17.
Adsoption of Mn, Fe, Cu, and Zn versus concentration In the mine drainage.
-------
metals here are in test solutions at various concentrations up to those typical of the Big Five mine
drainage. At higher total metal concentrations, the percent of Mn and Zn adsorbed decreases as Fe and
Cu are preferentially adsorbed onto the most available sites. At a pH of 4.5, it is possible that Fe (til) may
have precipitated as Fe(OH)3, though no typical orange coloration from such a precipitate was observed.
However, the other metals, Mn, Zn, and Cu, are very soluble at mine drainage concentrations and a pH of
4.5 (21), suggesting that the most likely metal removal process occurring is metal adsorption onto organic
material.
Pinlri Fv/iriencft for Adsorption Versus Sulfide Precipitation
In Table 15 the effluent concentrations for all the sampling episodes for the first four months of
operation of Cell B-Upflow, Cell B-Downflow, Cell E, and the Big Five drainage are given. In Figures 18
and 19 the removal data for Cell B-Downflow and Cell E are presented; and in Figure 20, the pH of the
effluents are piotted.
Data gathered since the flow of mine drainage into the new cells was started indicate that
saturation of organic adsorption sites in the fresh mushroom compost occurred within months after Cells fi-
ll pf low and B-Downllow began receiving flow. Figure 18 shows the ratio of metal concentrations in the
outflow to those in the inflow for Cell B-Downflow. Mn and Zn are almost completely removed from the
mine drainage during the first 30 days. Afterwards, the concentration of these two metals dramatically
increases, suggesting that Mn and Zn may be less preferentially adsorbed onto the substrate material after
enough adsorption sites become filled. The concentration of Cu and Fe, however, remains low tor over
100 days before starting to rise, implying that Fe and Cu may be more competitive for adsorption sites than
Mn or Zn. The rise in metal concentrations after about four months may indicate that adsorption sites in
the substrate are becoming saturated. The pattern of metal concentration ratios for Cell B-Downflow is
similar to the data trend from the sorption experiment as shown in Figure 17. Notice that the sulfate
concentration in Cell B-Downflow is essentially the same as in the mine drainage for the entire time period,
showing, along with the high Eh's and lack of measurable sulfide in the output waters, that no significant
sulfate reduction is taking place.
Metal concentrations In the outflow from Cell E-Downflow show a much different pattern. The CeU
E-Dowrrflow substrate was used for two years in the original Cell B. In this used substrate there were
substantial populations of sulfate reducing bacteria (10). Figure 19 shows the ratio of metal
concentrations in the outflow to those in the inflow for Cell E-Downflow. The difference between the
pattern of metal removal in Cell B-Downflow and Cell E-Downflow is substantial. Copper and Zn are
completely removed for the entire period. In the first 60 days, Fe is completely removed, Mn is removed
30 to 40 percent, and sulfate is removed 10 to 20 percent. The consistency in metal and sulfate removal
through the first 80 days followed by a trend of decreasing removal which is similar for sulfate, Fe, and Mn,
5-5
-------
Table IS. Metal and sulfate concentrations (mg/L), pH, Eh (mV), and water temperature (°C) of
the Big Five mine drainage (MD) and of the output waters from Cells B-Upflow, and
B-Downflow, and E-Downflow since September 1,1989. Flow rates, in liters/minute,
are given for the input flow to the cells as weft as for the output flow for
Cel E-Downfbw due to some loss of water over Its spillway.
SAMPLING DATE
14-Sep-89
22-Sep-89
03-OCI-89
19-Oct-89
MINE DRAINAGE
MD-1101
no water
MD-1201
MD-1218
Cu (mg/L)
0.647
sample
Q.660
0.624
Fe (mg/L)
36.8
taken
45.9
44.3
Mn (mg/L)
30.4
34.9
34.5
Zn (mg/L)
8.5
9.9
9.6
SO4 (mg/L)
1750
1670
1690
pH
2.9
2.9
3.2
3.0
Eh(mV)
655
630
605
660
Temperature (C)
13.5
14.5
13.3
13.0
OUTPUT B-UPFLOW
B1-1105
B1-1110
B1-1205
B1-1219
Cu (mg/L)
0.543
0.445
0.590
0.400
Fe(mg/L)
26.6
17.9
39.4
4.59
Mn(mg/L)
29.9
27.5
34.3
30.4
Zn (mg/L)
8.6
8.5
9.3
7.2
SO4
1720
1550
1730
1650
PH
3.3
3.3
3.5
3.7
Eh(mV)
545
580
585
510
Temperature (C)
11
13
9.6
11
Input flow (L/min)
1.7
1.1
1.1
1.0
OUTPUT B-DOWNFLOW
B2-1106
B2-1113
B2-1206
B2-1220
Cu(mg/L)
0.558
<0.05
<0.05
0.180
Fe(mg/L)
2.69
<0.5
22.9
0.603
Mn (mg/L)
14.8
6.29
7.7
37.6
Zn (mg/L)
3.79
1.0
0.76
9.9
SO4 (mg/L)
1730
1700
1720
1680
pH
5.1
6.4
6.5
5.5
Eh(mV)
475
475
330
530
Temperature (C)
11.5
14
12.5
8
Input flow (L/min)
0.83
0.61
0.61
0.72
5-6
-------
Table 15. - continued
SAMPLING DATE
14-Sep-89
22-Sep-89
03-OC1-89
19-Oct-89
05-NOV-89
OUTPUT E-DOWNFLOW
OE-1109
OE-1111
OE-1209
OE-1222
OE-1309
Cu (mg/L)
<0.05
<0.05
<0.05
<0.05
<0.05
Fe (mg/L)
0.627
0.304
0.530
0.603
<0.5
Mn (mg/L)
20.6
19.6
24.0
24.2
19.7
Zn (mg/L)
<0.05
<0.05
<0.05
<0.05
<0.05
SO4 (mg/L)
1340
1370
1350
1450
1480
pH
6.5
6.4
6.3
6.4
6.5
Eh (mV)
5
-5
15
140
-40
Temperature (C)
13
14
14
10.0
10
Input flow (L/min)
0.72
2.8
2.5
0.93
2.3
Output flow (L/min)
0.31
missing
0.72 0.41
0.41
SAMPLING DATE
21-NOV-89
03-Dec-89
20-DOC-89
13-J«n-90
27-Jan-90
MINE DRAINAGE
MD-1310
MD-1401
MD-1410
MD-1501
MD-1510
Cu (mg/L)
0.85
0.578
0.606
0.590
Fe (mg/L)
35.1
42.7
38.2
33.3
Mn (mg)/L)
31.4
29.5
31.1
30.7
Zn (mg/L)
9.3
9.0
9.5
9.0
SO4 (mgil)
1690
1700
1660
PH
2.9
3.0
2.85
2.9
3.0
Eh (mV)
710
680
695
668
690
Temperature (C)
13.5
7.3
12.0
13
10.3
OUTPUT B-UPFLOW
B1-1311
B1-1405
B1-1411
B1-1505
B1-1512
Cu (mg/L)
0.284
0.440
0.439
0.489
Fe (mg/L)
<0.5
33.7
31.8
33.3
Mn (mg/L)
27.7
29.8
30.5
29.9
Zn (mg/L)
6.85
9.0
8.2
8.9
SO4 (mg/L)
1760
1690
1670
pH
5.5
3.5
3.15
3.25
3.55
Eh (mV)
515
460
595
535
670
Temperature (C)
7
6.4
6.0
6.3
2.7
Input flow (L/min)
0.29
1.4
0.55
0.81
0.54
5-7
-------
Table 15. - continued
SAMPLING DATE
21-NOV-89
03-Dec-90
OUTPUT B-DOWNFLOW
B2-1312
02-1406
Cu (mg/L)
0.217
0.209
Fe (mg/L)
3.51
7.2
Mn (mg/L)
22.5
24.7
Zn (mg/L)
8.1
6.84
SO4 (mg/L)
1730
1700
PH
5
4.1
Eh (mV)
530
560
Temperature (C)
8
2.6
Input flow (L/min)
clogged
1.1
OUTPUT E-DOWNFLOW
OE-1314
OE-1409
Cu (mg/L)
<0.05
<0.05
Fe(mg/L)
1.42
2.65
Mn (mg/L)
22.5
26.0
Zn(mg/L)
0.220
<0.05
SO4 (mg/L)
1580
1700
pH
6.7
6.3
Eh (mV)
135
25
Temperature (C)
11
7.1
Input flow (L/min)
0.28
0.58
Output flow (L/min)
missing
0.43
5-8
2O-D0C-89
13>Jan-90
27-Jan-!
82-1412
B2-1506
B2-1513
0.077
0.442
2.05
11.7
20.0
26.9
4.41
8.1
1680
5.7
3.25
3.3
385
660
740
4.5
7.7
5.8
0.18
0.76
0.59
OE-1414
OE-1509
OE-1514
<0.05
<0.05
14.5
9.8
25.1
27.7
0.125
<0.05
1630
6.5
6.0
6.25
305
140
345
3.0
6.3
4.0
missing
missing
missing
missing
0.39
2.5
-------
CELL B - S REMOVAL DATA
«
£
O
-J
Ll_
£
o
-J
U_
h
3
O
Mn B-D
\
Fe B-D
Cu B-D
Zn B-D
S04 B-D
100 120 140
DAYS
Figure 18. Removal of contaminants in CeH B Downftow over the first four months of operation.
-------
CELL E REMOVAL DATA
£
o
IL
z
5
r 2
-»• a_l
e u.
H
1.0
0.8
0.6"
0.4
0.2-
MnE
FeE
CuE
Zn E
S04E
20 40 60 80 100 120 140
DAYS
Figure 19.
Removal of contaminants in Celt E over the first four months of operation.
-------
v MINE DRN
¦ CELL B-U
* CELL B-D
CELLE-D
0 20 40 60 80 100 120 140
DAYS
20. Effluent pH for Cells B Upflow, B Downflow, and E over the first four months of
operation.
i.—;
-------
suggests that metal and sulfate removal are linked. This, along with the expected saturation of adsorption
sites in the substrate due to its prior two year exposure to mine drainage, suggests that adsorption is not
the major metal removal process. The relatively high removal of sulfate along with the measurement in the
output waters of 0.5 millimole per liter of sulfide and generally low Eh's indicate that significant sulfate
reduction is occurring. This pattern of sulfate and metal removal is explained by sulfate reduction and the
precipitation of metal sulfides. Sulfate concentrations are expected to decrease significantly as sulfide is
produced. Copper and Zn sulfides are expected to precipitate most readily followed by Fe sulfides and,
finally. Mn sulfides are expected to precipitate least readily. This pattern follows the trend in Ktp's for
these metal sulfides: CuS < ZnS < FeS < MnS (21).
Fluctuations of pH during this period for the mine drainage and for Cells B-Upflow, B-Downflow,
and E-Downflow are shown in Figure 20. The pH for the mine drainage tends to be relatively consistent
around 2.9. For Cell B-Upflow, the pH remains relatively tow, around 3.5, reflecting an absence of an
effective neutralizing process. The pH for Ceil B-Downflow. however, shows a decreasing trend from over
6 to below 3.5, indicating the presence of a neutralizing process that becomes less effective over time.
Adsorption may not be a consistent neutralizing process due to available sorption sites becoming filled
over time. The removal of metals in Cell B-Downflow varies with the fluctuation of pH in the outflow water
and is probably the effect of pH on the adsorption of metals onto the substrate material. Hydrogen ions
are expected to be more competitive for adsorption sites than metal ions, causing less adsorption of
metals and higher dissolved metal concentrations at lower pH's. As noted above for the adsorption
experiment, it is unlikely that significant metal hydroxide was formed. This is because no typical orange
coloration from a ferric hydroxide precipitate was observed and because Mn, Zn, and Cu are very soluble
at mine drainage concentrations and pH's below 6.5 (21).
The output water for Ceil E-Downflow has a consistently high pH, between 6.0 and 6.7,
suggesting the presence of a consistent process acting to neutralize the mine drainage. Sulfate
reduction may be a more consistent neutralizing process than adsorption due to the pairing of hydrogen
ions with continuously produced sulfide ions. There does not appear to be a trend between pH and metal
removal similar to that between the removal of sulfate and metals, implying that pH is not as Important as
sulfide precipitation in removing metals from solution. Although the pH is relatively high, it is unlikely that
significant metal removal is the result of metal hydroxide precipitation due to the reducing conditions of
the system and high metal solubilities (21).
pH INCREASE OF THE EFFLUENT
The consistently high pH of the effluent from Celt E shown in Figure 20 cannot be due to removal
of iron by oxidation. As shown in reactions 1 and 2 below, precipitation of Fe(OH)3 releases 2 moles of H+
for every mole of Fe oxidized and precipitated as Fe(OH)3.
5-12
-------
Fe* + 1/4 02 + H+ —> Fe^ + 1/2H2O [1]
Fe* + 3H2O —> Fe(0H)3 + 3H+ [2]
On the other hand, generation of HjS can increase the pH.
S04" + 2CH2O + 2H+ bacteria >H2S + 2H2O + 2CO2 (pH<7.0) (3)
S0<- + 2CH2O -bas*ada_> HS- + 2HC03- + H* (pH>7.0) (4)
It appears that in a wetland in which oxidation and precipitation of iron is the predominant removal process,
the pH of the effluent should decrease. Indeed, a few case studies have shown this to be the case (73,
74,75). On the other hand, in a wetland in which sulfate reduction is the predominant process, the pH of
the effluent can increase. However, there are two situations that confuse this analysis.
The first situation in which pH might not be lowered even though Fe(OH)3 is precipitated is when
the pH of the mine drainage is above 5.5. This is because the water is neutral enough to be buffered by
the bicarbonate system. Consider the following two equilibrium reactions:
COjq + H2O <—> H2CQ3
H2OQ3 <—> H+ + HCO3-
If the pH of the water is high enough to maintain carbonic acid in solution instead of CO2 exsolving, then
the pH of the water can be maintained at slightly below 7 by atmospheric carbon dioxide dissolving in the
water. The necessary minimum pH is about 5.5. Brodie (68) finds that if the pH of a mine drainage is
above 5.5, then treatment by a surface flow wetland will be effective. In this situation, precipitation of
Fe(OH)3 is the predominant removal process. The pH of the treated effluent remains above 5.5.
The second situation not related to sulfate reduction that can raise the pH is buffering by the
wetland substrate. In natural peat wetlands, the humic acid material will maintain the soil at a pH of about 4
(50). At this low pH, sulfate reducers will have a difficult time surviving and removal of heavy metals
through sulfide precipitation will be retarded. Indeed, this retarded the removal efficiency of the Big Five
Cells in the first year of operation. Choosing a more basic substrate will promote these reactions. In
addition, if the soil has a large buffering capacity, then basic conditions will be maintained until the wetland
substrate has time to generate a neutral pH suitable for sulfate reduction. For the four original materials
used in the Big Five cells, the acid-base characteristics are given in Table 16.
Since the peat is primarily composed of humic acids, its initial pH shows it to be quite acidic. On
the other hand, the other three materials have initial pH's that are basic. For the manure and decomposed
wood product, the odor of ammonia was readily apparent. This was caused by microbial breakdown of the
amino acids in the material. If basic substrate material is used, then the pH of the effluent during the initial
three to four months of operation may be controlled by the buffering capability of the substrate.
5-13
-------
In many respects, use of a substrate whose soil pH is above 7 is almost essential to the success of
sulfate reduction as a removal process, the sulfate reducing microbes operate best in the pH range
between 5 and 9.5. If the substrate has a soil pH between these ranges, and has some buffering capacity,
then the ability of the suifate-reducing bacteria to create their own microenvironment will help to maintain
the substrate pH around 7. At the Big Five site, the well waters within the cells give a good indication of
the pH of the substrate waters. Before mine drainage was introduced into the cells, the pH of well waters
ranged between 5 to 7.5. The pH of the well waters in Cell A were the lowest ranging from 5 to 6.6. After
10 months of operation, the pH of the well waters in Cell A ranged from 5.8 to 7.6 and at this time the pH's
in Cell A were from 6.2 to 7.3. the substrate in Cell C is underlain by six inches of limestone cobbles.
However, after 10 months, the pH in the well waters in Cell C were lower than those in Cell A, ranging from
5.8 to 7.0. Some of this ability of Cell A to maintain a high pH has to be attributed to the sulfate- reducing
bacteria.
Table 16. Acid-base characteristics of some substrate materials
Substrate Material
Split
Initial pH
Buffering
Capacity*
Mushroom Compost
11
8.05
0.769
17
8.30
0.672
Grant Bog Peat
8
3.10
0.192
4
3.10
0.197
Aged Manure
1
8.55
1.18
11
8.05
1.05
Decomposed Wood
5
8.60
0.987
17
8.65
0.987
aMillimoles of HCI needed to titrate one gram of substrate to a pH of 2.5.
SULFUR FORMS IN THE SUBSTRATE
In the summer of 1988 when it was speculated that sulfate reduction was the important removal
process, Laudon initiated research on the forms of sulfur in the substrates of the Big Five wetland cells (8).
Analytical results on forms of sulfur in the substrate done by ASTM test D-2492 (78) suggested a
significant increase in sulfate sulfur from initiation in October 1987 to the first substrate sampling in
January 1988 (61). This seemed confusing since if metal sulfides were precipitated, they would form
pyrite or metal monosulfides. This discrepancy led Laudon to conduct an extensive sequential extraction
5-14
-------
procedure for the various forms of sulfur on the selected substrates.
The extraction sequence is shown in Figure 21 and was modified after procedures used by Tuttle
(79) and Wieder (80). Substrate cores were taken, sealed, and immediately transported under wetland
water to the laboratory. The sequential analysis was started within two hours of sampling. The forms of
sulfur determined included:
Acid volatile sulfur (AVS): Sulfur in metal monosuifide precipitates such as FeS, CuS, and ZnS.
Elemental Sulfur (S<>)
Sulfate Sulfur: Sulfur primarily in pore waters.
Pyrftlc Sulfur: Sulfur as FeS2.
Organic Sulfur: Sulfur bound in insoluble organic compounds.
A separate analysis is made of total sulfur by the Eschka Method (81). In addition, total soluble
sulfide in nearby well waters was determined on site by an electrochemical titration method using 0.0001
M AgNOa (7). The description of the sample sites is in Table 17. Table 18 contains the results of the
sequential analyses. Table 19 contains the results of the soluble sulfur titrations.
Original substrate materials and those samples collected in January of 1988 were air dried and
stored in partially sealed polyethylene bags. Because acid volatile sulfides (AVS) can oxidize rapidly, it
was assumed that a separate AVS fraction could not be recovered. In these cases AVS is reported as
AVS + So. Laudon also performed duplicate analyses and recoveiy tests and these are discussed in her
thesis (8). The sum of the fractions agreed with the total sulfur analysis to within 20 % in-all cases. The
relative standard deviations on total sulfur analyses were within 10 %. Large deviations occurred on
duplicates of the sequential extractions especially when the form of sulfur was only present in minor
amounts.
Figures 22 and 23 show, in Cells A and B respectively, the changes in sulfur content in the
substrates over the first 10 months of operation. The results show increases in the AVS in the substrates,
especially in those from Cell A which was the best performing cell in 1987-88. Certainly in this cell, the
removal mechanism is formation of metal sulfides by dissimilatory sulfate reduction. In Cells B and C there
is an increase in the amount of acid volatile sulfides but one could not assert that removal through sulfate
reduction was the predominant process.
Two other conclusions from the results of Laudorfs research should be pointed out. In Cell A, the
system that has the greatest sulfur increases, there is no apparent increase in organic or pyritic sulfur. In
Cell B, there may be an increase in these two sulfur forms. The lack of pyritic sulfur formation is contrary to
the expectation was that pyrite would form immediately or, at least, some of the acid volatile sulfide would
change to pyrite. Upon review of the literature (82), it is not unusual to form significant amounts of AVS
with little formation of pyrite. Acid volatile sulfides form under conditions that are strongly reducing and
remain that way. Also, there is little diagenesis of acid volatile sulfides to pyrite. Pyrite formation requires
the presence of elemental sulfur or polysulfides. These will be present under more oxidizing conditions
5*15
-------
ORIG.
AIR t
DRY
ESCHKA
SAMP.
DRY
WET
FUSION
5
TOTAL
S
HoS
6 N
Ba 2+
HCI
CAI Kl
- AVS
so;
ACETONE
EXTRACT.
*
+
CM
o
SOLN.
REDN.
H2^ ELEMENT
' SULFUR
REDN.
HoS
DISULFIDES
(FeS 2)
ESCHKA
ORGANIC
FUSION
SULFUR
Figure 21. Extraction sequence for the forme of sulfur determination.
5-16
-------
Table 17. Locations of sample used in the sequential sulfur extraction.
Abbreviation
Description
cs
Compost substrate (initial material used in cett A).
TAS
Top 6" of substrate collected near well A5.
#1(A2)
First core collected (near well A2).
#2(A5)
Second core collected (near well A5).
P/M/WP
Peat, manure, wood products (Initial material in cells B and C).
TSS
Top 6* of substrate collected near well B5.
B(B5)
Core colected in cell B near well B5.
C(C5)
Core collected in cell C near well C5.
Table 18. Forms oi S in substrate samples and NBS Coal 1635. All values in %S.
Sample
«¦¦¦¦ Masai
Date
Collected
«asiiBBMa ¦ mm
Stot
vh wa
SavS
iai>nn
Ssp
SFeS2
KraaBMHM
SoRG
SS04
TA5
1/88
0.83
0.45*
0.05
0.21
0.17
#1(A2)
7/88
1.61
0.31
0.49
0.24
0.35
0.07
#2(A5)
7/88 .
1.39#
0.83
0.16
0.09+
0.22+
0.08+
P/M/WP
1/87
0.61
<.02
0.05
0.34
0.14
0.16
TB5
1/88
0.59
-
0.09*
0.25
0.15
0.15
B(B5)
7/88
0.67#
.08
0.09
0.42+
0.20
0.02
C(C5)
7/88
0.73+
0.19
0.13
0.40
0.19
0.07
NBS Coal
1635
0.32
<0.03
0.02
<0.02
0.35
<0.02
*AVS + S°
^average of 3 values
+average of 2 values
SjOT ¦ Total Sulfur
SavS " Acid volatile sulfur
Ss° - Acetone soluble sulfur
SFeS2 ¦ Disulfides
SoRG ¦ Organic sulfur
Ss04 ¦ SO42- sulfur
Table 19. Results of H2S titration on well water samples.
Screened
Wed Interval mg/L S as HjS
A
A4
A5
B4
C5
3*
V
2"
V
3'
20
<1
79
49
9
5-17
-------
1
AVS+S
0.9
0.7
0.5
0.3
ORG S
FeS2
0.1
1988
JAN
JULY
Figure 22. Changes in sulfur content and forms within the substrate
in Ceil A over the first 10 months.
5-18
-------
0.9
0.7
0.5
FeS
0.3
ORG S
0.1
AVS + S
1987 OCT JAN 1988 JULY
Figure 23. Changes In sulfur content and forms within the substrate
in Cell B over the first 10 months.
5-19
-------
or in areas where there is periodic incursion of oxygen (82,83). Apparently in Ceil A, the vigor of sulfate
reduction coupled with the highly anaerobic conditions account for acid volatile sulfides being the
predominant product.
FORMS OF HEAVY METALS IN THE SUBSTRATE
Since research on sulfur forms points to the formation of metal sulfides in the substrate, the next
logical step is to investigate where the heavy metals are accumulating in the substrate. It was hoped that
this could be done by some spectroscopic method thus giving an atomic confirmation of the presence
and molecular form of the metal sulfides. Since sulfur is present in abundances greater than one percent,
x-ray diffractometry should detect pyrite or iron monosulfide crystals. To enhance the concentration of the
heavy metal sulfides, the substrate samples were screened and x-ray diffractometry was performed on the
fraction that was less than 200 mesh. Quartz, feldspar, illite, and gypsum were identified but no metal
sulfides or oxides could be conclusively identified. Apparently, if precipitates are formed, they are
amorphous. Mossbauer spectroscopy is an excellent method for detecting and measuring the
concentrations of pyrite and other iron oxides in solids (84). Often this spectroscopy can detect iron
minerals that are amorphous to x-ray diffractometry. Two fresh sediments were taken from Cell A, packed
in sealed sample containers, and analyzed by Mossbauer spectroscopy. There was some Fe(ll) mineral
present in small amounts, but no pyrite or FeS was detected. Again the precipitate is so fine grained that it
was amorphous to Mossbauer spectroscopy.
Because spectroscopy does not reveal how the heavy metals are bound to the substrate, it was
decided to perform sequential extraction studies on the heavy metals. These procedures would be
comparable to the studies of the forms of sulfur. As reviewed by Chao (85), the objective of a sequential
extraction is to chemically determine the forms of metal compounds In a soil or sediment. Bound forms of
metals typically tested for include:
o Easily extracted cations loosely bound to the sediment;
o Metals bound by organic complexes, usually humic acids;
o Metals adsorbed onto manganese (IV) oxides;
o Metals adsorbed onto iron (III) oxides;
o Metals precipitated as sulfides or oxides;
o Metals bound in silicate or other resistant minerals.
Because of the sequential nature of the extraction, a number of uncertainties are inherent in the
experiment. First it is obvious that errors accumulate and so one can only expect correspondences to
within 20 percent for duplicate analyses. One method for determining the success of an experiment is to
compare the sum of abundances in the extraction with a separate total metal analysis. If the two compare
5-20
-------
within 20 %, the experiment is considered successful. Related to the general experimental error is the
fact that the more steps in an extraction sequence, the greater the chance for systematic or accidental
error.
There also are chemical problems that are sometimes unavoidable. The extraction steps assume
discrete phases have been formed (86), especially for the oxides and sulfides. In the case of Big Five
substrates, the targeted phases are x-ray amorphous, and so the chances for discrete phases in the
substrate Is diminished. The experimental design assumes that the chemical agents in each step will
attack and release the metals from only that phase and also do this completely. Complete separation is
unlikely, especially for the organically bound rnetals (87, 88) and for the Mn and Fe oxides (89, 90, 91).
Finally, it is assumed that a metal released in a certain step will not be resorted when the solid is separated
from the extract. This is a controversial assumption especially with regard to the first steps of the extraction
sequence (92, 93).
In summary, sequential extractions to determine how metals are bound or contained in sediments
require careful interpretation. Perhaps the best way to interpret the results is that the tests show the
tendency of metals held by the substrate to be released. The first steps in the sequence use relatively
mild reagents and continue to more aggressive reagents in the final steps. If a metal is released in an early
step, it is relatively mobile and the possibility of it being released back into the environment is high. If a
metal remains until the final extraction steps, it Is in a more resistant site in the substrate and the possibility
°f release is diminished.
Two different sequential experiments were performed on substrates from the Big Five cells: a six
step sequence, and a five step sequence. The six step sequence was the first method tried. The first
steps in this sequence are quite aggressive in terms of pH changes. The acetic acid buffer is used to
release carbonates as well as easily extracted metals. The pyrophosphate step uses a relatively high pH to
dissolve humic acids and the metals associated with them (50, 87). The five step sequence uses
methods for releasing easily extractable metals and organically bound metals that are more mild in terms of
PH changes. However the sodium hypochlorite used to attack organic sites may oxidize metal sulfides.
Ateo, this sequence has no step specific to metal sulfides.
SjfcStftn Fraction Sflnnanrft
Figure 24 shows the steps in the six step sequence. This sequence was tried on substrate
samples taken from the top of Cell A in July 1988, about the same time samples tor the forms of sulfur
extraction were taken. They were chosen because they had accumulated the highest amounts of metals.
The samples were air dried and stored until the analyses were performed in March, 1989. Some of the acid
volatile sulfides may have oxidized during air drying and storage. Table 20 gives the results of the six step
sequential extraction experiments. Figures 25 through 28 compare the metal speciation in the original
mushroom compost to speciation in the average of the three Cell A samples.
5-21
-------
For manganese and zinc the experiments show that the metals were held in sites where they
could be easily released. Iron is distributed throughout all the sites but mostly is contained in an organic
form. Copper is strongly held. It persists until the step that would release crystalline oxides and sulfides.
These experiments indicate that the heavy metals are held to the substrate on sites where release is quite
possible. Oxidation during storage may account for some of this mobility. However if the substrate was
removed from Cell A and prepared for disposal, it likely would be subject to some oxidation. It may be that
the acetic acid in the first step causes the dissolution of poorly formed oxides or sulfides. For this reason,
the five step sequence was tried.
Five-Step Extraction Saotianca
Figure 29 shows the steps in the five step sequence. This sequence was also tried on the same
substrate samples taken from the top of Cell A in July 1988. The analyses were performed by Sellstone
(94) in the fall of 1989 as part of his thesis project. Again, some of the acid volatile sulfides may have
oxidized during air drying and storage. Table 21 gives the results of the five step sequential extraction
experiments. Figures 30 through 33 compare the metal speciation for the original mushroom compost to
speciation on the average of the three Cell A samples. Due to an analytical problem, the absolute
abundances for zinc in each step were lost.
The first step of this sequence uses no acid and assumes that magnesium replaces the easily
extracted metal on the substrate. As can be seen in Figures 30 through 33, now only Mn is appreciably
released in the first step. All the other metals are held until the last two steps. The idea that the acetic acid
in the first step of the first extraction sequence released metais from acid sensitive sites appears to be
reasonable.
Metal Extraction Summary
The metal extractions could not definitively show the sites where heavy metals were bound onto
the wetland substrates. Since there also is no spectroscopic evidence, it is not certain where metals are
bound after all the organic sites are occupied. The metal extraction experiments do give an indication of
the tendency of heavy metals to be released from the substrate. It appears that mobilization could be
initiated by miid acids, probably due to the formation of AVS. This suggests that the substrate could be
classified as a hazardous waste once it is removed from the wetland, we are presently considering
experiments in which deeper, less organic-rich wetlands will be tested to treat the same volume of mine
drainage. The aim of these experiments is to control the redox potential and rate of sulfate reduction in
order to maximize the ratio of pyrfte-to-AVS that is produced.
5-22
-------
Ul
•
N
U
SEQUENTIAL METALS EXTRACTION
METAL FORM
TOTAL METALS
EASILY EXTRACTED
ORIG.
HNOo & Ho0o
° \
SAMPLE
1.5 M HO AC
SOLN.
Na4P2°7
0.10 M, pHa10
SOLN.
0.2 M NH2OH:HCI
SOLN.
0.02 M HNO3
0.5 M NH2OH:HCI
0.5 M HCI
— *
SOLN.
4 M HNO,
50° C
SOLN.
hno3& h2o2
SOLN.
& CARBONATES
ORGANIC BOUND
Mn OXIDE BOUND
Fe OXIDE BOUND
CRYSTALLINE
OXIDES & SULFIDES
RESIDUE
Figure 24.
Six step sequential extraction for metal speciation In substrate samples.
-------
Table 20. Results of the Six-step Metal Extraction Sequence
on Substrates from the Top of Cell A.
Orig. Compost
Soil A4, 6"
Soil A6,
6"
Soil A3,
6"
Avg Soil
Step
pm
%
ppm
%
ppm
%
ppm
%
%
Manganese
1
228
68
562
86
837
85
1008
86
85
2
23
7
24
4
48
5
54
5
4
3
13
4
5
1
18
2
20
2
2
4
23
7
21
3
36
4
46
4
4
5
9
3
9
1
12
1
14
1
1
6
36
12
35
5
35
4
32
3
4
Totaia
334
-
656
-
987
-
1173
-
-
ActTota
359
93
609
108
1013
97
1327
88
-
Copper
1
5
6
4
3
5
2
6
3
2
2
5
7
22
16
46
14
46
26
18
3
0
0
26
19
12
4
27
15
10
4
9
12
26
19
100
30
29
16
24
5
34
47
39
29
135
40
43
24
34
6
20
28
18
14
36
11
27
15
13
Total
73
-
138
-
334
-
179
-
-
Act Tot
57
128
165
83
474
70
182
98
-
Iron
1
261
3
2900
14
3280
19
1030
5
13
2
273
3
8000
40
5900
35
8450
44
40
3
273
3
2190
11
300
2
1930
10
8
4
1260
15
1280
6
1540
9
1580
8
8
5
930
11
1280
6
1000
6
1280
7
6
6
5230
64
4600
23
4900
29
4850
25
25
Total
8220
-
20200
-
17000
•
19100
-
.
Act Tot
9640
85
23600
86
20900
81
23080
83
-
Zinc
1
53
42
711
59
1860
53
723
69
57
2
18
14
284
24
971
28
197
19
25
3
13
10
64
5
93
3
42
4
3
4
23
18
107
9
442
13
59
6
11
5
8
6
17
1
106
3
8
1
2
6
12
9
13
1
24
1
13
1
1
Total
127
-
1197
-
3490
-
1040
-
•
Act Tot
127
100
1398
86
3950
88
1180
88
-
aThe total * sum of the partial extractions.
The Act Total - results of a separate total analysis.
The percent next to Act Tot.-(Total/Act Total) x 100
5-24
-------
Ul
IS*
at
Mn IN CELL A SOIL FRACTIONS
370 ppm 1060 ppm
100
p
80
E
R
60
C
E
40
N
T
20
0
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT Ml ORGANIC
Fe OXIDE Bill SULFIDES
Mn OXIDE
RESIDUE
Figure 25. Six step manganese speciation In original mushroom compost
and in substrate from the top of Cell A after 10 months.
-------
Zn IN CELL A SOIL FRACTIONS
130 ppm 2180 ppm
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT.
ORGANIC I I Mn OXIDE
Fe OXIDE ffiSSBI SULFIDES
RESIDUE
Figure 26. Six step zinc speciation in original mushroom compost
and in substrate from the top of Cell A after 10 months.
-------
Fe IN CELL A SOIL FRACTIONS
0.96 % 2.25 %
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT.
ORGANIC
Fe OXIDE ¦¦ SULFIDES
Mn OXIDE
RESIDUE
Figure 27. Six &tep Iron speclation In the original mushroom compost
and in substrate from the top of Cell A after 10 months.
-------
Cu IN CELL A SOIL FRACTIONS
58 ppm
270 ppm
::::::::::::::::
¦ll
—JTTTTTTTT^
liiil
a
SUB
lilliiliiilHl
mnngjg!
iiiiiiiiiiliil
SSTOtil
Mil
it:::::::::"
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT.
ORGANIC I I Mn OXIDE
Fe OXIDE liliil SULFIDES
RESIDUE
Figure 28. Six step copper speciation in the original mushroom compost
and in substrate from the top of Cell A after 10 months.
-------
M
(O
SEQUENTIAL METALS EXTRACTION "llw
METAL FORM
OR G.
SAMPLE
hno3& h2o2
-v- TOTAL METALS
SOLN.
1 M MgCI2
SOLN.
0.7 M NaOCI
u
pH ¦ 8.5
v- EASILY EXTRACTED
0.1 M NH2OH:HCI
pH » 2.0
SOLN.
v- ORGANIC BOUND
Mn OXIDE BOUND
0.2 M OXALIC BUFFER
0.1 M ASCORBIC ACID
SOLN.
¦*- Fe OXIDE BOUND
hno3& h2o2
SOLN.
RESIDUE
Figure 29. Five step sequential extraction for metal speciation in substrate samples.
-------
Table 21. Results of the Five-step Metal Extraction Sequence
on Substrates from the Top of Cell A.
Orig. Compost Soil A4, 6" Soil A6,6" Soil A3, 6" Avg Soil
Step pm % ppm % ppm % ppm % %
Manganese
1
35
6
470
44
390
34
720
45
44
2
29
6
160
14
116
10
210
13
14
3
217
41
67
15
220
19
290
18
15
4
121
23
68
15
240
21
260
16
15
5
127
24
101
13
182
16
170
11
13
Total
529
1070
1153
1660
Act Tot
490
108
1160
92
1400
82
1600
103
-
Copper
1
0
0
0
0
0
0
0
0
0
2
0.9
0
2.1
1
1.0
0
0
0
0
3
0
0
9.6
2
0
0
0
0
0
4
9.9
15
36
8
13
9
13
10
9
5
47
89
410
90
135
91
123
90
90
Total
58
458
149
136
Act Tot
54
107
390
116
170
68
130
105
-
Iron
1
8
0
200
0
0
0
10
0
0
2
35
0
370
0
60
0
80
0
0
3
71
1
620
3
310
2
320
2
2
4
7540
56
11200
59
13200
60
14300
64
61
5
5840
43
7300
38
8200
38
7800
32
37
Total
13500
19700
21800
22500
Act Tot
12000
112
19000
107
20000
109
17000
132
-
Zinc
1
0
-
17
3
-
3
8
2
3
-
18
4
5
9
3
4
-
26
23
23
24
4
54
-
20
51
48
40
5
38
-
20
18
20
20
Total
-
Act Tot
'
"
'
•
5-30
-------
Mn IN CELL A SOIL 2nd EXTRACTION
490 ppm 1300 ppm
50
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT
Mi
ORGANIC
Mn OXIDE
Fe OXIDE Bill RESIDUE
Figure 30. Five step manganese spedation in the original mushroom compost
and in substrate Irom the top ol Cell A alter 10 months
-------
Zn IN CELL A SOIL 2nd EXTRACTION
23 ppm 340 ppm
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT.
ORGANIC
Mn OXIDE
Fe OXIDE liliiSi RESIDUE
Figure 31. Five step zinc speciation in the original mushroom compost
and in substrate from the top of Cell A after 10 months.
-------
Fe IN CELL A SOIL 2nd EXTRACTION
1.2 %
1.9 %
cn
w
w
P
E
R
C
E
N
T
70
60
-
50
-
S.
x SV
40
30
-
? ¦
: S
20
-
*
10
«
fllllllllll
X-V ^
0
. 1
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT. ¦¦ ORGANIC
Mn OXIDE
Fe OXIDE liSli RESIDUE
Figure 32. Five step iron speciation in the original mushroom compost
and in substrate from the top of Cell A after 10 months.
-------
Cu
IN CELL A SOIL
2nd
EXTRACTION
54 ppm
220 ppm
U»
CO
¦u
p
E
R
C
E
N
T
100
80
60
40
20
0
ORIGINAL SUBSTRATE TOPSOIL (10 MONTHS)
EXTRACT.
ORGANIC
Mn OXIDE
Fe OXIDE ¦¦ RESIDUE
Figure 33. Five step copper speciation in the original mushroom compost and in substrate from
the top of Cell A after 10 months.
-------
DECREASE OF SULFATE IN THE EFFLUENT
If by reactions 3 and 4 sulfate is reduced to sulfide, and It is precipitated as a metal sulfide; then the
concentration of sulfate should decrease in the effluent leaving a wetland cell. However, this decrease
can be difficult to verify. As Laudon reviewed in her thesis (7), it is quite easy to oxidize the precipitated
sulfides by using the bacteria that oxidize pyrite. Wieder (95) observed that Big Run Bog, West Virginia
was a source of SO4- to receiving streams during periods of low flow and a sink for SO*- during high flow
periods corresponding to the water saturation status of the wetland.
Based on seasonal studies of saitmarsh sediments, Cutter and Velinsky (96) interpreted that in
spring to early summer, photosynthesis injects 02 into the upper 15 cm of sediment that oxidizes Fe
sulfides and precipitates Fe oxides and elemental sulfur. Then in autumn, O2 infusion slows and sulfate
reduction predominates. Fe oxides and elemental sulfur are reduced and reprecipitated as Fe sulfides.
Finally in winter, oxidation and reduction rates slow but Fe sulfides continue to precipitate due to upward
diffusion of H2S and Fe(ll).
It appears that in natural wetlands, sulfate concentration oould be higher or lower than the average
value depending on climatic conditions. In the first year, it was not clear that a change in sulfate
concentration would occur.
Sulfate Reduction Evidence
When the Emerging Technologies Project started, sulfate analyses started to be performed at
Mines by the Eschka method (81). Since this method uses a sulfate precipitation and the concentration of
sulfate in the Big Five waters is quite high, the precision of duplicate analyses was well within 5 %. The first
analyses by this method were through late fall and winter in 1988-89. In Table 22, the sulfate
concentrations and pH of the mine drainage, Cell A, and Cell C are reported. It is obvious that the
difference in sulfate concentration between the mine drainage and the ceil effluents is minimal. However
in February, the pH in Ceil A effluent increased and the sulfate decreased beyond the analytical error.
Also in February, the sulfate concentration in Cell C effluent increased to a level definitely above the
concentration on the mine drainage. It appears that sulfate concentration can increase or decrease
depending on whether sulfide is stable in the wetland substrate. In the case of sulfate decrease, the pH
appears to increase.
In October, 1989, to further test the sulfur budget, total sulfide in Cell E pore waters was
determined by electrochemical titration at the time of monthly water sampling. The balance of sulfur is
shown in Figure 34. In this balance, it was assumed that all of Mn, Fe, Cu, and Zn were precipitated as
sulfides. If this is the case, only 1.08 millimole / liter (100 mg/L) of sulfate is needed to completely
precipitate all the metal sulfides. The Big Five mine drainage is a gypsum water as are most mine drainage.
There is an excess of sulfate in the water. Another important feature of this balance is that when sulfate
5.35
-------
concentration in the effluent is decreased, there is a definite presence of sulfide in the wetland pore
waters. Both these experiments confirm that if sulfate-reducing bacteria are operating and sulfide
precipitation is an important removal process, then the sulfate concentration should decrease.
When Cells B and E were started in September, 1989, changes in sulfate concentration as well as
heavy metals were closely monitored. In Figures 18 and 19 sulfate removal as output / input is plotted for
the first four months of operation. In Figure 20 pH is plotted. Note that in Cell B-downflow sulfate has not
decreased and the pH of the effluent has not risen above 4 even though metals are removed. As
discussed earlier in this section, removal in this cell was by adsorption onto organic phases. On the other
hand, sulfate is being removed from Cell E and the pH is around 6. In laboratory experiments, Machemer
(7) found the same pattern of concentration change when sulfate-reducing bacteria were vigorously
growing. Reduction in the concentration of sulfate is a reliable indicator of removal of metals by bacterial
reduction of sulfate and precipitation of sulfides.
Table 22. Sulfate concentrations and pH's in the
Big Five Cells in the Autumn and Winter of 1988-89.
WATER
NOV
DEC
JAN
FEB
SO-4
1750
1710
1690
1780
DRAINAGE
PH
2.9
3.0
2.8
3.0
(/>
O
»
1690
1710
1670
1660
CELL A
PH
3.3
3.4
3.3
5.1
SO-4
1740
1700
1720
2000
CELL C
PH
3.2
3.3
3.1
3.5
Changes in Sulfate finnr»ntration with Flow
Another indication that continuous bacterial reduction of sulfate is manifested by a reduction in
sulfate concentration is the change in chemistiy of effluent water with the flow through the wetland cell.
From 1987 through 1989, changes in chemical parameters in Cell A effluent were compared with flow into
the wetland cell (11). For metal concentrations, correlations were not obvious. However the
co icentration of sulfate in Cell A effluent did decrease as the flow decreased. A graph of the change is
shown in Figure 35. A linear regression analysis on the data points produced a correlation coefficient of
0.82. In addition, the Eh decreased as the sulfate decreased. The graph is shown is Figure 36. In this
case the correlation coefficient is 0.78.
5-36
-------
As discussed in SECTION 3, sulfate-reducing bacteria require reducing conditions. The two
chemical variables that are most directly related to sutfate-reducer activity do change in the expected
manner. Removal of heavy metals requires that sulfide, the product of sulfate reduction, comes in contact
with the mine drainage. Apparently in Cell A, this does not always happen. However, the results of Cell E
as shown in Figures 16 and 19 prove that proper design can make this removal linkage be effective.
SUMMARY
Through all the experiments performed on the Big Five Cells, the case for vigorous sulfate-
reducing activity being linked to metals removal has been reasonably shown. Of the experiments, the
most direct effect of sulfate reduction-decrease in sulfate concentration in the effluent-appears to be the
best indicator. Now that it has been shown that sulfate concentration should be consistently lower, this
can be used to monitor the removal efficiency of a wetland removal system.
5-37
-------
SULFUR BALANCE
CELL E
INFLOW
(ALL VALUES IN MMOL/L)
.8 AS SULFATE IN SOLN.
RETAINED
RELEASED
jAIR ??
v
1.07 AS
{METAL SULFIDES
14.6 AS
SULFATE IN SOLN
X
0.5 AS
SULFIDE IN SOLN.
Figure 34.
SuNur balance in Cell E in October, 1989.
-------
2.0
1.5
1.0
0.5
0.0
0
100
150
200
50
S04 DEC. (mg / L)
Figure 35. Change in sulfate concentration versus flow in Cell A.
5-39
-------
RA2 = 0.788
O 300"
S04 DEC. (mg/L)
Figure 36. Change in Eh versus decrease in sulfate concentration in Ceil A.
5-40
-------
SECTION 6
ECOPHYSIOLOGICAL CONSIDERATIONS OF PLANTS
AT THE BIG FIVE CONSTRUCTED WETLAND, IDAHO SPRINGS, COLORADO
INTRODUCTION
One component of the pilot-scale treatment system at the Big Five tunnel is emergent wetland
vegetation. The role and importance of the vegetation to the metal-removal efficiency of the system is not
entirely known. However there is documentation at other constructed wetland sites of improved removal
of some metals in systems with emergent vegetation, when compared to similar systems without
vegetation. Among the effects the wetland vegetation has upon the treatment system are:
o oxygenation of the substrate;
o provision of nutrients to microorganisms living within the substrate, both by exudation of
chemical substances from the plant roots and rhizomes that are used by certain microbes
(e.g. sugars), and by addition of biomass;
o alteration of the permeability and hence the flow dynamics of the substrate, in the upper
pari of the substrate;
o effects upon the water budget of the system, through evapotranspiration;
o aesthetic enhancement
The primary objectives of the study of the emergent vegetation at the Big Five system were: 1) to
further the understanding of the role that emergent vegetation has upon the dynamics of the treatment
system (e.g. addition of organic material, oxygenation, evapotranspiration); and 2) to examine the health
of the emergent plants by such means as monitoring of elemental uptake, visual appraisal, and
comparison of annual mapping.
QUANTIFICATION OF BIOMASS
One effect the vegetation has upon the treatment system is by addition of organic material as
biomass to the substrate, providing a nutrient source for microorganisms. The addition of such biomass
might extend the life of the substrate within a cell, by possibly providing more sites for complexation of
metals, as well as nutrients for microbial processes. An attempt was made to quantify the amount of
material added to the substrate by the plants. Because of limitations on the totai amount of plant material,
and because of a concern that destructive sampling might affect cell performance, methods that were
largely nondestructive were used to estimate biomass.
6-1
-------
The measurements were made on Cell C in summer 1989, as this cell had been undisturbed Irom
the beginning of the project, and had essentially total vegetation cover. This quantification was made by
the following method:
1) Visual appraisal that Typha (cattail) biomass comprised on the order of 75% of total biomass of
plants in the cell (based on the estimates of four different workers at the site)
2) Mapping and counting of the number of Typha plants within the cell.
3) Estimation of total Typha leaf area, by taking the mean of leaf area measurements for several
plants. This leaf area was then multiplied by average dry weight per area lor Typha leaves,
resulting in total dry weight of leaves.
4) Estimation of total below-ground biomass. This component probably has the uncertainty. Dry
weights of above/ below-ground biomass were compared for several plants, with the below-
ground including roots and rhizomes. Results indicated more dry weight biomass below ground
level, on the order of 1.75 to 2 times as much as above. Other reports (58) indicate commonly
50% of Typha biomass is below ground, with variation. As a conservative compromise, a figure of
1.5 times as much below-ground biomass as above ground was used.
This method resulted in the following estimate:
767 total Typha plants in Cell C X 33.5 g dry weight leaves per plant - 25.69 kg total leaves
Biomass below water level: 1.5X25.6 kg - 38.53 kg
Total Typha biomass: 64.22 kg
Total biomass estimate for all species (estimating Carex aquatilis, Carex utriculata, and Juncus
arcticus to comprise 25% of total) in Cell C: 85.62 kg.
This figure is an approximation of the dry weight amount of organic material contributed to Cell C in
the 1989 growing season. Analysis of the plant samples in 1989 showed an average dry weight of
31.07% carbon content. Using these figures, (total biomass X percent carbon), an estimated 26.6 kg
carbon was represented by the biomass of Cell C in 1989. For Cell C, with an area of 18.6 m2, this would
be 1.43 kg/ m* of carbon. For natural wetlands, an average figure is often near 1 kg/rtf. On an annual
basis, all of the above-ground, and perhaps half of the below-ground biomass will senesce and contribute
to the soil organic matter. This would represent approximately 70 percent of the total carbon in the
biomass entering into the organic carbon pool. For Cell C this would amount to about 1 kg / m2 of carbon.
This amount may be insufficient to meet most of the microbial carbon demands of sulfate-reducing
bacteria (65); and could, in fact, become a limiting factor on the system once the carbon in the origin.*.!
substrate is depleted.
6*2
-------
EFFECT OF PLANT RESPIRATION ON Eh VALUES
Another effect of the emergent vegetation upon the treatment system is from the diffusion of
oxygen from the roots into the surrounding substrate. The result is the creation of localized oxidizing
zones within the substrate, which should be conducive to removal of some metals by oxidation (primarily
iron). This is probably a contributing factor to the seasonality of metals removal reported in other
constructed wetlands (73). Batal (10) reports increased populations of iron- and manganese- oxidizing
bacteria during summer 1988, and decreased populations of sulfate-reducing bacteria during that
summer, at the Big 5 wetland. Data for 1989-90 in Table 7, however, show that the populations of iron-
oxidizing apparently have decreased steadily since 1988 in Cells A and C particularly in the 90 cm depth in
the substrate. Along with temperature, this release of oxygen from plant roots probably affects the
microbial populations particularly at shallow depths in the substrate.
A simple experiment was conducted to examine some of the effects of plants upon the oxidation
potential values in the substrate. Eh measurements were made under controlled conditions, where a
microprobe could be used to measure directly next to plant roots. Substrate from Cell A was placed in a
30-gallon glass tank in an indoor growing chamber and roots and rhizomes of Typha plants (cattails) were
transplanted into the tank. Big Five mine drainage water was used to keep the substrate below water
level, simulating conditions at the treatment site. After one month, several Typha plants had grown to a
height of 1-1.5 meters, and the first set of measurements was made. Over 50 Eh measurements were
made at varying depths within the tank, making one set of measurements within a centimeter of plant
roots, and one set 10-15 cm. away from plant roots (Table 23).
Table 23. RESULTS OF FIRST Eh MEASUREMENTS IN GROWING CHAMBER
Measurements under plants/ or
next to visible roots (millivolts)
Eh measurements away from
plants (millivolts)
Depth Mean High Low StdDev
Depth Mean High Low Std.Dev
1.5 cm +131 +240 +60 56
4 cm +94 +150 +20 41
12 cm +78 +170 -60 67
1.5 cm +110 +150 +40 38
4 cm +70 +130 -70 67
12 cm -53 +60 -200 126
6-3
-------
Not enough measurements were made for rigorous statistical testing. However there is a distinct
qualitative difference in the oxidation potential near plant roots where oxygen is diffusing from the roots to
the surrounding substrate, compared to measurements in substrate away from plant roots. In every case
of paired measurements (under/away from plant roots), Eh values were higher underneath the plants. At
the deeper depth, the mean difference was more than 100 mv, indicating small, localized oxygenated
zones due to the presence of the roots. The experiment was later repeated using two tanks, one
completely filled with Typha plants, and one with no cattails. Slightly greater differences were seen
between the Eh measurements of the two tanks, with a mean difference at 12 cm depth of 133 mv higher
in the tank with plants. These results show an oxygenation effect of the plants within and near the root
zones that should affect metal removal by bacterial processes.
Other research (60) has compared the amount of oxygen diffusion (termed radial oxygen loss)
from roots of five different wetland plants from a constructed wetland in Tennessee, and found Typha to
have the highest oxygen loss of those compared.
EVAPOTRANSPIRATION MEASUREMENTS
Transpiration from the wetland vegetation has an effect upon the water budget of the treatment
system. To estimate the magnitude of this component, measurements were made, at different times of
the growing season, of total evapotranspiration (ET) by measuring the difference in flow between input
and output to each cell. Evaporation was also estimated by use of evaporation pans, and the amount of
surface evaporation from the cell calculated. The difference between evaporation and total water loss was
taken as the transpiration component. These measurements were taken at regular intervals over 24 hours
to assess the diurnal variation. Again, data for Cell C were used, as it had virtually a complete vegetation
cover for the entire growing season.
During summer, the transpiration component was found to be greater than the evaporation
component, by almost an order of magnitude. Figure 37 shows one measurement of evapotranspiration
taken at peak growing season. For Cell C, the water toss in mid-afternoon was 0.13 gpm, or almost 20% of
the inflow of 0.66 gpm; surface evaporation accounted for only approximately a 1.5% loss of the inflow,
with the remaining 18% loss from transpiration. However, the water loss at pre-dawn measurements,
when transpiration was at a minimum, was only 2%. Averaged over a 24- period, the water loss was 8-9%
of total inflow.
An important consideration in such calculations is the incoming flow rate. Presumably, the
evapotranspiration will remain approximately the same, no matter what the flow rate, as long as the
substrate remains sufficiently moist, and the increased concentration of salts does not present an osmotic
barrier to the plants. With a flow rate of half as much, as has been used at the Big 5 system at times, this
6-4
-------
Diurnal variation in evapotranspiration, August 21/22, 1989
0.14
Inflow range .63 — .66 gal/min
High temp. 74° F
19 % of inflow
0.12
14.5% of inflow
0.10
m
13 % of inflow
0.08
10 % of inflow
0.06
6% of inflow
0.04
4% of inflow
0.02
2% of inflow
I
I
i
1
I
1
1
I
1
1
I
0.00
1 p.m. 3
Time
Figure 37. Diurnal variation in evapotranspiration from Celt C over a 24 hour period, August 21
/ 22, 1989.
-------
evapotranspiration percentage would double. Flow measurements by Lemke (pers. communication,
1989) show water loss as high as 50% from Cell C, at peak ET times and low flow rates. Conceivably, at
very low flow rates and peak ET rates, the evapotranspiration could account for most of the flow.
Evapotranspiration should be considered in terms of ET per area, rather than as percentage of flow rate.
At these times of high water loss, there will be an increased concentration of substances
dissolved in the drainage water flowing through the wetland cell. This raises the question of how this
affects the treatment system's ability to remove the materials desired when they are present in the
concentrated amounts, and whether this leads to periods during the day when treatment performance
goals may not be met. We do not believe that this should be a problem. In SECTIONS 3,4, and 7, it was
shown that removal efficiency apparently depends on the total amount of base metal in the inflow per day,
and not on the relative concentration.
ELEMENTAL ANALYSIS OF PLANTS
A program of vegetation sampling was conducted in 1987, 1988, and 1989. Among the
objectives of this sampling were:
o Monitoring of elemental uptake, to provide a baseline database for use in evaluating the health of
the vegetation in this and future treatment systems, and for predictions of longevity of the
vegetation in such systems,
o Monitoring of seasonal partitioning of various elements in the vegetation,
o Further understanding of the metal removal budget within the treatment system,
o Documentation of elemental concentrations of the vegetation growing in the system, because of
concern for possible entry of heavy metals into the food chain should large-scale systems
become operational.
Procedure
Destructive sampling of the three primary emergent plant species, Carex utriculata, Carex aquatilis
(both sedge species), and Typha iatifolia (cattail) was performed. Samples were washed with deionized
water, oven-dried, and sent to EPA laboratories in Oregon and Cincinnati for ICP analysis. For
comparison, in the 1989 growing season, samples were also collected at three 'control' sites, which are
mountain wetlands in Colorado not directly impacted by mine drainage. These sites were:
1) Big Meadow, in Rocky Mountain Paik; (C.utricuiata and C.aquatilis. only)
2) Deer Creek, near the Keystone resort (C.aquatilis only);
3) Shadow Mountain Lake, near the west side of Rocky Mountain Lake, which was one of the sites
from where the original plants for the Big Five system were obtained (all three species).
6-6
-------
Results
Carex aquatilis is the species that was collected at all the sites not impacted by acid mine drainage.
The abundance of Cu and Zn in C. aquatilis roots from the different locations is given in Figures 38 and 39
respectively. In these figures, DC stands for Deer Creek, BM stands for Big Meadow, and SM stands tor
Shadow Mountain Lake. For both elements, there is an increase of over a factor of ten in roots taken from
the Big Five Cells. One might argue that these metals may be strongly adsorbed to the surfaces of the
roots. Nevertheless, the high concentrations of these two metals in the water cause high abundances of
these metals in the roots.
With respect to how the plants have changed over time, Cell C has remained undisturbed since
construction in 1987. Plant samples were taken in October, 1987, when the vegetation was transplanted,
and again in October, 1988 and 1989. By October, the leaves of the plants had gone into dormancy.
Figures 40 and 41 show the abundance distributions of Cd and Pb in various plant parts from Cell C. Metal
concentrations in both the leaves and the roots for Typha, C. aquatilis, and C. utriculata are shown in these
figures, in general, the abundances of these two metals in all the plant parts have increased over the
course of two growing seasons. For the first two samplings, the concentrations were higher in roots than
leaves. However in 1989, the abundances in the two plant parts are about equal. This equality of
abundances is manifested in all three species, which suggests that high metal concentrations in the water
start to affect these wetland plants after two years. Also, shifting contaminants to senescing leaves at the
end of the growing season is a known strategy for plant survival in contaminated locations. In 1990, the
growth of plants in Cell C was just as vigorous as in previous years. Even though the abundances of
microcontaminants may be increasing in the plants, it has not as yet affected growth and production of
biomass. This is especially true of the Typha species.
Finally, root and leaf samples of Typha were taken in both July and August of 1988 to see if
abundances of metals increased during the growing season. In Figure 42, Fe and Mn abundances in
Typha leaves and roots are shown. Fe is known to precipitate at the root surfaces in Typha (60); while Mn
doesnt precipitate and is taken into the leaves. For both July and August, this situation is seen for Mn and
Fe. In all cases, abundances of the metals are higher in August than in July. However, the increases are
within or just beyond the bound of analytical uncertainty. Abundances of these microcontaminants show
small increases during the height of the growing season.
6-7
-------
100
E
Q.
Q.
HI
O
z
<
Q
CD
<
3
o
10
rrrrrmq
. \ \ \ \ \ \ \
////////
. V \ \ \ V N \
////////
t \ \ V \ S \ \
////////
. \ \ \ \ \ \ N
////////
>>>})> >\
X % \ \ X X X
///////
N \ \ V \ \ \
///////
X X X X X X X
///////
\ \ \ \ \ \ \
///////
I F f # F f I
\ V \ \ N N \ •
////////
X X X X X X X '
///////y
xxxx*\xxN
////////
\ \ \ \ \ v \ •
////////
\ \ \ \ \ \ N
/////
'///////
XXXXXXXX
'///////
\\\\\\\v
'///////
XXXXXXXX
'///////
r\ \ \ \ \ S \ N
•///////
XXXXXXXX
' / / y / / / /
\\\\\\\\
'///////
\\N\\\\V
'///////
SV\VS\\N
'//////
X X X X X X X '
///////
X X \ X \ N \ '
'///////
xxxxxxxx
///////
xxxxxxxx
////// /
xxxxxxxx
^ / / / / / ~ /
xxxxxxxx
'///////
xxxxxxxx
'////// v .
xxxxxxxx
•///////.
xxxxxxxx
'///////
xxxxxxxx
•///////,
xxxxxxxx
'///////,
xxxxxxxx
•///////
xxxxxxxx
'///////
xxxxxxxx
'///////
iXXXXXNXX
L / / / / / y /
xxxxxxxx
'///////
xxxxxxxx
///////,
xxxxxxxx
'///////
xxxxxxxx
'///////
xxxxxxxx
•///////
xxxxxxxx
• / / / / y / /
xxxxxxxx
'///////
xxxxxxxx
¦ / / / / y / /
xxxxxxxx
'///////
xxxxxxxx
•///// y /
xxxxxxxx
¦~/~~/~/
xxxxxxxx
^ /////// ,
xxxxxxxx
•///////
/////// ^
xxxxxxxx
///////«
xxxxxxxx
/~~///~<
xxxxxxxx
///////-
xxxxxxxx
/ / y / / / / -
xxxxxxxx
////~~~<
xxxxxxxx
/ / y / y y / -
xxxxxxxx
~
xxxxxxxx
///////•
xxxxxxxx
y y y y y y y .
xxxxxxxx
y y y y y y y ,
xxxxxxxx
/ / y / y y y ,
xxxxxxxx
y y y y y y y
xxxxxxxx
y y y y y y y
xxxxxxxx
y y y y y y y
xxxxxxxx
y y y y y y y
xxxxxxxx
y y y y y y y ,
xxxxxxxx
y y y y y y y ,
xxxxxxxx
y y / y y y y
xxxxxxxx
y y y y y y y
xxxxxxxx
y y y y y y y .
xxxxxxxx
y y y y y y y .
xxxxxxxx
y y y y y / y
xxxxxxxx
y y y y y y y
xxxxxxxx
/ y y y y y y .
xxxxxxxx
y y y y y y y
xxxxxxxx
y y y y y y y .
xxxxxxxx
xxxxxxxx
y y y y y y y (
xxxxxxxx
/~/////.
xxxxxxxx
y y y y y y y
xxxxxxxx
y y y y y y y
DC
BM
SM
C- 1 1ST C - 1 2ND
B- 1
LOCATION COMPARISON OF Cu UPTAKE
IN ROOTS
Figure 38.
Comparison of Cu in C. Aquatilis roots from wetland sites not impacted by mine drainage
and from Cels B and C. The locations are: Deer Creek (DC), Big Meadow (BM), Shadow
Mountain Lake (SM), the front of Cell B (B-1), and the front of Cell C (C-1).
-------
10
DC
BM
SM
C ¦ 1 1ST C - 1 2ND
B - 1
LOCATION COMPARISON OF Zn UPTAKE
IN ROOTS
Figure 39. Comparison ol Zn in C. Aquatilis roots from wetland sites not impacted by mine drainage
and from Cells B and C. The locations are: Deer Creek (DC), Big Meadow (BM), Shadow
Mountain Lake (SM), the front of Cell B (B-1), and the front of Cell C (C-1).
-------
E
Q.
Q.
LU
O
CO
<
T5
O
.1
T LEAVES T ROOTS CA LEAVES CA ROOTS CU LEAVESCU ROOTS
SPECIES & PLANT PARTS FROM CELL C
¦ Oct '87
B Oct '88
H Oct '89
Figure 40.
Uptake of Cd into roots and leaves of the plant species in Cell C from October 1987
through October 1989. The plants are: Typha (T). Carex aquatilis (CA). and Carex
utriculata (CU).
-------
T LEAVES T ROOTS CA LEAVES CA ROOTS CU LEAVES CU ROOTS
SPECIES & PLANT PARTS FROM CELL C
Oct '87
m Oct '88
H Oct '89
1000
Figure 41. Uptake of Pb into roots and leaves of the plant species in Cell C from October 1987
through October 1989. The plants are: Typha (T), Carex aquatilis (CA), and Carex
utriculata (CU).
-------
10000
Q.
Q.
LU
O
z
<
o
CD
<
o>
K>
1000
©8
d>
LJL 100
Figure 42. Uptake of Mn and Fe In Typha roots and leaves In July and August, 1988.
B-Ty-L B--Ty--R C--Ty--L C--Ty--R
(July) (August)
LOCATION & TYPHA PART
¦ Fe
¦ Mn
-------
OTHER CONSIDERATIONS
Changes in Vegetation
As one means of qualitatively assessing the health of the wetland plants, vegetation mapping was
done in 1988 and in 1989, to help chart the changes as the wetland matured. Figure 6 is the diagram of
the original distribution of plants in all three cells. The most obvious change with time was the spread of
the Typha plants. The cattails have been by far the most successful of the original transplanted species
(Carex aquatilis, Typha latlfolia, Carex utriculata, and Juncus arcticus), spreading from their original
transplanted locations to virtually all parts of the cells. In Cell C, numbers of cattail plants increased from
less than 500 in 1988 to more than 750 in 1989. A similar increase occurred in Cell B. There is no visual
evidence of dieback of the Typha.. It is estimated that the Typha plants account for perhaps 75-80% of
the total plant biomass.
Of the other species, the sedges (C. aquatilis and C. utriculata) show some sign of dieback in Cell
C. In 1989 some patches of the sedges were dead, and in 1990 other patches of the sedges were less
dense within Cell C compared to the previous year, although there are still healthy stands of sedges in
certain locations. The rush species, Juncus arcticus, increased in cover area in Cell C in 1989 compared
to 1988.
Physical Effects
Wetland vegetation has an effect upon the flow dynamics of the cells by altering the permeability
of the substrate, at least as deep as the roots penetrate. One effect might be the reduction of
compaction or sticking together of portions of the substrate, which could remove them from being
effective reaction sites in the treatment process. Some such zones were observed in the substrate of cell
B below the root zone, when it was removed in August 1989.
Another beneficial consideration of a plant cover might be the reduction of erosion by wind at
times when the water level drops below the level of the substrate.
Aesthetic Considerations
An important contribution of emergent vegetation in a constructed wetland system is its aesthetic
value. In a scaled-up treatment system , a vegetation-covered design would be more visually appealing
than a barren site. This could be especially desirable in the case of a treatment system for the Big Five and
Argo tunnels, as the system would likely be in a location of high visibility, e.g. close to the I-70 highway.
The presence of a plant community in at least part of the system would contribute to more favorable public
opinion of such a treatment option.
6-13
-------
Fttprts on Fauna
One question that should be considered is whether the uptake of metals by plants could affect
animals. There are two possible problems if animals forage on the plants:
o Ingestion could cause mortality;
o Ingestion by lower forms of animal life could be magnified up the food chain.
An analysis by John Emerick (personal communication) suggests that these problems will not happen.
The animals that forage in wetlands typically have a wide food range. Even if the metal abundances in the
plants were dangerous, these plants would be only a small part of the animals diet. Magnification or
disease would be circumvented by the diverse diet of the animals that would use the wetland.
CONCLUSIONS
Emergent vegetation can be an important component of a metals-removal system for treatment of
acid mine drainage, particularly with respect to addition of biomass for bacterial nutrition, oxygenation of
the substrate, and aesthetic considerations.
The result of the calculation of biomass contribution on an annual basis for Cell C was modest (-60
kg dry weight/year, for an area of 200 sq. ft.). For an unvegetated system, a simple procedure such as
adding hay to the treatment cells might provide an effective substitution. However, for a large scaled-up
design, such an addition could be a large and moderately expensive task; the natural addition of biomass
by the vegetation helps keep the concept of a low-cost, minimal-maintenance wetland.
Aesthetically, the appearance of a plant-covered wetland is more pleasing than one without
vegetation, and would be more desirable from a public relations aspect.
The zone of oxygenation from plant roots is only a fraction of the depth of the cells, probably less
than 25 cm below the surface. Oxygenation of the substrate by plants may not be wanted for all parts of
the system. For cells that are designed to be anaerobic for metals removal by sulfate reduction processes,
a plant cover might be counter-productive. For the specific circumstances at the Big Five site, it may be
that only the final cells would be designed to be aerobic, for final 'polishing' removal of iron and
manganese as hydroxides. (There are indications that plants may have a more important role in
constructed wetlands for treating waters that are not as acidic, e.g. in eastern U.S. coal country, where
aerobic conditions are more important for metal-removing goals.) Still, even if the vegetation was used for
the final 'polishing' cells in future, larger designs, these cells might comprise half of the surface area of the
design.
After three years of operation, indications are that the vegetatie \ can withstand the conditions of
this system; and that it can have a role in the treatment process. Through time, the vegetation is likely to
become increasingly Typha-dominated.
6-14
-------
SECTION 7
AREA REQUIREMENTS AND LOADING FACTORS
DISCUSSION OF UNITS
This research is necessarily interdisciplinary. Unfortunately, scientists from different disciplines
use different units to describe the same properties. This typically causes confusion and in the case of
wetlands research the confusion is compounded. English units are mixed with metric units.
Environmental scientists designing wetlands use the reciprocal of the properties that sanitary engineers
use in their design equations (39, 42, 55). This section is not an attempt to settle on the proper units.
Technical people work with the units that are comfortable to them and only grudgingly change. What
follows is only an attempt to put the units for various parameters in one place for easy reference.
The important parameters are volume, concentration, flow, and a loading factor. Units and
conversion factors will be discussed for each parameter.
Volume
For chemical analysis purposes, the standard volume is a liter (L). However, for wastewater
treatment situations handling large volumes, cubic meters (m3) is preferred. For many mine drainages,
volume estimates in gallons (gal) seem to be the easiest and consequently this unit sneaks into the
literature (see Figures 9 and 10).
1 L - 0.001000 m3 . 0.2642 U. S. gallons
1 m3 . 1000. L - 264.2 U.S.gallons
1 U. S. gallon - 3.785 L - 0.00003785 m3
Concentration
For this property, the units are fairly standard. Concentration in milligram/liter (mg/L) is the
accepted unit. For large volumes gram/cubic meter (g/m3) is sometimes used. For water solutions, parts-
per-million (ppm) is sometimes substituted for mg/L.
1 mg/L - 1 g/m3 - 1 ppm aqueous solution
Flow
Here units become less standardized because different disciplines use units comfortable to them.
Liters per second (l/s) is the primary units, but gallons per minute (gal/min), cubic meters per second
(m3/s), cubic feet per second (ft3/sec), and millions of gallons per day (mgpd) are all used to some extent.
1 L/s - 0.001000 m3/s - 15.85 gal/min - 0.03531 ft3/s
1m3/s- 1000 L/s - 15850 gal/min - 35.31 ft^/s - 22.82 mgpd
1 mgpd - 0.04381 m3/s ¦ 1.547 ft3/s ¦ 43.81 L/s - 694.4 gal/min
7-1
-------
L oadino Factor
The loading factor or flux, describes the amount of water or contaminant that enters an area of the
wetland per unit time. Since it is a combination of all the above units, the confusion is compounded. In
addition, environmental scientists usually consider loading factors in area per flow; while sanitary
engineers consider the reciprocal, that is in flow per area. Since flow is volume per time, flow per area can
also be expressed length per time. Sometimes the amount of contaminant is used instead of flow. In this
case a mass loading factor would be in area times time over mass or the reciprocal would be mass over area
times time.
For the environmental scientists, typical hydraulic loading factors would have the units of square
meters X second/ liter (m^/L/s), square feet X minute/ gallon (ft2/gal/min), hectares X days over 1000 cubic
meters (ha/1000 m^/day), and acres X days over millions of gallons (acre/mgpd). For the sanitary
engineer, typical hydraulic loading factors would have the units of cm per day (cm/day),. Typical mass
loading factors have the units of square meters X minute/ milligram (m2/mg/min), grams X days over square
meters (g/m2/day) and kilograms X days over hectares (kg/ha/day).
1 gnVm2/day - 10 kg/ha/day - 1/1.440 x 106 m2/mg/min
1 m2/L/s - 0.001157 ha/1000 m3/day - 0.6788 ft2/gal/min -
¦ 0.01082 acre/mgpd - 1/8640 cm/day
A curious arithmetic quirk sets up the following conversion measure: 200 ft^/gal/min - 294 m2/|_/s,
which is the reciprocal of 29.4 cm/day. The initial loading rate for the Big Five constructed wetlands was
200 ft2/gal/min.
A way to consider the reciprocal loading factors is as follows: If one gallon per minute flowed into a
200 ft* container, the water in the container would rise 29.4 cm in one day. If the container were larger in
area, then the water would not rise as high in one day. Table 24 gives conversions for typical loading
factors discussed in SECTION 7.
Table 24. conversions of loading factors used in this section.
ENVIRONMENTAL SCIENTISTS (Area / Flow) | SANITARY ENGINEERS (Length per Time)
ft2/gal/min acres/mgpd m2/L/sec
cm/day
29.4
8.6
7.3
2.0
0.68
0.59
200
680
800
2920
8560
10,000
3.2
10.8
12.8
46.5
136
160
294
1,000
1178
4300
12,600
14,700
7-2
-------
, . »*'ryiytTt
.• % i i•'.jr-
•• •• • •• • m ^m • ¦
II .*• .*Xi * a
mm
POROUS .Jf
ROCK
DAM tiTf'
ANAEROBIC ZON!
Figure 43.
A diagram of
a
typical subsurface flow wetland.
-------
SURFACE FLOW SYSTEMS AND SUBSURFACE FLOW SYSTEMS
An important division has been established in wetlands used for municipal wastewater treatment
and this distinction will be important to the design of wetlands used for mine drainages (42, 52). This is
between surface flow systems and subsurface flow systems. The surface flow system is also called the
free water surface system. Since free water is shown on the surface, Figure 2 is a diagram of a surface flow
system. The depth of water is from 5 to 30 cm. Soil or another suitable medium is used to support the
vegetation that emerges through the water. To insure plug flow conditions, as discussed in SECTION 4,
the basins or channels are long and narrow. Typically the length to width ratio is above 10 to 1. Almost all
mine drainage systems are designed for surface flow. Cells A and C in the present Big Five wetland
(Figure 5 andl 1) are surface flow systems.
A subsurface flow system, also called a vegetated submerged bed, has had greater development
in Europe (39, 52). There they have been described as root-zone method, hydrobotanical systems, soil
filter trench, biological-macrophytic, and marsh beds by wastewater scientists. The substrate includes
rock or crushed stone, gravel, and different soils used alone or in combination. The water flows laterally
through the substrate. Figure 43 is a diagram of a subsurface flow system. In municipal systems, plants
may be desireable because oxygen is conveyed down the plant, into the roots and rhizomes, and out into
the soil. The subsurface is saturated and consequently anaerobic, but the oxygen supplied by the
plants creates aerobic microenvironments next to the roots and rhizomes (60). In a subsurface system,
plug flow is maintained as long as the water flows through the substrate and not across the surface, thus
the length to width is quite smaller, usually around 2 to 1. Cells B-Upflow, B-Downflow, and E in the
present Big Five wetland configuration (Figures 11 and 13) are subsurface flow systems.
REVIEW OF LOADING FACTORS FOR MUNICIPAL WETLANDS
References 42 and 52 give provide good reviews of constructed wetlands treating municipal
wastewaters. Table 25 is a summary of the current estimates of loading rates for treatment of municipal
wastewaters taken from the review by Watson and others (42). A number of considerations have gone
into producing the numbers. First, it appears that surface flow systems are less efficient than subsurface
•
systems. Consequently loading factors are higher for subsurface systems. Also treatment objectives
vary considerably for municipal systems. Table 25 is twice divided by treatment. A first division is by basic
(handling of raw waste with no settling or clarifying), secondary (handling the effluent from primary
operations), or polishing stages (handling the effluent after it has gone through primary and secondary
processes). A second division by treatment objectives determines whether the wetland was built to
handle one stage of the treatment process or multiple stages.
7-4
-------
Table 25. Hydraulic Loading Rates for the Preliminary Design of Constructed Wetlands
for Treating Municipal Wastewaters taken from Watson et al. (42).
Treatment Objectives
Secondary Treatment Advanced Treatment/Multiple Objectives
Surface Flow I Subsurface Flow Surface Flow I Subsurface Flow
Use cm/day |acres/mgd| cm/day |acres/mgd cm/day |acres/mgd| cm/day | acres/mgd
Basic
treatment a | a | 2.3-6.2 | 40-15 | a | a | 2.3.1 | *30
Secondary
treatment 1.2-4.7 | 75-20 |4.7-18.7| 20-5 | *1.9 | *50 | *4.7 | *20
Polishing
treatment 1.9-9.41 50-10 |4.7-18.7| 20-5 | *3.1 | *30 | *4.7 | *20
a This use has not yet been demonstrated. Surface flow systems constructed to date are preceded
by at least primary treatment units (septic tanks, darifiers, lagoons, etc.).
Designers of municipal systems (42) stress that "current information is adequate to design
systems that substantially reduce targeted contaminants but inadequate to optimize the design and
operation 1or consistent compliance." Two suggestions are made to cope with the uncertainty. The first is
to use multiple cells in series, parallel, or combination to optimize treatment processes. The first
suggestion was also made for mine drainage situations by Howard and others (6). This parallel and series
design configuration can be called a modular design. Figure 44 is a diagram of the modular concept. An
important aspect of this type of design is that the route from a pilot system to complete treatment can be
made in stages. A complete, monolithic wetland never has to be constructed. Also, maintenance on the
modules would be much easier than on one large system. The second suggestion on municipal wetlands
is to take two approaches to the loading factor question. Design very conservatively with low loading rates
to avoid future problems. The alternative approach is to use higher loading rates and donl expect great
performance. In the latter case, plan for expansion so that upgrades can be made based on experience
and actual performance data.
In summary, study of municipal loading factors indicates that the initial Big Five Wetland loading
rate of 200 ft2/gal/min (29.4 cm/day) is highly optimistic for a wetland treating mine drainage comparable in
chemistry to the waters in Table 1. Even a subsurface system requires lower loading factors than 29
cm/day. Results from municipal systems also suggest that for surface flow systems receiving mine
drainage, a loading factor of 1000 m2/L/s (8.6 cm/day) may be a bit optimistic. Although treatment
objectives are quite different, applying municipal design guidelines on a mine drainage treatment system
7-5
-------
Inflow
A. Parallel Modules
I n f I o w—
B. Series Cells
Inflow
C. Combination Arrangement
Figure 44. Different possibilities for modular wetland configurations.
7-6
-------
would suggest a loading factor in the neighborhood of 4300 m2/L/s for a surface system. Experience
from municipal systems provides some good ideas on how to cope with the uncertainty in determining
wetland size.
REVIEW OF LOADING FACTORS FOR MINE DRAINAGE WETLANDS
Eartv Concepts on Loading Factors
There have been a number of estimates of the loading factor for a wetland receiving acid mine
drainage. From the first suggestions made in the early 1980's, until now, the area requirement per unit of
flow has increased significantly. The first rule-of-thumb, proposed for wetlands receiving coal mine
drainage in Appalachia, was 200 ft2/gal/min (33). In the design of the Big Five pilot wetland, the standard
cell size of 18.6 m2 (200 ft2) was chosen using this rule of thumb.
Girts and Kleinmann (39) in their review of constructed wetlands for treatment of acid mine
drainage found the following size and flow ranges: Sizes ranged from 93 to 6070 m2 with a mean of 1550
and a median of 929 m2. Flows ranged from 0.06 to 13 Us with a mean of 1.3 and a median of 0.5 L/s.
Loading factors ranged from 61 to 10700 m2/Us with a mean of 2390 and a median of 928 m2/L/s.
Watson et al. (42) noted that most wetland systems rely on surface removal and have minor interaction of
the drainage with the substrate.
Recently, the wetland research group at the Tennessee Valley Authority made estimates of
loading factors based upon the systems they have installed (55). The estimates are mass loading factors
and take into account the pH of the drainage, the amount of Fe, and the amount of Mn. The desired
discharge requirements for the constructed wetland were selected as 3 mg/L or less for iron and 2 mg/L or
less for manganese. The loading factor for Fe is 2 nr^/mg/min if the pH of the drainage is less than 5.5 and
0.75 m2/mg/min if the pH is greater than 5.5. The loading factor for Mn is 7 m2/mg/min if the pH of the
drainage is less than 5.5 and 2 m2/mg/min if the pH is greater than 5.5. Note that the loading factors take
into account the greater difficulty of Mn removal compared to Fe. The area calculation is made separately
for Fe and Mn and the largest area is chosen. Gross estimates on the average mine drainage can be made
from Table 1 of about 100 mg/L of Fe, 30 mg/L of Mn, and a pH of approximately 3. Using these, loading
factors are 100 x 2 x 60 - 12000 m^Usec for Fe and 30 x 7 x 60 - 12600 m2/L/sec for Mn. The selected
size would be 12600 m2/Usec (8560 ft2/gal/min) of drainage. Again, these estimates relate to surface
flow systems and not subsurface flow systems.
In summary, it appears that the early rule-of-thumb of 294 m2/L/s (200 ft2/gal/min) was highly
optimistic. If surface flow systems are considered, a better estimate of the loading factor would be some
where around 1000 m2/L/s. No estimates have been made on subsurface flow systems. Cell A in the Big
Five wetland had significant flow through the substrate and the results shown in SECTION 4 suggest that
a loading factor of about 800 rtf/L/s would be sufficient for removing heavy metals and raising the pH to
7-7
-------
between 5 and 6. As developed in SECTION 12, this value depends on the thickness of the anaerobic
zone.
Area Adjusted Loadings and Removals
In 1990, Hedin suggested a new method for sizing and performance of constructed wetlands
(96); and a number of research groups used his suggestion to evaluate their wetland operations (11,97,
98). Hedin's suggestion was based on the following analysis:
1. If one looks at wetland efficiency based on percent removal or on outflow/inflow (Figures 14,15,
and 16), this does riot take into account the flow of water going into a wetland. Presumably, a
wetland with a bad removal efficiency but receiving a high flow could be removing a large absolute
amount of contaminants.
2. If one looks at wetland loading only from the viewpoint of what is entering and the areal size, then
the loading factor doesnl contain a good measure of what was removed.
i
3. The best situation is to combine both measures into what would be called a wetland area-adjusted
loading and removal factor. The units for this factor are grams of contaminant removed / day /
square meter (abbreviated as gdm).
4. The calculation of the area-adjusted loading and removal factor is by the following formulas in
which Fe is the contaminant removed:
Fe in (g/day) -1.44 X inflow (L/min) X Inflow Fe Cone (mg/L)
Fe out (g/day) - 1.44 X outflow (L/min) X Outflow Fe Cone (mg/L)
Fe rem (g/day/m2)« Fe gdm - [Fein - Fe out]/area (rr*)
In the case where outflow - inflow, such as the Big Five Cells, then:
Fe odm »1.44 X Flow ( L/min \ f Inflow Fe cone - Outflow Fe Cone 1
area(m2)
Looking at the formulas it can be seen that an area-adjusted loading and removal factor (gdm)
does combine the loading factor with the removal efficiency. Design calculations using this method are
developed in SECTION 12 in the subsection Area/Flux Method.
Using gdm's as an analytical tool, Hedin suggested that what would happen in a constructed
wetland is at low gdm's Fe-rem ought to increase as Fe-in is increased (95). Then, when the removal
capacity of the wetland is met, the Fe-rem gdm would reach a plateau and not increase as the Fe-in gdm is
increased. The plateau value would be the maximum Fe gdm for that wetland. This analysis worked for
the Somerset Wetland, but produced mixed results for the Latrobe and Friendship Wetland. None of the
three wetlands achieved complete removal of iron.
7-8
-------
£
o
_l
Li.
Cu DEC. ( mg / L)
Figure 45. Decrease in copper concentration in Cell A versus Flow for 1989.
7-9
-------
2.5
^ 2.0-
5
O
mJ
Li.
1.5-
1.0-
0.5-
0.0
—, ! 1—
12 18 24
30
Fe DEC. (MG/L)
Figure 46. Decrease in iron concentration in Cell A versus Flow for 1989.
7-10
-------
2.5
2.0
Z
i 1.5
_J
g 1.0
o
-J
LL
0.5
0.0
0 2500 5000 7500 10000
S04 RMVL. (mg I DAY / SQ M )
Figure 47. Area adjusted removal factor ( gdm ) for sulfate versus flow in Cell A for 1989.
-------
Based on his analysis, Hedin made suggestions on iron loading factors for constructed wetlands.
He estimated that if a wetland is receiving mine drainage whose pH is 3 or less, then a wetland can remove
4 gdmof iron. If the pH of the mine drainage is 4 or more, the wetland can remove 10 gdm of iron. The
average mine drainage in Table 1 has a pH of 3 and a concentration of Fe of 100 mg/L. Assuming outflow
equals inflow and using a removal factor of 4 gdm, the loading factor calculates to be 2160 m2 / L / sec
(1466 «2 / gal / miri). Over 1989, for Big Five Cell A when it was operating in an assumed plug flow mode,
area-adjusted loading and removal factors lor iron ranged from -1.5 to 4.4 gdm; the average was 1.8 gdm.
For Big Five Cell E a subsurface system, the flow out averages about 0.4 L/min, Fe concentration
averages 40 mg/ L and is completely removed, and the wetland size is about 10 m2. The Fe gdm for Cell E
calculates to be 2.3 gdm.
Brodie (69) has recently analyzed the area-adjusted loading and removal rates for TVA wetlands.
He found the range for Fe gdm to be from 0.5 to 10; the average was 1.25 gdm. He suggested 10 gdm to
be the practical limit for Fe removal by wetlands. For Mn, the area adjusted loading and removal factors
were quite a bit lower, and he suggested 2 gdm to be the practical upper limit.
In the case of the Big Five Wetland, it was difficult to analyze the performance of the cells based
on area-adjusted loading and removal factors (11). As shown in Figures 35 and 36, sulfate and Eh
decreases are directly related to the flow. However, for the heavy metals this correlation doesnl always
work. Figures 45 and 46 show how decreases in the effluent concentration of copper and iron change
with flow in Cell A. For copper removal is complete at low flows and sporadic at high flows. For iron, there
is no obvious correlation. Since sulfate showed such a good removal trend with flow, a sulfate gdm was
calculated for Cell A over 1989. The result is show in Figure 47. Sulfate removal in gdm has no correlation
with flow, even though decrease in sulfate concentration did correlate. For Big Five Cell A, lack of
correlation of removal with flow was disturbing. Failure of the cell to provide consistent removal was even
more disturbing. Yet, excellent removal results were being provided by Cell E (Figure 16). Also when the
flows on Cell B-Upflow and B-Downflow were cut so the loading remained constant at around 800 ft2 / gal /
min, removal of Fe, Cu. and Zn was nearly 100 % (Figures 14 and 15). This led us to determine loading
factors for wetlands emphasizing sulfate reduction by a completely different method described below, that
considers reaction rates and the volume rather than the area of the wetland.
LOADING FACTORS FOR SULFATE REDUCING WETLAND CELL
The Limiting Reanant Concept
In our experiences,at the Big Five site, typical measures of loading factor do not seem to explain
the removal of metals even though heavy metals such as Cu and Zn are reduced by greater than 99 %
(12). We have discovered that a key factor in sulfate reduction is to insure that the optimum
microenvironment for suHate-reducers is maintained. The most important environmental factors are
7-12
-------
reducing conditions and a pH of around 7. Since the wetland cell is receiving mine drainage with pH below
3 and Eh of above 700 mV, the water can easily overwhelm the microenvironment established by the
anaerobic bacteria. This leads to the limiting reagent concept for determining how much water can be
treated, as an alternative to the use of typical loading factors.
Consider the foltowing precipitation reaction: .
Fe2+ + S" —> FeS
At high flows of mine drainage through the substrate, sulfide will be the limiting reagent, the microbial
environment will be under stress to produce more sulfide, the pH of the microenvironment will drop, and
removal will be inconsistent. At low flows of mine drainage through the substrate, iron will be the limiting
reagent, the excess sulfide will insure a reducing environment and a pH near 7, the microbial population
will remain healthy, and removal of the metal contaminants will be consistent and complete. Using this
concept, loading factors should be set to insure that the heavy metal contaminants are always the limiting
reagents. The question then is how much sulfide can a colony of sulfate-reducing bacteria produce per
cubic cm of substrate per day?
Studies by the U. S. Bureau of Mines wetlands group suggest that a reasonable figure for sulfide
generation is 300 nanomole sulfide / cubic cm / day (0.3 mole sulfide/m3/day) (65, 67). This number, the
volume of the wetland cell, and the metals concentrations in the mine drainage are used to set the flow of
mine drainage through the wetland cell. Using this concept in a subsurface wetland cell to determine the
loading factor has resulted in year round complete removal of Cu and Zn, a nearly complete removal of Fe,
and a rise in pH from 3 to 6 that is seen in Cells B-Upflow, B-Downflow, and E. Design of wetlands using
this method is discussed in SECTION 12 in the subsection Sulfate-Reducing Stoichiometrv Method.
Volume Loading Factors
This volume loading factor will be used extensively in SECTION 12 in the subsection on
Volumetric Loading Method. For now, consider how it was used to set the desired flow into the -
redesigned B Cells. The depth of the B Cells is about 1 meter, this makes the volume of substrate to be
about 8 m3. Using the volume loading factor, 2.4 moles of sulfide will be produced in the cell per day.
Using the limiting reagent concept, heavy metals flowing into the cell should not exceed 2.4 moles per
day. Big Five mine drainage has 40 mg/L Fe, 30 mg/L Mn, 10 mg/L Zn, and 1 mg/L of Cu for a total 81
mg/L of heavy metals. Using the atomic weight of manganese (55 g/mole) as the average atomic weight of
the metals, the total concentration of heavy metals in the drainage is about 1.5 millimoles/L.
Consequently, flow into the cell should be limited to 1600 L/day or about 1.1 L/min. This works out to a
traditional loading factor of about 430 m2/L/s or 290 ft2/gal/min. As a safety factor, over 1990, the flows at
the Big Five Ceils have been set so the loading factor is 800 ft2/gal/min. Note that with an area of Cell B of
9.3 m2, if all the heavy metals were removed and the flow rate was 1.1 L/min, the gdm of heavy metals
would be 14, since half the heavy metal concentration is iron, the Fe gdm would be 7.
7-13
-------
One important feature of this volume loading factor is that a poorly acting cell will recover if the
volume loading factor is cut back to below the value of 300 nanomole/cubic cm/day. Over the course of
the last year both the B Cells developed problems. Correcting the problem and adjusting the flow to
within the proper range allowed the cells to recover. By the end of the testing period, both cells were
removing heavy metals quite well.
Recent Example of the Use of the Volume Loading Factor
In a bench scale study just recently completed, garbage cans filled with substrate to a depth of
about 60 cm were used to determine whether using the sulfide generation figure of 300 nanomole sulfide
/ cubic cm of substrate / day could be used to set the conditions for treating severely contaminated
effluent that flows from the Quartz Hill Tunnel in Central City, CO. Contaminant concentrations for this
drainage are shown in Table 26. Using the limiting reagent concept described above and the amount of
substrate contained in the garbage can, flow could not exceed one milliliter / minute to insure that sulfide
would always be in excess. Contaminant concentrations from the outputs of three different bench scale
cells are shown in Table 26. For cell A the mine drainage was passed through the cell with no delay. For
cell B the substrate was soaked with city water for one week before mine drainage started passing through
the cell. For cell C, the substrate was inoculated with an active culture of sulfate-reducing bacteria and
soaked with city water for one week before mine drainage started passing through the cell. Preparations
on cells B and C were done to insure that there would be a healthy population of sulfate-reducing bacteria
before mine drainage flowed through the substrate. All cells were run in a downflow mode of the mine
Table 26. Constituent concentrations in mg/L in the Quartz Hill Tunnel mine drainage and
in effluents from the bench scale tests.
Days
Mn
Fe
Cu
Zn
SO4
. PH
Sample
Operated
<
Concentration in mgfl. —
>
Mine Drainage
24
80.
630
48
133
4240
2.4
Cell A
24
0.94
1.6
0.06
0.27
450
7.4
Cell B
24
0.91
1.9
<0.05
0.17
770
7.5
CellC
24
0.99
1.0
<0.05
0.16
412
7.4
Mine Drainage
43
80.
640
50
135
4300
2.5
Cell A
43
0.97
0.87
<0.05
0.18
1080
7.2
Cell B
43
0.64
0.96
<0.05
0.24
660
7.4
CellC
43
1.6
0.46
<0.05
0.14
1180
7.2
Mine Drainage
71
70.
820
70.
101
NA
2.6
Cell B
71
0.48
0.40
<0.05
O.'l
NA
8.0
CellC
71
1.6
0.40
<0.05
0.25
NA
7.9
drainage through the substrate. In all three cells removal of Cu, Zn, Fe, as well as Mn is greater than 99 %.
7-14
-------
The increase in pH is from about 2.5 to above 7. These results were consistently maintained for over ten
weeks of operation.
The substrate used was a mix of 3 / 4 cow manure and 1 / 4 planting soil. The results from cells B
and C show that the cow manure has an indigenous population of sulfate-reducing bacteria that are quite
active. Inoculation with an active culture of bacteria is not necessary in this case. Also, since the results
from cell A are comparable to those of cells B and C, the population of sulfate reducers can withstand
immediate exposure to severe mine drainage and still produce sufficient quantities of sulfide. The key to
good initial activity is to insure that the flow of mine drainage is low enough that its low pH does not disturb
the micro-environment established by the bacteria.
Another feature of the results shown in Table 26 is that Mn is removed in all three cells. Typically,
Mn is the most difficult contaminant in mine drainage to remove (3,4,5,6,7,8). It is usually presumed that
removal of Mn has to be achieved by raising the pH to above 7, and then introducing the effluent into an
aerobic wetland cell so that Mn will be oxidized to Mn02- Removal in an anaerobic cell must be as Mn(ll).
Analysis of possible species at a pH above 7, suggest that removal could be as MnS or MnCOs. In this
case, it is hypothesized that MnCOs is the precipitate because it is more insoluble than the sulfide. In
either case, a key to Mn removal in an anaerobic cell appears to be the ability to raise the pH of the effluent
above 7. If raising the pH to above 7 can be consistently achieved, then all the contaminants in mine
drainage can be removed in one anaerobic cell.
For the garbage cans, the volume is 0.114 m3, and the surface area is 0.204 m2 (2.2 ft2 ). For the
Quartz Hill drainage, the sum of heavy metal concentration is about 1060 mg/L or 19.3 millimole/L. With a
flow of 1.0 mL/min, the area adjusted removal rate is 7.6 gdm. For Fe, it is 5.9; for Mn, it is 0.50 gdm. In
one day, the generation of sulfide would be 0.034 moles, and the loading of heavy metals would be 0.027
moles. The areal loading factor is 14,700 m2/L/sec (10,000 ft2/gal/min) for a wetland thickness of about
60 cm. Comparing the figures, it appears that an area-adjusted removal factor of between 5 and 10 gdm is
the maximum and for mine drainages with a pH below 3, this is may be based on the amount of sulfide
generated in the wetland substrate.
An interesting hypothesis derived from this work is that downflow and upflow cells combined with
anaerobic processes may allow wetlands to be built with greater effective thicknesses. Experiments are
currently being planned (filipek, pers. communication, 1991) to both increase the thickness and
permeability and decrease the concentration of organic matter in the substrate. It is hoped that this
combination will allow increased formation of pyrite relative to acid-volatile sulfides throughout virtually the
entire substrate column.
7-15
-------
SUMMARY
In the past few years, suggested loading factors for constructed wetlands treating acid mine
drainage have become much more conservative than the 200 square feet/gallon/minute used in the early
1980's. For a highly effective wetland Hedin's area-adjusted removal rate estimates of 4 gdm for mine
drainages of pH less than 3 and 10 gdm for mine drainages greater than 4 appear to be appropriate.
However prudence would suggest building a safety factor of 2 into the design.
For downflow and upflow wetland cells using microbial sulfate reduction as the primary removal
process, a volume loading factor appears to work well. Using the value of 300 nanomole/ cubic cm / day as
the amount of sulfide generated and the concentration of heavy metals in the drainage, the flow should
be adjusted so that sulfide is always in excess. This volume loading factor has worked well on bench scale
and pilot scale tests.
7-16
-------
SECTION B
DESIGN CONSIDERATIONS
-------
SECTION 8
REGULATORY ISSUES
The regulatory issues associated with the construction of passive treatment systems for
acid/metal drainages can be categorized into 1) pre-construction, and 2) operation and decommissioning
components.
PRE-CONSTRUCTION ISSUES
Pre-construction issues involve those regulations that are not exclusive to constructed wetlands
and typically apply to many types of construction. Pre-construction regulatory issues for passive treatment
systems may include:
o Environmental Review (National Environmental Policy Act, 1969)
o NPDES Permit (Clean Water Act, 1972)
o Mining Reclamation Permit (Surface Mining Control and Reclamation Act, 1977)
o Air Quality/Construction Permit (Clean Air Act, 1967)
o Archeological Survey (National Historic Preservation Act, 1966)
o Protected Species or Habitat (Endangered Species Act of 1973)
o Floodplain and Wetland Considerations (Executive Orders 11988 & 11990, respectively)
o Water Rights (State Water Laws)
These regulatory issues and similar local laws that may vary from state to state should be
considered in siting any passive treatment system. As these and similar issues are not unique to the
permitting of a passive treatment facility, they will not be discussed further.
OPERATION AND DECOMMISSIONING ISSUES
After a passive treatment facility is permitted and constructed, the regulatory issues appear to be
more complex. Operation and decommissioning regulatory issues encompassing constructed wetlands
may include:
o RCRA - Resource Conservation and Recovery Act
Hazardous Waste Characteristics of Substrate
Bevill Amendment (Mining Exclusions)
o NPDES • Discharges are excluded from RCRA but not while being treated, stored, etc. Water
treatment sludges are not excluded from RCRA.
o Floodplains and Wetlands Considerations
o State Water Rights
o Endangered Species Act
o Reclamation Bond Release (post-mining land use)
Each of these issues is considered in the following narrative.
8-1
-------
RCRA WASTE
A major issue determining the design of the passive systems and the disposal of the saturated
wetlands substrate is whether the substrate would be classified as a RCRA hazardous waste. Two of the
RCRA classification criteria are expected to apply to this material: TCLP toxicity and reactivity. The loading
rate and metal suite at a particular wetland site will determine when (or if) the substrate would be
considered TCLP (toxicity characteristic leaching procedures) toxic according to the RCRA definition.
Under RCRA, a waste can be defined as reactive if it meets any of the following eight criteria (CFR
261.23):
(1) It is normally unstable and readily undergoes violent change without detonating.
(2) It reacts violently with water.
(3) It forms potentially explosive mixtures with water.
(4) When mixed with water, it generates toxic gases, vapors, or fumes in a quantity sufficient
to present a danger to human health or the environment.
(5) It is a cyanide or sulfide bearing waste that, when exposed to pH conditions between 2
and 12.5, can generate toxic gases, vapors, or fumes in a quantity sufficient to present a
danger to human health or the environment.
(6) It is capable of detonation or explosive reaction if it is subjected to a storing initiating
source or if heated under confinement.
(7) It is readily capable of detonation or explosive decomposition or reaction at standard
temperature and pressure.
(8) It is a forbidden explosive as defined in 49 CFR 173.51, or a Class A explosive as defined
in 49 CFR 173.53, or a Class B explosive as defined in 49 CFR 173.88.
Of these criteria, only number five potentially applies to the passive substrate material since
hydrogen sulfide gas would be generated at a pH of 2 (EPA SW846-Sect 7.3.4.1). However, at higher
pHs the material may be expected to be stable and its disposal may pose little, if any, threat to human
health or the environment.
The "Bevill Amendment" exclusion [40 CFR 261.4 (b)(7)] of "mining wastes" from RCRA
hazardous waste classification may or may not apply to the substrate within passive treatment systems.
The Bevill Amendment excludes the following materials from RCRA Subtitle C regulation, which pertains
to hazardous wastes:
Solid waste from the extraction, beneficiation, and processing of ores and minerals (including
coal, phosphate rock and overburden from the mining of uranium ore)...
For the purposes of 261.4 (b)(7), beneficiation of ores and minerals is restricted to the following
activities: Crushing; grinding; washing; dissolution;...ion exchange;... precipitation;...
8-2
-------
Waste from the treatment of an excluded waste may also be excluded from hazardous waste
regulations.
The exclusion provided by the Bevill Amendment is currently in effect. However, regulations on
mining wastes are being considered by the EPA that may eliminate the Bevill exclusion. Mining wastes will
then be regulated as a special category of solid waste.
The disposition of used passive treatment system substrate materials from a mining operation will
be a function of their chemical and physical characteristics. However, it appears that used substrate is not
really a "mining waste"; it is more likely to be classified as a "waste water treatment sludge" which is
regulated by RCRA Subtitle C.
When coal mining regulations are considered, used substrate may be considered an unsuitable
material requiring burial on-site. For example, Colorado coal mining regulations require the covering of
coal and "acid and toxic-forming materials" in accordance with the following regulation:
4.14.3 (1) COVER.
(a) A person who conducts surface coal mining operations shall insure that all debris, acid-forming
materials constituting a fire hazard are treated or buried and compacted or otherwise disposed of
in a manner approved by the Division an. are designed to prevent contamination of ground or
surface waters...
(b) Where necessary to protect against... formation of acid or toxic seeps, to provide an adequate
depth for plant growth, or otherwise meet local conditions, the Division shall specify an
appropriate amount of cover using non-toxic material or special compaction and isolation from
ground water contact.
From a geochemical standpoint, an argument can be made that the precipitated metals in a
constructed wetland could be viewed as a mineral resource and theoretically, metals could be recovered
from the used substrate using conventional metallurgical techniques. The residue after this "processing"
could be currently viewed as a "Bevill Waste", or a waste that would be regulated under Subtitle D and
assumed to be suitable for landfill disposal. The term "mineral resource" introduced above should not be
interpreted to mean "ore". The accepted definition of "ore" recognized by the mining industry is a mineral
resource that can be recovered, processed, and sold at a profit. It is not certain that the recovery of metals
from wetland substrate can be accomplished at a profit. However, the material handling and processing
cost associated with metals recovery may be less than the cost of disposal of the material at a hazardous
waste site.
Thus, the long term operational policy of a passive treatment system will influence the disposal
options for used substrate from the facility and the design of the facility itself. From the first operation of
the facility, the substrate is by definition a Subtitle C waste, which is unlikely to be hazardous at first. If the
substrate is allowed to become a Subtitle C hazardous material through metals accumulation, the passive
treatment facility would have to meet RCRA design criteria as discussed below. Thus, as the substrate
8-3
-------
START
FLOW TO
WETLAND
HAZ
WASTE
CRITERIA
REACHED
HAZARDOUS WASTE
CHARACTERISTICS ARE
EXCEEDED —.
BEVQl
MINING
WASTE
METALS
LANDFILL
DISPOSAL
RECOVERY
tiAZ
WASTE
DISPOSAL
SUBSTRATE
"WORN OUT
RECYCLE
CLEANED
SUBSTRATE
TIME
Figure 48 Changes in used substrate disposal alternatives with time and concentrations of metals.
-------
ages and is loaded with metals, metals recovery may be a logical option to avoid the high costs of disposal
of used substrate as a hazardous waste.
This regulatory situation results in several alternatives for the disposal or regeneration of used
substrate material as illustrated on Figure 48. The most likely operational/ disposal scenario is the
exhumation or in situ processing of the substrate prior to its becoming classified as hazardous and the
processing of the material to yield a "mining waste" residue and possibly a saleable by-product (metals).
This is supported by the discussion of the complicated and expensive design criteria for a RCRA Subtitle
C facility in subsequent paragraphs. In any event, if landfill disposal of residue or Bevill Waste is
performed, steps should be taken to isolate these materials from other wastes in case regulations
governing their disposal are changed at some future date.
If the substrate will be allowed to become a RCRA hazardous waste, the passive treatment system
would be designed to comply with the RCRA criteria for surface impoundments. The passive treatment
systems would be designed to prevent the migration of "leachate" within the wetland to the surrounding
soils. The containment system might consist of a 40 mil high density polyethylene (HDPE) liner covered
by 6 inches of coarse sand in turn covered by a second HDPE liner and 6 inches of coarse sand. The
wetland substrate would be placed on top of a geotextile overlaying these four layers. To collect any water
which escapes the first HOPE layer, the system would be sloped to the effluent end of the cells to convey
leachate in the first sand layer for collection in a nearly horizontal, perforated leachate collection pipe.
In accordance with RCRA surface impoundment requirements, the system designs would
incorporate berms adequate to protect the systems from the 25-year, 24-hour storm event. In addition,
the integrity of the passive systems would be inspected weekly in accordance with RCRA.
The proposed passive treatment systems could result in two types of material that would contain
high concentrations of heavy metals and would have to be managed accordingly. These materials would
consist of accumulated metal precipitates (primarily iron hydroxides) in the settling basins used as an
aerobic pretreatment step and metal laden substrate within the anaerobic wetlands.
The metal precipitates from the settling basins may be classified as a RCRA hazardous waste, and
as such, the operational cost of the facility should included disposal of these materials at a RCRA landfill.
Also, the anaerobic substrate from the wetlands would contain metals that were removed from the mine
discharges. The disposal option for this alternative assumes the substrate is classified as hazardous. This
option would entail placing the material in appropriate containers and transporting it to a RCRA hazardous
waste landfill.
WATER QUALITY DISCHARGE STANDARDS AND THE CLEAN WATER ACT
Wetlands constructed to improve the quality of the discharge function as waste water treatment
facilities. Their discharge which generates a point source load may have to be regulated under Section
8-5
-------
402 of the Clean Water Act National Pollution Discharge Elimination System Permit (NPDES).
Technology Based Limitations
The 1972 amendments of the Clean Water Act established a two-step program for the reduction
of the discharge of pollutants into the nation's waters. First, categories of industrial dischargers were
required to meet a level of pollutant control based on the across-the-board application of "best practicable
control technology currently available" (BPT) by July 1,1977. The second level of effluent limitations, to
be achieved by July 1,1983, was to be based on the "best available technology economically achievable"
(BAT).
Technology-based standards are now the foundation of the industrial effluent limitations program,
although recently both Congress and EPA again have placed renewed emphasis on water quality.
The BAT standards do not apply to conventional pollutants (BOD, TSS, fecal coliform, pH, and oil
and grease}. In 1977, Congress decided that full application of the BPT standards provided adequate
protection from conventional pollutants and that more stringent control of these pollutants would, in many
cases, yield only marginal benefits. Accordingly, in its 1977 "mid-course corrections" to the Act, Congress
enacted Section 301(b)(2)(E), which required the application of a more lenient "best conventional
pollutant control technology" (BCT), rather than BAT, for conventional pollutants by July 1,1984.
The Water Quality Act of 1987 extended the compliance deadlines for most technology-based
requirements to "as expeditiously as practicable" but not later than three years after the requirement is
established, and in no case later than March 31,1989.
BPT deals primarily with traditional pollutants of concern - BOD, oil and grease, pH, TSS, some
metals, etc. BAT, by contrast, deals primarily with toxics, (e.g., organics and heavy metals). In determining
what level of treatment constitutes BAT, EPA has more latitude to depart from the usual technologies
employed by the industry than when setting BPT standards. EPA may consider process controls, as well
as end-of-pipe treatment, and it may base its standards on transfer technology or pilot plant data, although
it must meet the "economically achievable* test in the statute.
The BAT determination does not involve even a limited cost-benefit analysis, although cost is one
of the statutory factors EPA is directed to consider. In essence, BAT represents the maximum feasible
pollution reduction. BAT treatment requirements are considered "economically achievable" so long as
their imposition would not force the closure of a large portion of the plants in a category or subcategory of
an industry. Cost is thus relevant, but there is no explicit weighing of the benefits against the costs. As a
practical matter, however, EPA witl be influenced by a showing that substantial additional costs will
produce only minor incremental pollution reductions.
From another viewpoint, it appears that the application of BPT/BAT limitations are not applicable to
a newly constructed wetland. Typically, BPT/BAT standards apply to existing water sources. A new facility
would have to meet new source performance standards or water quality based stream standard limitations.
8-6
-------
Consequently, it may best to concentrate on new source standards and not worry about BPT/BAT
limitations.
Water Quality-Based Limitations
Under the Clean Water Act, where technology-based limitations are insufficient to ensure that
water quality standards for the receiving streajn will be met, water quality-based limitations are incorporated
into a discharger's NPDES permit. Water quality standards consist of two elements: (1) use classifications,
and (2) water quality criteria.
The Clean Water Act requires each state to classify all of the waters within Its boundaries according
to intended use. In establishing the classifications, states are to consider the value of waters for public
drinking supplies, propagation of fish and wildlife, recreational purposes, and industrial, agricultural and
other purposes. EPA's regulations require that all classifications that do not provide for protection and
propagation of fish and wildlife and water recreation must be reexamined to determine whether new
developments warrant an upgrading to attain such protection.
Where a state has identified water quality-limited segments, it must adopt permit limitations that will
ensure the standards for the water quality of each segment are met. For heavy metals and other pollutants
whose effect on water quality is not complicated by biodegradation or other reactions over time, these
limitations are usually set in a straightforward manner calculated to ensure that the concentrations in
excess of those allowed by the standard are not exceeded at the point of discharge. Heavy metals
precipitated in passive treatment facilities will react over time only if the ambient environment changes,
thus a stable wetlands operational environment (constant substrate saturation) helps to assure discharge
compliance.
For pollutants such as BOD or ammonia, whose effect on water quality varies in a complex manner
over time, the setting of water quality-based limitations is much more complicated, requiring the use of
models, or alternatively, reliance on conservative assumptions which may restrict discharges much more
than is necessary to comply with the standards.
Setting water quality-based limitations is further complicated where more than one facility
discharges into the segment and the burden of effluent reduction must be allocated among dischargers.
States may grant variances from compliance with water quality standards on a case-by-case basis.
Such variances are considered by EPA to involve amendments to state water quality standards and,
therefore, must follow the same procedures and meet the same basic requirements, including approval by
EPA. Even in states where EPA administers NPDES programs, variances are not available directly from
EPA.
EPA will approve a variance on a showing of "substantial and widespread economic and social
impact." EPA recommends that a state adopt a variance, rather than change a designated use, if the state
believes the use can eventually be attained.
8-7
-------
Individual Control Strategic for Point Sources Causing Toxic Hot Spots"
The Water Quality Act ot 1987 created a new program intended to further the goal of achieving
water quality standards. This new program does not substantially change the pre-existing water quality
program, but establishes a tight timetable for achievement of state water quality standards. Within two
years after enactment of the Act (by February 4, 1989) each state must prepare and submit for EPA
approval a list of those waters within the state which will not meet water quality standards or maintain
beneficial uses due to point source discharges of toxic pollutants, despite the implementation of
technology-based limitations. For each such segment, referred to as "toxic hot spots," the state is to
identify the sources of the discharges causing the impairment and the amount of pollutants from each
source.
States must also develop and implement an "individual control strategy" for each point source
identified which, in combination with other controls on point and nonpoint sources, will result in
achievement of the applicable water quality standard within three years after the strategy is established.
EPA must approve or disapprove state lists and strategies within 120 days after February 4,1989.
If a state fails to submit information or EPA disapproves a strategy, EPA will, within one additional year,
implement the requirements for listing and strategies for such state.
The states are to adopt specific numerical criteria for all toxic pollutants which could be expected
to interfere with the designated uses of the water segment. Where numerical criteria are not available for a
pollutant, the state is to adopt criteria based on biological monitoring or assessment methods.
FLOODPLAINS AND WETLANDS CONSIDERATIONS
Executive Orders 11988 and 11990 require federal agencies to take action to avoid adversely
impacting floodplains and wetlands, respectively. Executive Order 11990 requires the minimizing of
wetlands destruction and the preservation of wetland values. These prders apply only to existing, natural
wetlands and not to constructed wetlands with one exception. Current federal policy on constructed
wetlands appears to hinge on whether the wetlands is constructed for water treatment or for other uses.
Water treatment type constructed wetlands are not as controlled or protected as wetlands constructed for
other purposes such as flood control or those created coincidentally with earthwork projects that intersect
the water table.
However, it is possible that constructed wetlands for water treatment may evolve, in the long term,
to exhibit many beneficial features of natural wetlands. At this point, the intent of Executive Orders 11988
and 11990 may be argued in the legal arena on a case by case basis if significant changes (such as the
replacement of substrate or the decommissioning of a constructed wetland) to these systems are
proposed. For example, new, more efficient technologies or the depletion of finite metal sources may
allow the decommissioning of constructed wetlands. It is likely that each site will be handled on an
8-8
-------
individual basis. It is possible, however, that the operators of a constructed wetland may be required to
replace it with a similar facility in the case of closure or may have to take extraordinary precautions in
maintaining it (substrate replacement) to preserve its natural benefits.
One federal permit required for construction of wetlands regulates dredge and fill activities in
navigable waters of the United States. This is better known as the 404 permit issued by the Army Corps of
Engineers. It is important to note that this permit regulates only the discharge of dredge and fill materials
¦within such waters. The deposition of hydroxide or sulfide precipitates in a passive treatment system may
be construed to be a form of "tilling" of a wetland and thus require a 404 permit for the lifetime of the
passive treatment system. In addition, decommissioning of a constructed wetland, either by disrupting its
source of water or excavating the substrate, may involve obtaining of a 404 permit.
EPA's Wetlands Protection Policy
Recent revisions (1991) of the definition of wetlands has reopened debate on this sometimes
emotional issue. Constructed wetland systems may be included within the current definition despite the
fact that these systems function as water treatment systems that are not naturally connected hydraulically
to sources of water. Constructed wetland systems may fail under this purview since the term wetlands
means "those areas that are inundated by surface or ground water with a frequency sufficient to support
and under normal circumstances does or would support a prevalence of vegetative or aquatic life that
requires saturated or seasonally saturated soil conditions for growth and reproduction". Wetlands typically
include swamps, marshes, bogs, and similar areas such as sloughs, potholes, wet meadows, river
overflows, mud flats, and natural ponds.
Whether or not a constructed wetland is included in the new category of wetlands, these systems
should not impact existing "natural wetlands", if it is avoidable.
STATE WATER RIGHTS
Since each state has its own body of laws governing ground and surface waters, it is not possible
to provide a complete discussion of this subject. A water "right", as legally defined, is not legal title to the
water, but the legal right to use it in a manner dictated by law. It may be difficult to determine whether water
exiting from an underground adit is a ground water withdrawal or the headwaters of surface water. For the
purposes of discussion, the later is assumed as a mine water discharge enters a larger surface body of
water. The right to that water may be appropriated to some downstream user. As discussed in
SECTIONSS 6 and 12, evapotranspiration can contribute a significant amount of water volume loss from a
constructed wetland. If evapotranspiration losses from a new constructed wetland facility are significant
enough to affect downstream water rights, the operator of the facility may be required by water law to
purchase or replace those rights. In surplus water years, the downstream effects of a constructed wetland
may not be felt. However, in a drought year, the wetland may be considered a "junior" user and may have
8-9
-------
to compensate a more senior, downstream user for signif icant evapotranspiration losses.
In another instance, a feed water source to a wetland (for example, a discharging adit) may be
controlled/throttled or eliminated by means such as underground bulkheads. The consequences of this
action on downstream senior water rights would also have to be considered.
ENDANGERED SPECIES ACT
The construction of passive treatment systems may consequently create an ideal wildlife habitat.
Eventually, an ecosystem may develop that could include endangered species in the flora or fauna
populations. This event could complicate typical maintenance operations, restrict the operation of the
facility and perhaps affect facility decommissioning. Occurrences of endangered species at passive
treatment sites will undoubtedly be handled on an individual basis. It is likely that relocation to more
protective sites may be preferred for endangered fauna. Endangered floral occurrences, which may be
more sensitive to relocation, may need to be protected in situ, with protective measures developed
specifically at each site in concert with regulatory agency guidance.
RECLAMATION BOND RELEASE (POST-MINING LAND USE)
The goals of constructed wetlands include the immobilization of metals in the substrate and the
positive adjustment of pH. As such, passive treatment systems function as water treatment plants. This
may be the basis of reported current federal policy which appears to preclude the employment of
constructed wetlands as a post mining land use. Although changes lo this policy are reportedly being
sought, il is likely that long term, historical, performance of passive treatment systems will be required
before agency policy change is observed. In summary, il appears that passive treatment systems
remaining on a mine site after closure of olher aspects of the operation may preclude "total" bond release.
A nominal portion of I he bond may be retained to provide funds for maintenance and
decommission/reclamation of the passive treatment system until its operation is no longer required.
SUPERFUND ACT
The Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) of
1980, commonly known as Superfund, was passed by Congress to address the nation's abandoned and
inactive hazardous waste sites. In the event that constructed wetlands are chosen as the preferred
alternative for treating wastes from a Superfund site, it is likely that some, if not all, of the regulations listed
earlier in this SECTION would apply.
CERCLA as it was passed in 1980, did not contain a specific requirement pertaining to the
compliance of on-site CERCLA actions with other laws. CERCLA §105, which authorizes EPA to prepare
the National Contingency Plan (NCP) for hazardous substance response, says only that the NCP shall
8*10
-------
include "methods arid criteria for determining the appropriate extent of removal, remedy, and other
measures." EPA, however, stated in the NOP (as revised in 1985) and in its policy memorandum on
CERCLA compliance with other environmental statutes, which was attached to the preamble to the 1985
NCP, that it would attain or exceed applicable or relevant and appropriate Federal environmental and
public health standards in CERCLA response actions unless one of five specifically enumerated situations
was present.
CERCLA §121, added by Congress in the Superfund Amendments and Reauthorization Act
(SARA) in 1986, in effect codifies EPA's existing approach to compliance with other laws. Section 121
establishes cleanup standards for remedial actions under §§104 and 106 of CERCLA. Remedial actions
must attain a general standard of cleanup that assures protection of human health and the environment,
must be cost effective, and must use permanent solutions and alternative treatment technologies or
resource recovery technologies to the maximum extent practicable. In addition, for any material remaining
on-site, the level or standard of control that must be met for the hazardous substance, pollutant, or
contaminant is at least that of any applicable or relevant and appropriate standard, requirement, criteria, or
limitation under any Federal environmental law, or any more stringent standard, requirement, criteria, or
limitation promulgated pursuant to a State environmental statute.
CERCLA §121 (e) provides that no Federal, State, or local permit shall be required lor the portion
of any removal or remedial action conducted entirely on site," when the action is selected and carried out
in compliance with the cleanup standards requirements in §121. EPA interprets "on-site" to include the
"areal extent of contamination and all suitable areas in very close proximity to the contamination necessary
for implementation of the response action." As a matter of policy, this definition would be implemented
with certain limitations. Generally, best professional judgement should be used to determine that the area
is within "very close proximity" to the contamination and is necessary for implementation of the portion of
the response action addressing the nearby contamination.
Finally, §121 (d)(4) provides that under six specific circumstances legally applicable or relevant and
appropriate requirements can be waived. However, the requirement that the remedy be protective of
human health and the environment cannot be waived.
8-11
-------
SECTION 9
SITE CONSIDERATIONS
SOURCES
The sources of acid drainage from a metal mining-affected environment can be diverse. They are
dependent on site geology, hydrogeology, mining methods and mining/milling waste disposal policies. In
the case of abandoned sites, acidic site runoff may be derived from non-point sources such as scattered
occurrences of waste rock or mill tailings mixed with indigenous soils.
Typically, acid drainage from mining sites manifests itself in the following features:
o Adit/tunnel portals. For these easily identified point sources, discharge rates may range from
seeps of fractions of a gallon per minute to several cubic feet per second; flow rates may be
responsive to local precipitation/runoff/snowmelt events; the portals may be caved or in unsafe
condition; caved roof occurrences at the portal or deeper into the adit/tunnel are likely to impound
water.
o Waste rock piles. These features may have been formed by the filling of ravines or large valleys; in
older facilities, precipitation infiltrating through waste rock piles may follow buried drainage fea-
tures but may exit as a non-point source into ground water; in recently-constructed facilities,
infiltration may follow planned drainage features within the piles and drainage may exit as point
sources.
o Impounded Mill Tailings. Many of the characteristics of waste rock piles, above, apply to tailings
storage facilities; differences in material permeability typically render drainage discharge rates
from tailings facilities underdrains less sensitive to precipitation events.
o Inundated Pits. Mined out pits often fill with water from runoff or ground water sources; site
hydrogeology and final reclaimed topography may result in a steady discharge through a low point
in the highwall of the pit.
o Shafts. These features comprise an unlikely source of acid metal mine drainage; however, some
shafts may have encountered artesian conditions which bring acidic water to near-surface aquifers
or drainages.
o Inclines. Like shafts, these features are an unlikely source of acid mine drainage.
o Seeps "associated" with any of the above features. Hydrologic connections may be difficult to
defend. These may be naturally-occurring seeps that may have been affected by nearby mining
activities or seeps that are naturally acidic and/or metal-bearing.
FLOW RATE VARIABILITY
Determining the typical or average flows that the wetland and conveyances will normally
experience is an important task to complete prior to beginning wetland design. Historical data, if not
already available, should be developed over at least a year to determine seasonal fluctuations in discharge
quantity and quality. As discussed in a subsequent section, mass loading rates will influence wetland
sizing.
9-1
-------
The wetland site flow data should be correlated to historical flow data from other sources/nearby
sites to expand the wetland source database. For example, the gathering of quality and flow data for
receiving streams and meteorologic data for the site watershed may allow correlation of variations in source
flows (from adits, etc.) to larger databases such as those from USGS water gaging stations. Thus,
projections of adit/source flow variability beyond those immediately observed for the site may be derived.
On the effluent end of the wetland, flow variability of the receiving stream may affect the loading
and sizing of the wetland. Changes in water use, water rights, and allocations should be considered in
determining the "base" flow of the stream. This "base" flow may play a role in sizing the wetland to assure
compliance with stream standards.
H a large variability In discharge flow rates is observed, it is typically an indication that surface water
is intruding directly into the hydrologic system. It may be reasonable to attempt to abate the intrusion by
source control measures which may stabilize the discharge rates.
Some sources, particularly long drainage tunnels/adits, are subject to dramatic but short-lived
increases in flow that may have catastrophic effects on downstream wetlands. The source of these flow
increases, referred to subsequently as "surge flow events", is likely to be the erosion and subsequent
catastrophic failure of underground roof fall related "dams" that impound significant amounts of acid water.
Drainage features whose flows exhibit close correlation to surface precipitation/runoff events appear to be
more likely to experience surge flow events. It is suspected that extraordinary surface runoff reporting to
underground mine workings could create stress on roof fall dams, increasing erosion and accelerating
catastrophic failure.
The following mine/tunnel characteristics should be evaluated in order to provide a relative
indication of surge event potential for the mine/tunnel systems providing water to wetland treatment
installations:
o Length of tunnel. The volume of water impounded behind a roof fall is proportional to a tunnel's
length, ignoring the workings connected to the tunnel.
o Extent of connected workings. A tunnel is typically connected to other workings. These
additional connected workings are capable of providing additional hydrostatic head and volume to
a surge event. The more connected workings there are, the higher the likelihood of a surge
event volume being impounded.
o Reports of water, either at the portal or underground. These reports, either in the literature or by
personal observations, contribute to the likelihood of water being impounded underground and
being released in a single event.
o Stopes or multiple shafts intersecting the surface. These mining-related characteristics provide
multiple pathways for surface water inflows into the mine workings.
A means of objectively assigning a relative risk of a surge event to mine/tunnel systems
considered for wetland treatment should be developed. Discharge features with high surge event risk
9-2
-------
may require controls such as underground structures/impoundments to throttle surge flow event flows so
that wetland treatment systems are not overwhelmed or damaged.
In essence, the flows expected at the portal of a mine/tunnel/ waste rock pile system will be a
function of the system's hydrology. As discussed in SECTION 2, the hydrology of mining features may be
broadly characterized into "diffuse" and "conduit" systems. Features with "diffuse" hydrology are less
likely to experience surge flow events and their discharge water quality does not change with climate.
Minimal flow values (drought conditions) will also be important to quantify prior to the wetland
design. Wetland substrate chemical stability appears to be sensitive to desiccation (99). As discussed in
SECTION 3, desiccation of substrate may subsequently result in the re-mobilization of precipitated metal
sulfides that oxidize during desiccation. Ivanov (100) observed that dewatered peat lost permeability as a
result of compression forces generated when the buoyant effect of the water was removed. These
changes in the substrate could dramatically impact the performance of the wetland.
Changes in minimal loading conditions also need to be considered. In areas where snowmelt
provides a significant portion of the acid drainage, flow rates and loading rates are typically not equivalent.
While flow rates to the wetland may decrease after the initial melt, loading rates may remain constant or
actually increase until spring runoff "flushes" accumulated contaminants resulting in a temporary loading
rate spike.
FLUID COLLECTION ALTERNATIVES
Metal hydroxide formation in conveyances due to oxidation of mine waters can interfere with
efficient transport of water to the wetland. Therefore, fluid collection strategies should attempt to reduce
formation of hydroxides by limiting seepage exposure to oxygen. This can be accomplished by inter-
cepting fluids as close to their source as possible to minimize air contact. Previously flooded and
subsequently rehabilitated underground workings with acid drainage exhibited a reduction in metal
hydroxide accumulations in areas with mine atmospheres deficient in oxygen. Otherwise, underground
workings would tend to "self seal" with hydroxide accumulations.
Fluid collection alternatives that conform to a reduced oxygen exposure strategy follow. They
include underground impoundments, portal impoundments, rock/pipe galleries and open ponds.
Underground Impoundments
Besides reducing acid water exposure to oxygen, underground full-face bulkheads provide
several advantages in collecting fluids to be diverted to constructed wetlands:
o Impounded water may provide driving head energy necessary to convey acid drainage to wetland
sites far removed from the tunnel/adit portal location.
o Bulkheads can provide protection from surge event flows by throttling such flows through
pipeline conveyances equipped with valves.
9-3
-------
o Bulkheads allow the utilization of the mine workings themselves as reservoir space. This could
permit the temporary suspension of drainage to allow periodic maintenance of conveyances or
other wetland components.
Bulkhead design considerations include:
o The site geology, particularly the faulting patterns in the underground workings
o Stability of the underground opening, for safety during construction and reduction of grouting
requirements,
o Maximum anticipated static pressures and potentially dynamic heads developed from surge
events; dynamic pressures can be reduced through the construction of a raised air chamber
upstream from the bulkhead,
o Bulkhead locations that provide geostatic pressures above the bulkhead site equal or greater to
the potential combined static and dynamic (surge) pressures,
o Multiple bulkheads to insure containment of fluids within underground workings; i.e, prevent
impounded water from rising to form a new point source discharge on the surface,
o Impacts on adjacent mines or surlace facilities,
o Allowable leakage around bulkheads, fracture flow around bulkheads,
o Contact with acid mine drainage water,
o Allowance for pipes or other conveyances that permit the passage of water through them either
for maintenance or as a standard operating condition,
o Long-term durability; i.e., 50-plus year design life
Underground bulkheads have probably been used in some form to control water since
underground mining began. Bulkheads installed in deep South African gold mines appear to provide the
most contemporary design and construction experience applicable to the construction of bulkheads to
collect acid drainage.
Data on the construction of underground bulkheads in deep gold mines in South Africa (101)
indicate that:
o Leakage through wall rock fractures adjacent to the bulkhead, which is related to the pressure
behind the bulkhead, is the primary design criterion rather than plug/wall resistance to thrust. Data
indicate that the bulkhead should have at least one foot of thickness for every 25 to 40 psi of static
water pressure exerted on the plug.
o The plug may be constructed of concrete formed by emplacement of a cement grout into dean,
strong angular rock fragments (up to boulder size) that have been previously packed between
forms.
o If indicated by drilling and water pressure tests, the rock surrounding the bulkhead should be
grouted through boreholes ring-drilled to a minimum depth of 20 feet deep. Grout pressures
should be at feast equal to the local iithostatic pressure.
9-4
-------
o Plugs may not need to be hitched into the wall rock, tapered nor constructed with reinforcing bars.
The plug/wall rock friction from normal rock irregularities may be sufficient to maintain plug stability
as long as the leakage/pressure criteria above are met.
o Leakage adjacent to the finished bulkhead is usually observed in the floor and back (roof), even at
low pressures. The flow paths are created at the concrete/wall rock contact due to concrete
weeping in the curing process, air pockets and entrained mud. These leakages can be sealed by
standard grouting techniques.
Testing for rock in-situ permeabilities and flow paths should occur prior to, or commensurate with,
the plug design. Rock grouting should be an integral part of plug installation. The harsh chemical
environment that may exist in the tunnel or the rock should be considered for all materials to be placed. A
logical extension of underground bulkheads includes underground excavations (including horizontal and
inclined boreholes) that collect acid drainage.
Portal Impoundments
Portal impoundments include bench or weir type installations that do not completely fill the
mine/tunnel opening. Portal impoundments should only be considered for mining features where risk of
surge flow events is low or cost prohibits the rehabilitation of the workings to allow the safe construction of
a full face bulkhead.
If an existing portal is collapsed, a natural impoundment condition may exist. This condition
should be approached with extreme caution; earth and rock in a collapsed portal should not be
considered adequate substitutes for an engineered, constructed portal impoundment. The region
beyond the collapsed zone should be dewatered with caution before the collapsed material is totally
excavated for construction of impoundment facilities.
Portal impoundment design should consider most criteria applicable to the full-face bulkhead with
the obvious exception of geostatic pressures. Typical portal impoundment facilities should be
constructed of acid-resistant reinforced concrete.
Portal impoundments are typically constructed near the entrance to the underground workings.
Therefore, designers should consider additional measures to reduce the exposure of impounded water
• to oxygen. One method is installing brattice curtains to reduce air movement/oxidation of water. Brattice
curtains of acid-resistant material or masonry walls with underflow conveyances could be hung
from/attached to the roof and walls of the mine opening. In addition, mine timbers could be placed behind
the brattice; the timbers would slowly rot and consume oxygen. These oxygen-depletion measures will
decrease, but probably not totally prevent, hydroxide sludge accumulations.
Rnrk/Pioe Galleries
These installations may be constructed as an integral part of waste rock or tailings facilities or
retrofitted as a toe/embankment extension for completed facilities lacking underdrainage. Designs should
typically include "french drain" type components such as gravel/rock seepage zones contained in
9-5
-------
permeable geofabric envelopes, gas traps, perforated acid-resistant piping or rigid geomembrane/ g-
eofabric composites that eliminate gravel/rock requirements.
Designers should assume that any pipe, either with full pipe flow or open channel flow, will
eventually clog with iron hydroxide precipitate even if oxygen exposure is eliminated. Valves, low points
in the pipe route and bends will clog even faster. Thus, cleanout provisions should be included in every
design.
In past experience, the use of limestone or other acid-neutralizing rock in the construction of
rock/pipe galleries resulted in rock armoring with hydroxide precipitates and the subsequent failure of the
system as a collection/treatment device. Recent advances reported by Brodie and Britt (117) indicate that
the total exclusion of oxygen from a limestone rock gallery prevents the formation of precipitate armoring.
Brodie's term for such a facility is an "Anoxic Limestone Drain" or ALD. Total oxygen exclusion features of
a typical ALD include geosynthetic/plastic and clay soil covers and gas traps. The consumption of
limestone may pose a long-term maintenance consideration. Brodie estimated that typical ALD's installed
by the TVA have operational lifetimes on the order of decades. Thus, the design of a rock/pipe gallery
should consider the eventual replacement of limestone if that particular rock type is used. As discussed in
SECTIONS 3 and 4, limestone may be used as a component of anaerobic wetland substrate. Oxygen
exclusion in a rock/pipe gallery is a design challenge with few obvious solutions. Impoundment of fluids
within an embankment can cause slope failures. The employment of inverted pipe traps and minimum
soil/geomembrane covers over galleries appear to be the best apparent strategies for minimizing oxygen
exposure to water in rock/pipe galleries. However, traps may be maintenance problems because they may
be difficult to clean if they become plugged.
Open Ponds
These water collection features may include inundated mine pits or excavated wide channels with
little, if any, gradient. The exposure of the drainage to air is unavoidable in this circumstance; it is likely that
metal hydroxides will form, posing a long-term maintenance problem. Further, the orange/red precipitates
are not aesthetically pleasing. On the other hand, if the mine drainage is issuing from a tunnel that
historically has a high surge potential, some structure such as an open pond may be necessary to prevent
catastrophic destruction of the wetland from surge flow.
The wetland designer should consider converting the open pond situation to an in-place wetland
if land use or regulatory restrictions allow it. From a geomorphologic viewpoint, lakes and ponds naturally
tend to become wetlands as sediments and vegetation accumulate in the lake bottom. Thus, conversion
of open water to shallow wetlands may: 1) provide a more stable hydrologic environment, 2) increase site
aesthetics and 3) provide water quality improvements.
If total conversion of an open pond collection feature is not practical, shoreline wetland features
that may include wetland treatment cells should be considered.
9-6
-------
SECTION 10
CONSTRUCTABILITY - METHODS AND MATERIALS
SUBSTRATE FROM NATURAL SOURCES
Substrate materials may consist of mixtures of organic and inorganic soils and typically include
animal waste in the form of manure. Substrate component materials have included:
Data suggest that wetland removal performance is closely linked to how the acid metal drainage
flows through the substrate materials. Substrate materials may be selected initially based on local
availability and reasonable cost, then amended (if necessary) to produce a composite substrate material
for a particular application.
A substrate material that has demonstrated good performance for both metals removal and flow
characteristics at the Big Five Tunnel Project is mushroom compost (3, 6). Mushroom compost is a mixture
of manure and brewery waste. Some physical characteristics of the mushroom compost substrate follow
Hydraulic Conductivity
Hydraulic conductivity is an important wetland design parameter because the hydraulic perfor-
mance of the wetland is particularly sensitive to this variable. Hydraulic conductivity, "K", is a cornerstone
variable in Darcy's Law and other hydrologic relationships typically utilized to predict the hydrologic
performance of wetlands.
For the Big 5 Tunnel Project, hydraulic conductivity (permeability) values of wetland substrates
were measured in laboratory and field permeameters by Lemke (9). Laboratory methods have been
developed and documented in the Peat Testing Manual (102) and Fetter (103); methods include
o depleted mushroom compost (50% manure/50% barley mash waste)
o peat moss
o aged manure
o decomposed wood products
o limestone
o planters mix soil (topsoil)
o straw
(9):
Specific Gravity of Solids
Bulk Density, Wet Substrate
Porosity by Volume
Bound Water by Weight
Permeability (down-flow)
Grain Size Distribution
1.66 to 1.78 grams/cubic centimeter
1.23 to 1.33 grams/mi lliliter
25% (typical)
3.7%
10-3 to 10-5 centimeters/second
56.5% passing a No. 10 sieve;
15% passing a No. 200 sieve
71.5% (typical)
Ash Content
10-1
-------
constant and falling head test procedures for permeameters with down-flow and up-flow configurations.
Permeameter configurations are presented in Figures 49 and 50. Laboratory procedures are addressed
in SECTION 14.
Laboratory permeameters are typically clear plastic tubing about two inches (5 cm) in diameter fed
by constant head or falling head measurement plumbing. Laboratory measurements may typically employ
distilled or de-ionized water or other fluids. There may be some debate as to whether this is
representative of actual field conditions; i.e., acidic, metal bearing drainages.
In-situ measurement techniques for permeability in constructed wetland installations typically
involve falling or rising head testing at shallow depths (102). However, dewatering of substrate may result
in compression of lower layers, reducing permeability and yielding erroneously lower values than what may
actually exist in-situ (100).
Field permeameters constructed with 30 gallon (120 liter) capacities have provided data that
correlate well with laboratory scale permeameters (9). However, caution must be employed in the
construction of field permeameters to insure that hydroxide precipitation in feed plumbing does not
interfere with hydraulic conductivity determinations. A typical field permeameter is depicted in Figure 51.
Lemke reported wetland substrate hydraulic conductivity values (mushroom compost) from 10-3
to 10*5 cm/sec for "fresh" and "used" (submerged for about one year), respectively. Laboratory permeabi-
lities for wetland substrate appeared to vary with the following parameters:
o Age (Humification)
o Wetted condition
o Size distribution
o Flow direction (up-flow versus down-flow)
Lemke (9) observed that permeability was a function ol wetted condition. Used mushroom
compost that had been allowed to dry typically had permeability values nearly equal to unused mushroom
compost that was obtained in a dry condition. Diying of used mushroom compost appeared to allow the
recovery of about one order of magnitude of permeability when compared to the wet condition. Various
mechanisms may be responsible for this phenomenon; a likely one is the adherence of small particles to
larger particles upon drying. Studies have shown that permeability values for dry mushroom compost
decrease as soon as submergence begins and approach typical "wet" permeability values after about
three months (99).
Winneberger (104) gives a detailed discussion of the changes in permeability of soils as they are
inundated with water. Winneberger points out that in 1922, a soil scientist named E.V. Winterer observed
that inundated soils go through three phases of permeability change. During Phase I (about 10 days),
infiltration rates decrease rapidly. This initial decrease is explained as the result of "slaking" of the soil; i.e.,
changes in the affinity of the soil surfaces for water and the decrease of the cohesive forces that hold the
soil particles together (Winterkorn, 1942 and Christiansen, 1942, referenced in (104)].
10-2
-------
u
WIRE SCREEN
UPPER RESERVOIR
WATER TI6H
SEAL
77777771
EZZZZZ
SIPHON
WATER SUPPLY
OVERFLOW
UPPER TROUGH
SPECIMEN (Supported)
WIRE SCREEN SUPPORT
FUNNEL (Supported)
BEAKER with PERCOLATE
AH = TOTAL HYDRAULIC HEAD DIFFERENCE
ACROSS SPECIMEN.
Figure 49. A diagram of a typical downtlow laboratory scale permeameter.
-------
CuntUN ««un ierrl
WMzy/'w/,.
mmrnrn
V2ZZZZZ3
PT7777
Overflow
POROUS PLATES
FUNNEL (SUPPORTED)
BEAKER WITH PERCOLATE
TOTAL HYDRAULIC HEAD DIFFERENCE
ACROSS SPECIMEN.
Figure 50. A diagram of a typical upflow laboratory scale permeameter
10-4
-------
tfl
.^V"V"V*V*V*V*V'
issis;A
INLET
OVERFLOW
GRAVEL
•».•» *
. i\ •»"< »V\ .
WVV*N>V\
•' r -v <*; «v .7 O ** «¦; «*.¦'• -*.*/• /¦ /«/ »v
^ \ ¦/. •'» ;S ¦ s ¦/» ¦ s >s ¦ % • •« • s¦% • s • % • \ ¦ •%•%•%¦
•"» • S ;S • \ • \ >\• % • % «\«\ •> •> • % »\• S ¦% • S •
5$
•#(
ft
I Vim
f®
,°a0,
OUTLET
BENCH SCALE WETLAND MODULE
DOWNFLOW MODE
Figure 51. A diagram of a bench scale permeameter.
-------
In Phase II, which lasts about 25 days, the permeability increases again, due to removal by
dissolution of air trapped in soil pores. Phase III is a slow (several month) decrease in permeability due to
biological activity [Allison, 1947 in (104)].
According to Winneberger (104), "With numerical variations, such a (three-phase) curve has been
observed of essentially all inundated soils whether tested in a laboratory or in the field." Further, the three
phases appear to compress in time when sewage or waste water is applied due to the addition of high
concentrations of bacteria and their food source.
In addition to the phenomena described above, decreases in substrate permeability with
age/use/submergence (9) may be due to the biochemical decay and disintegration of organic fractions in
the substrate. Theories on the formation of coal (105, 106) suggest that stagnant flow conditions in
naturally occurring wetlands promote the preservation of organic matter. Conversely, "dilution" of
stagnant conditions resulting from water movement through the wetland appears to promote organic
decay or humification.
TABLE 27. HUMIFICATION EFFECTS ON COEFFICIENT OF SEEPAGE
VALUES OF DIFFERENT KINDS OF PEAT FROM REF (100)
Coefficient of Seepage (cm/s)
Type of Peat in Natural Deposit and
I Average
I Limits of
Degree of Humification
| Value
| Variation
Fen peat (Hypnum-sedge, sedge, sedge-Sphagnum)
Slightly humified (25-30%)
0.005
0.002 - 0.01
Moderately humified (40-55%)
| 0.0008
0.0002 • 0.002
Bog peat
Very slightly humified (up to 10%)
0.015
0.01 - 0.025
Slightly humified (10-20%)
0.004
0.002 - 0.007
Moderately humified (35-45%)
0.0005
0.00025 - 0.001
Much humified (55-65%)
5 x 10"5 |
2x 10-®-8x 10 s
Thus, inundation in a reducing environment appears to slow, but does not halt the humification
of wetland substrate. Ivanov (100) observed that "coefficient of seepage" (permeability) is a function of
the degree of humification and not a function of peat type. Table 27 presents permeability values for
two types of peat as a function of humification.
Day, et. al. (102, Appendix C) contains a classification of degree of humification. The distinction
among the 10 grades of humification appears to be a function of "muck" content, plant remains,
10-6
-------
consistency and the color of water extracted when the sample is hand-squeezed. Thus, degree of
humification appears to be a relatively subjective characterization.
Further, the accumulation of metallic precipitates in substrate void spaces will also tend to
decrease permeability with continued use. Thus, it appears likely that the permeability of substrate will
change (decrease) as a constructed wetland is operated and may become a maintenance consideration.
Permeability of granular materials is universally accepted to be a function of particle size
distribution. The smaller the particles and the more evenly distributed (as opposed to uniform-sized
particles) a granular material (substrate) is, the less permeable it will be. Data from Hough (107) as
presented in Table 28 support this.
Permeability values appear to be a function of flow direction for some wetland substrate
materials (9). Permeability measurements in a down-flow configuration are significantly different from up-
flow measurements. The phenomenon is probably due to suspension of fine-grained particles in the
upward flowing fluid compared to clogging of flow passages with finer-grained particles in a down-flow
configuration. These permeability differences have been noted in bench-scale (30 gallon container)
experiments, and in larger pik>t-scale(100 square feet) wetland configurations over short time frames.
However, after the pilot-scale upflow cell had been operating for 10 months, permeability decreased to
about the same as in the downflow cell. Flow velocity through the substrate materials in up-flow
conditions appears to be the Key criterion that needs consideration.
Particlfl Si?ft Distribution
Hydraulic conductivity is directly related to particle size distribution of substrate materials.
Particle size distribution requirements for substrate are typically easier for a design engineer to specify in
construction bid documents than hydraulic conductivity.
The discussion of particle size distribution is another case in which the distinctive vocabularies
of different disciplines (in this case civil engineers and geologists) can lead to confusion. For example,
to describe a collection of particles of primarily a single size, civil engineers use the term "uniform". For
the same collection of particles, geologists use the term "well-sorted", based on hydrological sorting
processes, or "well-graded", because the particles all lie within the limits of a single "grade".
At the other extreme is a collection of particles of various sizes, in which all sizes are
approximately equally represented. Civil engineers use the term "evenly graded" or "well-graded" for
this collection. By this term, they signify that a uniform mixture, of "gradation" of particle sizes exist.
Geologists use the term "graded" in the opposite sense, so that they would term this same mixture of
particles "poorly sorted" or "poorly graded".
The geological perspective is the three-dimensional context in which the particles were
deposited in a sedimentary environment, based on the different settling rates of different-sized
particles. In geological terms, a "graded bedding" is a sedimentary deposit "in which each layer displays
a gradual and progressive change in particle size, usually from coarse at the base of the bed to fine at the
10-7
-------
TABLE 28. TYPICAL VALUES OF PERMEABILITY COEFFICIENTS
FROM HOUGH (107) AND FETTER (103)
Material
Particle
Inches
Size Range
Millimeters
•Effective*
Size
Permeability
Coefflcient-k
Civil Engineering
Terminology (107)
Dmax.
Dmin.
DfTMX
Dmin.
Dajin.
Diomm.
Cm./sec.
Derrick STONE
120
36
-
-
48
100
One-man STONE
12
4
-
-
6
30
Clean, fine to
coarse GRAVEL
3
1/4
80
10
1/2
10
Fine, uniform GRAVEL
3/8
1/16
8
1.5
1/8
5
Very coarse,
clean, uniform SAND
1/8
1/32
3
0.8
1/16
3
Uniform, coarse SAND
1/8
1/64
2
0.5
-
0.6
0.4
Uniform, medium SAND
-
-
0.5
0.25
-
0.3
0.1
Clean, well-graded
SAND & GRAVEL
-
-
10
0.05
-
0.1
0.01
Uniform, fine SAND
-
-
0.25
0.05
-
0.06
40 x 10-4
Well-graded, silty SAND
& GRAVEL
-
-
5
0.01
•
0.02
4 x 10-4
Silty SAND
-
-
2
0.005
-
0.01
10-4
Uniform SILT
-
' -
0.05
0.005
-
0.006
0.5 x 10-4
Sandy CLAY
-
-
1.0
0.001
-
0.002
0.05 x 10-4
Silty CLAY
-
-
0.05
0.001
-
0.0015
0.01 x 10-4
CLAY (30 to 50%
clay sizes)
-
-
0.05
0.0005
-
0.0008
0.001 x 10-4
Colloidal CLAY
(-2u £ 50%)
-
-
0.01
10 Ao
-
40A°
10-9
Hydrogeologlsts
Terminology (103)
Conductivity
(cm/sec)
Clay
-
-
10*9-10-6
Silt, sandy silts,
clayey sands, till
-
-
10-6-10-4
Silty sands, fine sands
-
-
10*5-10-3
Well-sorted sands,
glacial outwash
-
-
10-3-10-1
Well-sorted gravel
•
•
10*2-1
10-8
-------
top" (Glossary of Geology, Bates and Jackson). Each layer of this deposit would be "well-sorted" (or
"uniform" in civil engineering terminology). Only if a slice of this deposit, through all layers, were stirred
or homogenized in some way, would that slice become "well-graded" in civil engineering terms.
In this report, we will use the term "sorted" whenever possible. If the term "graded" is used, it will
be in the civil engineering context.
The relationship of particle size distribution to hydraulic conductivity has been extensively
studied. It is generally accepted that the more uniform the particle size is in a collection of materials, the
higher the void ratio and the higher the hydraulic conductivity when compared to a well-graded or non-
uniform collection of particles. For example, a group of uniform spheres of a certain size diameter T will
have a higher hydraulic conductivity than a group of spheres which have evenly distributed diameters
between diameter T and a significantly smaller diameter 1". The small diameter particles fit within the
voids between the larger particles, thus reducing overall volume of voids and lowering the permeability.
Thus, if a high permeability is required, a material with relatively uniform size particles of a size T
would be specified. If a low permeability is required, a poorly-sorted size distribution would be specified
between particle diameter sizes T and t. Table 28 presents relative size distributions (standard soil
classifications) and estimated permeabilities. The first part of Table 28, which gives hydraulic
conductivity values for various particle size collections, is from a civil engineering text (107). The second
part is from a hydrogeology text (103).
Particle size measurements for substrate are typically performed by either dry sieving (ASTM
D2977) or wet sieving (102). These methods were developed for "peat" materials. While some
substrate materials do not satisfy the definition of peat, these testing protocols appear to provide
sufficient data for substrate particle size characterization. As with all laboratory methods, revisions
should be considered to fit individual materials and situations.
Day, et al (102) summarize the dry sieving method:
A representative test specimen of air-dried peat is separated into four designated fractions by
means of an 8-mesh and 20- mesh sieve. The fractions are: (1) foreign matter, removed
manually from the 8-mesh sieve; (2) coarse fibre, retained on the 8-mesh sieve; (3) medium
fibre, passed through the 8-mesh sieve but retained on the 20-mesh sieve; and (4) fines,
passed through the 20-mesh sieve retained in the bottom pan. The mass percentage of each
fraction is calculated on the as-received basis.
Day, et ai, (102) summarize the wet sieving method:
A shaken mixture of peat and distilled water is wet sieved through the standard soil sieves. The
residue on each sieve is oven-dried at 105°C and its mass weighed to determine the
percentage of each of the four particle sizes.
Due to the likelihood that finer particles may adhere to larger particles as long as the substrate
material is dry, wet sieving methods are probably more representative of actual particle size distribution
10-9
-------
within the inundated wetland. Slaking effects noted by Winneberger could be evaluated for dry
substrate samples by observing variations in permeability with "presoaking" time periods. Presoaking
times on the order of weeks should be considered. Also, if excessive substrate depths are being
considered, triaxial cell permeability determinations should be conducted.
There is no "recommended'' size distribution for "typical" wetland substrate. The desirability of a
given substrate material will be a function of how closely it provides a desired hydraulic conductivity and
how stable the size distribution remains with prolonged submergence. Table 29 (9) presents
permeabilities determined for mushroom compost that was used in Cell B at the Big Five Wetland. At the
bottom of Table 29 is the size distribution of the substrate used in Cell A of the Big Five Wetland Site (9).
From a practical standpoint, it may be desirable to mechanically sieve or physically amend a
substrate material to provide a size distribution that will ultimately yield a given hydraulic conductivity.
Lemke (9) examined mechanical separation of finer size fractions and amendments to adjust
permeability of used mushroom compost with measurable but marginal changes (increases) in
permeability.
TABLE 29. PERMEABILITIES (VARYING SCALE & FLOW PATH)
AND SIZE FRACTIONS FROM CELL A MUSHROOM COMPOST FROM REF (9).
PERMEABILITIES
SCALE
FLOW PATH
K (cm/sec)
Lab
Downflow
3.50 x 10-3
Intermediate
Downflow
3.14 x 10-3
Pilot
Downflow
2.96 x 10-4
Lab
Upflow
6.65 x 10-2
Intermediate
Upflow
1.44 x 10*2
Pilot
Upflow
1.38 X 10-2
SIZE FRACTIONS
SIEVE
RETAINS
% DRY WT
CUMULATIVE %
10
2 mm
25.8
28.5
20
0.85 mm
21.1
46.9
100
0.15 mm
32.6
79.5
200
0.075 mm
4.5
84.0
PAN
<0.075 mm
16.0
100.0
10-10
-------
Substrate particle size distribution (arid hydraulic conductivity) at a given vertical zone within the
wetland may change with time due to several factors:
o Addition of coarser particles from the development of plant root systems and dead stalks.
o Disintegration of organic fractions due to decomposition/ humification, which is inevitable with
extended submergence
o Gravity migration of heavier or larger particles toward the bottom of the wetland.
o Migration of lighter particles in the direction of flows.
o Precipitation of mineral species in interstitial voids,
o Compaction of the substrate under its own weight.
Bio-Compatibility
Substrate bio-compatibility is an important design consideration. If sulfate-reducing bacteria do
not thrive in the media, the wetland may not meet performance expectations even though it may be
properly designed in accordance with other criteria.
The presence of naturally-occurring sulfate-reducing bacteria in candidate substrate materials is
a strong indication that bacteria will thrive in acidic/metallic drainage. Tests for the presence of sulfate
reducing bacteria are discussed by Batal, et. al. in (10). Many animal manures have naturally-occurring
sulfate reducing bacteria populations. At a minimum, laboratory scale tests with substrate candidates
and effluent samples should be conducted.
A1:1 ratio of substrate to effluent sample (30 grams solid : 30 grams liquid) has been shown to
be sufficient in developing significant sulfate-reducing bacteria populations in as little as two weeks (99)
of incubation. Qualitative indicators such as the occurrence of black precipitates in test containers are
used to denote the presence of sulfate reducing bacteria.
Offensive Smells and B O D. Considerations
If the substrate material contains manure, offensive smells, high Biological Oxygen Demand
(B.O.D.) and ammonia may be observed. These problems can be minimized if the substrate materials
are composted well before use in a wetland. Composting methods have been well developed and
documented in the literature (108). Composting is an accepted method of encouraging the biochemical
degradation of the organic fraction of solid waste material; having a humus-like substance as an end
product. Aeration and moisture control of composting materials are important parameters for the
successful use of the technique. Typical composting time varies from two to ten weeks, depending on
initial carbon to nitrogen and carbon to phosphorus ratios in the feed stock.
If composting of manure-rich substrate materials is not practical, polishing steps provided by
additional constructed wetlands downstream from the metal-removing wetlands may be required. The
design and construction of wetlands for municipal waste water treatment are well documented in the
literature (109).
10-11
-------
Organic Content
Minimum organic content requirements to "fuel" biochemical reactions have not been
established. A discussion of organic content stoichiometry in relation to sulfate reduction is presented
in SECTION 12. Lemke (9) measured organic contents of 28.5% and 29.0%, respectively, for "fresh"
and "used" mushroom compost materials. Statistical uncertainty or organic additions from plant sources
may have been the cause of the apparent net gain in the substrate's organic content with one-year's
use/submergence.
However, Lemke noted that the size distribution of organic material did change with use; the
medium size fractions experienced an increase in organic content, suggesting organic material disinte-
gration or humification.
Carbonate Rnnrrag
Carbonate is required in the substrate to provide buffering capacity and a source of anions for
the removal of manganese as a carbonate. Limestone is the preferred carbonate source due to its
typical low cost and wide availability.
Size distribution of limestone amendments, when required, should be consistent with overall
substrate characteristics. The heavier density of limestone (2.6 grams per cubic centimeter [g/cc]) when
compared to substrate (1.6 g/cc [9]) could induce the settlement of limestone particles to the bottom of
the wetland. A finer grind of limestone (tempered by permeability considerations) should be considered
to counteract this effect and provide more surface area for geochemical reactions.
Minimum carbonate content requirements in substrate materials should be established based
on the stoichiometfy of each wetland system and field trials.
Substrata from Svnthatlr ftm.r
The US Bureau of Mines has developed a "synthetic" substrate material it has named "Bio-Fix"
beads (110). The material consists of sphagnum peat moss surrounded by pervious long chain
polymers that make the peat moss more durable and more reusable. The beads function in the same
way as ion-exchange resins and thus are limited to a narrow operating range in pH and loading rate.
At a significant sacrifice in a system's passive treatment nature, the beads may provide a suitable
downstream polishing step to a constructed wetland. The beads are reusable; i.e., metals can be
stripped from them using dilute sulfuric acid.
Proper sizing of the polishing ceil using Bio-Fix beads may allow the stripping of beads to
coincide with scheduled periodic wetland maintenance, thus preserving the passive aspect of
remediation of water quality problems with wetlands. However, since the USBM estimates that
approximately 80 bed volumes at relatively dilute metals concentrations will load the system and require
the subsequent stripping of the Bio-Fix media, this is not an extended enough time to allow the truly
"passive" operation of a wetland using this substrate. A typical substrate loading would be completed in
about 240 days.
10-12
-------
CONTAINMENT STRUCTURES
From a cost perspective, substrate containment structures for constructed wetlands should take
advantage of local conditions; i.e., on-site materials should be employed if practical and designs should
consider the availability of specialized materials.
There are four basic construction alternatives available to contain substrate. These are:
o Natural soil-lined excavation: While typically economical, this approach allows some seepage
losses to local ground water, which may not be an acceptable condition, even if the seepage
meets quality criteria.
o Geosynthetic lined excavation: While seepage losses are minimized, substrate maintenance
considerations may further increase cost. For example, the liner needs to be protected if the
substrate has to be periodically removed/replaced/rejuvenated.
o Structural concrete: This high capital cost alternative provides a durable, maintainable facility.
System flexibility may be sacrificed, however. Some components will require acid-resistance or
acid protection.
o Prefab coated metal/fiberglass tanks: These might be considered for situations where waters
have dilute metal concentrations that allow "column" configurations and adequate hydraulic
head is available to drive the system. Metal surfaces should be coated with rubber or other inert
material to provide protection from corrosion.
MODULAR UNIT CONCEPTS
A modular unit design philosophy should be considered to allow system operational flexibility.
Here, a balance must be struck between the concept of a single large wetland treating all the acid/metal
drainage effluent and a multitude of smaller cells that each receive a proportion of the effluent.
The basic problem with large modules is that they may be difficult to control, slow to respond to
changing conditions and difficult to adapt if redesign is necessary. The same problem might apply to
balancing flows and conditions among many small modular units. Clearly, there should be a typical
"range" of module sizes that fulfill the criteria.
Overall wetland size will be governed by design criteria and methods developed in SECTION
12. However, overall effluent flow rate and the minimum "manageable" flow rate will probably be the
governing criteria to module size.
Minimum module size will probably be governed by the overall flux criteria and a unit flow rate of
one gpm through the module. One gpm is probably the lowest practical flow rate value that can be
managed without sophisticated flow metering/controlling devices. Thus, if a design flux rate of 800
square feet per gpm is required for metals removal, a minimum module size might be 800 square feet or
a square with 28.3 foot side lengths.
Assuming one large single module, the maximum module size is governed by the overall system
requirements. Here, site considerations may take precedence, assuming that acid/metal drainage
10-13
-------
effluent flow rates are held relatively constant by controls Installed at the source or within the collection
systems.
Among the criteria that will influence the size of the wetland cell are:
o Ease of performing maintenance functions, such as cleaning of pipes and adding organic
materials to the substrate and aspects of long term reliability.
o Hydrology of flow distribution. Large flow rates may need to be allocated among several wetland
cells to allow reasonable distribution pipe sizes.
o Site configuration. The topography and hydrology of the site may limit wetland cell dimensions.
o Need for additional removal process cells. As discussed in SECTION 3, aerobic processes can
release what was removed by anaerobic processes. Separate cells may be necessary to isolate
processes.
o Need for a substrate materials handling area. New substrate materials may need to be stored
and mixed. Old substrate materials may need to be stored or dried prior to disposal. These
operations may have to be done on the site.
An "optimum" cell size is difficult to define because each of the above
criteria may be satisfied by different configurations. It seems logical to at least divide the flow into from
five to ten equal and parallel streams. More than a dozen cells might be difficult to control. If treatment of
acid drainage must be highly reliable; i.e., system availability must be near 100 percent, the minimum
number of cells might be two, each large enough to handle the entire flow from the source. Thus, if
maintenance needs to be performed on one cell, the other functions as a backup.
The long term reliability of wetland performance has yet to be established since it appears that
few man-made systems have been operated for more than five years. As evidenced by peat bog
accumulations in many parts of the world, the stability of natural systems has been measured in centuries
(111), given constant climatS conditions and constant rates of land subsidence and water recharge.
10-14
-------
SECTION 11
CONVEYANCES/FLOW CONTROL
Ideally, wetland treatment systems should be located as close as possible to the collection
systems or sources of acid/metal discharges. Site restrictions such as land ownership or land use,
however, may prohibit the utilization of otherwise ideal wetland sites. Thus, acid/metal drainage may need
to be conveyed some finite distance to the constructed wetland site.
PIPES AND PIPELINES
Pipes are the logical first-choice conveyance method, in keeping with the "oxygen-exclusion"
philosophy developed in SECTION 9 to limit the formation of clogging metal-hydroxide precipitates,
conveyances should be designed to be fully enclosed. Piping of water certainly satisfies this criteria; full-
pipe flow satisfies it to a greater extent. However, open channel flow within an enclosed pipe may offer
advantages as discussed subsequently in this section.
Besides meeting the typical pipe design criteria related to flows and pressures, pipeline materials
should be acid/chemical resistant and, if they will be exposed to the elements, ultraviolet light resistant.
Many plastic and fiberglass pipe materials satisfy these criteria. Stainless steel also satisfies the criteria, but
exorbitant costs will probably limit its use to relatively short reaches. As periodic pipe cleaning will probably
occur, the materials/linings should be selected to withstand such treatment.
To insure year-round operation in sub-freezing climates, pipes should be buried below'the frost
line. In rocky terrain, this requirement may increase installation costs, but system maintenance headaches
from freezing pipes will be minimized. Pipe burial also provides security from other forms of surface
damage.
If elevated mine water temperatures are the norm and flows are constant, pipe insulation may
substitute for below frost line burial.
The most critical aspect of pipe utilization is the maintenance consideration of metal-hydroxide
precipitation accumulation. Pipeline configurations that convey feed waters to wetlands should avoid low
points, valves and sharp bends, as these features may induce hydroxide accumulations. Pipes should be
installed to allow for easy inspection and cleanout. If precipitate accumulation is a serious problem, several
approaches might be considered to alleviate It, none of which is totally satisfactory:
1) Settling of precipitates in a holding pond prior to entering the pipeline. The pond may require
periodic cleaning, which lessens the passive nature of the installation and increases operating
cost. Sludge disposal needs to be considered.
11-1
-------
2) installing a parallel/backup pipeline to allow uninterrupted flows from the collection system to the
wetland during periodic cleaning. The increase in capital cost for the extra pipe would need to be
compared to other alternatives.
3) Design for partially filled conduits (open channel flow). While the flows are exposed to oxygen,
the excess cross sectional area that is available for precipitation buildup may prolong the time
interval between required cleanings. However, utilization of hydraulic head is very limited in open
channel flow situations.
For pipeline conveyances, the diameter of the pipe should be selected to promote full-pipe flow
and to provide scouring velocities (2.5 feet/sec) to limit the accumulation of metal-hydroxide precipitate.
Excessive headlosses may preclude the employment of scouring velocities to maintain precipitate-free
pipes.
Anywhere thai precipitation is anticipated within the wetland system, traditional flow controls like
valves should be avoided if possible. The precipitate can foul valve mechanisms such as gates or valve
seats. If valves must be used, they should be lull flow" types such as ball valves or valves designed for
slurry pipeline application.
Headloss through a "partially-closed valve" should be avoided as a flow control technique. Again,
this is due to the accumulation of precipitate in the high turbulence areas of the valve body, especially for
flows below scouring velocity. Flow rates can be adjusted by modifying headlosses with unrestricted full
pipe flow methods such as a llexible tube whose discharge elevation is varied. An example of this type ol
flow control is shown as Figure 52.
From experience in handling mine drainages over the last five years, the following guidelines
indicate when precipitate buildup in pipes will be a problem:
o The most troublesome precipitate is ferric hydroxide. Any water with over 1 rng/L of dissolved iron
can potentially cause a problem.
o Analysis of stability diagrams of iron species in water (20) reveal that mine drainage with pH > 3.25
is certain to cause problems because, above pH 3.25, ferric hydroxide is the stable form. Waters
with pH < 2.75 will cause fewer problems because dissolved ferrous iron is the stable form.
However, these low pH waters will be quite corrosive.
o Mine drainages whose flows fluctuate because of the invasion of shallow subsurface water will
cause problems because the mixing of the waters will drive the pH above 3.25. Problems will
appear a week or two after invasion.
o With waters below pH 2.75, a coating of precipitate will eventually adhere to the wall of pipes-even
plastic pipes. Once this occurs, monthly maintenance is a necessity. The key maintenance points
are partially closed valves.
11-2
-------
OPEN CHANNEL FLOW CONVEYANCES
Open channel conveyances like ditches and flumes are one alternative to pipe flow. Alternately,
as suggested earlier, open channel flow in partially-filled conduits should also be considered.
A distinct disadvantage of open channel flow conveyances is the sacrificing of hydraulic head that
may be available to force feed water though wetland substrate with a tow hydraulic conductivity. However,
open channels such as ditches and flumes may offer maintenance advantages. First, these conveyances
are easy to inspect; they do not require elaborate monitoring appurtenances such as test spools or
inspection ports. Second, if open channel conveyances are sized to allow for the accumulation of
hydroxide precipitate, the time period between conveyance cleaning may be extended to perhaps
decades. For example, if an open ditch will carry water from the collection system to the wetland, the
bottom width and depth of the ditch should be "over-designed" to allow for the accumulation of precipitate
without compromising the flow capacity of the conveyance. The same applies to over-sized pipes carrying
drainage flows in an open-channel flow mode.
Freezing conditions may preclude open-channel flow conveyances. If the installation of open-
channel covers is considered, one might as well opt for a buried pipe conveyance, in open-channel or full
pipe flow mode.
Continuously-primed siphons (CPS) may be used within a compartmentalized wetland to control
short circuiting or to provide low-tech passive flow or level control. Such devices, as shown on Figure 53,
may be used to balance or distribute flows among various wetland components. The upper "U" of the
CPS should be protected against freezing; the only other constraint to operation is that the elevation of
the lower "U's" must be equal.
Maintenance of the open channel is an important design consideration. Open channel routes
should provide for heavy equipment access during routine maintenance operations such as mucking out
precipitates or other conveyance cleaning tasks.
11-3
-------
Qin
Qout
FLEXIBLE HOSE
Qout= K • dH/L
(DARCY'S LAW)
VARY Q BY VARYING dH
BY RAISING/LOWERING
FLEXIBLE HOSE OUTLET
MEASURE L FROM TOP
OF SATURATED ZONE
z;
JSZ_
SUBSTRATE
PERMEABILITY = 'V
AREA, A
°r\°n° °o °oo O ° Oo°
oMdO°(Vo oOO°o0oO°
•°«
bo
GRAVEL BED/PLENUM
K gravel >= K
NOT TO SCALE
Figure 52.
A cross-section view of a wetland cell flow control system.
-------
SIPHON
CELL WALL
A
NOT TO SCALE
Figure 53. A cross-section view of a constant-prime siphon
11-5
-------
11-6
-------
SECTION 12
WETLAND DESIGN METHODOLOGIES
Besides the basic hydrologic design approach (stornVrunoff routing) that is necessary to assure
that a wetland can handle design flows, other methodologies may be applied to satisfy geochemical/
bacteriological constraints.
Design size/configuration of wetlands may be based on:
o Area/flux - this is based on Darcy's Law.
o Precipitated metal mass loading - independent of void ratio, substrate should be capable of daily
loadings of about 300 nanomoles of metals per cubic centimeter. pH values associated with mass
loading above this value may overwhelm sulfate-reducing bacteria.
o Precipitated metal volumetric loading - filling of void spaces in the substrate.
o Self sustaining capability • surface area is large enough to allow dying plants to replenish organic
material to support a suitable void ratio.
o Water balance - evapotranspiration can contribute to wetlands metal removal efficiency in warmer
climates.
o Sulfate reducing stoichiometry and its effects on substrate carbon content.
A brief discussion of each design methodology follows.
AREA/FLUX METHOD
The application of Darcy's Law is the physical foundation of wetland design, as the typical wetland
flows can be characterized as laminar flow through porous, saturated media (substrate).
Darcy's Law (see equation below) relates the flow (Q) to the cross sectional area (A) perpendicular
to the fluid flow direction, the hydraulic gradient (i) and the permeability of the media (K) as follows:
Q»K*i*A - K * (dH/L) * A
where: Q is flowrate (cubic centimeters per second)
K is hydraulic conductivity (centimeters [cm] per second)
dH is value of constant head, cm, needed to maintain a sustained flowrate, Q
L is the length of the specimen, cm
A is cross section area perpendicular to the flow path (square cm)
Ivanov (100) reported that permeability varies with depth in the wetland, but since flow directions
are perpendicular to the wetland surface, the effects of minor variations are masked and the flow of water is
12-1
-------
restricted by the smallest permeability value in the substrate column. Typically, permeability would be
expected to decrease with depth, as the pore/void spaces in the substrate are influenced by increasing
static pressure from overlying substrate. Ivanov reported that humification of the substrate also signifi-
cantly affects substrate permeability as shown in Table 27.
The hydraulic gradient is a variable that is a function of the depth of substrate, L, and the loss of
driving head from friction, dH, as the flow passes through the substrate media.
The value of L for wetlands typically ranges from two to five feet, the nominal substrate depth; the
value of K for substrate ranges from 10-2 to 10-5 cm/sec for upflow or downflow cells; the area, A is the
surface area of the wetland; i, the hydraulic gradient across the substrate, is typically assumed to be no
greater than 1.0 because ponding on the surface of the wetland should be avoided to preserve anaerobic
conditions. However, hydraulic considerations in a "closed" system could require a hydraulic gradient
greater than 1.0 while anaerobic conditions are preserved by other means.
Table 30 presents a mathematical application of Darcy's Law using typical wetland design
parameters. The "spreadsheet" presentation of Darcy's equation allows the evaluation of many possible
incremental changes in the variable parameters of flow, permeability, substrate depth and surface area.
Lotus 123 ™ was used for the spreadsheet calculations. Cell formulas are included on the Tables to allow
designers to develop similar tools.
Table 30 includes several derived parameters, including flow flux, F, otherwise called the hydraulic
loading factor (SECTION 7). Many wetland researchers employ this flow flux as a key indicator of wetland
performance. A dimensional analysis of flux units as presented in Table 30; i.e., square feet per gallon per
minute (sf/gpm), reveals that the flux unit is the reciprocal of velocity, as discussed in SECTION 7. On the
other hand, the unit for permeability is velocity (cm/sec). In Table 30 at a hydraulic gradient of 1.0, flux
units are actually the inverse of permeability. For example, 800 sf/gpm can be converted mathematically to
the value 1/8.5x10 5 cm/sec (1 /K).
Caution should be exercised in comparing flux values for aerobic and anaerobic wetlands. The
surface area cited in aerobic system flux discussions is parallel to fluid flow through the substrate and not
used in Darcian calculations. The area value utilized in Oarcian analysis of aerobic systems is typically
calculated by multiplying the depth of the wetland by the breadth (109). The area value in anaerobic
system flux calculations (using upflow or downflow) and Darcian analysis is the same; i.e., the surface area
of the wetland, which is perpendicular to fluid flow direction.
Other derived parameters in Table 30 include uniform shape (square and circle) wetland
dimensions of side length, S, and circle diameter, D, respectively. These provide a rough perception of
wetland cell dimensions.
An explanation of flow flux rate nomenclature is appropriate at this point for clarification. For the
purposes of discussion, flux rates associated with high flow velocities through substrate are defined as
12-2
-------
Delta* »
(SEE MOTE 1)
TABLE 30 ESTIMATE OF PRESSURE DROP ACROSS AND UPFLOW OR DOWNFLOW
USING DARCVS LAW
ft 800 sq-ft
'¦variables
1 gpot
WETLAND CELL
90 X 0
NOTES
Initial value
•
•
•
•
S
0
0
0
F
K
L
dH
dH
i
A
SQUARE
CIRCLE
FLOW
FLOU
FLUX PERHEA8LTY
DEPTH
HEAD10SS
HOLOSS
GRADIENT
AREA
SIDE L
D1A
gpm
eft
af/gp*
cm/sac
fMt
fMt
inches
ft/ft
sq-ft
ft
ft
¦an
am
ana
••¦a
am
ana
SMS
ssss
SSSS
nn
aan
1.0
2.2E-03
800
. 3.5E-04
6.0
1.46
17.5
0.24
800
28
32
2.0
4.5E-03
800
3.2E-04
6.0
1.62
19.4
0.27
1600
40
45
3.0
6.7E-03
800
2.8E-04
6.0
1.80
21.6
0.30
2400
49
55
4.0
8.9E-03
800
2.6E-04
6.0
2.00
24.0
0.33
3200
57
64
5.0
1.1E-02
800
2.3E-04
6.0
2.22
26.6
0.37
4000
63
71
6.0
1.3E-02
800
2.1E-04
6.0
2.46
29.6
0.41
4800
69
78
7.0
1.6E-02
800
1.9E-04
6.0
2.74
32.9
0.46
5600
75
84
8.0
1.8E-02
800
1.7E-04
6.0
3.04
36.5
0.51
6400
80
90
9.0
2.0E-02
800
1.5E-04
6.0
3.38
40.6
0.56
7200
85
96
10.0
2.2E-02
800
1.4E-04
6.0
3.76
45.1
0.63
8000
89
101
11.0
2.5E-02
800
1.2E-04
6.0
4.17
50.1
0.70
8800
94
106
12.0
2.7E-02
800
1.1E-04
6.0
4.64
55.6
0.77
9600
98
111
13.0
2.9E-02
800
9.9E-05
6.0
5.15
61.8
0.86
10400
102
115
14.0
3.1E-02
800
8.9E-05
6.0
5.72
68.7
0.95
11200
106
119
15.0
3.3E-02
800
8.0E-05
6.0
6.36
76.3
1.06
12000
110
124
16.0
3.6E-02
800
7.2E-05
6.0
7.07
84.8
1.18
12800
113
128
17.0
3.8C-02
800
6.SE-05
6.0
7.85
94.2
1.31
13600
117
132
18.0
4.0E-02
800
5.8E-05
6.0
8.73
104.7
1.45
14400
120
135
19.0
4.2E-02
800
5.3E-05
6.0
9.70
116.3
1.62
15200
123
139
20.0
4.5E-02
800
4.7E-05
6.0
10.77
129.3
1.80
16000
126
143
21.0
4.7E-02
800
4.3E-05
6.0
11.97
143.6
1.99
16800
130
146
22.0
4.9E-02
800
3.8E-05
6.0
13.30
159.6
2.22
17600
133
150
23.0
5.1E-02
800
3.4E-05
6.0
14.78
177.3
2.46
18400
136
153
24.0
5.3E-02
800
3.1E-05
6.0
16.42
197.0
2.74
19200
139
156
8
C
0
E
6
H
1
J
K
L
M
CELL FORMULAS
COLUMNS INITIAL VALUES (ROW K)
VARIABLE VALUES (ROW IS, ETC.)
• 1
C ~814*0.002228
0 ~IC14/B14
E 0.0003S
6 6
H ~C14*G14/(K14*(E14/30.48»
1 ~N14*12
J +C14/((E14/30.48)*K14)
K 800
L SSQRT(KH)
N !SQRT<4*K14/fPl)
~8K*S8$7
~115*0.002228
~K15/B15
~€15*SES7/100
~G15*SG$7
+C15*G15/(K15*(E15/30.48))
~H15*12
~C15/(
-------
"high flux rates". Flux rates associated with low flow velocities through substrate are defined as "low flux
rates". Numerically, the opposite is true. Thus, a flux rate of 400 sf/gpm is a high flux rate when compared
to a flux rate of 800 sf/gpm because the flow velocity associated with 400 sf/gpm is higher than the flow
velocity associated with 800 sf/gpm for an identically-sized wetland.
Flow flux rate adjusted by metals concentration may provide a benchmark criterion for wetland
performance with respect to sulfate reducing bacterial efficiency. However, the mechanism responsible
may not be metals concentration alone; it appears that the viability of sulfate reducing bacteria is sensitive
to substrate pH. pH values of less than 5 standard units cause stress in suifate-reducing bacteria (99).
High fluxes, especially those with acidic characteristics, may result in a dilution of more-neutral waters in
the substrate which are favorable to suifate-reducing bacteria and a reduction in metals reduction
efficiency.
Dissolved metals concentration in acidic drainage is closely related to pH. Typically, the lower the
pH of the water, the more metals in solution. Accordingly, some workers have focused on metals
concentration effects on wetland performance, as discussed in SECTION 7, in the subsection Area
Adjusted Loadings and Removal
Therefore, wetland performance from a chemical perspective may be a function of total metals
concentration/pH flux because sulfate reducing bacteria may be overwhelmed by the metals toxicity/pH
changes and not perform optimally. For example, in the Big Five Wetlands (99), sulfate reducing bacteria
performance appears to decrease when exposed to flux rates less than 800 sf/gpm at about 80 mg/liter
total dissolved metals.
From a physical/hydrologic perspective, the lower limit of flux rate will be a value necessary to
preclude substrate desiccation. The maximum achievable flux rate may be a function of the practical
working permeability of the substrate. For example, the nominal flux is inherently low for a substrate with a
tow permeability; this flux may satisfy wetland design criteria if the metals content is excessively high (with
accompanying low pH) in the feed water. For substrate with a high predicted permeability, flux may need
to be physically controlled by varying area and/or unit flow rates to the cell to satisfy chemical criteria as
described in the preceding paragraphs.
MASS LOADING METHOD
U.S. Bureau of Mines studies (65, 67) have shown that the reaction rate of suifate-reducing
bacteria may have a limiting effect on wetland performance. For the US Bureau of Mines Friendship Hill
wetland, a maximum sulfate reduction activity of nanomoles per cubic-centimeter per day (nmole/cc-day)
was found. According to the discussion in SECTION 7, the loading of metals delivered to the wetland
substrate should not exceed this value of 600 (nmole/cc-day). The Table spreadsheet "slideruie/
nomograph" can be used to select the size a particular "example" wetland using a particular metals loading
value, which is Qm in the Table.
12-4
-------
TABLE 31 ESTIMATE OF PRESSURE DROP AND METAL LOADING ACROSS
UPFLOW OR DOWNFLOW WETLAND CELLS OF VARIOUS DEPTHS
Delta*
(SEE NOTE 1)
0 gpn
100 X 0.1 ft
METAL LOADING
CRITERIA
0 aq-ft total
*»tal*->
100 ing/liter
Q OF K L dH dH i A
FLOW FLOW FLUX PERMEABLTY DEPTH HEADLOSS HOLOSS GRADIENT AREA
V Ln
VOLUME nanoanlea/
NOTES
V*
cfa
sf/gpa
ca/aec
feet
feet
Inches
ft/ft
»q-ft
cc'a
sua
day
wmrnm
(t-w/v)
MMMM
Initial value
1.0
2.2E-03
800
3.SE-04
1.0
0.24
2.9
0.24
800
2.3E+07
9.9E+09
437
1.0
2.2E-03
800
3.5E-04
1.1
0.27
3.2
0.24
800
2.56+07
9.9E+09
398
1.0
2.2E-03
800
3.5E-04
1.2
0.29
3.5
0.24
800
2.7E+07
9.9E+09
364
1.0
2.2E-03
800
3.5E-04
1.3
0.32
3.8
0.24
800
2.96+07
9.9E+09
336
1.0
2.2E-03
800
3.5E-04
1.4
0.34
4.1
0.24
800
3.26+07
9.9E+09
312
1.0
2.2E-03
800
3.5E-04
1.5
0.36
4.4
0.24
800
3.46+07
9.9E+09
292
THESE CONFIGS
1.0
2.2E-03
800
3.5E-04
1.6
0.39
4.7
0.24
800
3.6C+07
9.9E+09
273
WOULD WORK
1.0
2.2E-03
800
3.5E-04
1.7
0.41
4.9
0.24
800
3.9E+07
9.9E+09
257
BASED ON
1.0
2.2E-03
800
3.5E-04
1.8
0.44
5.2
0.24
800
4.16+07
9.9E+09
243
METAL LOADING
1.0
2.2E-03
800
3.SE-04
1.9
0.46
5.5
0.24
800
4.36+07
9.9E+09
230
1.0
2.2E-03
800
3.5E-04
2.0
0.49
5.8
0.24
800
4.5E+07
9.9E+09
219
1.0
2.2E-03
800
3.5E-04
2.1
0.51
6.1
0.24
800
4.8E+07
9.9E+09
208
1.0
2.2E-03
800
3.5E-04
2.2
0.53
6.4
0.24
800
5.0E+07
9.96+09
199
1.0
2.2E-03
800
3.51-04
2.3
0.56
6.7
0.24
800
5.21+07
9.9E+09
190
1.0
2.2E-03
800
3.5E-04
2.4
0.58
7.0
0.24
800
5.46+07
9.9E+09
182
1.0
2.2E-03
800
3.5E-04
2.5
0.61
7.3
0.24
800
5.76+07
9.96+09
175
1.0
2.2E-03
800
3.5E-04
2.6
0.63
7.6
0.24
800
5.96+07
9.9E+09
168
1.0
2.2E-03
800
3.5E-04
2.7
0.65
7.9
0.24
800
6.1E+07
9.9E+09
162
1.0
2.2E-03
800
3.5E-04
2.8
0.68
8.1
0.24
800
6.36+07
9.9E+09
156
1.0
2.2E-03
800
3.5E-04
2.9
0.70
8.4
0.24
800
6.6E+07
9.9E+09
151
1.0
2.2E-03
800
3.5E-04
3.0
0.73
8.7
0.24
800
6.8E+07
9.9E+09
146
TYPICAL
1.0
2.26-03
800
3.5E-04
3.1
0.75
9.0
0.24
800
7.0E+07
9.9E+09
141
DEPTHS
1.0
2.2E-03
800
3.5E-04
3.2
0.78
9.3
0.24
800
7.36+07
9.96+09
137
COLUMNS-->
8
¦c
0
E
6
H
I
J
K
M
M
0
CELL FORMULAS
PARAMETER COLUMNS INITIAL VALUES (ROU 14)
VARIABLE VALUES (ROU 15, ETC.)
COMMENTS:
Q.FLCU, gpa B 1
0.FLOU, efa C +814*0.002228
F, FLUX, sf/gpa D +C14/B14
K, ca/aec E 0.0003S
L, DEPTH, ft G 1
dH, HOLOSS,ft N +C14*G14/(K14*(E14/30.48))
dH, HOLOSS,in I +H14*12
1.GR0NT,ft/ft J +C14/((E14/30.48)*K14)
A,AREA,aq-ft K 800
V,VOLUME,cc'a M +K14*G14*1000/0.03531
La,LOADG,nm/d N +B14*54S0.4*SNS8/(S5*0.000001)
LOAD/VOLUME 0 +N14/M14
~S14+S8S7
~815*0.002228
-HC1S/81S
+€15*S€#7/100
~G15+WS7
+Cl5*G15/(K15*(E15/30.48))
+H15*12
+C15/( (E15/30.48)*1C15)
+K15+SKS7
+K15*G15*1000/0.03531
+815*5450.4*SNS8£(55*0.000001)
~N15/M15
CONSTANT Q,F,A.K
VARY L
OBSERVE Oa < 300
NOTE 1: DELTA VALUES ARE INCREMENTAL VALUES ADOED TO 08 MULTIPLIED BY THE INITIAL
VALUES TO GENERATE THE TABLE
12*5
-------
500
450
400
350
300
250
200
150
100
50
0
~
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
DEPTH, FEET
HEADLOSS, IN. x 10 + LOADING NMOLE/CC/D 300 NM/CC/DAY
Changes in loading and head losses as depth varies in substrate. The values plotted are
from Table 31.
-------
TABLE 32 MODIFICATION OF TABLE 31 TO DETERMINE MINIMUM
PERMEABILITY
Deltas — -->
0
gpM
90
(SEE NOTE 1)
•
•
Q
0
F
K
FLOW
FLOW
FLUX PERMEABLTY
NOTES
a*
cfs
sf/gpa
cat/sec
sass
vast
nu
mi
Initial valua
1.0
2.2E-03
800
3.5E-04
1.0
2.2E-03
800
3.2E-04
1.0
2.2E-03
800
2.8E-04
1.0
2.2E-03
800 '
2.6E-04
1.0
2.2E-03
800
2.3E-04
1.0
2.2E-03
800
2.1E-04
1.0
2.2E-03
800
1.9E-04
1.0
2.2E-03
800
1.7E-04
1.0
2.2E-03
800
1.5E-04
1.0
2.2E-03
800
1.4E-04
THESE
1.0
2.2E-03
800
1.2E-04
CONFIGURATNS
1.0
2.2E-03
800
1.1E-04
WOULD WORK
1.0
2.2E-03
800
9.9E-05
1.0
2.2E-03
800
8.9E-05
1.0
2.2E-03
800
8.0E-05
THESE
1.0
2.2E-03
800
7.2E-05
CONFIGURATNS
1.0
2.2E-03
800
6.5E-05
WOULD NOT
1.0
2.2E-03
800
5.8E-05
WORK,HEADLOSS
1.0
2.2E-03
800
5.3E-05
IS GREATER
1.0
2.2E-03
800
4.7E-05
THAN L
1.0
2.2E-03
800
4.3E-05
1.0
2.2E-03
800
3.8E-05
1.0
2.2E-03
800
3.4E-05
1.0
2.2E-03
800
3.1E-05
1.0
2.2E-03
800
2.8E-05
COLUMNS—>
8
C
D
E
0 ft
0 sq-ft
total
awtals-
METAL LOADING
CRITERIA
L dH dK ( a
DEPTH HEADLOSS HDLOSS GRADIENT AREA
feet fact inches ft/ft sq-ft
0.36
0.40
0.45
0.50
0.55
0.62
0.68
0.76
0.85
0.94
1.04
1.16
1.29
1.43
1.59
1.77
1.96
2.18
2.42
2.69
2.99
3.32
3.69
4.10
4.56
H
4.4
4.9
5.4
6.0
6.7
7.4
8.2
9.1
10.1
11.3
12.5
13.9
15.5
17.2
19.1
21.2
23.6
26.2
29.1
32.3
35.9
39.9
44.3
49.3
54.7
I
0.24
0.27
0.30
0.33
0.37
0.41
0.46
0.51
0.56
0.63
0.70
0.77
0.86
0.95
1.06
1.18
1.31
1.45
1.62
1.80
1.99
2.22
2.46
2.74
3.04
J
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
K
100 HQ/liter
V Lm
VOLUME nanomolas/
cc's day
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E-MJ7
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E*07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E-HJ7
3.4E+07
3.4E+07
3.4E-KJ7
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.46*07
N
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E*09
9.9E+09
9.9E+09
9.96*09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
M
On
(L«/V)
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
CELL FORMULAS
PARAMETER COLUMNS INITIAL VALUES (ROU 14)
0.FLOW, flpa B 1
Q.FLCW, cfs C 14*0.002228
F,FLUX,sf/gpa D *<14/814
K, aa/uc E 0.00035
L, DEPTH, ft G 1.5
dH, HOLOSS.ft H ~C14*G14/(K14*
-------
M
09
3.2
3.0
2.8
£
2.6
2.4
2.2
'<*
2.0
z
1.8
*
1.6
1.4
•s
1.2
£
1.0
$
£
*
0.8
0.6
0.4
0.2
wAA/Wvv v
0.00002 0.00006 0.0001 0.00014 0.00018 0.00022 0.00026 0.0003 0.00034
PERMEABILITY, CM/SEC
HYDRAULIC GRADIENT 7 i=1, PONDING AT i>1
Figure 55. Changes in hydraulic gradient with varying permeability. The values plotted are from
Table 32.
-------
TABLE 33 MODIFICATION OF TABLE 31 TO A DEPTH OF 3 FEET
Deltas >
(SEE NOTE 1)
NOTES
Initial valua
THESE
CONFIGURATNS
WOULD WORK
THESE
CONFIGURATNS
WOULD NOT
UORK.HEADLOSS
IS GREATER
THAN L
0 gps
90 X 0 ft
0 *q-ft
METAL LOADING
CRITERIA
total
mtals->
Q
FLOW
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
COLUMNS--*
Q
FLOW
Cft
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-0S
2.2E-03
2.2E-03
2.2E-QS
2.2E-03
2.2E-03
2.2E-03
2.2E-03
2.2E-03
C
F K
FLUX PERMEABLTY
af/gpn ca/sac
400
400 '
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
L dN dH i A
DEPTH HEADLOSS HOLOSS GRADIENT AREA
faat fe«t inches ft/ft sq-ft
3.5E-04
3.2E-04
2.8E-04
2.66-04
2.3E-04
2.1E-04
1.9E-04
1.7E-04
1.5E-04
1.4E-04
1.2E-04
1.1E-04
9.9E-05
8.9E-05
8.0E-05
7.2E-05
6.5E-05
5.8E-05
5.3E-05
4.7E-05
4.3E-05
3.8E-05
3.4E-05
3.1E-05
2.8E-0S
E
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
G
CELL FORMULAS
PARAMETER COLUMNS INITIAL VALUES (ROW 14)
Q.FLOkf, gpm I 1
0.FLOW, eft C 4014*0.002228
F,FLUX,sf/gpa 0 -HC14/B14
K, cai/sae E 0.00035
L, DEPTH, ft G 3
dH, HOLOSS,ft H *C14*G14/(K14*(E14/30.48))
dH, HOLOSS,in I +H14*12
1.GRONT,ft/ft J ~C14/((E14/30.48)*K14>
A,AREA,sq-ft K 400
V,VOLUME,ce's M +K14*G14*1000/0.03531
LB.LOAOG.rWd N +B14*5450.4*SNS8/(55*0.000001)
LOAD/VOLUME 0 +N14/M14
1.46
1.62
1.80
2.00
2.22
2.46
2.74
3.04
3.38
3.76
4.17
4.64
5.15
5.72
6.36
7.07
7.85
8.73
9.70
10.77
11.97
13.30
14.78
16.42
18.24
H
17.5
19.4
21.6
24.0
26.6
29.6
32.9
36.5
40.6
45.1
50.1
55.6
61.8
68.7
76.3
84.8
94.2
104.7
116.3
129.3
143.6
159.6
177.3
197.0
218.9
I
0.49
0.54
0.60
0.67
0.74
0.82
0.91
1.01
1.13
1.25
1.39
1.55
1.72
1.91
2.12
2.36
2.62
2.91
3.23
3.59
3.99
4.43
4.93
5.47
6.08
J
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
K
VARIABLE VALUES (ROW 15, ETC.)
+B14*S8$7
+815*0.002228
•HC15/B15
~E15*SE$7/100
~G15+86S7
+C15*G15/(K15*(E15/30.48)}
~H15*12
+C15/((E15/30.48)*K15)
~K15+SW7
<*15*615*1000/0.03531
~815*5450.4*$NS8/<55*0.000001)
~N15/M15
100 mg/liter
VOLUME nanonolas/
ce's day
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
3.4E+07
H
COMMENTS:
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E-KJ9
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
9.9E+09
N
CONSTANT Q, On, V
VARY K,
HALVE AREA, A
DOUBLE DEPTH, L
OBSERVE i
-------
TABLE 34 MODIFICATION OF TABLES 31 AND 33 ALLOWING
FLOW AND FLUX TO VARY
Deltas »
1
9P*
100 X 0
ft
400
(SEE NOTE 1)
•
•
•
•
a
Q
F
K
L
dH
dH
i
A
FLOW
FLOW
FLUX PERMEA8LTY
D6PTH
HEADLOSS
HOLOSS
GRADIENT
AREA
NOTES
OP"*
cfa
af/gpM
en/sac
fe«t
feet
inches
ft/ft
aq-ft
ttrt
sm
¦ ¦¦¦
IT
mmmu
ma
mmmm
UM
asra
Initial value
20.0
4.5E-02
400
1.9E-04
3.0
2.68
32.2
0.89
8000
21.0
4.7E-02
400
1.96-04
3.0
2.68
32.2
0.89
8400
22.0
4.96-02
400
1.96-04
3.0
2.68
32.2
0.89
880Q
ALL
23.0
5.16-02
400
1.9E-04
3.0
2.68
32.2
0.89
9200
CONFIGURATNS
24.0
3.36*02
400
1.96-04
3.0
2.68
32.2
0.89
9600
UOULD WORK
2S.0
5.66-02
400
1.96-04
3.0
2.68
32.2
0.89
10COO
26.0
5.86-02
400
1.9E-04
3.0
2.68
32.2
0.89
10400
27.0
6.06-02
400
1.9E-04
3.0
2.68
32.2
0.89
10800
28.0
6.2E-02
400
1.9E-04
3.0
2.68
32.2
0.89
11200
29.0
6.SE-02
400
1.96-04
3.0
2.68
32.2
0.89
11600
30.0
6.76-02
400
1.96-04
3.0
2.68
32.2
0.89
12000
31.0
6.96-02
400
1.96-04
3.0
2.68
32.2
0.89
12400
32.0
7.1E-02
400
1.96-04
3.0
2.68
32.2
0.89
12800
33.0
7.46-02
400
1.9E-04
3.0
2.68
32.2
0.89
13200
34.0
7.6E-02
400
1.96-04
3.0
2.68
32.2
0.89
13600
3S.0
7.86-02
400
1.96-04
3.0
2.68
32.2
0.89
14000
36.0
8.06-02
400
1.96-04
3.0
2.68
32.2
0.89
14400
37.0
8.26-02
400
1.96-04
3.0
2.68
32.2
0.89
14800
UPPER LIMIT
38.0
8.5E-02
400
1.96-04
3.0
2.68
32.2
0.89
15200
DESIGN FLOWS
39.0
8.7E-02
400
1.96-04
3.0
2.68
32.2
0.89
15600
•
•
I
<
•
•
1
•
•
«
•
V
40.8
8.96*02
400
1.96-04
3.0
2.68
32.2
0.89
16000
41.0
9.1E-02
400
1.96-04
3.0
2.68
32.2
0.89
16400
42.0
9.46-02
400
1.96-04
3.0
2.68
32.2
0.89
16800
43.0
9.66-02
400
1.96-04
3.0
2.68
32.2
0.89
17200
44.0
9.8E-02
400
1.96-04
3.0
2.68
32.2
0.89
17600
COLUMNS—>
B
C
D
6
G
H
I
J
K
CELL FORMULAS
PARAMETER COLUMNS
INITIAL VALUES (ROW 14)
VARIABLE VALUES
(ROU 15,
ETC.)
O.FLOU. gpa
B
20
+B14+SBS7
O.FLOU, cfa
C
+814*0.002228
+815*0.002228
F,FLUX,*f/9pa
D
+K14/B14
+K15/B15
K, cm/fc
E
0.00019
+615*SE*7/100
L, OEPTH, ft
G
3
~G15+SGS7
dH, HOLOSS,ft
H
+C14*G14/(IC14*(614/30.48) )
+C15*G15/)
dH, NOLOSS,in
I
+N14*12
+H15*12
i.GRONT,ft/ft
J
+C14/((E14/30.48)*K14)
+C15/<(615/30.4S)*K15)
A,AREA,sq-ft
K
400
+K15+SW7
V, VOLUME,CC'S
N
+K14*Gl4*1000/0.03531
+K15*G15"1000/0.03531
La,LQADG,na/d
N
+B14*5450.4*SNS8/(55*0.000001)
+815*5450.4*SNS8/(55*0.000001)
LOAD/VOLUME
0
+N14/H14
~N15/M15
total
metals->
METAL LOADING
CRITERIA
100 ag/litar
V Ln
VOLUME nanomoles/ On
cc's day (Lw/V)
6.SE+08
7.1E+08
7.5E+08
7.86+08
8.2E+08
8.5E+08
8.8E+08
9.2E+08
9.5E+08
9-96+08
1.06+09
1.16+09
1.1E+09
1.1E+09
1.2E+09
1.2E+09
1.26+09
1.36+09
1.36+09
1.3E+09
1.46+09
1.46+09
1.46+09
1.56+09
1.56+09
M
COMMENTS:
2.0E+11
2.1E+11
2.2E+11
2.3E+11
2.4E+11
2.5E+11
2.6E+11
2.7E+11
2.86+11
2.96+11
3.0E+11
3.1E+11
3.2E+11
3.3E+11
3.46+11
3.5E+11
3.66+11
3.7E+11
3.8E+11
3.9E+11
4.0E+11
4.1E+11
4.2E+11
4.3E+11
4.4E+11
H
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
292
CONSTANT F,Qn,LtKadH.i
VARY Q I A
OBSERVE ALL
CRITERIA SATISFIED
NOTE 1: DELTA VALUES ARE INCREMENTAL VALUES AOOED TO OR MULTIPLIED BY THE INITIAL
VALUES TO GENERATE THE TABLE
USER INPUT INITIAL AND DELTA VALUES
12-10
-------
19
18
17
16
15
14
13
12
11
10
9
8
56.
44
40
28
32
36
24
20
FLOW, GALLONS PER MINUTE
Changes in flow with area keeping all other variables constant. The values plotted are
from Table 34.
-------
TABLE 35 MODIFICATION OF TABLE 34 TO ALLOW FOR VARIATION
OF THE PERMEABILITY WITH DEPTH
METAL LOADING
CRITERIA
Deltas >
(SEE NOTE 1)
NOTES
Initial valua
0 9P»
99 X 0.1 ft
0 sq-ft
WATER PONDING
AT L-4.1 FEET
total
iMtalo
100 mg/liter
COLUMNS-->
*
•
•
•
Q
0
F
K
L
dH
dH
i
A
V
La
FLOW
FLOW
FLUX PERMEASLTY
DEPTH
HEADLOSS
HDL0SS GRADIENT
AREA
VOLUME (WWKlM/
Qa
9P"
cfs
•f/gpa
cot/sac
faat
faat
Irtchaa
ft/ft
•q-ft
ec's
day
(L m/V)
ssas
mamm
mmmm
mrnmrn
nsa
ana
rata
aan
wmwm
ammm
»»»»
aaaa
20.0
4.5E-02
400
1.9E-04
3.0
2.68
32.2
0.89
8000
6.8E+08
2.0E+11
292
20.0
4.5E-02
400
1.9E-04
3.1
2.80
33.6
0.90
8000
7.0E+08
2.0E*11
282
20.0
4.5E-02
400
1.9E-04
3.2
2.92
35.0
0.91
8000
7.3E+08
2.0E*11
273
20.0
4.5E-02
400
1.86-04
3.3
3.04
36.5
0.92
8000
7.5E+08
2.0E*11
265
20.0
4.5E-02
400
1.8E-04
3.4
3.16
38.0
0.93
8000
7.7E*08
2.0E*11
257
20.0
4.5E-02
400
1.8E-04
3.5
3.29
39.5
0.94
8000
7.96+08
2.06+11
250
20.0
4.5E-02
400
1.8E-04
3.6
3.42
41.0
0.95
8000
8.2E+08
2.0E+11
243
20.0
4.5E-02
400
1.8E-04
3.7
3.55
42.6
0.96
8000
8.4E+08
2.0E+11
236
20.0
4.SE-02
400
1.0E-04
3.8
3.68
44.2
0.97
8000
8.6E+08
2.0E*11
230
20.0
4.5E-02
400
1.7E-04
3.9
3.81
45.8
0.98
8000
8.8E+08
2.0E+11
224
20.0
4.SE-02
400
1.7E-04
4.0
3.9S
47.4
0.99
8000
9.1E+08
2.06*11
219
20.0
4.SE-02
400
1.78-04
4.1
4.09
49.1
1.00
8000
9.3E+08
2.0E+11
213
20.0
4.5E-02
400
1.7E-04
4.2
4.23
50.8
1.01
8000
9.5E+08
2.06+11
208
20.0.
4.SE-02
400
1.7E-04
4.3
4.38
52.5
1.02
8000
9.7E+08
2.0E*11
203
20.0
4.5E-02
400
1.7E-04
4.4
4.53
54.3
1.03
8000
i.oe*09
2.06*11
199
20.0
4.5E-02
400
1.6E-04
4.5
4.68
56.1
1.04
8000
1.0E+09
2.0E+11
194
20.0
4.5E-02
400
1.6E-04
4.6
4.83
57.9
1.05
8000
1.06*09
2.0E+11
190
20.0
4.5E-02
400
1.68-04
4.7
4.98
59.8
1.06
8000
1.1E*09
2.08*11
186
20.0
4.5E-02
400
1.6E-04
4.8
5.14
61.7
1.07
8000
1.1E+09
2.0E+11
182
20.0
4.SE-02
400
1.6E-04
4.9
5.30
63.6
1.08
8000
1.1E*09
2.0E+11
179
20.0
4.5E-02
400
1.66-04
5.0
5.46
65.5
1.09
8000
1.1E+09
2.0E+11
175
20.0
4.5E-02
400
1.5E-04
5.1
5.63
67.5
1.10
8000
1.2E+09
2.0E+11
172
20.0
4.51-02
400
1.5E-04
5.2
5.80
69.6
1.11
8000
. 1.2E+09
2.0E+11
168
20.0
4.5E-02
400
1.5E-04
5.3
5.97
71.6
1.13
8000
1.2E+09
2.0E+11
165
20.0
4.5E-02
400
1.5E-04
5.4
6.14
73.7
1.14
8000
1.2E+09
2.06+11
162
B
C
0
E
G
N
1
J
K
N
N
0
CELL FORMULAS
PARAMETER COLUMNS INITIAL VALUES (ROU 14)
VARIABLE VALUES (ROU 1S, ETC.)
COMMENTS:
0.FL0U. gpa
B
20
~S14+S8S7
0,FLOW, eft
C
~•14*0.002228
~•15*0.002228
F.FLUX.sf/gpa
0
~<14/114
~K15/B1S
K, ca/sac
E
0.00019
+€15**£S7/100
L, DEPTH, ft
6
3
~G15*SC$7
dH, NOLOSS.ft
H
+C14*G14/(K14*(E14/30.48))
~C15*G15/(K15*(E15/30.48))
dH, HOLOSS.in
I
~H14*12
~H15*12
i.GRONT,ft/ft
J
~C14/((E14/30.48)*K14)
~C15/((E15/30.48)*K15)
A,AREA,iq-ft
K
8000
~IC15+S»7
V,VOLUM6.CC'*
M
+K14*G14*1000/0.03531
~K15*G15*1000/0.03531
L*,L0A0G,rw/d
N
~814*5450.4*SNS8/<55*0.000001)
~815*5450.4*SNS8/(55*0.000001)
LOAD/VOLUME
0
~N14/M14
~N15/M15
NOTE 1: 0ELTA
VALUES ARE INCREMENTAL VALUES ADDED TO OR MULTIPLIED BY THE INITIAL
VALUES TO GENERATE THE TABLE
12-12
• a USER INPUT
INITIAL AN0 DELTA VALUES
VARY K, 1i LOSS FROM
COMPRESSION PER
ADOTNL 0.1 FT OF L
-------
0.8 ' 1 1 1 I I I 1 ' ' ' I I
3 3.4 3.8 4.2 4.6 5 5.4
SUBSTRATE DEPTH, FT
~ PERMEABILITY X 1E4 + HYDRLC GRADIENT, i i-1, PONDING AT i>1
Figure 57. Decrease in permeability by one percent for each 0.1 feel of depth of substrate. When
the hydraulic gradient is greater than one, ponding on the surface occurs. The values
plotted are from Table 35.
-------
TABLE 36 MODIFICATION OF TABLE 35 TO ALLOW FOR CHANGES IN
FLOW AND PERMEABILITY WITH VARYING DEPTH
METAL LOADING
CRITERIA
Delta* »
(SEE NOTE 1)
-0.1
99 X 0.1 ft
0 aq-ft
total
ratals-
100 mg/liter
•
•
•
•
a
0
F
K
L
dH
dH
1
A
V
FLOU
FLOU
FLUX PERMEABLTY
DEPTH HEADLOSS HOLOSS
GRADIENT
AREA
VOLUME nanomolts/
Qm
NOTES
gpm
cfa
af/gpa
c*/s«c
fa«t
fMt
inches
ft/ft
aq-ft
cc'a
day
CLlVV)
mm
mmmm
mmmm
mmmm
on
¦m
mmmm
sua
«»«•
mmmm
BIBS
•BBS
BBSS
Initial valua
20.0
4.5E-02
400
1.9E-04
3.0
2.68
32.2
0.89
8000
6.8E+08
2.06+11
292
19.9
4.4E-02
402
1.9E-04
3.1
2.78
33.4
0.90
8000
7.06+08
2.0E+11
281
19.8
4.4E-02
404
1.9E-04
3.2
2.89
34.7
0.90
8000
7.36+08
2.0E+11
271
19.7
4.4E-02
406
1.8E-04
3.3
2.99
35.9
0.91
8000
7.5E+08
2.06+11
261
19.6
4.4E-02
408
1.8E-04
3.4
3.10
37.2
0.91
8000
7.76+08
1.96+11
2S2
19.5
4.3E-02
410
1.BE-04
3.5
3.21
38.5
0.92
8000
7.96+08
1.96+11
244
19.4
4.3E-02
412
1.8E-04
3.6
3.31
39.8
0.92
8000
8.2E+08
1.96+11
236
19.3
4.3E-02
415
1.BE-04
3.7
3.42
41.1
0.93
8000
8.4E+08
!.96+11
228
19.2
4.3E-02
417
1.6E-04
3.8
3.53
42.4
0.93
8000
8.6E+08
1.96+11
221
19.1
4.3E-02
419
1.7E-04
3.9
3.64
43.7
0.93
8000
8.86*08
1.96+11
214
19.0
4.2E-02
421
1.7E-04
4.0
3.75
45.1
0.94
8000
9.1E+08
1.96+11
208
18.9
4.2E-02
423
1.7E-04
4.1
3.87
46.4
0.94
8000
9.3E+08
1.9E+11
202
18.8
4.2E-02
426
1.7E-04
4.2
3.98
47.8
0.95
8000
9.5E+08
1.9E+11
196
18.7
4.2E-02
428
1.7E-04
4.3
4.09
49.1
0.95
8000
9.7E+08
1.96+11
190
18.6
4.1E-02
430
1.7E-04
4.4
4.21
50.5
0.96
8000
1.06+09
1.86+11
185
18.5
4.1E-02
432
1.6C-04
4.5
4.32
51.9
0.96
8000
1.06+09
1.8E+11
180
18.4
4.1E-02
43S
1.6C-04
4.6
4.44
53.3
0.97
8000
1.0E+O9
1.86+11
175
18.3
4.1E-0S
437
1.6E-04
4.7
4.56
54.7
0.97
8000
1.1E+09
1.86+11
170
18.2
4.1E-02
440
1.6E-04
4.8
4.68
56.1
0.97
8000
1.1E+09
1.8E+11
166
18.1
4.0E-02
442
1.6E-04
4.9
4.80
57.6
0.98
8000
1.1E+09
1.8E+11
162
18.0
4.0E-02
444
1.6E-04
5.0
4.92
59.0
0.96
8000
1.1E+09
1.8E+11
157
WATER PONDING
17.9
4.0E-02
447
1.5E-04
5.1
S.04
60.4
0.99
8000
1.2E+09
1.8E+11
154
AT L"5.4 FEET
17.8
4.0E-O2
449
1.5E-04
5.2
5.16
61.9
0.99
8000
1.26+09
1.86+11
150
0»17.6 GPM
17.7
3.9E-02
452
1.5E-04
5.3
S.28
63.4
1.00
8000
1.26+09
1.8E+11
146
17.6
3.9E-02
455
1.5E-04
5.4
5.40
64.9
1.00
8000
1.2E+09
1.76+11
143
COLUMNS--*
B
C
D
E
6
N
I
J
K
n
N
0
CELL FORMULAS
PARAMETER COLUMNS INITIAL VALUES (ROU U>
0.flou. arm
Q,R0W, eft
F,FLUX,«f/gpa
K, cm/Mc
1. DEPTH. ft
dM, NOLOSS,ft
dH, NOLOSS,in
l,GXONT,ft/ft
A,AREA,sq-ft
v.vocuMe.cc't
Ua.LQADG.rm/d
LOAD/VOLUME
•
C
0
E
6
K
1
J
X
M
N
0
20
<414*0.002228
+K14/B14
0.00019
3
+C14*G14/(K14*(E14/30.48))
~M14*12
+C14/((E14/30.48)*K14)
8000
+K14*G14*1000/0.03531
~¦14*5450.4*SNS8/(55*0.000001)
~N14/M14
VARIABLE VALUES (ROU 15, ETC.)
+B14+S8S7
~•15*0.002228
+K13/B15
+615*8687/100
+G15+SGS7
+C15*G15/)
~H15*12
+C15/((E15/30.48)*K15>
~K15+SKS7
~K15*S1S*1000/0.03531
~815*5450.4*SNS8/(55*0.000001>
~N15/M1S
COMMENTS:
NOTE 1: OELTA VALUES ARE INCREMENTAL VALUES ADDED TO OR WLTIPLIEO ST THE INITIAL
VALUES TO GENERATE THE TABLE
12-14
* a USER INPUT INITIAL AND OELTA VALUES
VA*r K. IX LOSS
FROM CGMPRSH PER
0.1 FT OF L,
DECREASE FLOU
-------
300
290
280
270
260
250
240
230
220
210
200
190
180
170
160
150
140
130
120
no
100
~
Figure
3.4
3.8
4.2
4.6
5.4
SUBSTRATE DEPTH, FEET
FLOW, GPM X 100 + LOADING, nm/cc/day O PERMEABILITY X 1E6
58.
Changes in flow, permeability, and loading with increasing depth. Note the restriction that
the hydraulic gradient has to be £ 1. The values plotted are from Table 36.
-------
The key input/operating parameters for the example site are:
o 100 mg/liter of dissolved metals in feed water stream
(iron/lead/copper/zinc/cadmium/manganese)
o 20 to 40 gallons per minute of flow
o Substrate initial permeability of 3.5x10-* cm/sec
o Flux may vary from 400 to 800 sf/gpm
o Depth may vary from 3 to 6 feet based on topography
o Site topography is not constraining (area available is large)
o 300 nmoles/cc-day is the maximum mass loading rate - assume that metals average 55 grams/gram-
mole (a conservative assumption). This is related to mg/liter of metal in the drainage.
o To preserve anaerobic conditions, no ponding on the surface of the wetland is allowed; i.e.,
hydraulic gradient is less than 1.0
o For each 0.1 foot of depth increase above 3 feet, permeability drops by one percent.
In Table 31, the flow rate, flux, area (by definition) and permeability are held constant; depth is
varied to observed the configurations required to provide less than 300 nmoles/cc-day of loading (Qm) in
the last column). The spreadsheet assumes 100 percent removal of metals, which should be verified in
lab or field scale tests. Figure 54 graphically presents some of the variables changing in Table 31. Table
31 shows that depths greater than 1.5 feet would satisfy this criteria. The Table indicates that the depth
range of 3 to 6 feet, as chosen for our example wetland above, and labeled "Typical Depths" in the first
column satisfies the loading rate criteria.
In Table 32, the flow rate, flux, area, depth, and mass loading rate are held constant and
permeability is varied to determine the lower bound of permeability values that would yield a gradient of
less than 1.0. Figure 55 graphically presents some of the variables changing in Table 32. Table 32 shows
that a substrate with a K value of less than 8.9 x 10~5 cm/sec would produce ponding with a substrate
thickness (depth) of 1.5 feet (that is headloss is greater than depth).
In Table 33, the flow rate, mass loading and wetland volume are again held constant; but the
depth is doubled to 3 feet. The mass loading is kept constant by halving the surface area to 400 square
feet. Permeability is again varied to determine the lower bound of permeability values that would yield a
gradient of less than 1.0. Table 33 shows that a substrate with a K value of less than 1.9 x 10*4 cm/sec
would produce ponding.
In Table 34, flux, miss loading rate, depth, permeability (at the lowest acceptable value) and
gradient are held constant; flow rate and surface area are varied. Figure 56 graphically presents some of
the variables changing in Table 34. Table 34 shows that for flows of 20 gpm to 44 gpm, wetland areas from
8,000 to 16,000 square feet, respectively, would be required.
12-16
-------
In Table 35, flow rate, flux and surface area are held constant with values consist with a 20 gallon
per minute cell. Figure 57 graphically presents some of the variables changing in Table 35. Depth and
permeability are varied to show the one percent decrease in substrate permeability for every 0.1 foot of
depth. Table 35 shows that water ponds on the surface at a depth of 4.1 feet when the permeability of the
substrate reaches 1.7 x 10~4 cm/sec (gradient equals 1.0). The mass loading rate proportionately
decreases with the increase in substrate depth and volume.
In Table 36, the conditions of Table 35 are used and the flow rate is incrementally lowered to
observe what depth increases might be gained in this situation. Figure 58 graphically presents some of
the variables changing in Table 36. Table 36 shows that if the flow could be reduced to 17.6 gpm, the
wetland depth could be increased to 5.4 feet. The design example could be carried further by increasing
the ceil area to carry the required minimum 20 gallons per minute and verifying that all design parameters
are satisfied or optimized.
Tables 31 through 36 and Figures 54 through 58 illustrate the utility of the Lotus 123 ™
spreadsheet format for estimating wetland configurations. However, other computer codes that are more
user-interactive may be preferable to some designers. The purpose of the illustration was to show a
comprehensive design methodology that incorporates hydrology and substrate geochemistry.
VOLUMETRIC LOADING METHOD
Bacterial action results in the precipitation of metal sulfide minerals. Consequently, the pore/void
spaces within the wetland substrate will become filled. Lemke (9) estimated that void space in mushroom
compost accounted for approximately 25 percent of the substrate volume. This design method assumes
that the filling of void spaces within the substrate is a key limiting factor to wetland performance as a
acid/metal drainage treatment system.
This method allows the estimation of the ultimate size of a wetland (substrate volume) based on
the following assumptions:
o The substrate has a lab- or field-measurable void ratio of which a significant percentage is available
for the precipitation of metal sulfides,
o An average daily mineral loading rate can be estimated in terms of mass and volume of metals
sulfides.
Table 37 is an example spreadsheet that employs this substrate volume sizing methodology.
Mass loading rates from water balance and metals concentration perspectives were assumed to be
constant over the life of the wetland.
Note that the Table 37 spreadsheet includes an estimate of the concentrations of key metals
accumulated in the substrate at various design lives. These data are useful in determining temporal points
after which the substrate might be considered a hazardous waste. Since final metals concentration
12-17
-------
TABLE 37 PROJECTED WETLAND VOLUME REQUIREMENTS BASED ON VOID SPACE
AVAILABILITY FOR METAL SULFIDE AVAILABILITY PRECIPITATE FORMATION
METAL LOWING RATES
IRON 65 Ng/L1t*r
COPPER 6 Mg/Llt«r
OTHER HEAVY METALS 4 Hg/Lft«r
TOTAl LOADING 75 Hfl/Llt«r
SOURCE FEED RATE
50 GPN
15 X VOIDS
PERCENT VOIDS
AVAILABLE FOR
METAL SULFIDE
PRECIPITATION
MINERAL LOADING RATES
FtS
COVELLITE, CuS
OTHER SULFIDES (XxS)
TOTAL LOADING
0.27 CU-FT/DAT
0.02 CU-FT/DAT
0.01 CU-FT/DAY
0.30 CU-FT/OAY
DRY SUBSTRATE
TOTAL
CONCENTRATIONS BY WEIGHT
WETLAND
SUBSTRATE WETLAND
UETLAND
VOID
LIFE OF
LIFE OF
UETLANO
AT DESIGN
LIFE
AREA
DEPTH
VOLUME
VOLUME
VOLUME
UETLAND
UETLAND
DRY MASS
IRON COPPER
OTHER
(ACRES)
(FT)
(CU-FT)
(CU-YDS)
(CU-FT)
(DAYS)
(YEARS)
(KG)
(X)
(X)
(X)
0.25
3
32670
1210
4901
16515
45
627,264
46.6X
4.3X
2.9X
0.25
4
43560
1613
6534
22020
60
836,352
46.6X
4.3X
2.9X
0.25
5
54450
2017
8168
27525
75
1,045,440
46.6X
4.3X
2.9X
0.25
6
65340
2420
9801
33030
90
1,254,528
46.6X
4.3X
2.9X
0.5
3
65340
2420
9801
33030
90
1,254,528
46.6X
4.3X
2.9X
0.5
4
87120
3227
13068
44040
121
1.672,704
46.6X
4.3X
2.9*
0.5
5
108900
4033
16335
55050
151
2,090,880
46.6X
4.3X
2.9X
0.5
6
130680
4840
19602
66060
181
2,509,056
46.6X
4.3X
2.9X
1
3
130680
4840
19602
66060
181
2,509,056
46.6X
4.3X
2.9X
1
4
174240
6453
26136
88080
241
3,345,408
46.6X
4.3X
2.9X
1
5
217800
8067
32670
110099
302
4,181,760
46.6X
4.3X
2.9X
1
6
261360
9680
39204
132119
362
5,018,112
46.6X
4.3X
2.9X
2
3
261360
9680
39204
132119
362
5,018,112
46.6X
4.3X
2.9X
2
4
348480
12907
52272
176159
483
6,690,816
46.6X
4.3X
2.9X
2
5
435600
16133
65340
220199
603
8,363,520
46.6X
4.3X
2.9%
2
6
522720
19360
78408
264239
724
10,036,224
46.6X
4.3X
2.9X
A
B
C
D
E
F
G
H
I
J
K
CELL FORMULAS
PARAMETER COLUMNS
AREA, ACRES A
DEPTH, FEET >
VOL, CU-FT C
VOL, CU-YDS D
VOID VOL, CU-FT E
UETLAND LIFE, DAYS F
UETLANO LIFE, YEARS G
WETLAND DRY MASS, Kg H
ULTIMATE f CONC., X I
ULTIMATE CU CONC., X J
ULTIMATE Other CONC., X K
INITIAL VALUES (ROW 24)
0.25
3
~A24*43560*B24
+C24/27
~SJS10*C24/100
~£24/SOS17
~F24/365
+C24*19.2 (SEE NOTE 1)
(*SCS8*SFS9*1440*3.785*SF24/SH24)/1000000
(+SCS9*$FS9*1440*3.785*SF24/SN24)/1000000
(~SCS10*$FS9*1440*3.785*SF24/SH24)/1000000
NOTE 1: (AFTER LEHKE, 1989) WETLANO
SUBSTRATE BULK DENSITY - 1.28 g/cc; S.G.
SOLIDS OF SUBSTRATE * 1.7 YIELDS 53X
DRY SOLIDS BY UEIGNT AND DRY DENSITY OF
19.2 Kg/CF
12-18
-------
in the substrate is a function of void space, all the configurations in Table 37 have identical metals
concentrations; "total" metals amounts would be different for each unique wetland volume/life span.
To determine wetland "life" (at a given substrate depth and surface area) whereby the metal-laden
substrate would not be considered a hazardous waste, void space is incrementally varied until the
hazardous material criteria are met. If the void space percentage required to make the material
"hazardous" is greater than the maximum void space physically available, the life of the wetland will
probably not be governed by hazardous material criteria.
VOLUMETRIC BIOMASS ACCUMULATION METHOD
Wetland substrate mass/volume may be incrementally increased through accumulation of dead
vegetation. Biomass accumulation rates are a function of climate. For example, researchers have
estimated that biomass accumulation rates in tropical primeval coal swamps approached one foot every 10
years (105). In "forest mires of the temperate zone" (100), biomass accumulations have been measured
to be on the order of one foot every 300 years.
There is a dearth of data on biomass accumulations in constructed wetland facilities for metal mine
drainage treatment in typical temperate climates. However, assuming that new substrate from vegetation
is added to the wetland at a rate of one foot every 100 years, the availability of void space for metal-
hydroxide or sulfide precipitation in the wetland may become self-perpetuating. As shown on Table 38,
the self-perpetuating threshold design area for a flow of 50 gpm appears to be about 1.275 acres,
whereby additional operating life due to bioaccumulation equals the original life of the wetland.
It is assumed that the added substrate from the vegetation die-off would have the same
characteristics as the original substrate. It is also assumed that there is allowance for increasing thickness
of the constructed wetland system. The extension on operating life would occur only if the submergence
of the biomass was maintained. That is, the hydraulic level controls would have to be adjusted
infrequently to allow for increased water depth.
The 300 year design life for a three-foot deep wetland in Table 38, however, is probably
unrealistic. Biomass accumulation rates significantly greater than one foot every 100 years would have to
be achieved in order for a wetland designed with respect to other criteria to be self perpetuating from a
volumetric perspective.
As discussed below, typical wetland substrate organic content may need to be maintained to
satisfy stoichiometric criteria. Even with plant accumulations, maintenance actions such as periodic
additions of beneficial materials like hay or other organic-rich supplements may be required. Organic
addition is likely to be a site-specific maintenance consideration.
12-19
-------
TABLE 38 EFFECT OF BIOMASS ACCUMULATION AND SULFIDE PRECIPITATION
AS SOURCES AND SINKS OF VOID SPACE ON WETLAND CELL DESIGN LIFE
METAL LOADING RATES
IRON 65 Mg/Liter
COPPER 6 Mg/Lit«r
OTHER HEAVY METALS 4 Mfl/Lit*r
TOTAL LOADING 75 Hfl/Llt«r
SOURCE FEED
RATE
50 GPM
15 X VOIDS AVAILABLE FOR METAL SULFIDE PRCN
MINERAL LOADING RATES
FaS
COVELLITE, CuS
OTHER SULFIDES (XxS)
TOTAL LOADING
0.20 CU-FT/DAT
0.02 CU-FT/DAY
0.01 CU-FT/DAY
0.23 CU-FT/DAY
UETLAND SUBSTRATE WETLAND UETLAND VOID
(1)
LIFE OF LIFE OF ADOITNL ADDITNL ADDTNL ADDITNL TOTAL
AREA DEPTH
VOLUME
VOLUME
VOLUME
WETLANO
UETLAND
DEPTH
BIOMASS VOID VOL LIFE
LIFE
(ACRES) (FT)
(CU-FT)
(CU-YDS)
(CU-FT)
(DAYS)
(YEARS)
(FT)
(CU-FT)
(CU-FT)
(YRS)
(YRS)
¦¦¦¦¦¦ siimss
¦¦¦¦¦¦¦
mmammmmm
manmwmmm
nan
aras
mm
umumm
¦ssaaai
0.25
3
32670
1210
4901
21435
59
0.59
6395
959
11
70
0.2S
4
43560
1613
6534
28580
78
0.78
8527
1279
15
94
0.25
5
54450
2017
8168
35725
98
0.98
10659
1599
19
117
0.25
6
65340
2420
9801
42870
117
1.17
12790
1919
23
140
0.75
3
96010
3630
14702
64304
176
1.76
57557
8634
103
280
0.75
4
130680
4840
19602
85739
235
2.35
76742
11511
138
373
0.75
5
163350
6050
24503
107174
294
2.94
95928
14389
172
466
0.75
6
196020
7260
29403
128609
352
3.52
115114
17267
207
559
1
3
130680
4840
19602
85739
235
2.35
102323
15348
184
419
1
4
174240
6453
26136
114319
313
3.13
136431
20465
245
558
1
5
217800
8067
32670
142899
392
3.92
170539
25581
307
698
1
6
261360
9680
39204
171478
470
4.70
204646
30697
368
838
1.275
3
166617
6171
24993
109317
299
2.99
166339
24951
299
599
1.275
4
222156
8228
33323
145757
399
3.99
221786
33268
399
798
1.275
5
277695
10285
41654
182196
499
4.99
277232
41585
498
998
1.275
6
333234
12342
49985
218635
599
5.99
332678
49902
598
1197
A
B
C
0
E
F
G
H
I
J
K
L
CELL FORMULAS
PARAMETER
AREA, ACRES
DEPTH, FEET
VOL, CU-FT
VOL, CU-YDS
VOID VOL, CU-FT
WETLANO LIFE, OAYS
UETLAND LIFE, YEARS
ADOITIONAL DEPTH, FT
ADOED BIOMASS, CU-FT
ADOED VOIDS, CU-FT
ADOITIONAL LIFE, YEARS
TOTAL LIFE, XV IS
COLUMNS
INITIAL VALUES (ROU 14)
0.2S
3
~A24*43560*B24
~C24/27
+$J$10*C24/100
+E24/SOS17
+F24/365
«G24/100
~H24*A24*43560
+I24*J<10/100
(~J24/SOS17)/365
~IC24+G24
NOTE 1: ASSUMES 1 FOOT OF
IIOACCUMULATION PER 100 YEARS
12-20
-------
TABLE 39 PROJECTED WETLAND VOLUME REQUIREMENT BASED ON THE
STOICHIOMETRY OF THE SULFATE-REDUCING BACTERIA REACTION
SOURCE
FEED RATE
GPM SO
METAL CONCENTRATIONS
LOADING RATE, IRON 65 mg/lit«r
LOADING RATE, COPPER 5 ng/lltar
LOADING RATE, OTHER 5 ng/lit«r
TOTAL LOADING 75 *g/Htar
ORGANIC NATTER
CONTENT IN
ORIGINAL
SUBSTRATE
30 X
METAL LOADINGS
LOADING RATE, F«
LOADING RATE, CU
LOADING RATE, OTHER
TOTAL LOADING
METAL LOADING FORMULAS
17.7 Kg/DAY <-- .785*1WO'GS/10OOOOO
1.4 Kg/DAT <-- *SU8*3.785*1440*69/1000000
1.4 Kg/OAY <-- *SBS8*3.785*1440*010/1000000
20.4 Kg/DAY <— §SUM(J14..J16)
<1> (2) (3)
WETLAND WETLAND WETLAND ORGANIC METAL LIFE OF LIFE OF ADOITNL ADDITNL ADOITNL TOTAL AREA
AREA OEPTH VOLUME DRY MASS MASS LOAD RATE WETLAND WETLAND BIOMASS DEPTH LIFE LIFE FLUX
(AC) (FT) (CU-FT) (Kg) (Kg) (Kg/DAY) (OAYS) (YEARS) (Kg) (FT) (YRS) SF/GPM
12
12
12
12
A
2.5E+06 7.5E*05
3.3E*06 1.06*06
4.2E«06 1.3E+06
5.0E*06 1.56*06
5.0E+06 1.5E+06
6.7E+06 2.0E+06
8.4E+06 2.5E+06
1.0E*07 3.0E+06
3 1568160 3.0E*07 9.0E*06
4 2090880 4.0E+07 1.2E*07
5 2613600 5.06*07 1.5E*07
6 3136320 6.06*07 1.8E+07
8 C 0 E
1
3
130680
1
4
174240
1
5
217800
1
6
261360
2
3
261360
2
4
348480
2
5
435600
2
6
522720
20
7,365
20
66*04
0.07
1.6
22
871
20
9,821
27
8E+04
0.10
2.2
29
871
20
12,276
34
1E*05
0.12
2.7
36
871
20
14,731
40
11*05
0.15
3.3
44
871
20
14,731
40
2E+05
0.15
6.6
47
1,742
20
19,641
54
3E*05
0.20
8.8
63
1,742
20
24,552
67
4E*0&
0.24
11.0
78
1,742
20
29,462
81
3E*0S
0.29
13.1
94
1,742
20
88,386
242
96*06
0.88
237
479
10,454
20
117,848
323
1E*07
1.17
316
638
10,454
20
147,310
404
1E*07
1.47
394
798
10,454
20
176,772
484
2E*07
1.76
473
958
10,454
G
N
I
J
K
L
N
CELL FORMULAS
PARAMETER
AREA, ACRES
SUBSTRATE DEPTH, FT
UETLNO VOL, CU-FT
UETLNO DRY MASS, Kg
ORGANIC MASS, Kg
TOTAL LOADING, Kg/DAY
UETLANO LIFE, DAYS
WETLAND LIFE, YEARS
ADDITIONAL BIOMASS, Kg
ADDITIONAL OEPTH, FT
ADDITIONAL LIFE, YEARS
TOTAL LIFE, YEARS
AREA FLUX, SF/GPM
COLUMNS FORMULAS (ROU 24)
A 1
B 3
C +A24*43560*B24
D *C24*19.2
E *SCS16*D24/100
F *SCS17
0 («E24/(SF24*2.5))*0.50
H *€24/365
1 *A24*43560*0.093*H24*0.75
J (I24/19.2)/(A24*43560)
K («I24/(SF24*2.5)*0.5)/36S
L *K24«M24
M *A24*43560/SBS8
NOTE 1: (AFTER LEMKE, 1989) UETLAND
SUBSTRATE BULK DENSITY « 1.28 g/ce; S.G.
SOLIOS OF SUBSTRATE ¦ 1.7 YIELOS 53X
DRY SOLIDS BY UEIGHT AND DRY OENSITY OF
19.2 Kg/CF
NOTE 2} LIFE OF UETLANO IS BASED ON 2.5
GRANS ORGANIC MATTER ARE NECESSARY FOR
EVERY GRAN OF METAL REDUCTION AND ONLY
SOX OF CARBON IN ORIGINAL SUBSTRATE MASS
IS CHEMICALLY AVAILABLE (IN THE LONG
TERM) OUE TO HUNIFICATION.
NOTE 3: ACCUMULATION BASED ON DATA IN
NEDIN, ET. AL. IN HAMMER, 1989; NPP OF
CARBON IS .093 (Kg/yr)/SF OF MUCH 75X IS
AVAILABLE TO SULFATE REOUCERS.
12-21
-------
SULFATE-REDUCING STOICHIOMETRY METHOD
Hedin, Hyman and Hammack (65) present a discussion of sulfate-reducing bacteria stoichiometry
and its relationship to carbon content in the substrate. Stoichiometricly, one mole of sulfate is reduced to
hydrogen sulfide for every two moles of carbon oxidized. Further, one mole of sulfate is required to
precipitate iron as "FeS" as discussed in SECTION 5. Thus, two moles of carbon are required for every
mole of ferrous iron in the wetland feed water. Most other metals (Cu, Zn, Cd, Ni and Hg) tend to follow the
1:1 metal to sulfur stoichiometry. Thus, they also require two moles of carbon for every mole of metal.
Assuming "organic matter" [O.M.] in substrate has the chemical formula CH20 (molecular wt ¦ 30 grams),
60 grams of organic matter would be required per mole of metal precipitation.
However, the formation of a mole of pyrite (FeSj) requires an additional two moles of carbon.
Pyrite is more stable in acid solution than FeS, so it would be desirable to optimize conditions for pyrite,
rather than acid-soluble FeS. The formation of pyrite in a wetland environment may need to be induced by
other process mechanisms. Nevertheless, for every mole of pyrite formed, four moles of carbon are
required stoichiometricly.
The other heavy metals (with the exception of Mn) have larger atomic weights the Fe-from 1.14
times (for Cu) to 3.6 times (for Hg) the weight of Fe. Also, the Fe concentration is generally at least an
order of magnitude larger in acid drainage than that of the next most concentrated metal. Therefore, as a
first approximation, the concentrations of all the heavy metals, in mg/l, can be added. Then, the atomic
weight of iron and the stoichiometry of the pyrite reaction can be used to estimate substrate "life" based
on available carbon/organic matter. Biomass accumulations from a carbon source perspective can also be
used to estimate whether a particular wetland configuration will become self-perpetuating.
Table 39 presents constructed wetland life estimates based on the following assumptions:
o Wetland substrate dry density is 1.77 g/cc (9).
o The concentrations, in mg/l, of Fe, Cu, Zn, Cd, Ni, and Hg are added and assumed to be equal to
the concentration of a hypothetical metal with atomic weight of 55.
o Four gram-moles of carbon are required for each gram-mole of this hypothetical metal.
o Thus, 0.83 grams of carbon are needed to reduce 1 gram of dissolved metal to sulfide.
o 40 percent of organic matter is "stoichiometricly available" as carbon
(12*4 gram C /120 grams OM - 0.4), thus 2.5 grams of OM are required per gram of metal.
o 50 percent of the carbon in the original substrate is rendered unavailable to sulfate reducing
bacteria due to the cumulative long-term effects of humification.
o Carbon accumulation from plants occurs at a rate of 1 kg/mZ-yr or 0.093 Kg/square foot per year
(65 and SECTION 12).
o 75 percent of the accumulated plant carbon is available for sulfate precipitation (65)
Table 39 results suggest that a one acre, four-foot-deep wetland with a nominal metal loading of
75 mg/liter in a flow rate of 50 gpm should last about 27 years. Additional biomass accumulations could
12-22
-------
add about two years to the life of the facility for a total life of 29 years. The associated area flux value of 871
sf/gpm is within nominal limits for this variable.
Table 39 further suggests that at the given loading rates, a self-perpetuating wetland system
might develop for a nominal 12-acre site. The associated flux value of 10,454 sf/gpm is significantly higher
than the limits for this variable used in Tables 31-37. Cost trade-offs between annual operating costs and
front-end capital costs should be examined to establish minimum overall system costs. In addition, if
lifetime estimates of a wetland are extended to these high values, The question of how long the acid mine
drainage will last should also be studied.
The availability of "bacteria-usable" carbon in the substrate needs to be further addressed to allow
reasonable use of this wetland design methodology. Further, the amount of carbon that is actually used
by sulfate-reducing bacteria is unknown. Also, the sulfate-reducing bacteria may compete with other
bacteria for organic carbon.
Lemke (9) measured "organic matter" contents of about 30 percent in mushroom compost.
Lemke did not measure carbon content that is stoichiometricly available to sulfate-reducing bacteria. Data
presented in SECTION 6 indicate that 31% of Typha (cattail) plant mass is carbon. The chemical makeup
of "typical" organic matter in candidate substrates should be established before this methodology is used
to design constructed wetlands (see SECTION 14).
The stability of the carbon in the substrate with respect to bacterial utilization is another area
worthy of further study. Data from natural systems may suggest the extent that humification modifies
organic matter to the point that sulfate-reducing bacteria cannot use it.
EVAPOTRANSPIRATION LOSSES
Evapotranspiration losses from natural wetlands have been measured at levels several times that
of natural pan evaporation (58). This is consistent with evapotranspiration results from the Big Five
constructed wetland reported in SECTION 6.
In design considerations of wetlands with plants, an evapotranspiration rate of 1.8 times that of
standard pond evaporation might be assumed without raw data. Using this assumption in a sample water
balance analysis (see Table 40), net evaporation/evapotranspiration losses from a one acre wetland
ranged from 4 gpm to 6 gpm in the summer months to negligible amounts in the winter months. Similar
analyses were performed for wetland areas up to 3.5 acres in size with proportionate results; the maximum
net evapotranspiration loss from a 3.5 acre wetland was about 20 gpm.
This amount could provide a significant reduction in net wetland output in the summer months
and may impact water rights. However, the water budget of a typical wetland will probably not result in a
zero-discharge facility. As discussed in SECTION 5, substrate desiccation is to be avoided due to the
likelihood of oxidation of precipitated sulfides and re-mobilization of heavy metals. Thus, in certain
circumstances, evapotranspiration effects can work against the goals of'a constructed wetland.
12-23
-------
TABLE 40 NET EVAPOTRANSPIRATION LOSSES AT A HYPOTHETICAL
CONSTRUCTED WETLAND SITE
INFLOWS (SEE NOTE 1):
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC TOTALS
RAINFALL(IN) 1.29 1.S0 2.26 4.00 4.47 4.35 2.56 2.21 1.83 1.76 1.68 1.49 29.40 INCHES
EVAPORATN(IN) 0.81 1.23 2.18 3.26 4.79 5.84 6.54 5.95 4.06 2.62 1.18 0.77 39.23 INCHES
OUTFLOWS:
WETLAND FACTOR 1.8 (SEE NOTE 2)
, NET LOSSES (•}, GAINS (~) IN GALLONS PER MINUTE
WETLAND WETLAND
ACRES
0.5
1
1.5
2
2.5
3
3.5
SO-FT
21,780
43,560
65,340
87,120
108,900
-0.1
-0.1
•0.2
-0.2
-0.3
-0.2
•0.4
-0.7
-0.9
-1.1
-0.5
-1.0
-1.6
-2.1
•2.6
•0.6
•1.2
-1.8
-2.3
-2.9
-1.3
•2.6
-3.9
-5.2
•6.5
-1.9 -2.9
-3.9 -5.8
-5.8 -8.7
-7.7 -11.6
-9.7 -14.5
-2.7
-5.3
•8.0
-10.7
-13.4
-1.7
-3.4
-5.2
-6.9
-8.6
130,680 -0.3 -1.3 -3.1 -3.5 -7.8 -11.6 -17.4 -16.0 -10.3 -5.6
152,460 -0.4 -1.6 -3.7 -4.1 -9.1 -13.6 -20.3 -18.7 -12.1 -6.5
OCT
NOV
DEC
AVERAGE
muMwu
-0.9
-0.1
0.0
-1.1
-1.9
-0.3
0.1
-2.2
•2.8
-0.4
0.1
-3.2
-3.7
•0.6
0.1
-4.3
-4.6
-0.7
0.2
-5.4
-5.6
-0.8
0.2
-6.5
-6.5
-1.0
0.2
-7.6
NOTE 1: TYPICAL RAINFALL AM) PAN EVAPORATION DATA FROM BLACK HILLS, SO
NOTE 2: WETLAND FACTOR IS THE MULTIPLYING FACTOR APPLIED TO PAN EVAPORATION RATES
DUE TO PLANT EVAPOTRANSPIRATION. TYPICAL WETLAND FACTOR IS 2.0 OR LESS.
12-24
-------
SECTION 13
DESIGN CONFIGURATIONS
It has been said that a constructed wetland that treats acid/metal drainage is a "bioreactor with a
green toupee", referring to the organic substrate where most of the bioreactions occur and the collection
of plants that grow on the surface of the wetland. As discussed in SECTIONS 3 and 6, studies have
shown that plant uptake does not contribute significantly to water quality improvements in wetlands.
However, plants can replenish the wetland with organic material and add aesthetic appeal.
The design methods discussed in SECTION 12 will yield key design parameter values such as
surface area and substrate volume and depth as well as system hydrology constraints such as flux and
minimum acceptable hydraulic conductivity of the substrate.
The final configuration of a constructed wetland will in most respects rely on the land space
available and the topography of the site. Given that mining sites are typically found in rugged topography
where level land has been already allocated to other land uses, wetland sites may have to be developed
by excavation and earthwork. For example, excavated terraces or stabilized/reclaimed tailings or waste
rock dump surfaces might be considered as potential wetland sites.
Maintenance access will also influence configuration. As will be addressed in SECTION 15,
maintenance functions such as long term rejuvenation of substrate organic content and the possible
removal/ replacement of expended substrate need to be considered.
If the design criteria are satisfied, the bacteria that populate the wetland substrate should flourish
and the performance of the wetland as a bioreactor should meet design effluent concentrations. It is up to
the individual design engineer to incorporate "safety factors" where appropriate, based on parameter
uncertainties. For example, if flow rate fluctuations are expected, wetland surface area and volume
requirements (and other accompanying parameters) may need to be increased proportionately.
GENERAL CONFIGURATIONS
The fact that the bacteriologic processes flourish anaerobically and in the absence of large living ^
plants offers the design engineer more flexibility/creativity in the selection of a wetlands configuration.
Without plants, a wetland can be configured in two general ways: as a conventional/natural wetland
system or as a "stacked plate" system.
In both configurations, the water to be treated essentially makes "one pass" through the
dissolved-metals-removing wetland. Unless additional wetland polishing to remove B.O.O. or ammonia is
13-1
-------
FEED
WATER
CELL
CELL
CELL
CELL
CELL
CELL
EFFLUENT
WATER
Figure 59
A schematic plan view of a conventional wetland configuration.
-------
FEED
WATER
WETLAND CELL
WETLAND CELL
WETLAND CELL
WETLAND CELL
WETLAND CELL
DISCHARGE
Figure 60. A schematic cross-section view of a stacked wetland configuration.
-------
UNTREATED Mil
OVERFLOW
0!
JZ
OPTIONAL COVXB
CONSTANT man BE8BRVDIB
' V/INSULATED FLOATING TANK COVES
EFFLUENT
DISCHARGE
U
4k
PERFORATED PIPE
GRAVEL HEP
SUBSTRATE'
PERMEABLE GBOFABBIC
,%? A7 A7A7 ft? t'J- 131 'SSXSSXSSSS* i7 i"i7 M &J HJ-tJ-lZil
Y///////J
IMPERVIOUS CONSTRUCTION MATERIAL
(CONCRETE. GBOMEMBRANE/EARTH. COMPACTED CLAY)
\
UNTREATED WATER
OVERFLOW
EFFLUENT
DISCHARGE
\
VARIABLE HEAD
DISCBARGE PIPE
OPEN VALVE
(TYPICAL)
Figure 61. A schematic construction detail of a downflow wetland cell.
-------
required, the treated water only passes through a wetland cell once. In the following discussions,
additional polishing steps are assumed to be unnecessary. Howeyer, minor amounts of wetland effluent
may pass through several down-gradient cells to replenish evaporation losses and insure substrate
saturation.
Natural/Conventional Configuration
Given a site situation with abundant land area, the design engineer may employ a "conventional"
configuration that appears to be a natural system as shown in Figure 59. A conventional configuration
could be used regardless of plant usage policies in the wetland. Brodie of the TV A (55,68,69) has been
quite successful at building these types of systems, although the systems are shallow and promote
aerobic bacterial reactions and fluid flow over the surface. Anaerobic "conventional" wetland systems
have deeper ceil depths and vertical fluid flow.
From a construction materials perspective, a conventional configuration would allow the use of a
broad range of materials such as earthen berms, geomembranes, compacted clay with reinforced
concrete and piping as required. From a visual perspective, such a configuration might resemble a natural
wetland or a series of tiered settling ponds.
The addition of plants in the wetlands could provide a more natural appearance. Even if plants are
not directly introduced at the completion of wetland construction, it is likely that volunteer plants will
establish themselves with the passage of time. Thus, unless some plant control measures (such as
geomembrane/geotextile covers over open wetland areas) are installed, one might as well plant
acceptable flora species to avoid the ultimate invasion of noxious ones.
Stacked Plate Configuration
Given a site situation with inadequate land area for the design flows and the decision not to use
plants, the design engineer may employ a "stacked plate" configuration as shown in Figure 60. From a
construction materials perspective, a stacked configuration would compel the use of rigid materials such
as reinforced concrete or coated/lined compartmentalized tanks. From a visual internal perspective, such
a configuration might resemble a multi-storied automobile parking garage. Externally, the facility might
appear to be a concrete water tank.
In such a configuration, anaerobic conditions could be*easily maintained. However, substrate
maintenance may require extraordinary procedures that render the configuration impractical.
Excavated underground workings near the portal of a discharging tunnel may provide the
necessary "land area" for a constructed wetland without plants in a stacked plate configuration. The
apparently simpler legalities and hidden aesthetics of developing underground excavations to house
piantless wetland facilities could result in capital cost savings when compared to the land acquisition costs
and the securing of conveyance rights of way for alternative surface sites. Further, the "buried" nature of
the facility is unlikely to produce public opposition fueled by "not in my backyard" attitudes.
13-5
-------
In areas of exorbitant land acquisition costs or restrictive site topography, use of a stacked plate
wetland configuration should be considered.
DETAILED CONFIGURATIONS
Flow Directions
Anaerobic wetland bench and pilot scale testing has employed both up-flow and down-flow
configurations with nearly equal success. The development of up- and down- flow systems was prompted
by a desire to maintain a relatively high hydraulic conductivity of mushroom compost substrate in order to
vary loading rates.
Work with field permeameters (99) has shown that the physical operation of up-flow systems
presents numerous problems that include flow control and short circuiting. In pilot scale systems (cells B
north and south at the Big Five Site), the differences in hydraulic conductivity between up-flow and down-
flow decreased significantly (99). From a hydraulic perspective, up-flow systems require a driving head at
least double the thickness of the substrate layer and as permeability decreases, short circuiting becomes
more likely.
Thus, while up-flow systems appear to be useful on bench and pilot scale experiments, they
appear to be of limited application in full scale systems and will not be discussed further.
Conceptual nail Pastrm
A conceptual wetland down-flow cell detail is shown in Figure 61. The key aspects of the cell
include:
o Surface coverings (above perforated inlet pipes) to promote anaerobic reactions and protect
pipes in seasonal subfreezing climatic conditions (assuming no plants).
o Compartmentalization to provide for flow control/maintenance
o Subsurface collection of effluent in a gravel bed and pipe network.
o Paired cells adjacent to a single central feed pipe to minimize feed water exposure to oxygen.
o Separation of substrate and gravel bed/plenum with permeable geofabric.
o Overflow weir or pipe for untreated effluent to pass to the next down-gradient cell.
o Covered central "constant head reservoir that contains a burled feed pipe for providing source
head and flow control within each cell.
o Floating reservoir cover to reduce oxygen contact with feed water.
o Variable height discharge line for differential head control; i.e., flow control.
o Impervious construction materials for cell containment could include concrete,
geomembrane/earth berms, compacted clay.
13-6
-------
o Full-section valves for cell operational control/maintenance. The valve positions would normally
be fully open or completely closed. The valves could be used for flow throttling on a short-term
basis only.
o Groups of cells would be cascading in as much as site topography allows; i.e., the elevation of the
underflow from an up-gradient cell pair would be slightly higher than the overflow from a down
gradient cell pair. Thus, an "idle" ceil will automatically remain inundated from upgradient sources.
Figure 62 shows the grouping of several pairs of wetland cells in a conceptual wetland treatment
facility. Note that the overflow streams of untreated water from up-gradient cell pairs are routed to the
central reservoir feeding the down-gradient cell pairs.
Furthermore, the underflow from up-gradient cells is routed (via constant prime siphons with
check valves, if necessary) to the top of down-gradient cells. This design feature is a means of keeping
down-gradient cell substrate mass saturated if the facility operates unattended and feed flow rate
decreases. If the substrate mass is allowed to desiccate, sulfide precipitates could become oxidized and
would be released when flows to the cell returned. Desiccation could also result in substrate permeability
losses from compression of lower layers of the cell.
If the wetland facility is inspected on a regular basis, manual adjustments in cell discharge rates
could balance the flows among the cell pairs; constant-prime siphons would not be required. Still, gravity
flow configurations should be included in the installation for routing the underflow from up-gradient cells
to the tops of down-gradient cells.
Figure 63 shows a cross section through a group of cell pairs that highlights the anti-desiccation
features of the conceptual design.
The dimensions of individual cells will be estimated using methods discussed in SECTION 12.
The key criteria to successful cell operation should be the maintenance of flows through the cells and the
uninterrupted saturation of the substrate.
13-7
-------
CONSTANT HEAD
RESERVOIR (TYP)
CONSTANT
HEAD
RESERVOIR
UNDERFLOW
CONSTANT
HEAD
RESERVOIR
WETLAND
CELL
UNDERFLOW
OVERFLOW
WETLAND
CELL
OVERFLOW
FEED WATER FROM
ACID/METAL SOURCE
UNDERFLOW
NOT TO SCALE
NOTE: BYPASS PLUMBING
OMITTED FOR CLARITY.
FEED WATER FROM SOURCE
CAN BE DIVERTED TO
ANY CELL PAIR'S
CONSTANT HEAD RESERVOIR.
CLEAN EFFLUENT
DISCHARGE
Figure 62.
A schematic plan view of a downflow wetland configuration.
-------
FEED WATER
CELL 1 OVERFLOW
SUBSTRATE
CHECK
VALVE
SUBSTRATE
o OqQq °o °q°o0o0o0o0o°o
CELL 2 OVERFLOW
Li
CELL 2
SUBSTRATE
O °n°<> °Q ° f>°o eO °n000a00
CELL 3
1
CONSTANT PRIME SIPHON
CONSTANT HEAD
TANK W/OVERFLOW
NOT TO SCALE
Figure 63.
A schematic cross-section view of a downflow wetland installation.
-------
13-10
-------
SECTION 14
INSTRUMENTATION/PERFORMANCE EVALUATION
Effluent water quality is the ultimate indication of anaerobic wetland performance in removing
dissolved metals and neutralizing acid conditions. However, because the technology /nay not be
accepted as a totally "proven" method by some governmental agencies, documentation of all
performance parameters is recommended.
Prior to the startup of a constructed wetlands system, a Sampling and Analysis Plan (SAP) and a
Quality Assurance Project Plan (QAPP) should be written specifically for the site. These plans should
assure the consistent gathering and analysis of samples from the wetland and provide documentation on
wetland performance. The testing results could provide information useful in optimizing wetland
performance as operating experience is gained.
WETLAND CELL INSTRUMENTATION
For the entire site, a continuous chart recorder should monitor flow rate from the acid/metal
source. Flow meter selection should be influenced by the probability that metal hydroxide precipitates
may form in the conveyance section and disrupt flow measurements. Non-contacting flow meters such as
ultrasonic and magnetic units are recommended.
Periodic sampling of source water quality should be conducted in concert with the monitoring of
effluent water quality. Automated samplers that can retrieve composite samples should be considered
after wetland performance has stabilized.
At a minimum, the following data should be periodically obtained from each wetland cell:
o Differential head ("dH", Figure 52) between the water level on top of the substrate and the
underflow (effluent) pipe outlet. This measurement may be obtained with manometers or
pressure transducers. It is recommended to monitor at least one cell in an installation with a
continuous chart recorder for the first few years of operation,
o Average depth of substrate in the cell ("L", Figure 52), determined by survey after construction,
checked periodically for signs of substrate compression,
o Effluent flow rate ("Qout", Figure 52), measured with a calibrated flow meter. At a minimum, it is
recommended to monitor the cumulative effluent flow rate from all cells with a continuous chart
recorder and totalizer. Ideally, each cell would be continuously monitored for these data, so that
the organic matter content of the substrate in the cell could be correlated to the cumulative flow
and metal loading that the cell treated. As the quality of the effluent will probably not be corrosive,
flow meter construction materials need not be as chemical resistant as those used in measuring
source flow rates. Assuming no seepage* losses, evapotranspiration losses can be estimated
using source and effluent flow rate values.
14-1
-------
o Substrate samples for laboratory testing. The frequency of sampling and testing will be driven by
the performance of the system. Also, substrate disposal considerations may dictate some of the
sampling methodology. Recommended sampling procedures and test parameters are discussed
in the next two subsections.
Sampling points should be standardized to reduce sampling bias. From an academic perspective,
piezometers may be installed at various substrate levels to monitor the progress of bacterial treatment as
the water passes through the substrate. Water quality results may be correlated with other substrate
parameters such as organic matter or carbon content.
SAMPLING SUBSTRATE FOR PHYSICAL TESTING
Baseline Sampling of Candidate Substrate Materials
Candidate substrate materials should be sampled using methods that provide truly
"representative" samples. Statistical evaluation of important substrate characteristics such as permeability
and carbon/organic matter content should be considered. As with all soils, sample mass/volume should
be a function of material size distribution; i.e., the larger the maximum particle size, the larger should be
the mass of the sample. The number of samples might be governed by confidence interval requirements
or other statistical parameters.
Sampling In-Situ Materials
It is desirable to obtain relatively undisturbed samples of in-situ substrate to adequately
characterize it. There are a number of mechanical devices described in the literature on testing of peats
and organic soils. Some "undisturbed" sampling procedures have been attempted for substrate with
soupy consistency at the Big Five Wetland (114) such as closing the head end of the sampling tube
and/or creating a suction at the back end of the sampling tube, but there has been little or no undisturbed
sample retrieval success.
A thin wall sampling tube, such as specified in the American Society of Testing Materials (ASTM)
Method D 1587, with extremely sharp (possibly serrated) leading edges may be used to delineate a
relatively undisturbed sample. Recovering the sample, however, would be difficult due to its
unconsolidated, nearly liquid nature. The use of a small heat exchanging device has been recommended
to freeze the lower three or more inches of sample within the tube, effectively creating a temporary plug
with which to extract the practically undisturbed sample. In practice, the heat exchanging device would be
inserted and activated after the thin wall sampling tube had been driven a desired distance.
The heat exchanging device might consist of copper tubing coiled to just fit over the outside of
the thin walled sampling tube as shown in Figure 64. The coil would be protected by an outer tubing and a
layer of insulation to prevent freezing of the sample tube to a large volume of surrounding material. The
outer tubing with coil would be put in place after placement of the sampling tube.
14-2
-------
TUaiMQ (HIAT IXCHAMQM)
#/•">
OUTIA TUB!
¦ISOLATION
LI
Figure 64. A prototype substrate sampling device.
14-3
-------
The heat exchange source could be a tank of compressed gas, such as carbon dioxide or
nitrogen, which would be connected to the tubing. The gas would be released and allowed to expand at
the bottom of the heat exchanging device. To assist in the gas expansion, the diameter of the coil would
be abruptly enlarged (for example, from 0.125 inches to 0.375 inches) in the area required to be frozen.
After extraction, the tube would be kept in an upright position and the ends of the thin wall sample
tube would be covered with flexible plastic caps or similar devices.
The preservation of the sample after extraction could be completed in the following ways:
o Allowing for expansion, the entire sample could be frozen, then shipping short distances in
insulated containers would be possible;
o With caps on both ends, the samples could be moved short distances (from the site to the lab) in a
rack that kept the sample tubes oriented vertically. Freezing could be accomplished at the lab, if
desired. Alternately or in addition to an airtight cap, a hot paraffin plug might be placed on top of
the sample to further preserve sample integrity.
If it can be shown that freezing does not disrupt sample material characteristics, freezing the entire
sample as soon as possible should be considered. As discussed in the next subsection, this approach
offers a variety of advantages in testing procedures.
PHYSICAL TESTING
The following physical properties are considered important for classification and comparison of
this research with other constructed wetlands research, furtherance of understanding of the treatment
processes, and selection and development of values for design parameters:
o Material Classification
o Hydraulic Conductivity
o Moisture-Density/Compaction Relations
o Moisture-Density vs. Hydraulic Conductivity
The primary characteristic of a substrate material necessary for a number of other determinations is
the specific gravity of the solid components in the substrate. Given that the "muck" in the wetlands is
actually a mixture or slurry of solid substrate and acid mine water, knowing the specific gravity of the solids
would allow estimations of moisture contents by measuring the specific gravity (bulk density) of samples.
The specific gravity of the solids and bulk density of the mixture provide data from which the following
characteristics can be calculated:
o Porosity
o Moisture content
o Volume of solids in the mixture (for a given vol. or wt.)
o Weight of solids in the mixture (for a given vol. or wt.)
o Void ratio
14-4
-------
o Compaction
o Concentration of solids by weight
o Concentration of solids by volume
Field measurements immediately after sample collection may allow the estimation of sample gas
content.
If the entire sample is frozen, adjustments for the density of water as a liquid and water as a solid
would have to be made. However, the freezing of the entire sample would allow ease of sample splitting,
either axially or laterally. It may be possible to make polished sections of frozen samples to evaluate
compaction mechanisms, gas content and other phenomena.
Specific Gravity of Solids
The following alternative methods require evaluation:
o Landva, et. el. (112, p. 48), subscribe to a method proposed by Akroyd:
Essentially it involves the placing of the pulverized peat sample in a flask or bottle,
covering it with de-aired filtered kerosene, and applying a high vacuum until air bubbles cease to
be emitted from the sample. The container is then filled with kerosene and permitted to reach a
constant temperature. The specific gravity (Gs) may be calculated from the equation:
Wt. of Dry Sample
Gs ¦ X Spec. Grav. of Kerosene
Wt. of Kerosene Displaced
o There is no specific ASTM method for determining the specific gravity of peat. ASTM method C
128 applies to fine aggregate and uses water as the displaced medium but does not specify the
de-airing and filtering as stated above.
The above methods both involve the same principle proposed by Archimedes. The first method
is probably more accurate, and should be performed on a number of samples. The second method may
be more appropriate to determine the effects of entrained air on the apparent specific gravity of the solids.
A comparison of results between the two methods may be informative.
Bulk Density of Substrate/Water Mixtures
ASTM method D 4531 (113) was evaluated as it may apply to the bulk density determinations of
either frozen or unfrozen samples recovered from constructed wetlands. Method D 4531 utilizes two
different approaches, one which assumes a consolidated core material whose volume can be determined
by direct measurement and a second approach that requires dipping a less easily measured sample in hot
paraffin and the measuring of volume by submersion.
The first approach is probably more applicable than the second for both frozen and unfrozen
samples. Assuming that the weight of the sample tube has been predetermined and that the volume of
14-5
-------
the sample in the tube can be accurately measured, a bulk density of the mixture could be calculated.
Corrections for frozen conditions would have to made as necessary. In addition, some adjustment for gas
content would have to be made as the mixture actually consists of solid, liquid and gas components.
Alternately, ASTM method D 4380 (113) may be used if a sample may be disturbed. This method,
developed to test the density of bentonitic slurries, utilizes a mud balance and a measured volume to
determine the bulk density of the mixture. De-gassification of samples would assist in determining true
solids/liquid content of samples using this method.
Material Classification/Aah Content
Most contemporary references agree that "peat" consists of material with ash contents of less
than 25% of dry weight. Considering the sources of the materials in a typical substrate might be
composted manure, soil and other organics, (which probably contain significant mineral ash forming
material), it is likely that the mushroom compost would not be classified as a peat. Nonetheless, since the
organic content is a key substrate performance parameter due to its role in the sulfate-reducing bacteria's
health, this parameter would be a useful material characteristic to measure. The organic content of a
sample is calculated as follows:
Organic Content - 1 - (Ash Content, expressed as a decimal).
For the determination of ash content, Landva, et. al. (112, p. 45) recommend the ASTM method D
2974 modified by lowering the muffle furnace temperature to 440 degrees C (from 550 degr. C) and firing
the sample for five hours. As this recommendation was made primarily to preserve the characteristics of
mineral clays in samples, the modification is not required if very little clay is present in the material.
Typically, ash content of substrate samples may be determined using method D 2974, unmodified.
It is interesting to note that Landva, et. al. (112, p. 44) draw a relationship between ash content
and specific gravity for organic samples. However, the anticipated range of values for typical substrate
materials might be expected to be different from reported values for peat.
Carbon Content (Ultimate Analysis!
The Ultimate Analyses procedure is typically applied to coal samples but can be applied to any
material with combustible components such as wetland substrate. An ultimate analysis (ASTM 0-3176) of
a substrate sample is the determination of the ash and the elements of carbon, hydrogen, nitrogen, sulfur,
and oxygen as products of complete combustion. To insure that organic carbon content is reported in the
ultimate analysis results, samples should be digested in mild acid (HCI) and then thoroughly rinsed to
remove mineral carbonate materials, an inorganic source of carbon that would not be available to sulfate-
reducing bacteria.
14-6
-------
Material Classification/Size Distribution
There are three general ASTM methods of classifying peats which may be applied to substrate
materials by their size distributions and other factors:
D 2607 This method is a classification system based on five major material types according to generic
origin and fiber content (Sphagnum, Hypnum, etc). Fibers are defined as materials retained on an
ASTM No. 100 (0.15 mm) sieve, consisting ol stems, leaves, or fragments of bog plants, but
containing no particJes larger than 0.5 inches (12.7 mm). It excludes fragments of other materials
such as stones, sand and gravel.
O 2977This method separates peat material into arbitrary fractions based on particle size. Physical
separation of peat material according to particle size provides a useful indicator of the properties of
the peat specimen such as pore space, decomposition, etc. it also provides a means of
determining the amount of foreign matter. The four fractions are:
o Foreign matter consisting of ash-forming material such as rocks and shells is
removed manually from the 8 mesh sieve,
o Coarse fiber is retained on the 8 mesh sieve,
o Medium fiber is retained on the 20 mesh sieve and,
o Fine fibers and fines are passed through the 20 mesh sieve and retained in the
pan.
D 4427 This classification method standardizes naming peat products on the basis of fiber content (see D
2607, above), ash content (D 2974), acidity (D 2976), absorbency (D 2980) and botanical
composition based on inspection.
A modified version of method D 4427 is recommended by substituting method O 2977 for size
distribution in lieu of D 2607 as recommended in D 4427 unmodified. It is believed that the modified
version offers more data of a physical nature compared to unmodified D 4427 which provides
biological/geological data.
Volume Weights. Water Holding Capacity Air Capacity of Saturated Peat
Ivanov (100) discusses "bound" and ."free" water in peats and draws a relationship between the
volume of bound water and the active or effective porosity and the hydraulic conductivity of peat.
Research has indicated that the "immobilized" water in peat varies between 300% to 400% of the weight
of solid matrix. In effect, the presence of bound water may increase the relative velocity of water through a
wetlands system, thus reducing the residence time and adversely affecting chemical processes.
Preliminary estimates of residence time, using approximate data in Ivanov and the dimensions of
the Big Five wetlands cell (114), indicate that actual residence time might be about 60% of the estimated
residence time if bound water volume is ignored. Reed, et. al. (52) appear to have ignored the effects of
bound water in calculations and projections.
14-7
-------
Ivanov (100) reported that bound water estimates were derived from radioactive tracer data. A
method that may closely approximate the amount of bound water in a sample is ASTM D 2980. This
method sets up standardized conditions for measuring the volume and weight of water-saturated peat.
From these data, saturated volume weights, water-holding capacity on a weight and volume basis, dry
substrate volumes, and air volumes can be determined. The method provides useful information in
evaluating substrate materials. If large air spaces are present, high water penetration and aeration exist. If
air spaces are smaller, water retention is increased. Water retention would be greater in humified substrate
materials (smaller air spaces), whereas water penetration would be greater in unhumified substrate. This'
test method may provide useful design data as to when the substrate needs to be periodically replaced to
provide the optimum water treatment capacity.
Lemke (9) reported a 3.6 to 3.8 percent bound water content (as measured by drying at 105 °C)
in dry mushroom compost; a free water content (as measured by drying at 25 0 C) of 69.5 percent was
found for mushroom compost used for one year in Cell A.
Measurement of a sample's bound water may provide useful data for the comparison of candidate
substrate materials with those that have shown to be adequate. Further, the data may be useful in
developing modifying existing design methodologies. Perhaps sulfate reducing bacteria reaction rates
may be a function of bound/free water contents in substrate materials.
Hydraulic Conductivity Lahoratorv Mflthods
The goal of hydraulic conductivity measurements is to be able to simulate in the laboratory actual
flow conditions that might occur in the field wetlands environment. Thus, laboratory measurements would
allow the evaluation of critical design parameters that might be encountered in a scale up of bench or pilot
scale experiments. Standard ASTM methods provide a stable frame of reference from which to formulate
an appropriate method that would approach the above goal.
ASTM method 0 4511 yields the hydraulic conductivity of essentially saturated peat under
constant head conditions (see Figure 49). This method is similar to ASTM method D 2434 for the perme-
ability of granular soils under constant head. The principal distinction between the two methods is the way
of measuring of pressure differential, which is a function of the range of expected hydraulic conductivity
values.
Method D 2434 uses a fully saturated permeameter with the sample confined between two
porous plates and a differential manometer to measure a wide range of values, from about 10-2 to 10-1°
cm/sec.
Method 0 4511 uses a partially saturated permeameter with the sample confined between two 40
mesh screens and a measured height difference between the top of the feed reservoir and the bottom of
the permeameter to measure a range of values greater than 10*5 cm/sec.
Both methods assume laminar flow through a porous media as an ideal test condition so that
Darcy's law (see SECTION 12) can apply.
14-8
-------
The laboratory configuration prescribed by method D 4511 is probably more applicable to the field
wetlands configuration due to the use of the 40 mesh screen rather than porous plates. The screen might
be considered analogous to the layer of coarse rock in the bottom of the existing pilot plant. However, the
partially saturated condition of method D 4511 may need to be modified to a fully saturated condition to be
more representative of field conditions. Such a modification is shown on Figure 65.
Further, method D 4511 attempts to provide flow conditions with the void spaces saturated with
water and no air bubbles in the voids. The likelihood of gas being generated in the substrate (from sulfate-
reducing reactions) suggests that gas concentrations in the D 4511 permeameter should be allowed to
remain as close as possible to in situ concentrations during hydraulic conductivity measurements to better
simulate field conditions. Measurements with and without ambient gas concentrations may provide useful
design data.
ASTM reports that the K of peat is very sensitive to changes in bulk density and that even under
light compression, K can change by several orders of magnitude. This observation would probably apply
to typical substrate materials. From a laboratory and field perspective, then, care must be observed to be
able to adjust for changes in bulk density. Given that the sampling method proposed will provide relatively
undisturbed samples, minor changes in bulk density might be desirable to achieve more correctable
results among samples; i.e., compare the hydraulic conductivity for many samples at the same bulk
density.
The effects of Stokes law of settling may allow the minor adjustments to bulk density suggested
above. In order to achieve this, the permeameter configuration in 0 4511 might be changed to allow water
inflow from the bottom upwards instead of the top down (see Figure 50). If the upper 40 mesh screen
were allowed to float freely to a level of a desired bulk density, and the flow through the permeameter
temporarily adjusted into the turbulent range to generate "quicksand" conditions, the sample material
might adjust to the new available bulk volume to provide a "standard" or repeatable bulk density. Thus,
changes in hydraulic conductivity at a given location within the wetlands cell could be correlated to either
bulk density or actual material characteristics or both. Other methods of adjusting bulk density might be
subject to more error and sample disturbance.
The concept of bound water might be considered at this point. The amount of bound water may
be a direct function of size distribution and bulk density. The modified D 4511 permeameter configuration
provides a means of testing these hypotheses.
Hydraulic Conductivity. Field Mn Situ! Methods
These is no ASTM method of measuring field hydraulic conductivity of saturated soils although
methods are available for measuring in situ permeability of rock (ASTM D 4630). An applicable method
that should be considered is either a falling head or constant head permeability test (102).
Falling head tests would be conducted in a boring completed in a wetland cell by either raising the
water level in the boring above the static level of the wetland cell level and observing the fall of the water
level in the boring as a function of time, or by pumping water from the boring so the water level in the
14-9
-------
WIRE SCREEN
UPPER RESERVOIR
WATERTIGHT \ OVEMflOW
SEAL V//H/K Minim
SIPHON
WATER SUPPLY
OVERFLOW
UPPER TROUGH
SPECIMEN (Supported)
WIRE SCREEN SUPPORT
FUNNEL (Supported)
-BEAKER with PERCOLATE
AH « TOTAL HYDRAULIC HEAD DIFFERENCE
ACROSS SPECIMEN.
Figure 65. A downflow laboratory permeameter modified for full saturation.
-------
boring is below the static level of the cell and then observing the recovery of water in the boring as a
function of time. Given the fairly liquid nature of the substrate in the cell which might be sucked into the
casing, the first approach is preferred; that is, observing the fall in water level from some level above static.
The boring would be cased with perforated PVC pipe, with the perforations located in the zone of
interest. Due to boundary conditions, a small diameter, say 1 inch (2.54 cm), boring is preferred. Various
shape factors and other parameters influence the final calculations, but the data acquisition would be fairly
rapid considering the site configuration.
Constant head tests could be conducted in a similar manner and are subject to the same
limitations and comments as above. The primary difference, as the name implies, is that the head of water
in the casing is kept constant by the addition of a known volume of water. Hence, for practical purposes,
this test can only be performed with the head of water in the casing above the static level of the wetland
cell.
To maintain a constant head of water in the casing, the use of a volumemeter or similar device is
required.
14-11
-------
14-12
-------
SECTION 15
SYSTEM OPERATION AND MAINTENANCE
SUBSTRATE MAINTENANCE
The substrate is a Key component in any functioning constructed wetland treating acid/metal
drainage. Like any mechanical/ chemical/biological treatment system, declining effluent characteristics will
indicate that system maintenance is required. Design methodologies and physical testing of engineering
parameters have established a foundation for determining which aspects of the wetland operation may
need attention/correction.
Whether substrate materials can "wear out" remains a central issue in wetlands design, operation
and maintenance. From the perspective of carbon content available for bacterial utilization, the substrate
has a "useful life" imposed by chemical stoichiometry adjusted by the effects of humification.
Summarizing, wetland substrate will be subjected to stresses that tend to decrease its
performance with use. These stresses include:
o Precipitation of metal sulfides in void spaces
o Consumption of organic matter/substrate caition
o Humification of organic matter
o Loss of permeability which may be related to organic matter conditions and/or compaction from
settlement
The effects of the stresses are typically irreversible, but mitigation measures could provide minor
extensions of substrate usefulness. Mitigation measures include:
o Maintenance of a plant community on the surface of the wetland to provide a source of organic
matter/carbon
o Periodic additions of organic matter as a solid on the surface of the wetland (straw added using
mulch-spreading machines such as those employed in reclaiming disturbed land)
o Periodic removal of finer-grained materials that could lower substrate permeability; precipitated
metals in amorphous form may be removed simultaneously
o Continuous maintenance of substrate saturation, even if the wetland cell is "idle". Permeability
restoration by allowing the substrate to dry out is only temporary and will result in the oxidation of
precipitated sulfides and the remobilization of metals when flows are re-established. Furthermore,
Ivanov (100) points out that drying accelerates organic material breakdown and decreases
substrate permeability due to compression,
o Prohibition of machinery/personnel on substrate (this may require additional construction to allow
periodic sampling without impacting substrate)
15-1
-------
Ultimately, perhaps after decades of operation, the substrate may require total replacement.
Given that the wetland system has been designed properly, substrate removal should not impact other
subsystems of the wetland cells such as underdrains, gravel layers and system plumbing. Substrate
replacement in a "stacked" configuration would need to be considered in the initial design of the facility.
Assuming that the substrate material is totally, humified and mucky, slurry pumping/dredging
technology might be considered in its removal and replacement. For a conventional configuration
wetland, substrate could be removed from inundated cells with a shallow-draft dredge and replaced with
the same craft in almost a continuous process. For a stacked configuration, flushing mechanisms and
bottom or side drains and launders/flumes could be employed to remove/replace spent substrate.
Dewatering and disposal of the spent substrate material or removed fines will be governed by site-
specific substrate characteristics and governmental regulations (see SECTION 8). However, consider-
ation should be given to the resource potential of the metals contained in the substrate. Bog iron ore in
naturally-occurring wetland metal deposits was developed as a valuable resource by the early American
iron industry. Sulfide minerals could be concentrated using flotation methods for ultimate recovery by
smelting or other processes.
Data suggest (41,65) that partial substrate replacement by mixing old substrate with new materials
should be attempted with great care because the mixing process would infuse oxygen into the substrate.
The alteration of anoxic conditions in the substrate would result in the oxidation of metal sulfides and the
re-mobilization of metals. However, mobilized metal-laden water could be temporarily routed to a still-
functioning wetland cell (perhaps by pumping) or to a metals recovery system. Alternately, the
rejuvenated cell could be allowed to "lay fallow" without fresh inflows to allow the sulfide-reducing bacteria
to re-establish anoxic conditions and reduce any oxidized metals present in the substrate. Thus, the total
removal of substrate could be delayed or circumvented entirely. Also, as seen in Cell A, substrate
disturbance may have irreparably altered the permeability of the new/old substrate mix.
Land disposal of substrate without metals recovery may follow two process options in accordance
with hazardous or solid waste handling regulations:
o The substrate would have to be maintained in a saturated condition, otherwise oxidation of
sulfides could produce new acid drainage,
o Dried substrate would have to be maintained in a dry environment, perhaps secured beneath an
impermeable soil/geomembrane cap.
MAINTENANCE OF CONVEYANCES AND FLOW CONTROLS
Pipeline Maintenance
The precipitation of metal hydroxides and the corrosion of metal components are likely to
comprise the typical maintenance problems associated with the operation of a constructed wetland.
15-2
-------
Periodic inspections of the installation should include measurements associated with flow rates
and pressures. Headlosses in pipes should be calculated to determine if metal hydroxides are reducing
the cross sectional areas of pipes. The net effect of this phenomenon would be increases in pipeline flow
velocities and headlosses and accompanying decreases in flow rates.
Plugging of measuring points may result in erroneous pressure data readings. Thus, measuring
points (pipeline taps) should be designed to allow flushing with fresh water prior to measurements. This
requirement applies to continuous recording pressure and velocity probes as well. These probes need to
be periodically inspected, cleaned and calibrated to insure that they are operating properly.
To obtain a tangible observation of internal pipeline conditions, test spool sections should be
considered. Spool sections can be removed and measured to obtain qualitative data on pipeline
conditions such as corrosion or precipitation plugging. Thus, internal pipeline conditions can be
physically documented. A typical spool arrangement is shown on Figure 66.
Pipelines should be cleaned as needed before hydroxide deposits significantly or completely fill
the pipe cross section. Theoretically, scouring effects from higher flow velocities could maintain a stable
pipe cross sectional area. However, low pressure heads available at a site may limit the effects of scouring.
Maintenance of design flows should not rely entirely on scouring effects unless the factors controlling
scour are well understood.
Pipeline "pigs" are normally used to clean the interiors of pipes with detrimental accumulations.
Pigs are commercially available and should be used periodically on a site-specific schedule based on
experience. The pig's construction material should be compatible with the acidic conditions that will be
encountered.
Provision should be made to divert metal hydroxide sludges resulting from pipeline cleaning
operations away from wetland cell distribution systems. Holding/evaporation ponds or containment berms
may be required to complete this periodic maintenance task.
As shown in Figure 67, valves should be installed with clean-out plugs at tee- or wye- intersec-
tions to allow periodic cleaning with brushes or similar tools.
Surface Convevance/Wetland Containment Maintenance
Surface conveyance maintenance will probably consist of the removal of metal hydroxide deposits
form the conveyance invert and possibly the removal/replacement of substrate material lining the surface
conveyance. Other maintenance tasks may include repairs from burrowing animals or damages from storm
events. Earthmoving equipment such as backhoes or small bulldozers would typically be used to
complete these types of repairs.
Burrowing animals can impact wetland operations. Other workers (69) have addressed mitigation
measures against burrowing animals. These include covering embankment or channel surfaces with chain-
link fence and/or rip-rap and installing drop-pipe spillways.
15-3
-------
BOLT-
>=!
K\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\v
GASKET
J
ui
V///////////
w /y //////// /:
1 vvv\\u\\\\\\\\\\\\\\\\\\\\\\\\\\N
FLANGE FITTING
PIPE
^WWWNWWWWWWWWWWWN^
f
-TEST SPOOL
SECTION
NOT TO SCALE
Figure 66.
A typical spool arrangement used to monitor internal pipeline conditions.
-------
TEE
VALVE
CLEAN-OUT
PLUG?
WYE
VALVE
NOT TO SCALE
Figure 67. A diagram of the pipe cleaning valve arrangement used at tee-intersections.
15-5
-------
15-6
-------
SECTION 16
COST ESTIMATING
Typical costs of constructing wetlands have been reported in the literature (2, 69,109). Due to
the wide variation in siting and construction conditions, no "typical" unit costs for constructing wetlands
are presented herein. This section provides an outline of typical cost components associated with the
construction and operation of a constructed wetland.
Actual cost estimates for a specific project can be generated by summing the cost values of
individual components. Unit construction cost data are available from private-sector publications such as
the Means Cost Data (115) or Blue Book Equipment Costs (116).
CAPITAL/CONSTRUCTION COSTS
The major components of capital cost for constructed wetlands include:
o Preliminary Engineering and Testing
o Environmental Baseline Studies, Permits
o Land Acquisition
o Rights of Way Access
o Final Engineering Design and Construction Specifications
o Construction
The entire scope of the project should be included in the capital cost estimate, from the design
and construction of a collection system, to the acquisition of land for conveyances and the wetland itself
and finally to the construction of the designed facility.
Preliminary Enaineerinn and Testing
Given that a source of acid/metal-laden water has been identified and characterized, preliminary
engineering and testing are appropriate. Preliminary engineering and testing comprise those rudimentary
activities that help to determine the feasibility of utilizing wetlands technology for a particular site. Typical
activities include but are not limited to:
o Lab scale "bottle testing" of different combinations of source water and candidate substrates.
These tests take several weeks to complete,
o Physical testing of candidate substrate materials as detailed in SECTION 14. These tests can be
completed while bottle testing is underway,
o Field scale testing with 30 gallon (120 liter) "mini-cells" to establish candidate substrate long-term
permeability, viable loading rates and flux capacities. These tests should be continued until "mini-
cell" geochemistry and hydrology is completely understood. This may require six to 12 months of
monitoring and testing.
16-1
-------
o Preliminary wetland designs. If loading constraints require a land area that is not practical and a
"stacked" configuration has been eliminated due to cost considerations, wetland treatment may
not be feasible compared to other alternatives. This activity should be conducted as soon as
reliable field scale testing data are available,
o Pilot scale testing with cells at least 10 square meters in size using preliminary wetland designs.
Pilot scale testing would verify the "mini-cell" results and investigate the feasibility of typical
construction components such as conveyances, valves, source collection systems and scaled-up
substrate masses. Pilot scale tests should be conducted for at least a year. Pilot scale facilities
could become a "module" in the eventual full-scale system,
o Hydrologic investigations to establish source flow rate variation, if any. This effort may involve
rehabilitation of underground workings to evaluate potential bulkhead locations. The effort may
require up to a year of flow rate monitoring unless flow rate/hydroiogic data are available.
The costs of conducting these activities will vary significantly for each site. Typically, both
preliminary engineering and testing and final engineering design (SECTION 9) may amount to from six to
ten percent of a project's total capital cost (115).
Environmental Baseline Studies. Permits
Baseline environmental studies will be required prior to the siting and permitting of a constructed
wetland. The extent of the studies will be a function of the volume of existing data. For example,
regulatory agencies may not require background air quality studies if an adequate database already exists,
and complex air quality modeling may not be necessary if there are no significant thermal emissions or
large material stockpiles, it should be assumed that air quality, water quality and geosciences aspects of
the project will require some level of investigation. In addition, cultural resources, vegetation and wildlife,
land use and socio-economic aspects of the project will need to be addressed.
In summary, key environmental issues include:
o air quality
o climatology
o water quality and water use (surface and groundwater)
o soils and geology
o vegetation, wildlife, threatened and endangered species
o used substrate "waste" characterization
o land use and visual impacts
o cultural resources
o socio-economic impacts
SECTION 8 addresses regulatory/permit aspects of constructed wetlands. Cost estimates for
environmental baseline studies and permits should consider time for professional representatives of the
project to meet with regulators.
16-2
-------
Land Acquisition
For sites with significant private land holdings, land acquisition costs can amount to a significant
percentage of overall project capital cost. For example, land acquisition costs for the Areata, CA wetland
installation amounted to nearly 15 percent of the total project cost (109). At the same site, land acquisition
represented about 32 percent of the construction cost and was greater than preliminary engineering (plan
of study), permit and right of way access combined.
Nominal increases in construction and operations costs by considering "stacked" wetland
configuration may be more than offset by savings in land acquisition costs.
Rights of Wav Access
If the wetland site is a significant distance from the acid water source, significant rights of way
access may be required. The routing of pipes, open channels and power lines (assuming that recording
instruments may require a non-battery power source) may be required through lands not necessarily
included in the wetland proper. The further the source water is from the wetland installation, the higher
these costs are likely to be.
Rights of way access costs may consist of crossing fees imposed by entities such as railroads or
private land owners but more than likely will involve professional fees to negotiate and finalize agreements
for the acquisition of rights of way.
Final Engineering Design and Construction Specifications
Final engineering design involves the employment of methodologies developed in SECTION 12
and standard engineering approaches to produce the details of wetland construction.
Minor field investigations to determine the geotechnical and hydrologic characteristics of the site
are included in this effort. The field investigations may include the installation of additional ground water
monitoring wells (to supplement those installed in the environmental studies) and the gathering of
geotechnical data associated with available soils on site or at nearby soil borrow sites. These data are
typically acquired by drilling of geotechnical borings.
Final designs are used to generate construction specifications or plans which include text and
shop drawings, detailing what will be built, how it will be built and, if necessary, how a contractor would be
paid to build it. Construction specifications are typically organized by construction task.
An Engineer's Cost Estimate is usually included with construction specifications.
As stated earlier, preliminary and final engineering can amount to from six to ten percent of total
project costs (115). Attempts to economize in engineering may result in higher operating costs when
systems do not perform as intended and must be subsequently re-designed and re-built.
Construction
A construction cost estimate is typically included with the documents provided by the project
engineering staff. Construction cost items are typically distributed among discrete tasks and may be
estimated on a lump-sum or unit price basis.
16-3
-------
Typical construction tasks may include:
o Mobilization and Demobilization - This task includes the movement of personnel, equipment,
supplies and other incidentals to the project site. This effort may include obtaining building
permits, securing construction utilities (power and water) and other items which must be
completed prior to the initiation of meaningful work at the site. Contractor's overhead costs and
profits are often included in this pay item,
o Site Preparation - This task includes clearing the site of vegetation and the removal of stumps and
roots (grubbing).
o Source Control Construction - This task may include the diversion of multiple sources to a single
wetland or the installation of measures to maintain relatively constant flows. This task may be a
complete project in itself if significant underground construction/rehabilitation is required,
o Earthwork - This task includes the removal and stockpiling of topsoil, rough and final grading,
embankment foundation preparation, installation of fills, berms, excavation of basins,
o Basin Lining - This task may include the excavation and compaction of impervious soils such as
clay in the bottoms of wetland basins or cells. Geomembrane may be used for lining material;
geomembrane placement requires a smooth, prepared base to reduce the probability of leaks,
o Concrete - This task may include the forming and pouring of concrete for water distribution
structures and flow controls. In a stacked configuration, concrete work may comprise the majority
of the project construction effort,
o Plumbing - This task includes the excavations for and the installation of pipe conveyances and
flow controls.
o Substrate Conditioning and Installation • This task may include:
the blending of substrate material to provide homogeneity
the removal/addition of key substrate materials to produce design substrate
characteristics (permeability, size distribution, organic content, carbonate content)
pre-soaking of substrate
inoculation of substrate with sulfate reducing bacteria
placement of substrate in cells
o Vegetation • This task may include the cultivation of wetland plants off-site and the
installation/transplantation of vegetation to the surface of substrate-filled cells,
o Instrumentation - This task may include the construction of sampling points and the installation of
flow meters, auto-sampling devices and water level indicators and their associated chart recording
devices/telemetry.
o Construction Management - This task is typically conducted by the owner or owner's
representative and often is directed by the design engineer. The purpose of construction
management is to document that the wetland was constructed in accordance with the
16-4
-------
specifications. Construction management insures that no shortcuts in installation procedures are
taken nor substitution of materials made that could compromise the design of the facility.
Typically, construction management costs comprise approximately six percent of the total
construction cost.
OPERATING COSTS
Operating costs are clearly distributed between two major categories, nominal maintenance and
inspection and major overhauls.
Nominal maintenance tasks typically should include:
o Periodic, scheduled inspections
o Sampling events, maintenance of instruments and chart recording devices
o Cleaning of conveyances
o Flow adjustments and balancing flows among cells
The costs of nominal maintenance should include the preparation of periodic reports,
management of sampling and testing quality assurance activities and the costs of sample analysis.
Major overhauls may include periodic substrate rejuvenation or total replacement. Cost
components may include:
o Removal of existing substrate materials
o Removal of some aspects of cell plumbing/water distribution system
o Removal and preservation of plants
o Treatment of removed substrate materials (drying, stabilization processing, metals recovery)
o Containerization of removed substrate materials
o Transportation and disposal of removed materials
o Purchase of rejuvenation materials or replacement substrate raw materials
o Transportation of raw materials to the site
o Preparation of materials prior to installation
o Installation of prepared, rejuvenated or new substrate materials
o Replanting of vegetation
o Temporary treatment or rehandlirtg of untreated effluents (may include pumping or temporary
impoundment of source water)
Costs associated with the above tasks may be partially offset by revenues derived from metals
recovery. Metals recovery from substrate materials could provide two distinct advantages:
o Spent substrate materials may not be considered hazardous and may be disposed in a municipal
landfill or used in another beneficial use (soil amendment)
o Recovered metals may be processed to yield a saleable product.
16-5
-------
16-6
-------
REFERENCES
1. Mine Drainage and Surface Mine Reclamation, Volume 1: Mine Water and Mine Waste. U. S.
Bureau of Mines, Information Circular 9183,1988. 413 pp.
2. Donald A Hammer, ed. Proceedings of the International Conference on Constructed Wetlands
for Wastewater Treatment. Lewis Publishing Co., Ann Arbor, Michigan, 1989. 831 pp.
3. Wildeman, T. R. and L. S. Laudon. The use of Wetlands for Treatment of Environmental
Problems in Mining: Non—Coal Mining Applications. In: D. H. Hammer, ed., Proceedings of the
International Conference on Constructed Wetlands for Wastewater Treatment, Lewis Publishing,
Ann Arbor, Ml, 1989. pp. 221-231.
4. Howard, E. A., Emerick, J. C., and T. R. Wildeman. The Design, Construction and Initial Operation
of a Research Site for Passive Mine Drainage Treatment in Idaho Springs, Colorado. In: D. A
Hammer, ed. Proceedings of the International Conference on Constructed Wetlands for
Wastewater Treatment. Lewis Publishing, Ann Arbor, Ml, 1989. pp. 761-764.
5. Batal, W„ Wildeman, T. R., Laudon, L. S., and N.Moodnoordin. Bacteriological Tests from the
Constructed Wetland of the Big Five Tunnel, Idaho Springs, Colorado. In: D. A. Hammer, ed.
Proceedings of the International Conference on Constructed Wetlands for Wastewater
Treatment. Lewis Publishing, Ann Arbor, Ml, 1989. pp. 550-557.
6. Howard, E. A., Wildeman, T. R., Laudon, L. S., and Machemer, S. D. Design Considerations for
the Passive Treatment of Acid Mine Drainage. In: Proceedings of the Conference "Reclamation,
A Global Perspective". Alberta Land Conserv. and Reclam. Council Report No. RRTAC 89-2. p.
651-660.
7. Machemer, S. D., and Wildeman, T. R. Organic Complexation Compared with Sulfide Precipitation
as metal removal Processes from Acid Mine Drainage in a Constructed Wetland. Jour, of
Contaminant Hydrology. In review.
8. Laudon, L. S. Sulfur Mineralization in a Wetland Constructed to Treat Acid Mine Drainage,
Master's Thesis No. 3660, Colorado School of Mines, Golden, Colorado, 1988. 58 pp.
9. Lemke, P. R. Analyses and Optimization of Physical and Hydraulic Properties of Constructed
Wetlands Substrates for Passive Treatment of Acid Mine Drainage. Master's Thesis No. 3823,
Colorado School of Mines, Golden CO, 1989. 78 pp.
10. Batal, W. H. Dynamics of the Microbial Population in the Substrate of the Constructed Wetland at
the Big Five Tunnel, Idaho Springs, Colorado. Master's Thesis No. 3813, Colorado School of
Mines, Golden, CO, 1989. 176 pp.
11. Wildeman, T. R., Machemer, S. D., Klusman, R. W., Cohen, R. R., and P. Lemke. Metal Removal
Efficiencies from Acid Mine Drainage in the Big five Constructed Wetland. In: Proceedings of the
1990 Mining and Reclamation Conference, J. Skousen, J. Scencindiver, and D. Samuel eds.,
West Virginia Univ. Publications, Morgantown, WV, 1990. pp. 417-424.
12. Machemer, S. D., Lemke, P. R., Wildeman, T. R., Cohen R. R., Klusman, R. W., Emerick, J. C.,
and E. R. Bates. Passive Treatment of Metals Mine Drainage through use of a Constructed
Wetland. In: Proceedings of the 16th Annual Hazardous Waste Research Symposium, (J. S.
EPA, Cincinnati, OH, 1990. EPA Document No. EPA/600/9-90-037, pp. 104-114,1990.
R-1
-------
13. Boushka, V. Geochemistry of Coal. Elsevier Publishing Co., New York, 1981. 284 pp.
14. Valkovic, V. Trace Elements In Coal, Vol. 1. CRC Press Inc., Boca Raton, FL, 1983. 210 pp.
15. Manahan, S. E. Environmental Chemistry, 4th Ed. Willard Grant Press, Boston, MA, 1984. 612 p.
16. U. S. Envir. Protect. Ageniy. Development Document for Effluent Limitations Guidelines and
Standards for Coal Mining, Doc. No. EPA 440/1-82/057,1982. 408 pp.
17. Prodan, P. F., L. M. Mele, and J. P. Schubert. Runoff Water Quality and Hydrology at Coal Refuse
Disposal Sites in Southern Illinois. In: Proceedings , 1982 Symposium Surface Mining,
Hydrology, Sedimentology, and Reclamation, D. H. Graves, ed. U. of Kentucky, Louisville, KY,
1982, pp. 57—63.
18. Caruccio, F. T., J. C. Ferm, J. Home, G. Geidel, and B. Baganz. Paleoenvironment of Coal and its
Relation to Drainage Quality. EPA—600/7—77—067, U. S.Environmental Protection Agency,
Cincinnati, OH, 1977,107 pp.
19. Code of Federal Regulations. Title 40, Protection of the Environment: Parts 100—149, 1988.
pp. 530—531 and p.608.
20. Wildeman, T. R. Drainage from Coal Mines: Chemistry and Environmental Problems. In: Geology
in Coal Resource Utilization, D. C. Peters, ed. Amer. Assn. Petroleum Geologists, Tulsa, OK,
1991. In press.
21. Stumm, W., and J.J. Morgan. Aquatic Chemistry, 2nd Ed. John Wiley & Sons, New York, 1981.
780 pp.
22. Ehrlich, H. L. Geomicrobiology. Marcel Dekker, Inc., New York, 1981. 393 pp.
23. Taylor, B. E., M. C. Wheeler, and D. K. Nordstrom. Stable Isotope Geochemistry of Acid Mine
Drainage: Experimental Oxidation of Pyrite. Geochim. et Cosmochim. Acta, v. 48:2669—2678,
1984.
24. van Everdingen R. O. and H. R. Krouse. Interpretation of Isotopic Compositions of Dissolved
Sulfates in Acid Mine Drainage. In: Mine Drainage and Surface Mine Reclamation: U. S. Bureau of
Mines Information Circular 9183,1988. pp. 147—156.
25. Garrels, R. M., and F. T. Mackenzie. The Evolution of Sedimentary Rocks. W. W. Norton, New
York, 1971. 397 pp.
26. Wildeman, T. R.. Chemistry of the Argo Tunnel. Quarterly of the Colorado School of Mines, v.
78(4): 31—37, 1983.
27. Wildeman T. R.. A Water Handbook of Metal Mining Operations. Report 113, Colorado Water
Resources Research Institute, Fort Collins Co., 1981. 84 pp.
28. Klusman, R. W., and S. D. Machemer. Natural Processes of Acidity Reduction and Metal removal
from Acid Mine Drainage. Geology in Coal Resource Utilization, D. C.Peters, ed. Amer. Assn.
Petrol. Geologists, 1991. In press.
R-2
-------
29. Standard Methods for the Examination of Water and Wastewater, 16th ed. Amer. Public Health
Assn.,Washington, DC, 1985. pp. 265—269.
30. Lindsay, W. L. Chemical Equilibria In Soils. John Wiley & Sons, New York, 1979. 449 pp.
31. Watztaf, G. R. Chemical Stability of Manganese and Other Metals in Acid Mine Drainage Sludge.
In: Mine Drainage and Surface Mine Reclamation, Volume I. U. S. Bureau of Mines Circular 9183,
1988. pp. 83—90.
32. Wildeman, T. R., D. L. Cain, and A. J. Ramirez. The Relation Between Water Chemistry and
Mineral Zonation in the Central City Mining District, Colorado. In: National Symposium of Water
Resources Problems Related to Mining. Proceedings No. 18, American Water Resources Assn.
1974. pp. 219—229.
33. Sims, P. K., M. A. Drake, and Tooker. Geology and Ore Deposits of the Central City Mining
District, Gilpin County, Colorado. Prof. Paper 359, U. S. Geological Survey, 1963. 231 pp.
34. Shuster, E. T., and W. B. White. Seasonal Fluctuations in the Chemistry of Limestone Springs: A
Possible Means for Characterizing Carbonate Aquifers. Journal of Hydrology, v. 14: 93—128,
1971.
35. Jacobsen, R. L., and D. Langmuir. Controls on the Quality Variations of Some Carbonate Spring
Waters. Journal of Hydrology, v. 23:247—265,1974.
36. Caruccio, F. T., The Nature of Acid Mine Drainage Reactions and their Relation to Overburden
Analysis. In: Ecology and Coal Resource Development, M. K. Wali, ed. Pergamon Press, New
York, 1979. pp. 775—781.
37. Snyder, D. T., and F. T. Caruccio. The Partitioning of Flow Components of Acidic Seeps from
Surlace Coal Mines and the Identification of Acid Producing Horizons within the Backfill. In: Mine
Drainage and Surface Mine Reclamation, Volume I. U. S. Bureau of Mines Circular 9183,1988.
pp. 59—66.
38. Guertin, deF, J. C. Emerick, and E. A. Howard. Passive Mine Drainage Treatment Systems: A
Theoretical Assessment and Experimental Evaluation. Cooperative Agreement No. 202—317
Report, Colorado Mined Land Reclamation Division, Denver, CO, 1985.
39. Girts, M. A., and R. L. P. Kleinmann. Constructed Wetlands for Treatment of Acid Mine Drainage:
A Preliminary Review. In: National Symposium on Mining Hydrology, Sedimentology, and
Reclamation. Univ. of Kentucky Pres%, Lexington, KY, 1986. pp. 165—171.
40. Wieder, R. K., and G. E. Lang. Fe, Al, Mn, and S Chemistry of Sphagnum Peat in Four Peatlands
with Different Metal and Sulfur Input. Water, Soil, and Air Pollution, v.29:309—320,1986.
41. Hedin, R. S„ D. M. Hyman, and R. W. Hamack. Implications of Sulfate—Reduction and Pyrite
Formation Processes for Water Quality in a Constructed Wetland: Preliminary Observation. In Mine
Drainage and Surface Mine Reclamation, Volume I, U. S. Bureau of Mines Circular 9183,1988.
pp. 382—388.
R-3
-------
42. Watson, J. T., S. C. Reed, R. Kadlec, R. L. Knight, and A. E. Whitehouse. Performance
Expectations and Loading Rates for Constructed Wetlands. In: Proceedings of the International
Symposium of Constructed Wetlands, D.A. Hammer, ed. Lewis Publishing, Ann Arbor, Ml, 1989.
pp. 319-352.
43. Wixon, B. G., and B. E. Davies. Frontier Technology for Environmental Cleanup. In: Frontier
Technology in Mineral Processing, J. F. Spisak and G. V. Jergensen II, eds. American Institute of
Mining, Metallurgy, and Petroleum Engineering, New York, 1985. pp. 33—42.
44. Kalin M. Ecological Engineering and Biological Polishing: Methods to Economize Waste
Management. Canadian Mineral Processors Annual Operators Conference, Ottawa, Ontario,
1988.
45. Kepler, D. A. An Overview of the Role of Algae in the Treatment of Acid Mine Drainage. In: Mine
Drainage and Surface Mine Reclamation. U. S. Bureau of Mines, Information Circular 9183,1988.
pp. 286—290.
46. Maynard, J. B. Geochemistry of Sedimentary Ore Deposits. Springer—Verlag, New York, 1983.
305 pp.
47. Mason, B., and C. B. Moore. Principles of Geochemistry, 4th Ed. John Wiley & Sons, New York,
1982. 244 pp.
48. Wieder, R. K. Determining the Capacity for Metal Retention in Man—made Wetlands Constructed
for Treatment of Coal Mine Drainage. In: Mine Drainage and Surface Mine Reclamation. U. S.
Bureau of Mines, Information Circular 9183,1988. pp. 375—391.
49. Emerick, J. C., W. W. Huskie, and D. J. Cooper. Treatment of Discharge From a High Elevation
Metal Mine in the Colorado Rockies Using an Existing Wetland. In: Mine Drainage and Surface
Mine Reclamation. U.S.Bureau of Mines Information Circular 9183,1988. pp. 345—351.
50. Stevenson, F. J. Humus Chemistry. Wiley— Interscience, New York, 1982.443 pp.
51. Eger, P., and K. Lapakko. Nickel and Copper Removal from Mine Drainage by a Natural Wetland.
In: Mine Drainage and Surface Mine Reclamation. U. S. Bureau of Mines Information Circular
9183, 1988. pp. 301—309.
52. Reed, S. C., E. J. Middlebrooks, and R. W. Crites. Natural Systems for Waste Management and
Treatment. McGraw-Hill Co., New York, 1988. 308 pp.
53. Kerndorf, H., and M. Schnitzer. Sorption of Metals on Humic Acid. Geochim. et Cosmochim.
Acta, v. 44:1701—1758, 1980.
54. Postgate, J. R. The Sulphur—Reducing Bacteria. Cambridge Univ. Press, New York, 1979. 151
PP-
55. Brodie, G. A., D. A. Hammer, and D. A. Tomljanovich. Constructed Wetlands for Acid Drainage
Control in the Tennessee Valley. In: Mine Drainage and Surface Mine Reclamation, U. S. Bureau
of Mines, Information Circular 9183,1988. pp. 325—331.
56. Jenne, E. A. Trace Element Sorption by Sediments and Soils—Sites and Processes. In:
Molybdenum in the Environment, Volume 2, W. R. Chappell and K. K.Petersen, eds. Marcel
Dekker, Inc. New York, 1977. pp. 425—524.
R-4
-------
57. Holcombe, L. A. Adsorption and Desorption in Mine Drainages. Thesis No. 1944, Colorado
School of Mines,Golden, CO, 1977. 101 pp.
58. Sencindiver, J. C., and D. K. Bhumbla. Effects of Cattails (Typha) on Metal Removal from Mine
Drainage. In: Mine Drainage and Surface Mine Reclamation, U. S. Bureau of Mines, Information
Circular9183, 1988. pp.359—368.
59. Hiel, M. T., and F. J. Kerins. The Tracy Wetlands: A Case Study of Two Passive Mine Drainage
Treatment Systems in Montana. In: Mine Drainage and Surface Mine Reclamation, U. S. Bureau
of Mines, Information Circular 9183,1988. pp. 352—358.
60. Copeland—Michand, S. R., and C. J. Richardson. Relative Radial Oxygen Loss in Five Wetland
Plants. In: Proceedings of the International Conference on Constructed Wetlands, D. A. Hammer,
ed. Lewis Publishing, Ann Arbor, Ml, 1989. pp 501-507.
61. Wildeman, T. R., L. S. Laudon, R. L. Olsen, R. W. Chappell, and R. P. Rehm. Using Chemical
Analyses from the Contract Laboratory Program in Modeling: Problems and Suggestions. In:
Geochemical Modeling of Aqueous Systems II, D. C. Melchior and R. L. Bassett, eds. Volume No.
416, Amer. Chem. Society Symposium Series, pp 321-329.
62. Annual Book of ASTM Standards. Volume 5.05, Test D—3177. American Society for Testing
and Materials, Philadelphia, PA, 1988.
63. Klute, A., and C. Dirksen. Hydraulic Conductivity and Diffusivity: Laboratory Methods. In: Methods
of Soils Analysis, Part 1, 2nd Edition, A. Klute ed. American Society of Agronomy, Madison Wl,
1986. pp. 687—734.
64. Annual Book of Standards, Volume 4.08. American Society for Testing and Materials,
Philadelphia, PA, 1988.
65. Hedin, R. S., Hammack, R., and D. Hyman. Potential Importance of Sulfate Reduction Processes
in Wetlands Constructed to Treat Coal Mine Drainage. In: Proceedings of the International
Conference on Constructed Wetlands, D. A. Hammer, ed. Lewis Publishing, Ann Arbor, Ml,
1989. PP 508-514.
66. Hammack, R. W., and R. S. Hedin. Microbial Sulfate Reduction for the Treatment of Acid Mine
Drainage: A Laboratory Study. In: Proceedings of the Conference "Reclamation, A Global
Perspective". Alberta Land Conserv. and Reclam. Council Report No. RRTAC 89-2. pp. 673-680.
67. Mclntire, P. E., and H. M. Edenbom. The use of Bacterial Sulfate Reduction in the Treatment of
Drainage from Coal Mines. In: Proceedings of the 1990 Mining and Reclamation Conference, J.
Skousen, J. Scencindiver, and D. Samuel eds., West Virginia Univ. Publications, Morgantown,
WV, 1990. pp. 409-415.
68. Brodie, G. A. Treatment of Acid Drainage using Constructed Wetlands, Experiences of the
Tennessee Valley Authority. In: Proceedings, 1990 National Symposium on Mining, D. H. Graves
ed., U. of Kentucky Publications, Lexington, KY, 1990. pp. 77-83.
69. Brodie, G. A. Workshop on Design, Construction, and Operation of Wetlands Systems to Treat
Acid Drainage. Proceedings of the 12th Annual NAAMLP Conference, Breckinridge, CO, 1990.
In Press.
R-5
-------
70. Tchobanoglous, G., and E. D. Schroeder. Water Quality, Addison-Wesley Publishing Co.,
Reading, MA, 1985, 768 pp.
71. Olsen, R. L. Remedial Investigation Report - Clear Creek/Central City site. EPA Contract No. 68-
01-6939, Camp, Dresser, and McKee, inc. Denver, CO, 1987.
72. Brodie G. A., Hammer, D. A., and D. A. Tomjanovich. Constructed Wetlands for Acid Drainage
Control in the Tennessee Valley. In: Mine Drainage and Surface Mine Reclamation: U. S. Bureau
of Mines Information Circular 9183,1988. pp. 325—331.
73. Stark, L. R., and others. The Simco #4 Wetland: Biological Patterns and Performance of a
Wetland Receiving Mine Drainage. In: Mine Drainage and Surface Mine Reclamation: U. S.
Bureau of Mines Information Circular 9183,1988. pp. 332—344.
74. Emerick, J. C., Huskie, W. W., and D. J. Cooper. Treatment of Discharge from a High Elevation
Metal Mine in the Colorado Rockies using and Existing Wetland. In: Mine Drainage and Surface
Mine Reclamation: U. S. Bureau of Mines Information Circular 9183,1988. pp. 345—351.
75. Hiel, M. T., and F. J. Kerins. The Tracy Wetlands: A Case Study of Two Passive Mine Drainage
Treatment Systems in Montana. In: Mine Drainage and Surface Mine Reclamation: U. S. Bureau
of Mines Information Circular 9183,1988. pp. 352—358.
76. Dietz, J. M., and R. F. Unz. Effects of a Sphagnum Peat on the Quality of a Synthetic Acidic Mine
Drainage. In: Mine Drainage and Surface Mine Reclamation: U. S. Bureau of Mines Information
Circular 9183, 1988. pp. 310—316.
77. Wikjeman, T. R., Laudon, L. S., and E. A. Howard. Construction of a Wetland Demonstration Site
for a Metal-mine Drainage. In: Mine Drainage and Surface Mine Reclamation: U. S. Bureau of
Mines Information Circular 9183,1988. p. 401.
78. American Society for Testing and Materials, Annual Book of ASTM Standards. Standard Test
Method for Forms of Sulfur in Coal, v. 5.05, Method D-2492-84,1988.
79. Tuttle, M. L„ Goldhaber, M. B., and D. L. Williamson. An Analytical Scheme for Determining Forms
of Sulphur in Oil Shales and Associated Rocks. Talanta, v. 33,953-961,1986.
80. Wieder, R. K., Lang, G. E., and V. A. Granus. An Evaluation of Wet Chemical Methods for
Quantifying Sulfur Fraction in Freshwater Wetland Peat. Limnol. Oceanogr., v. 30, No. 5,1109-
1115, 1985.
81. American Society for Testing and Materials, Annual Book of ASTM Standards, v. 5.05, Method D-
3177, pp 418-422.
82. Davison, W., Lishman, J. P., and J. Hilton. Formation of Pyrite in Freshwater Sediments:
Implications for C/S Ratios. Geochim. Cosmochim. Acta, V. 49, pp. 1615-1620,1985.
83. King, G. M., Howes, B. L., and J. W. H. Dacey. Short-term End Products of Sulfate Reduction in a
Salt Marsh: Formation of Acid Volatile Sulfides, Elemental Sulfur, and Pyrite. Geochim.
Cosmochim. Acta, V. 49, pp. 1561-1566,1985.
84. Melchior, D. C., Wildeman, T. R., and D. L. Williamson. A Mossbauer Study of the Transformation
of the Iron Minerals in Oil Shale during Retorting. Fuel, V. 61, pp. 516-522,1982.
R-6
-------
85 Chao, T. T. The Use of Partial Dissolution Techniques in Geochemical Exploration. J. Geochem.
Exploration, v. 36, pp. 101-135, 1984.
86 Kheboian, C., and C. F. Bauer. Accuracy of Selective Extraction Procedures for Metal Speciation
in Model Aquatic Sediments. Anal. Chem., v. 59, pp. 1417-1423,1987.
87. Lavkulich, L. M., and J. H. Wiens. Comparison of Organic Matter Destruction by Hydrogen
Peroxide and Sodium Hypochlorite and its Effect on Selected Mineral Constituents. Soil Sci. Soc.
Amer. Proc., V. 34, pp. 755-758,1970.
88. Shuman, L. M. Sodium Hypochlorite Methods for Extracting Microelements Associated with Soil
Organic Matter. Jour. Amer. Soil Sci. Soc., v. 47, pp. 656-660,1983.
89. Chao, T. T., and L. Zhou. Extraction Techniques for Selective Dissolution of Amorphous Iron
Oxides from Soils and Sediments. Jour. Amer. Soil Sci. Soc., v. 47, pp. 225-232,1983.
90. Shuman, L. M. Separating Soil Iron- and Manganese-Oxide Fractions for Microelement Analysis.
Jour. Amer. Soil Sci. Soc., v. 46, pp. 1099-1102,1982.
91. Shuman, L. M. Effect of Removal of Organic Matter and Iron- or Manganese-Oxides on Zinc
Adsorption by Soil. Soil Science, v. 146, pp. 248-254,1988.
92. Rendell, P. S., Batley, G. E., and A. J. Cameron.- Adsorption as a Control of Metal Concentration
in Sediment Extracts. Environ. Sci. Technol., v. 14, pp. 314-318,1980.
93. Belzile, N., Lecomte, P., and A. Tessier. Testing Readsorption of Trace Elements During
Chemical Extractions of Bottom Sediments. Environ. Sci. Technol., V. 23, pp. 1015-1020,1989.
94. Sellstone, C. J. Sequential Extraction of Fe, Mn, Zn, and Cu from Wetland Substrate Receiving
Acid Mine Drainage. Master's Thesis No. 3851, Colorado School of Mines, Golden, CO, 1990,88
PP-
95. Hedin, R. S., and R. W. Nairn. Sizing and Performance of Constructed Wetlands: Case Studies.
In: Proceedings of the 1990 Mining and Reclamation Conference, J. Skousen, J. Scencindiver,
and D. Samuel eds., West Virginia Univ. Publications, Morgantown, WV, 1990. pp. 385-392
96. Stark., L. R., Stevens, S. E., Webster, H. J., and W. R. Wenerick. Iron Loading, Efficiency and
Sizing in a Constructed Wetland Receiving Mine Drainage. In: Proceedings of the 1990 Mining
and Reclamation Conference, J. Skousen, J. Scencindiver, and D. Samuel eds., West Virginia
Univ. Publications, Morgantown, WV, 1990. pp. 393-401.
97. Kepler, D. A. Wetland Sizing, Design, and Treatment Effectiveness for Coal Mine Drainage. In:
Proceedings of the 1990 Mining and Reclamation Conference, J. Skousen, J. Scencindiver, and
D. Samuel eds., West Virginia Univ. Publications, Morgantown, WV, 1990. pp. 403-408.
98. Reynolds, J. S., Bolis, J. L., Machemer, S. D„ and T. R. Wildeman. Sulfate Reduction in A
Constructed Wetland. Preprints Div. Environ. Chem., v. 31, No. 1,1991, In press.
99 Wildeman, T.. Personal Communications, Colorado School of Minds, Colorado, October, 1990.
100 Ivanov, K.E.. Water Movement in Mirelands. trans, from Russian by A. Thompson and H. A.
Ingram. Acedemic Press, London, 1981.
R-7
-------
101 Garrett, W.S., and L.T. Campbell Pitt. Design and Construction of Underground Bulkheads and
Water Barriers. 7th Commonwealth Mining and Metallurgical Congress, Johannesburg, Vol. 3,
1961, pp 1283-1301.
102 Day, J.H., et. al., eds. Peat Testing Manual. Technical Memorandum 125, National Research
Council of Canada, Associate Committee on Geotechnical Research, Ottawa, 1973.
103 Fetter, C. W. Jr. Applied Hydroqeology. Charles E. Merrill Publishing, Columbus, Ohio, 1980.
104 Winneberger, J. H. T. Septic-Tank Systems, A Consultants Toolkit, Butterworth Publishers,
Boston, MA, 1984.
105 Bateman, Alan M., Economic Mineral Deposits. New York: John Wiley and Sons, 1950.
106 Schmidt, R., "Coal in America, An Encyclopedia of Reserves, Production and Use," McGraw Hill
Inc., 1979, Pages 14 and 15.
107 Hough, B.K. Basic Soils Engineering. New York: Ronald Press, 1969.
108 Pavoni, J. L., J.E. Heer, Jr., and D.J. Hagerty. Handbook of Solid Waste Dfrnosal. New York: Van
Nostrand Rienhold Company, 1975.
109 U.S. Environmental Protection Agency. "Constructed Wetlands and Aquatic Plant Systems for
Municipal Wastewater Treatment",' Doc. No. EPA/625/1-88/022,1988.
110 Bennett, P.G. and T.H. Jeffers. "Removal of Metal Contaminants K-rom a Waste Stream Using
Biofix Beads Containing Sphagnum Moss" in Proceedings of the Western Regional Symposium
on Mining and Mineral Processing Wastes, Berkeley, CA, May 30 to June 1,1990.
111 Dapples, E.C. and M.E. Hopkins. "Environments of Coal Deposition", This Geological Society of
America Special Paper No. 114, Boulder, CO, 1969.
112 Jarrett, P. M. 1983, Testing of Peats and Organic Soils. A symposium on Soil and Rock, Toronto,
Canada, June, 1982, ASTM Special Technical Publication 820. Philadelphia. PA.
113 Annual Book of ASTM Standards, Volume 4.08, Construction: Soil and Rock: Building Stones:
Geotextiles. ASTM, Philadelphia, 1987.
114 Gusek, J. J. and J.T. Gormley. Memorandum: "Observations arid Recommendations for Repairing
and Instrumentation, Big Five Wetlands Experimental Site", Pers. Comm., Jan, 1989.
115 Mahoney, W. D, ed. Means Heaw Construction Cost Data. 1990 R.S. Means Company, Inc.,
Kingston, MA, 1990.
116 Rental Rate Blue Ronk Equipment Guide Book Co., Palo Alto, CA.
117 Brodie, G. A., C. R. Britt, T. M. Tomaszewski, and H. N. Taylor. "Use of Passive Anoxic Limestone
Drains to Enhance Performance of Acid Drainage Treatment Wetlands" in Proceedings of the
1991 National Meeting of the American Society of Surface Mining and R >clamation,W. R. Oaks
and J. Bowden eds. American Soc. Surf. Mining and Reclam., Princeton , WV, pp. 211-228.
R-8
------- |